Energy efficiency of fluorescent lamps and their ballasts

The efficiencies of technical devices and processes are normally rated as percentages. Just with light this does not really match, since with respect to the perception of brightness the human eye is differently sensitive to different colours. Therefore the sensitivity of a standardised average eye has already been integrated into the unit for assessing the brightness of light sources. This unit is called lumen (plainly the Latin word for light). Hence, the efficiencies of lamps and luminaires need to be given in lumens per watt. This and only this indication is appropriate to measure and compare which technical device generates most light per unit of drawn electrical power. Theoretically an efficiency of 683 lumens per watt (lm/W) can be achieved. This, however, is only valid for mono-chromatic green light with a wavelength of 555 nm, where the human eye has its greatest sensitivity. So the greenest assumable lamp is indeed green. Irrespective of any political opinion, however, it remains more than questionable whether we really want to illuminate streets, squares, halls, offices, supermarkets or even living rooms in this way. For white light – or what we consider white when mixing all colours from 380 nm to 780 nm wavelengths – yields a theoretical maximum of 199 lm/W. Setting this equal to 100% brings fluorescent lamps already considerably closer to the desired 100% ideal than a modern diesel engine is. Speaking in these terms, an incandescent lamp could merely be compared to a vintage steam locomotive. The European Commission set out to support such trends towards such efficient lighting techniques and in June 1999 released the first draft of a directive with the objective to accelerate the transition of the Community industry towards the production of electronic ballasts and the overall aim to move gradually away from the less efficient magnetic ballasts and towards the more efficient electronic ballasts which may also offer extensive energy-saving features, such as dimming. This sounds as if it went without saying that an electronic ballast is:

  • always dimmable
  • and always the more energy efficient choice.

Back to the latter item in section 7.4. The misconception of the former has already been clarified in section 2.3. Adding to this:

To top

Do away with old rumours!

Fig. 7.1: Power intake of a 58 W fluorescent lamp with class B1 magnetic ballast during inrush, cathode pre-heat and warm-up phases: Moderate reactive overcurrent, no excess active power above rating at all!
Fig. 7.1: Power intake of a 58 W fluorescent lamp with class B1 magnetic ballast during inrush, cathode pre-heat and warm-up phases: Moderate reactive overcurrent, no excess active power above rating at all!

Let us first tidy up an old rumour which has it that fluorescent lamps consume a vast lot of electricity during start-up or warm-up – nobody specifies this precisely – and should therefore rather be kept in operation instead of turning them off when not needed for a shorter period. This rumour refers to magnetic ballast operation, since it is older than the invention of electronic ballasts, and is, of course, a balderdash, while the conclusive advice is largely correct: How could you ever draw such a high current out of a properly designed and fused system that within a few seconds a substantial amount of energy comes to be consumed? But still, even senior experienced electricians propagate this misconception, even though already their apprentices should be able to calculate that this is impossible. In fact, during pre-heating, when nearly all of the lamp apart from the filaments is shorted, the current is about 35% higher than the rating. Yet this is almost entirely reactive current. The reactive power during pre-heat actually rises about 90% and during cold operation about 30% above that of normal operation. The active – and thereby costly – share of the power approaches its rated value only slowly from below (Fig. 7.1). The truth about the story is, however, that it is not economical to switch fluorescent lamps very often because this contributes much to their ageing. The Wuppertal Institut für Umwelt Klima Energie once indicated a rough value of 10 minutes, above which switching off a fluorescent lamp would pay off. However, the ageing effect is crucially dependent on the optimal pre-heating conditions, as they are granted by an electronic starter and also claimed for by a warm-start electronic ballast, while in practice there is not much left of the latter, but rather, the magnetic ballasts with glow starter just as well as the immediate-start electronic ballast place a severe strain on the lamp (see section 2.1.2) In Table 7.1 both cases, the »glow start« and the »magnetic ballasts with electronic start« were opposed against each other, and the evidence is: The value indicated by Wuppertal Institut is quite precisely correct, assuming that it refers to the mean value across operation of different lamp power ratings and across operation with glow starters and electronic starte

Table 7.1: From some 10 to 25 minutes switching off is worthwhile
Table 7.1: From some 10 to 25 minutes switching off is worthwhile

Further, it is confirmed once again that when operating electrical appliances energy costs by far dominate all other shares of lifecycle costs. This can be observed time after time again. Exceptions are installations or devices with a very low loading or very short operating times (e. g. a private-owned car). But the lamp at 2.50 € consumes electricity for about 125 € during its 15,000 h life. Hence, although one »glow start« costs the lamp more than 8 h of its usable lifetime, switching it off already pays of for about 10 minutes for a 58 W lamp and from about 25 minutes onwards for an 18 W lamp, with variances depending on lamp price and electricity price. At these respective points in time the energy saved through switching off costs just as much as the additional wear imposed upon the lamp, or inversely, if you do not switch off, the energy invested costs the same as the lamp life saved. Using electronic starters, however, switching off is already worthwhile for one or two minutes! A warm-start electronic ballast cuts down the poor glow starter’s switch-off payback period by no more than just a few minutes; the usually used cold start electronic ballast even exceeds it by double the number of minutes!

To top

Old EU Directive

The EU first of all classified fluorescent lamp ballasts by the overall power intake of the ballast and lamp circuit, targeting at gradually phasing out the less efficient models. For instance, the classes and limits for linear lamps are displayed in Table 7.1. The clue about class A1 is that these values refer to dimmable electronic ballasts. A ballast is classified A1 if it fulfils the following requirements:

  • at 100% light output setting the ballast fulfils at least the demands of class A3;
  • at 25% light output setting the total input power does not exceed 50% of the power at the 100% light output setting;
  • the ballast must be able to reduce the light output to 10% or less of the maximum light output.
Fig 7.2: Split of system input power across a 58 W T8 lamp and its ballast
Fig 7.2: Split of system input power across a 58 W T8 lamp and its ballast

Now it would have looked somewhat odd to see the losses decreasing from class D all through class A2 but then to come across the inconsistency of an increase again towards the »;upper class« A1. So an appropriate definition was invented that says the rated power is that measured at 25% light output, since a dimmable system will not always be run at full power. This is just as logical as saying a car's engine does not always need to supply its maximum power, so if the car's top speed is 200 km/h, let's rate the engine power necessary to drive the car at 50 km/h as the nominal engine power.

Supplementary to this comes the curious fact that electronic ballasts are promoted with lower heat losses inside the ballast being one of the chief arguments, while named Directive allows higher losses in an electronic ballast than in a magnetic one. For instance, in Table 7.1 we learn that a 58 W lamp together with a magnetic ballast must not exceed a consumption of 64 W to comply with the requirements of class B1. This allows for a loss level of 6 W. However, when we shift to class A3, the lamp power drops to 50 W and the systems power to 59 W, allowing for a loss level of 9 W for the allegedly better ballast (Figure 7.2).

Fig 7.2: Split of system input power across a 58 W T8 lamp and its ballast
Fig 7.2: Split of system input power across a 58 W T8 lamp and its ballast

It is argued that on account of the high operating frequency the lamp efficiency was better (cf. section 7.8 regarding this) and therefore the luminous density nearly the same, only 4% less. First of all, the criteria neglect these 4%, since the Directive values and classes specify electric power only, not light output. Second, the EU realized later that the price premium for an electronic ballast was very high (Table 2.1, Section 2.3), while converting from a poor to a good type of magnetic ballast proved much more cost efficient (Table 7.2). Mind that these calculations were done without any consideration of interest rates for the invested capital!

Of course is has to be borne in mind that discounts of up to 80% from these prices may be achieved by industrial customers – not so much by electrical contractors. But then it should also be considered that mentioning the 4% difference in light output is not yet telling the full truth, since this difference does not refer to the rated power but to the deviating actual power intake of a lamp with a good magnetic ballast operated at rated voltage. A deliberate usage of the very generous tolerance margin, which in principle would not any longer be required for today's precise production methods, makes this possible. Still, even with this ballast design the same lamp is about 4% brighter than the same lamp with an electronic ballast, as will be seen in the next section. The 5 W difference between a class B1 magnetic ballast and a class A3 electronic ballast for a 58 W lamp, which the values of Table 7.2 are based on, thereby dwindles away to leave hardly any more than 2 W. So the indicated payback periods remain valid even for the high rebates when real electrical values measured at equal light outputs are compared.

Therefore named EU directive so far aimed at phasing out merely the classes C and D, which was done in November 2005 and May 2002, respectively, and which indeed is not a pity. Then, the market and technologies available so far will be investigated and assessed once again and further steps decided according to the results. So this is by far not a displacement plan for magnetic ballasts, as had been the initial intention and is still often believed even within the lighting industry. After all there would have been little sense in doing so, since, as the directive itself mentions at a different point, the improvement steps so far defined can be achieved with a cost premium around 2 € per lamp, while all improvements necessitating a conversion to electronic ballasts comes at an additional cost of 20 € per lamp.

Table 7.3: Table 1 of Directive 2005/32/EC
Table 7.3: Table 1 of Directive 2005/32/EC

Of course is has to be borne in mind that discounts of up to 80% from these prices may be achieved by industrial customers – not so much by electrical contractors. But then it should also be considered that mentioning the 4% difference in light output is not yet telling the full truth, since this difference does not refer to the rated power but to the deviating actual power intake of a lamp with a good magnetic ballast operated at rated voltage. A deliberate usage of the very generous tolerance margin, which in principle would not any longer be required for today's precise production methods, makes this possible. Still, even with this ballast design the same lamp is about 4% brighter than the same lamp with an electronic ballast, as will be seen in the next section. The 5 W difference between a class B1 magnetic ballast and a class A3 electronic ballast for a 58 W lamp, which the values of Table 7.2 are based on, thereby dwindles away to leave hardly any more than 2 W. So the indicated payback periods remain valid even for the high rebates when real electrical values measured at equal light outputs are compared.

Therefore named EU directive so far aimed at phasing out merely the classes C and D, which was done in November 2005 and May 2002, respectively, and which indeed is not a pity. Then, the market and technologies available so far will be investigated and assessed once again and further steps decided according to the results. So this is by far not a displacement plan for magnetic ballasts, as had been the initial intention and is still often believed even within the lighting industry. After all there would have been little sense in doing so, since, as the directive itself mentions at a different point, the improvement steps so far defined can be achieved with a cost premium around 2 € per lamp, while all improvements necessitating a conversion to electronic ballasts comes at an additional cost of 20 € per lamp.

To top

New EU Directive

Table 7.4: Excerpt from Table 17 of Directive 2005/32/EC
Table 7.4: Excerpt from Table 17 of Directive 2005/32/EC

But by and large it became time to decide about further steps. Therefore the EU repealed the Directive 2000/55/EU and replaced it with the Commission Regulation for implementing the »;Ecodesign« Directive 2005/32/EC (ErP Directive – Energy related Products) in the area of lighting components in April 2010. However, other than frequently heard even from lighting experts, this new Directive does not incur any plans to abolish magnetic ballasts!

This new Directive is entering into force in three stages: One year after entry into force preliminary limit values become valid. Three years after entry into force they become one level stricter, and eight years after entry into force these levels will be replaced with yet stricter final limits. This way industry shall be given sufficient time for a conversion. At least this is the principle behind it. The practical implementation is somewhat more lenient. The most substantial novelties are:

  • As an »;Ecodesign« directive it does not only provide electrical values but also e. g. maximum limits for the mercury content and minimum values for the lifetime expectancy of lamps.
  • Minimum values for complete luminaires are included – although the only »;Requirement« is that »;all luminaires … shall be compatible with ballasts complying with the first / second / third stage requirements«.
  • Minimum efficiencies (light output efficacies) are introduced for all common fluorescent and gas discharge lamp types – i. e. for the lamps alone without consideration of the ballast.
  • Apart from this, there are separate limit values for the energy efficiencies of ballasts, measured as the ratio of the lamp power rating divided by the sum of the lamp power rating plus the ballast power loss.
  • A most substantial difference at this point is that the Table 17 of this new implementing regulation (in part reproduced here as Table 7.4) distinguishes between three different power values of lamps: a nominal power, which is, so to say, only the name of the respective lamp, a rated power for mains frequency operation and a rated power for HF operation. The »;nominal« power is usually identical with the 50 Hz rated power unless the latter is not an integer figure but has a decimal. Then the decimal is omitted. For instance, an FD-38-E-G13-26/1050 lamp according to ILCOS (International Lamp Codification System) with a power rating of 38.5 W for 50 Hz and 32.0 W for HF has a nominal power of 38 W and is hence called a »;38 W (T8) lamp«. In the old Directive the difference between the nominal 38 W and the 32 W HF rating appeared like a 6 W advantage for the HF (electronic) ballast, which it has never ever been. The new approach is to measure, calculate and assess the energy efficiency of a »;magnetic ballast for a 38 W T8 lamp« based on an output of 38.5 W and the energy efficiency of an »;electronic ballast for a 38 W T8 lamp« based on an output of 32 W, rather than comparing the inputs only.
  • For dimmable electronic ballasts and other remote controllable lamp operating devices there are maximum stand-by losses.
  • Moreover, the power intake – of the lamp as well as the power loss in the ballast – is now to be measured at the point where the light output equals the light output rating of the respective lamp at 25°C ambient temperature. This is a substantial improvement against the present approach to classify only the power intake of the entire system and ignore any possible differences in light output between the uses of different ballasts on the same lamp. Thereby an impartial treatment of both magnetic and electronic ballasts is now granted. The application of two different measures but without respect to the light output comes to an end.

At this point unfortunately the widespread misunderstanding mentioned above arose. The pitfall is that the old designations A1, A2, A3, B1 and B2 continue to be used. A1 continues to stand for dimmable ballasts. Two new classes A1 BAT and A2 BAT (»;best available technology«) have been introduced, whereas, again, the former is reserved for dimmable ballasts. However, none of these class designations relates to the old Directive 2000/55/EU, but they are redefined within the new Directive 2005/32/EC. As described above, this is done by means of the ballast energy efficiencies as a percentage value of the real electrical output power by real electrical input power ratio. Now no class is linked to any certain ballast technology any longer, as has been the case so long, such as A for electronic, B (and formerly also C and D) for magnetic, except that A1 and A1 BAT are by definition dimmable ballasts. But their efficiencies are defined in terms of the other classes, as used to be the case before.

The lamp efficiencies, however, are not divided into classes. This would have gone way too far, since there are so many different types around. These limits must be taken directly out of one of the countless tables, starting with Table 1 (reproduced as Table 7.3 here) splitting double-capped lamps into T8, T5HE and T5HO types. This table reveals rather clearly how far T5HO lamps fall behind not only T5HE but also behind T8 lamps. So T5 lamps are in no way generally more efficient than T8 lamps, as is frequently assumed and alleged (also see section 7.7). This becomes evident at the very first look at the new documentation. The changes in detail are, as far as energy efficiencies are concerned:

Lamp efficiencies

  • First stage requirements: One year after the entry into force of the new regulation T5 and T8 lamps shall have at least the rated luminous efficacies as specified in Table 1 of said regulation (see Table 7.3 here), all measured at 25°C ambient temperature. This appears to be a bit unfair against T5 lamps, though, because for some good reasons they are optimized for an ambient temperature of 35°C.
  • Second stage requirements: Three years after the entry into force the requirements for T8 lamps from the first stage will be expanded to all double capped fluorescent lamps. So this may mean that the T5HO lamp has to go, unless it undergoes some substantial improvement so as to match the requirements for T8 lamps!
  • Third stage requirements: Eight years after the entry into force fluorescent lamps are not faced directly with any additional efficiency requirements. It only says they »;shall be designed to operate with ballasts of energy efficiency class at least A2 according to Annex III.2.2«, but this can be said of any common fluorescent lamp already now. Note: It does not say, »;The ballast / system shall meet the energy efficiency requirements of class A2 according to 2000/55/EU«, which would have been something entirely different!

Ballast efficiencies

  • First stage requirements: One year after the entry into force of the new regulation the minimum energy efficiency index class shall be B2 (according to Table 17 of 2005/32/EC!) for ballasts covered by Table 17, and A1 for dimmable ballasts covered by Table 19 (of 2005/32/EC, not of 2000/55/EU, which it supersedes! See Table 7.4 here). Dimmable ballasts shall comply with the requirements for class A1 according to Table 19 in 2005/32/EC. Parallel with the old Directive, this implies that the ballast's efficiency shall match the requirements of class A3 when set to full power and shall use no more than 50% of its full power when set to 25% light output.
  • Second stage requirements: Three years after the entry into force there is no change to non-dimmable ballasts for fluorescent lamps. Limits for high-pressure discharge lamps are upgraded, and the stand-by consumption of dimmable ballast goes from 1 W down to 0.5 W maximum, as used to be the case in the old Directive.
  • Third stage requirements: Eight years after the entry into force the minimum efficiencies of ballasts are:<br/> η = 71% for ballasts up to 5 W (nominal power),<br/> η = 91% for ballasts from 100 W upwards,<br/>

This calculation of η is called EBbFL in 2005/32/EC. As described above, this approach yields different efficiency values for the same lamp, depending on whether it is being operated with a magnetic or an electronic ballast if different power ratings are given for either of these. The required efficiencies turn out to be a little bit lower for electronic ballasts, which is obvious when one enters slightly lower values of PLamp into the formula.

This calculation of η is called EBbFL in 2005/32/EC. As described above, this approach yields different efficiency values for the same lamp, depending on whether it is being operated with a magnetic or an electronic ballast if different power ratings are given for either of these. The required efficiencies turn out to be a little bit lower for electronic ballasts, which is obvious when one enters slightly lower values of PLamp into the formula.

So also this new document makes no statement whatsoever about any prohibition of magnetic ballasts. Otherwise what sense would there be in defining new values for classes B1 and B2? Rather, there used to be quite an imbalance to the advantage of electronic ballasts in the old scheme according to Directive 2000/55/EU, which will now have to go in the foreseeable future. While it is always argued among experts that one of the advantages of electronic ballasts was the lower internal power loss, even the old Directive 2000/55/EU stated the very opposite! For instance, it says there referring to a 58 W T8 lamp:

  • Lamp power with magnetic ballast: 58 W,
  • systems power with magnetic ballast (class B1 – old): ≤ 64 W.
  • This allows for a power loss of ≤ 6 W inside the magnetic ballast.
  • Converted to the new calculation method, this would yield a minimum efficiency requirement of<br/>η ≥ 58 W / 64 W ≈ 91%,<br/>matching even the new class A2, rather than just B2, which would already satisfy stage 1 of the new regulation! The EBbFL requirement of stage 3 is only:<br/>η = EBbFL ≥ 89.6%,<br/>so it is also easily fulfilled by the good old magnetic ballast!

But at the same time it also says in the old 2000/55/EU document:

  • Lamp power with electronic ballast: 50 W,
  • systems power with electronic ballast (class A3 – old): ≤ 59 W.
  • This allows for a power loss of ≤ 9 W inside the electronic ballast!
  • Converted to the new calculation method, this would yield a minimum efficiency requirement of<br/>η ≥ 50 W / 59 W ≈ 85%,<br/>passing B2 (new) but failing B1 (new), therefore just about compliant with stage 1. The EBbFL requirement of stage 3 is<br/>η = EBbFL ≥ 89.1%<br/>here, hence also failed! In other words: The old Directive allocates a higher class to a poorer ballast and vice versa!

The new classification requires the energy efficiency of a 58 W ballast for a T8 lamp to be 84.7% in class A3 or 86.1% in class B1, respectively. It is a bit confusing why the new class B1 requires a higher efficiency than class A3. In fact it also allocates a higher class to a poorer ballast here. This is the case not with all, but with a number of ballasts and may be a remnant of the old definitions for classes B1 and A3, whenever it is better concealed there (see above). After all this is nothing to worry too much about because these requirements are only a transition to the continuously calculated method of the final stage No. 3. However, it does become evident that a magnetic ballast of class B1 according to the present (old) classification has far lower losses than required by the present (old) class A3; moreover, it even complies with the new A2 requirements! An electronic ballast according to the old class A3, however, just about manages to comply with the new class A3. This does not really look like a prohibition of magnetic ballasts but rather the opposite!

Table 7.5: Comparison of electrical data and light outputs with small fluorescent lamps
Table 7.5: Comparison of electrical data and light outputs with small fluorescent lamps

To top

Avoiding avoidable losses in small fluorescent lamps

Fig. 7.3: One and the same ballast is designed for 4 different single lamps and (not listed here for reasons of space) 3 possible tandem configurations
Fig. 7.3: One and the same ballast is designed for 4 different single lamps and (not listed here for reasons of space) 3 possible tandem configurations

Advertisements in favour of electronic ballasts occasionally claim that in magnetic ballasts »;up to 30%« of the luminaire's total power intake is absorbed as losses. First of all, it remains to be noted that a statement like »;up to«, very popular though it may be, is also totally inappropriate to make any statement at all, unless simultaneously complemented by indicating the mean and the maximum values (see this, p. 289). The same here: the greatest relative losses occur with the smallest lamps. This can be traced back to a law of nature once called »;Paradox of the Big Machine«. In a 58 W lamp, for instance, it is only 13% (see section 7.4). Moreover, the piece numbers of smaller lamps are also smaller, and so their overall contribution to the total losses is all the smaller. So the indication »;up to 30%« tells nothing at all. While on the other hand, this is even understated. For instance, when measuring the power shares on a TC-S lamp rated 5 W and operated with a conventional magnetic ballast, a lamp power magnitude of 5.6 W may be found, along with once again the same magnitude of ballast losses, so in this case you may very well speak of 50% losses. Generally, however, the operating voltage drop across smaller, i. e. shorter fluorescent lamps of the same type family is lower than with the longer types of the same series. Thereby, for longer lamps a larger share of the voltage drops across the lamp and a smaller share across the ballast. At the same time the current rating is a bit lower with the longer lamps, while the ballast remains the same (Fig. 7.4).

Bild 7.4
Bild 7.4: Ein- und dasselbe Vorschaltgerät eignet sich für 4 verschiedene Einzellampen sowie (hier aus Platzgründen nicht aufgeführt) für 3 mögliche Tandemschaltungen

Doch damit nicht genug, denn bei den TC-S-Lampen mit 5 W, 7 W und 9 W ist die Brennspannung so gering, dass sich an der regulären Netzspannung von 230 V zwei Lampen in Reihe betreiben lassen. So verdoppelt sich die Brennspannung effektiv natürlich wieder. Da auch für diese so genannte Tandemschaltung wieder das gleiche Vorschaltgerät zum Einsatz kommt wie für den Einzel-Betrieb, liegen die Ströme und somit die Lampenleistungen jeweils etwas unter dem Nennwert. Um den Fehler zu minimieren, sind die KVG so ausgelegt, dass die Lampenleistungen in der Einfachschaltung etwas über den Nennwerten liegen. Insgesamt hat dies zur Folge, dass das KVG umso weniger belastet wird, je mehr Last daran hängt. Mehr Lampenlast führt gleichzeitig zu absolut fallenden Verlusten, spart also relativ gesehen gleich doppelt (Bild 7.6).

Gleichzeitig verbessern sich auch noch die Wirkungsgrade der Lampen, wenn diese nicht mit voller Leistung betrieben werden, und umgekehrt sind die Wirkungsgrade bei Überlast der Lampe schlechter. Eine Vergleichsmessung durch ein renommiertes, unabhängiges Licht-Institut (Tabelle 7.6), wobei außer den elektrischen Daten auch die Lichtströme gemessen wurden, brachte dies zu Tage. Die 9-W-Lampe mit KVG landete darin auf dem letzten Platz, da die TC-S-Lampen mit 5 W und 7 W gar nicht erst angetreten waren. Nach den Ergebnissen in Bild 7.6 hatte sich dies erübrigt. Dass anderenfalls die 5-W-Lampe im wahrsten Sinne des Wortes das Schlusslicht gebildet hätte, stand danach schon fest.

Fig. 7.5: TC-D lamp 18&nbsp;W, energy efficient magnetic ballast and electronic ballast for this (top) and energy efficient magnetic ballast for a commonplace T8 lamp of equal power rating (bottom)
Fig. 7.5: TC-D lamp 18 W, energy efficient magnetic ballast and electronic ballast for this (top) and energy efficient magnetic ballast for a commonplace T8 lamp of equal power rating (bottom)

But the tandem circuit is also applicable to T8 lamps with a power rating of 18 W. Although in this case different ballasts are meant to be used for single and tandem configuration, the results are similarly profitable. Here, too, the finding is that the power loss in the class B1 ballast attributable to two lamps is even lower than that in the class B1 ballast for only one lamp (Fig. 7.6).

Now there are some more lamp types with a rating of 18 W available on the market, e. g. the TC-D lamp. But this one has a much higher operational voltage drop and can therefore not be operated in tandem mode. But since the voltage drop across the lamp under normal operating conditions is greater, the voltage drop across the ballast is smaller. So the required reactive power rating of the ballast is also selected accordingly smaller – and thereby the whole ballast is. But this is not yet all. When the lamp voltage is greater, the lamp current is also smaller and reduces the required reactive power level again (see section 4.2). Therefore a magnetic ballast for a TC-D lamp can be built extremely small, even when desigen according to efficiency class B1 – even smaller than a commensurate electronic ballast (Fig. 7.7)! So especially a luminaire with a TC-D lamp and a high-efficiency magnetic ballast saves space, production costs and energy in one go.

Table 7.6: Compilation of measurements on 18&nbsp;W fluorescent lamps with magnetic and electronic ballasts – a 240&nbsp;V magnetic ballast having been measured in the middle erratically assuming it was a 230&nbsp;V model; measurement therefore repeated in the row below
Table 7.6: Compilation of measurements on 18 W fluorescent lamps with magnetic and electronic ballasts – a 240 V magnetic ballast having been measured in the middle erratically assuming it was a 230 V model; measurement therefore repeated in the row below

The latter finds its confirmation when you add another light output measurement. For this reason the single and tandem operation modes of class B1 magnetic ballasts for 18 W and 2*18 W, respectively, were compared to a single and twin operation mode on an electronic class A2 ballast rated 18 W or 2*18 W, respectively. The result is compiled in 3 blocks of 7 measurements of the light flux F each, displayed in Table 7.6:

  • One single T8 lamp,
  • two T8 lamps in tandem or twin mode, respectively, one TC-D lamp,

with the following ballasts and data:

  • Electronic ballast at the lower voltage tolerance limit 90% (207 V),
  • electronic ballast at rated voltage (230 V),
  • electronic ballast at the upper voltage tolerance limit 110% (253 V),
  • magnetic ballast at the lower voltage tolerance limit 90% (207 V),
  • magnetic ballast at rated voltage (230 V),
  • magnetic ballast at the upper voltage tolerance limit 110% (253 V)
  • magnetic ballast at the voltage magnitude where the light output equals that of the same lamp with electronic ballast at 230 V.

For measuring the T8 lamp in single-mode, a single-lamp electronic ballast was used instead of using the twin-mode one and connecting only one lamp, which would have been possible but would have yielded wrong results. The most crucial results can be found in Table 7.6, represented as the light efficiency in lumens per watt electrical power intake of the whole lamp and ballast system. The light efficiency cannot be given in per cent because regarding brightness the human eye is differently receptive to light of different colours. Therefore the sensitivity of a standardised average eye is already integrated into the unit for brightness. This unit is called lumen (simply the Latin word for light). So the efficiency of lamps and lumiaires has to be given in lumens per watt. So this and only this unit is adequate to assess which technical device provides the greatest brightness per power intake. Of course the share of ballast losses in the total power intake can be given as a percentage – as done in the last column of the table. However, with the electronic ballasts the required measurement of the lamp power, the ballast output power to the lamp so to say, was not possible due to the high output frequency. Therefore the efficiency ηLamp of the lamp alone could not be calculated. Nevertheless, the following results can be read and conclusions drawn from the table:

1. The advantages of the tandem configuration and of the TC-D lamp already found in the pre-measurement with respect to reactive power find their confirmation.

2. The magnetic ballast power loss increases highly over-proportionally to the systems operating voltage. At 253 V the power loss is usually double as high as at 207 V. Together with the slight increase of lamp efficiency ηLamp the voltage reduction practice results as an efficient means of loss reduction for all magnetic ballast configurations.

3. Inversely as with 58 W lamps (see section 7.4), the lamps are about 4% brighter with electronic than with magnetic ballasts. With the twin electronic ballast compared to the magnetic tandem configuration the difference is even 8%. The operating voltage on the tandem has to be turned nearly up to 244 V before the same brightness as with the electronic twin ballast is achieved.

Therefore when assessing the light efficiency two different approaches have to be considered:

4. Either the luminaires are operated at rated voltage in either case. The comparison will then be closer to what will usually happen in practice, though it is not objective. We are then talking about a systems power of 19.13 W with electronic ballast versus a systems power of 24.47 W with magnetic ballast. A payback time for the well over 5 W saved cannot be given, as the impact of the price premium for an electronic ballast upon the price for a complete lighting installation is subject to substantial variances. However, with an energy price of 10 ct/kWh it takes 1872 operating hours to save the first Euro. This cornerstone can be used for the according conversions: At 5 ct/kWh it takes 3744 hours, at 20 ct/kWh it takes 936 hours to save 1 Euro.

5. Or you calculate objectively. Nobody will increase the line voltage in order to achieve precisely the same brightness with the used / planned magnetic ballast as with the electronic ballast not used, but the lighting planner might include a few more lamps if the decision for magnetic ballasts has been taken. This would have practically the same effect as if the same number of lamps were connected to a line voltage of 241.7 V, which would be equivalent to the difference between 19.13 W and 26.18 W systems power, say 7 W. So the real, effective »;savings cornerstone« is then 1418 operating hours per Euro saved at 10 ct/kWh.

6. Moreover, it becomes obvious that the limits of the EU directive, which is 24 W systems power in class B1 and 19 W in class A2, are in principle not complied with, neither by the magnetic nor by the electronic ballast. Only by being rather lenient accounting to metering inaccuracy the EEI classes can still be seen as just about fulfilled.

But by all means this mode of operation does not represent the optimal combination. The power loss in a 36 W ballast is not double the loss in an 18 W ballast (»;Paradox of the Big Ballast«), about the triple advantage of the tandem mode not even to speak. Rather, the respective conclusions to above items 4 to 6 for the twin or tandem modes of two 18 W lamps will be:

7. Comparing the operation at rated voltage in either case, the difference between magnetic and electronic ballast operation is now only more 2 W per system, whereas a system now comprises two lamps and one ballast. So with an electricity price of 10 ct/kWh it takes 5000 operating hours to save one Euro. Or, selecting a different example: At uninterrupted permanent duty with 8760 h/a and an electricity price which is usually quite inexpensive for such use, e. g. 5.7 ct/kWh, the electronic ballast saves precisely one Euro per year.

8. With equivalent brightness, that is, assuming corrected voltage for the magnetic ballast (although, as mentioned earlier, hardly anybody will ever do this in practice) the difference is 6.6 W per system. With an electricity price of 10 ct/kWh one saves one Euro in about 1500 operating hours.

9. Although the directive provides a separate line with limits for two lamps being operated on one ballast, the values per lamp are identical to those for the single-mode operations as under item 6. Very much unlike with the configuration described under item 6, however, the limits are by far kept here: The electronic ballast remains well over 1.5 W below the class A2 limit, the magnetic ballast even falls 3.5 W below the B1 limit.

On the TC-D lamp the following can be observed:

10. The efficiency is about 5% to 10% poorer than that of the T8 lamp. This may be due to the compact design which leads to a part of the light generated hitting the lamp itself.

11. Here the use of the electronic ballast results in an uncommonly high saving of 28% on equal voltage or 34% at equal light output, respectively. It by far fulfills the requirements for class A2, while the magnetic one does not really match the limit for class B1. The magnetic one may have been designed a bit too small in favour of facilitating the design of very small luminaires (Fig. 7.7 top right), and in electrical engineering skimping on active material (magnetic steel and copper) always comes at the price of reduced efficiency. It has to be considered, however, that these two measurements possibly cannot really be compared because they could not be carried out on the same lamp. The TC-D lamp for magnetic ballast operation is equipped with an integrated starter and therefore has only two connections (Fig. 7.7). The starter is wired internally. The version for electronic ballast operation requires four pins.

12. Unlike the other electronic ballasts used in this test, the one for this lamp is not equipped with an electronic power stabilisation to offset variances of the input voltage.

Fig. 7.6: Split of total luminaire power intake for different TC-S lamp configurations with the same ballast
Fig. 7.6: Split of total luminaire power intake for different TC-S lamp configurations with the same ballast

Therefore when assessing the light efficiency two different approaches have to be considered:

4. Either the luminaires are operated at rated voltage in either case. The comparison will then be closer to what will usually happen in practice, though it is not objective. We are then talking about a systems power of 19.13 W with electronic ballast versus a systems power of 24.47 W with magnetic ballast. A payback time for the well over 5 W saved cannot be given, as the impact of the price premium for an electronic ballast upon the price for a complete lighting installation is subject to substantial variances. However, with an energy price of 10 ct/kWh it takes 1872 operating hours to save the first Euro. This cornerstone can be used for the according conversions: At 5 ct/kWh it takes 3744 hours, at 20 ct/kWh it takes 936 hours to save 1 Euro.

5. Or you calculate objectively. Nobody will increase the line voltage in order to achieve precisely the same brightness with the used / planned magnetic ballast as with the electronic ballast not used, but the lighting planner might include a few more lamps if the decision for magnetic ballasts has been taken. This would have practically the same effect as if the same number of lamps were connected to a line voltage of 241.7 V, which would be equivalent to the difference between 19.13 W and 26.18 W systems power, say 7 W. So the real, effective »;savings cornerstone« is then 1418 operating hours per Euro saved at 10 ct/kWh.

6. Moreover, it becomes obvious that the limits of the EU directive, which is 24 W systems power in class B1 and 19 W in class A2, are in principle not complied with, neither by the magnetic nor by the electronic ballast. Only by being rather lenient accounting to metering inaccuracy the EEI classes can still be seen as just about fulfilled.

But by all means this mode of operation does not represent the optimal combination. The power loss in a 36 W ballast is not double the loss in an 18 W ballast (»;Paradox of the Big Ballast«), about the triple advantage of the tandem mode not even to speak. Rather, the respective conclusions to above items 4 to 6 for the twin or tandem modes of two 18 W lamps will be:

7. Comparing the operation at rated voltage in either case, the difference between magnetic and electronic ballast operation is now only more 2 W per system, whereas a system now comprises two lamps and one ballast. So with an electricity price of 10 ct/kWh it takes 5000 operating hours to save one Euro. Or, selecting a different example: At uninterrupted permanent duty with 8760 h/a and an electricity price which is usually quite inexpensive for such use, e. g. 5.7 ct/kWh, the electronic ballast saves precisely one Euro per year.

8. With equivalent brightness, that is, assuming corrected voltage for the magnetic ballast (although, as mentioned earlier, hardly anybody will ever do this in practice) the difference is 6.6 W per system. With an electricity price of 10 ct/kWh one saves one Euro in about 1500 operating hours.

9. Although the directive provides a separate line with limits for two lamps being operated on one ballast, the values per lamp are identical to those for the single-mode operations as under item 6. Very much unlike with the configuration described under item 6, however, the limits are by far kept here: The electronic ballast remains well over 1.5 W below the class A2 limit, the magnetic ballast even falls 3.5 W below the B1 limit.

On the TC-D lamp the following can be observed:

10. The efficiency is about 5% to 10% poorer than that of the T8 lamp. This may be due to the compact design which leads to a part of the light generated hitting the lamp itself.

11. Here the use of the electronic ballast results in an uncommonly high saving of 28% on equal voltage or 34% at equal light output, respectively. It by far fulfills the requirements for class A2, while the magnetic one does not really match the limit for class B1. The magnetic one may have been designed a bit too small in favour of facilitating the design of very small luminaires (Fig. 7.7 top right), and in electrical engineering skimping on active material (magnetic steel and copper) always comes at the price of reduced efficiency. It has to be considered, however, that these two measurements possibly cannot really be compared because they could not be carried out on the same lamp. The TC-D lamp for magnetic ballast operation is equipped with an integrated starter and therefore has only two connections (Fig. 7.7). The starter is wired internally. The version for electronic ballast operation requires four pins.

12. Unlike the other electronic ballasts used in this test, the one for this lamp is not equipped with an electronic power stabilisation to offset variances of the input voltage.

Figure 7.7: 18 W fluorescent lamps in single and tandem mode comparison
Figure 7.7: 18 W fluorescent lamps in single and tandem mode comparison

But by all means this mode of operation does not represent the optimal combination. The power loss in a 36 W ballast is not double the loss in an 18 W ballast (»;Paradox of the Big Ballast«), about the triple advantage of the tandem mode not even to speak. Rather, the respective conclusions to above items 4 to 6 for the twin or tandem modes of two 18 W lamps will be:

7. Comparing the operation at rated voltage in either case, the difference between magnetic and electronic ballast operation is now only more 2 W per system, whereas a system now comprises two lamps and one ballast. So with an electricity price of 10 ct/kWh it takes 5000 operating hours to save one Euro. Or, selecting a different example: At uninterrupted permanent duty with 8760 h/a and an electricity price which is usually quite inexpensive for such use, e. g. 5.7 ct/kWh, the electronic ballast saves precisely one Euro per year.

8. With equivalent brightness, that is, assuming corrected voltage for the magnetic ballast (although, as mentioned earlier, hardly anybody will ever do this in practice) the difference is 6.6 W per system. With an electricity price of 10 ct/kWh one saves one Euro in about 1500 operating hours.

9. Although the directive provides a separate line with limits for two lamps being operated on one ballast, the values per lamp are identical to those for the single-mode operations as under item 6. Very much unlike with the configuration described under item 6, however, the limits are by far kept here: The electronic ballast remains well over 1.5 W below the class A2 limit, the magnetic ballast even falls 3.5 W below the B1 limit.

On the TC-D lamp the following can be observed:

10. The efficiency is about 5% to 10% poorer than that of the T8 lamp. This may be due to the compact design which leads to a part of the light generated hitting the lamp itself.

11. Here the use of the electronic ballast results in an uncommonly high saving of 28% on equal voltage or 34% at equal light output, respectively. It by far fulfils the requirements for class A2, while the magnetic one does not really match the limit for class B1. The magnetic one may have been designed a bit too small in favour of facilitating the design of very small luminaires (Fig. 7.7 top right), and in electrical engineering skimping on active material (magnetic steel and copper) always comes at the price of reduced efficiency. It has to be considered, however, that these two measurements possibly cannot really be compared because they could not be carried out on the same lamp. The TC-D lamp for magnetic ballast operation is equipped with an integrated starter and therefore has only two connections (Fig. 7.7). The starter is wired internally. The version for electronic ballast operation requires four pins.

12. Unlike the other electronic ballasts used in this test, the one for this lamp is not equipped with an electronic power stabilisation to offset variances of the input voltage.

To top

How to make magnetic ballasts more efficient than electronic ones

Fig. 7.8: Output current curve of an electronic ballast (H.-G. Hergesell, Paderborn Airport), recorded with 3 different power analysers
Fig. 7.8: Output current curve of an electronic ballast (H.-G. Hergesell, Paderborn Airport), recorded with 3 different power analysers

It is not so sure whether the undoubted, measurable rest of the efficiency improvement beyond the 4% difference of light output with electronic ballasts really bases on the high frequency – or perhaps rather on the current waveshape fed into the lamp? It was tried to find this out by means of another measurement at a special independent lighting institute. The idea behind this was another found statement that the efficiency of a fluorescent lamp is not optimal at rated current but better at lower current, as is the case with a lot of electrical equipment, incandescent lamps exempted. If this is valid for the TRMS or arithmetic mean value of same current, then it also goes for each and every instantaneous value along the curve. So, with sine current, efficiency drops within the range around the peak, since most of the light is generated during this time span. If the output current of an electronic ballast were rectangular, then there would be no efficiency drop at any point of the curve – and energy efficiency would be better, because this constant value would be considerably lower than the peak value of a sine wave. Indeed the output current looks more like a rectangle than like a sinus (Fig. 7.8).

Fig. 7.9: Test samples for the measurements documented in Fig. 7.10 and Table 7.7
Fig. 7.9: Test samples for the measurements documented in Fig. 7.10 and Table 7.7
  • One stone-old ballast from an installation that had already been knocked down in 1987, still being rated 220 V and of course not efficiency classified and thereby falling into class D according to Table 7.1.
  • One new »;superslim« magnetic ballast, inevitably falling into class C, since in electrical engineering restrictions of space nearly always come at the price of restricted efficiencies.
  • One new magnetic ballast efficiency class B2.
  • One new magnetic ballast efficiency class B1.
  • One mint condition electronic ballast rated efficiency class A3.

Now on each of these 5 samples all required parameters were measured, always using the same lamp: Active and reactive power across the whole system, active power (loss) across the ballast, and of course the light output of the lamp. All of the results have been compiled in Table 7.7) but, the graphic evaluation of this verbose table included in the download version of this text. Among others the 4% difference of luminous density in favour of magnetic ballasts (»;5000 lm at rated voltage versus »;4700 lm with electronic ballast) finds its confirmation here, but beyond this, the graphic evaluation of this verbose table provides much more ease of interpretation (Fig. 7.10). Unfortunately, on account of the high output frequency at the terminals of the electronic ballast, it was not possible to measure its output power. This is not a tragedy, though, since the most important data, system input power and light output, could be measured. The following can be concluded from the results:

  • On the electronic ballast neither system input power nor light output vary with varying voltage. So the device under test fully compensates variances of the supply voltage within the tested range, which is usually seen as an advantage – and one commonly expected from electronic ballasts. A deliberate variation of power input and thereby of light output via the feeding voltage, however, is therefore not feasible.
  • Of course the energy efficiency comparison turns out best for the electronic ballast at 230 V, but at 200 V the A3 electronic one is only more about the same as the class B1 and even the class B2 magnetic ballasts, and at 190 V the electronic one performs poorer! So at 190 V supply voltage the B1 and even the B2 turns out as an A2, and the B1 should even be classified as A2, since the efficiency of the A3 model has not altered, but while those of both the B1 and the B2 and B1 models have exceeded it!
  • The information of the light output with electronic ballasts being about 4% reduced against that of efficient magnetic ballasts at rated input voltage (not necessarily rated input power – see next bullet point) finds its confirmation.
  • The rated lamp power is not always reached precisely at rated voltage. Other than the old ballast, the later magnetic ballast models of all classes reach their rated power only considerably above the rated system voltage. At 230 V, however, the electric lamp input power still falls considerably below the 58 W rating. After all that has been said so far, such design, e. g. deliberate utilisation of the permitted minus tolerance, must be seen as a reasonable approach.
  • Still, this does not yet mean that the electric values are now totally comparable to those of an electronic ballast! With classes C, B2 and B1, the light output is around 5000 lm, while the electronic ballast tested here provided only 4720 lm.
  • So the improved magnetic ballast models under test only feed about 53.5 W into the lamp instead of the rated 58 W, and still, the lamp shines 4% brighter than with the electronic ballast! Hence, for reasons of objectivity, in order not to compare apples with pares, the electronic ballasts' light output at 230 V would rather need to be compared to those values metered on the improved magnetic models at about 220 V actual voltage.
  • At this point of operation the actual lamp inputs were only more around 50 W – matching the rating given for an electronic ballast. This makes the deviating lamp ratings for operation with magnetic versus electronic ballast operation appear relative and raises doubts about the improvement of efficiency at high frequencies. The confinement to this statement is the lack of measured electric output power at the electronic ballast. However, the systems' power intakes with electronic A3 and magnetic B1 ballasts at the points of equal light outputs deviated from each other only more by exactly 2.1 W, interpolating between the two measured points at 220 V (4662 lm) and 230 V (4952 lm) to the 4720 lm the lamp performs with electronic ballast.
  • By switching from a poor class C magnetic ballast to a class B1 model the efficiency at rated lamp power is improved by 10% from 70.3 lm/W to 77,4 lm/W, since the share of ballast losses among the total input power drops from 22.9% to 15.0%. The price premium for the more efficient magnetic ballast therefore pays off in nearly all applications, short payback periods guaranteed.
  • Contrary to this, the persistent use of very old poor efficiency ballasts – especially if still designed for 220 V line voltage rating – leads to a significant lamp overload with highly over-proportional increase of losses and reduced lamp life but only little increase of light output.
  • By reducing the operating voltage from 230 V to 190 V, the efficiency e. g. of a lamp with a class C ballast is improved from 73.0 lm/W to 84.1 lm/W, that is by well over 15%. When a class B1 ballast is used, the light efficiency still rises from 80.6 lm/W to 89.1 lm/W and hence still by about 10.6%. So the reduction of the feeding voltage also pays off, especially in cases where poor magnetic ballasts are not replaced with better ones. However, this shall not be an excuse for further operating »;old scrap« any longer, for also with high-efficiency magnetic ballasts the fairly simple and usually rather inexpensive voltage reduction technique provides pretty short payback periods. The upgrade from anything to a B1 ballast really is the bargain, and some greater or smaller voltage reduction may come on top of it as a perfection.
Fig. 7.10: Efficiency of various ballasts with the same lamp at varying voltages according to&nbsp;Table 7.7
Fig. 7.10: Efficiency of various ballasts with the same lamp at varying voltages according to Table 7.7

The high variance of efficiency even with moderate voltage reduction on a lamp circuit with whatever type of magnetic ballasts has three main reasons:

  • Copper loss and approximately also iron loss in the ballast rise by the square of the current. Therefore the power lost in the ballast drops over-proportionally when current is reduced (see Table 7.7).
  • Lamp voltage increases when lamp current decreases (Fig. 1). Therefore lamp power decreases under-proportionally with decreasing supply voltage, while lamp efficiency moderately increases and simultaneously ballast losses dramatically drop.
  • On account of this, current drops over-proportionally to the voltage reduction and accelerates the former effects.

To offset the lower absolute light output, about 150 magnetic ballast luminaires operated at 190 V would have to be used to replace 100 electronic ballast luminaires. Now since the 150 magnetic ballast luminaires are simultaneously the more energy efficient solution, a cost premium would be acceptable in replacing the electronic with magnetic ballasts in order to save energy, inverting the usual approach. Still, this need not necessarily be any more expensive. Cases have been reported where the solution with 100 electronic ballasts has been bid higher. So the payback time may assume a negative value! Adding the cost for voltage reduction, it is still very short. In two example cases from Switzerland 50 open longitudinal 58 W luminaires were bid alternatively with electronic ballasts at 2575 CHF and 50 commensurate luminaires with magnetic ballasts, regardless of efficiency class, at 1700 CHF. So no premium was charged at all for a better efficiency class of the magnetic ballast, but it was very well possible to get 150 lamps equipped with these at a lower price than 100 luminaires with electronic ballasts. Whether the price premium in such cases really improves the electrical contractors' businesses or whether the electronic ballast merely adds to the turnover but cuts revenues is yet another question to be critically scrutinized in each individual case.

In May 2000, being informed about this, the EU made an amendment to their document that any other measure judged appropriate to improve the inherent energy efficiency of ballasts and to encourage the use of energy-saving lighting control systems should be considered.

Table 7.8: Power savings and light losses at reduced operating voltage
Table 7.8: Power savings and light losses at reduced operating voltage

Indeed, in Germany there are at least three manufacturers of dedicated voltage reduction plant that is meant to operate fluorescent lighting at reduced voltages. Refurbishment in existing installations is easy as long as dedicated power lines for the lighting have been installed. Occasionally voltage reducers are also offered for the general supply but these have to be treated with care. Many power consuming devices have the inverse behaviour as fluorescent lamps with magnetic ballasts. Incandescent lamps, whenever living a lot longer, yield a dramatic loss of energy efficiency. Induction motors as well as practically all electronic devices, including decent electronic ballasts with constant light output regulation, have an increased instead of decreased current intake with reduced line voltage. Ohmic losses in the mains and especially inside the motor increase instead of decreasing. Also electronic ballasts of the type tested here, with constant regulated light output, react in this way and therefore cannot be influenced by varying the line voltage. With fluorescent lighting, however, the loss of luminous density can be offset by installing additional lamps – or simply taken for granted, which often is acceptable.

On the other hand, the undervoltage extends the lamp life by about 33% ... 50%, the voltage reduction plant manufacturers claim. The trade association of German lamp manufacturers points out that also the opposite can happen because the optimum filament temperature is not reached. So far it can only be concluded from the conflicting statements that this issue has not yet been experimentally investigated. Life time tests of longlife devices take a long time by definition.

To top

Warranty

In some cases the reduction uses only the permissible ±10% mains voltage tolerance range at the junction box, which would bring it to 207 V. Some use another 3% permissible drop within the installation, coming down to 199 V. Others go as low as 185 V. A further reduction is not feasible, since lamps – at least those without serial compensation – just cease to work then. This is an energy saving function, no dimming technique, since the brightness regulation range is not very wide. Various additional functions are available, control in steps or continuous, day time and temperature dependent (for street lighting) and others. Lamps are always started at full voltage and stepped down only when they have reached their normal operating temperature. The technique could also be used to operate old luminaires still equipped with ballasts rated 220 V on the new uniform European 230 V line voltage, especially since lamp efficiency drops and ballast loss rises dramatically at overvoltage and lamp life is shortened. But normally the old ballasts will have a very poor efficiency anyway and will be worthwhile replacing. On the other hand, the undervoltage extends the lamp life by about 33% ... 50%, the voltage reduction plant manufacturers claim. However, ZVEI, the trade association of German lamp and ballast manufacturers, points out that also the opposite can happen, because the optimum filament operating temperature is not reached but after a fierce contention both sides agreed on a permission for the voltage reduction plant manufacturers to publicise a lamp life gain of 22%.

So far it can only be concluded from the conflicting statements that this issue has not yet been experimentally investigated. Life time tests of longlife devices take a long time by definition. Moreover, ZVEI point out that undervoltage operation, as far as it falls below the permitted tolerance limit of 207 V, represents an operation outside the manufacturer's specification and therefore voids warranty. This is correct but rather relates to the fact that the affected ratings, also those for the compensation capacitors, as explained in section 5, have not been revised any more for decades. However, if the saving technique saves just 5 W all together through improved lamp efficiency and reduced ballast losses, then the lamp saves its own price within 10,000 hours of operation. If the lamps at average live as long as this, you may very well lose your warranty, and you still do make a bargain. Your warranty does under no circumstances include more than the purchase cost of a failed lamp, if any, or a ballast, respectively, but to assume a magnetic ballast might fail on account of undervoltage is as absurd as believing your car might fail because you don't always drive full speed.

By all means it is remarkable in this context and finally creates some clarification that the company Aura Light officially calculates with a life-time expansion effect and grants a 50,000 h warranty for their Longlife lamps. As a prerequisite, a device has to be used that ignites the lamps at full voltage and then gradually reduces in small steps when the operating temperature has been reach

Alternative solutions

A few other solutions may in certain situations achieve the same effect with an even lower or no price premium at all:

  • In some luminaires, 2 smaller fluorescent lamps may be connected in series on 1 magnetic ballast (and 2 starters), as described in sections 4.2 and 7.3.
  • Magnetic ballasts are also available with 240 V rating. Using these on a 230 V supply will normally not cause any problems, even less if electronic starters are used. The current is slightly reduced, accompanied by the over-proportional saving effects as described for lower input voltage, but with an even better stability of light because the full voltage is applied. As described earlier in this section, the operation of the modern magnetic ballasts at rated voltage did not match the point of operation with the electronic ballast in the test. Rather, although the electric lamp input power already fell 4% below the rating with the tested magnetic ballasts, the light output was still 4% above that of the electronic one. So the operation of these magnetic ballasts at 4% undervoltage provides a much closer equivalence to the electronic ballast than at rated voltage.

To top

Conclusions

For a concise insight into the economic potentials, here comes a review of all the saving quotes. By reducing the voltage from 230 V to 190 V (by 17.4%) reductions according to Table 7.9 are achieved:

 

Table 7.9: With magnetic ballast class D magnetic ballast class C magnetic ballast class B2 magnetic ballast class B1 electronic ballast class A3
ballast losses drop by 65.9% 70.2% 70.0% 69.5% ≈0%
electrical lamp power drops by 31.2% 38.3% 37.0% 38.3% ≈0%
total electrical systems power drops by 38.6% 44.9% 42.1% 46.7% ≈0%
light output drops by 27.1% 36.5% 35.1% 36.2% ≈0%
Subsequently the overall efficiency improves by 18.6% 15.2% 12.2% 10.6% ≈0%

 

 

 

Fig. 7.11: Demonstration model for a direct comparison
Fig. 7.11: Demonstration model for a direct comparison

It has to be borne in mind, though, that at 230 V and with the class B1 magnetic ballast the lamp already supplied 4.7% more light than was the case with the electronic ballast (at any voltage between 190 V and 230 V). Therefore the true light loss is not 36.2% but only 31.5%. So, to be precise, 46% more lamps would need to be installed to obtain the same light flux. Their costs need to be balanced against the savings with energy and lamp replacement. Final customers or their contractors will need to calculate this in each individual case. In general you may select to install some 20% to 30% more lamps as a compromise, alone because with the more even distribution of light a lower total light level may suffice. To calculate this in each individual case is the lighting planners' task.

It is remarkable in this context that the human sensitivity for brightness, as is the case for noise level, is logarithmic. Differently from noise, however, the applied assessment dimensions are linear, so a measured enhancement of luminous density by a factor 10 is perceived as a doubling of brightness, 100 times more light is felt to be triple, 1000 times more seems just 4 times brighter and so on. In the end of a day a number of test persons were not able to say whether certain lamps were operated at 190 V or at full line voltage. One company constructed a demonstration panel for this purpose (Fig. 7.11), in which 2 luminaires, each with 2 fluorescent lighting tubes rated 58 W (in lead-lag circuit) are operated, one luminaire at full line voltage and one at 190 V or even just 185 V. So visitors can convince themselves: You actually see no difference in brightness even here where both variants are inevitably viewed simultaneously side by side (Fig. 7.12)! A power saving of 23.5% costs only 4.8% loss of light. What remains to be subtracted from this saving is the power loss inside the voltage reducer (Fig. 7.13) but which is only 13 W in the case of this small unit, i. e. 1 W per each of the maximum 13 lamps that could be connected. 2 luminaires, each with 2 fluorescent lighting tubes rated 58 W (in lead-lag circuit) are operated, one luminaire at full line voltage and one at 190 V or even just 185 V. So visitors can convince themselves: You actually see no difference in brightness even here where both variants are inevitably viewed simultaneously side by side! What you do see is a difference between the leading and the lagging lamp in each luminaire. They seem to have a slightly different colour. This basically should not be the case, so if anything then this shouts for an adjustment of the serial capacitance rating (see section 4).

Fig. 7.12: Brighter or not brighter, that is hardly a question any more here: On the left 20,520 lx at 111 W, on the right 21,560 lx at 145 W
Fig. 7.12: Brighter or not brighter, that is hardly a question any more here: On the left 20,520 lx at 111 W, on the right 21,560 lx at 145 W

After all, when the EU Directive was finally published in September 2000 it read:

»This Directive aims at reducing energy consumption … by moving gradually from the less efficient ballasts, and to the more efficient ballasts which may also offer extensive energy saving functions.«

To top

Energy savings with special T8 lamps

It has been right three times already during recent years that the company Philips attracted attention releasing two new T8 lamps and one additional gadget. The first one was simply a compatible retrofit solution with a lower power rating for luminaires equipped with both magnetic or electronic ballasts – but which prove to be not quite as simple as that. The second one represents a solution adequate for magnetic ballasts only, whereas the starter has to be replaced with a very special version of an electronic starter.

Philips TL-D Master Eco

A new series of T8 lamps was recently released that comes with reduced power ratings at the same sizes, e. g. just 51 W instead of 58 W for the 1.5 m lamp. The manufacturer claims these lamps could be used as a fully compatible replacement for the existing lamps without replacing a ballast, be there magnetic or electronic ones installed in an existing luminaire. Now customers start to wonder and to ask whether these lamps can actually save energy. Well, this depends on what you mean by that and what you expect.

In order to reduce the lamp input power with a given ballast you need to vary the lamp impedance. Assuming as a first approach the lamp were an ohmic load, which is far from true, it still remains acceptable to view it as a greatly active load, while the ballast is approximately inductive. The serial interconnection of these two elements theoretically yields two values of lamp impedance (here assumed to be resistive – Fig. 7.15). In the case of the traditional 58 W lamp the lower one is the one used. In order to shift to 51 W the lamp impedance needs to be further reduced. Unfortunately this yields a higher current, while the losses in a magnetic ballast increase with the square of the current. So a slight reduction in lamp power comes at the price of a steep increase in losses and hence gnaws on the system efficiency from both sides. While it would have been attractive to use the upper one of the two theoretically possible points of operation, this is practically impossible. A fluorescent lamp with a voltage drop of more than something around half of the line voltage will not start on same line voltage.

Fig. 7.15: In order to reduce the lamp input power with a given ballast the lamp impedance has to be reduced
Fig. 7.15: In order to reduce the lamp input power with a given ballast the lamp impedance has to be reduced
Table 7.10: Catalogue data of T5 HE and T5 HO lamps with electronic ballasts compared to the measured data of T8 lamps with magnetic ballasts described in detail in Section 7.4
Table 7.10: Catalogue data of T5 HE and T5 HO lamps with electronic ballasts compared to the measured data of T8 lamps with magnetic ballasts described in detail in Section 7.4

It would have been fine if the lamp alone performed a significantly higher efficiency to offset the higher ballast losses, but another measurement (Table 7.9) revealed that this is unfortunately not the case. In fact the lamp efficacy is even poorer, and all the more is the system efficacy. Sure the absolute power intake is lower, but you receive a minor electrical energy saving at the price of a major loss of brightness.

With an electronic ballast the balance looks considerably better. The system efficiency is then at least the same. So when using this type of lamp the difference between magnetic and electronic mode of operation is considerably greater than with the ordinary type. Still, the morals is: If you can afford to sacrifice some light the best approach is to use good magnetic ballasts, regular T8 lamps and a voltage reduction unit. Then and only then you receive a major electrical energy saving at the price of a minor loss of light output.

To top

Philips TL-D Master Power Saving Set

Fig. 7.16: Special lamp
Fig. 7.16: Special lamp

Next, yet another novelty was released by the same manufacturer just in time for the great salient 2010 Light & Building fair in Frankfurt, again a compatible replacement for a T8 lamp, which this time even reduced the power rating from 58 W to 37 W. The lamp (Fig. 7.16) is sold under the smart sounding name »Power Saver Set«, whereas »set« already indicates that it is sold along with a special electronic starter (Fig. 7.17) and is hence usable with magnetic ballasts only. This starter differs, however, from the – though still too little – known electronic starter (see section 3.1), while it is a special version of same.

But how is it possible that a starter, wired in parallel with the lamp and normally not conducting any current at all while the lamp is on, reduces the overall active power intake of the whole system? The only difference can be that this starter – also after successful start – does still conduct some current. It seemed somewhat unlikely that some sort of »power plant« had been integrated into it, but a plain and simple measurement (Fig. 7.18) revealed

Fig. 7.17: Special starter for the special lamp
Fig. 7.17: Special starter for the special lamp

When switching the lamp on in a circuit with this special starter and a magnetic ballast classified EEI = B1 you observe just nothing for the first. The starter works just like a »common« electronic starter and ignites the lamp after optimal pre-heat time at the optimal point of the phase (at current peak) without any repeated flashing, blinking and flickering. The power intake lies within the usual range, and the plot of the current also looks quite familiar. Then, after one minute, the active power starts to drop. After two minutes it stabilizes at around 46 W (Fig. 7.19). The current curve still looks unobtrusive.

Only when looking at the starter current individually (centre of Fig. 7.19 and Fig. 7.20), one comes to realize: A phase-angle control was integrated into the starter which after one minute begins to open up a shunt current parallel to the lamp, which stabilizes this current at about 240 mA after two minutes (Fig. 7.21). If the starter is removed while the lamp is working the lamp power flips back to »common« values – and the brightness even rises beyond this level (see below). After re-positioning the starter the process starts again from the beginning (Fig. 7.19).

The phase angle control may here contain a little symmetry error: The TRMS current included a DC constituent of some 60 mA or 11% (Fig. 7.19 top right) in this example measurement, which did not arise from the mains, for there is nothing like that to be found on the line voltage (Fig. 7.19, row 1, column 3). Now the light flickers, since the two semi-waves of the lamp current differ in magnitude. Replacing the lamp or the starter with a different one of the same type, turning either of them round to swap polarity sometimes helps, but usually it does not.

Fig. 7.18: Test configuration »Power Saver Set« Master TL-D 58&nbsp;W → 37&nbsp;W
Fig. 7.18: Test configuration »Power Saver Set« Master TL-D 58 W → 37 W

The promised reduction of the lamp power, however, is kept. While the power displayed in Fig. 7.19, bottom, second take, is not the system power intake but a fantasy value because the lamp current was measured alone here and was not multiplied by the associated lamp voltage but by the overall voltage (see wiring sketch in Fig. 7.19, bottom right). Further measurements from a lighting lab

  • This special lamp already represents an improvement in its own right, with the special starter or without, but without it a brightness of 5648 lm and a respectable system efficacy of 90.3 lm/W are achieved. Compared to this, earlier measurements on a 58 W standard lamp (see section 7.5) with the same magnetic ballast had yielded 4952 lm and 80.6 lm/W (top of Table 7.11).
  • The lamp’s power intake at rated voltage lies significantly below 58 W, similarly as observed in said earlier measurement.
  • Also the power loss in the class B1 magnetic ballast is nearly the same in either case, being around 8 W.
  • Then it goes on following the same principle as described in section 7.6.1 for the 51 W lamp:
  • Using the special starter reduces the overall system power intake from 62.5 W down to 43.9 W. This is not trivial, since it represents a reduction by approximately 30%!
  • But at the same time the luminous output drops from 5648 lm down to 3761 lm, by 33.4%. Hence, the system efficacy drops from 90.3 lm/W to 85.6 lm/W. So once again, you lose more light than what you save on power.

So what would be more obvious than the idea of operating the new lamp without the dedicated starter (with a »common« electronic starter) but using a voltage reduction device? For as explained earlier, this technique incorporates the feature of enhancing rather than decreasing the overall luminous efficacy! This was also shown in a further measurement:

  • If you reduce the system operating voltage down to precisely 192.8 V, then the obtained brightness matches exactly the value measured with the special dimming starter at rated voltage. This pushes the system up to the fabulous efficacy of 97.6 lm/W!
  • The efficacy of the lamp alone even rises above 105 lm/W!
  • The overall power intake of the entire system is now only more 38.5 W. This would match operating the 37 W lamp on rated line voltage if the ballast power loss were only more 1.5 W. The actual ballast power loss, however, is around 6 W in this mode of operation.
  • This represents a 37% savings against a »standard« system on »ordinary« voltage, hence just 63% of the power intake at no less than 76% of the »ordinary« brightness.
Fig. 7.19: Measurements on a system made up from a magnetic ballast class. B1, special 37&nbsp;W lamp and special starter (top – total current; centre – the current share through the starter; bottom – the current share through the lamp)
Fig. 7.19: Measurements on a system made up from a magnetic ballast class. B1, special 37 W lamp and special starter (top – total current; centre – the current share through the starter; bottom – the current share through the lamp)

Now it only more remains to be clarified how it is possible that the ballast current, representing the total current after all, drops from 597 mA to 512 mA when opening up a supplementary current flow (notably at constant feeding voltage) through the phase angle control paralleling the lamp. In fact not only the active power drops due to this, but the overall apparent power of the whole system is reduced from 137 VA down to 118 VA, the reactive share drops from 122 VA to 109 VA and, subsequently, the ballast power loss decreases from 8 W to 6 W. This is an advantage against the initially mentioned solution with the 51 W lamp, where the ballast power loss did not drop but rose. The following remains to be considered in order to understand this:

Fig. 7.20: Current curves in the lamp, in the starter and total current under steady-state conditions
Fig. 7.20: Current curves in the lamp, in the starter and total current under steady-state conditions

The total current is defined more strongly by the ballast than by the other two components, hence it is highly inductive in nature. The control angle starts cutting »from behind« and does not assume a great amplitude, so it takes effect exclusively during the second half of any current semi-wave, i. e. in an area where the current has the inverse polarity as the line voltage has. This represents the reactive share of the total current. Because the dimmer shorts the load within this section of the phase the current can »gather a bit more impetus«. At first sight one would suppose that in this way the reactive share of the current should rise more than the active share would drop, hence the overall apparent current would increase. As can be seen, however, this is not the case, since the current’s zero crossing is deferred even further back now (to the right in Fig. 7.20) than it already is on account of its inductive character. Due to this, the time span during which the current flowing »in the wrong direction« is slowed down again by the applied voltage is expanded. The change of polarity (zero crossing) of the current, from where on it starts to »speed up« again, is delayed even more than is the case anyway. Hence a peak value of the same magnitude as before cannot be achieved any more, and its TRMS value drops – unexpectedly, as one might think, but yet somehow logically.

Fig. 7.21: Warm-up process of cold lamp up to stabilisation of phase-angle control
Fig. 7.21: Warm-up process of cold lamp up to stabilisation of phase-angle control

Well, the remaining question now is why this lamp is exclusively sold along with the special dimmer starter. Why can you not purchase and operate the significantly better lamp individually? Supplying more light even as a mere substitute for conventional lamps, it could achieve fantastic efficacies together with a voltage reducer. If you limit yourself to a reduction within the permissible tolerance of -10%, i. e. 207 V, you would reap the same light output as with a standard lamp – but at 15% less power input. 12% of this advantage is due to the better lamp and 3% to its efficacy improvement at lower voltage.

The »dimmer-starter«, however, though it does reduce the power intake, also reduces the efficacy. At 70.3% of the rated power it provides only 66.6% of the original brightness. The attractive aspect of the solution with the dedicated starter is, as was the case before with the 51 W lamp, that this energy savings measure can be carried out practically without any investment costs. These two solutions target exclusively at the replacement market. You only need to switch to a more expensive lamp when a lamp replacement is due anyway, and as a rule the necessity for a starter change comes along with the lamp replacement anyway. After replacement it will never ever need to be replaced again. But nothing comes for free. If you are displeased to reduce your efficiency along with your power reduction, you will require a voltage reduction device on top. Then, indeed, efficiencies may go beyond compare, be it in the new build sector or in the replacement area – and this with a technique which costs money only once and will provide salient longevity and robustness ever after.

To top

The Smart Starter by Philips

Fig. 7.22: The programmed Smart Starter by Philips
Fig. 7.22: The programmed Smart Starter by Philips

In 2012 Philips released their third novelty in this area, this time only a starter, the so-called Smart Starter, that was meant for use with any commonplace T8 fluorescent lamp between 32 W and 58 W and a magnetic ballast. It is an electronic starter, but programmed to a certain time interval, by default 8am to 6pm on weekdays. Then the respective lamp is switched on, the rest of the week switched off automatically. Different times are available on request, a flexible programming unit is planned to be released if demand arises. A sample was procured (Fig. 7.22) and tested. Indeed the internal clock was found to be very precise, and the starter has no measurable residual, ancillary or stand-by consumption (Fig. 7.23).

Fig. 7.23: Measurement of Smart Starter by Philips, power turned on 5 minutes before the programmed switch-off time at 6pm
Fig. 7.23: Measurement of Smart Starter by Philips, power turned on 5 minutes before the programmed switch-off time at 6pm

It is fitted with an internal long-life battery. Now if the free-programming option comes, this may be a very useful, plain and simple tool for a substantial energy savings share in certain applications. The payback period is accordingly very short. The device addresses shops and there particularly the deepfreezes where often the light remains on because a separate switch is missing, causing additional cooling demand on top of the unnecessary electricity consumption and wear of the lights. A good idea for future releases might be to power the internal clock by the mains voltage, that is usually applied all of the time, and buffer it with a supercap just in case this should not be given.

To top

Are T5 lamps more efficient?

Reports about new lighting systems and renovations of lighting installations regularly quote the »;new more efficient T5 lamps«, as if it went without saying that the efficiency of a T5 lamp is by default higher than that of a T8 lamp. Albeit, a look at the catalogue data already reveals that this, if at all, only applies to the so-called T5HE lamps optimized for High Efficiency. Those classified T5HO, optimized for High Output, perform significantly poorer than a commonplace triphosphor T8 lamp (Table 7.12).

Table 7.12: Catalogue data of T5 HE and T5 HO lamps with electronic ballasts compared to the measured data of T8 lamps with magnetic ballasts described in detail in Section 7.4
Table 7.12: Catalogue data of T5 HE and T5 HO lamps with electronic ballasts compared to the measured data of T8 lamps with magnetic ballasts described in detail in Section 7.4

In the cases of electronic ballasts the input power and light output remains stable independently of the input voltage, while the input power to a magnetic ballast system of course varies greatly with input voltage. So a point can be found (at 217 V) where the measured lamp power in a 58 W T8 lamp driven by a magnetic ballast is exactly 49 W and thereby matches the rating of an existing T5 lamp with a light output of 4300 lm. But at this point, namely of equal power inputs to the T5 and T8 lamps, the light output of the T8 lamp is already ≈4600 lm – even though it was operated at mains frequency here and the T5 lamp, of course, at high frequency, as specified. This casts serious doubts over the practical effect of the theoretical efficiency improvement at high frequency operation. Or over the »;more efficient T5 lamps«. Or both.

Table 7.13: Values and classes of linear fluorescent T5 lamps with ballasts (values for B, C, D classes missing because these lamps are specified for use with electronic ballasts only)
Table 7.13: Values and classes of linear fluorescent T5 lamps with ballasts (values for B, C, D classes missing because these lamps are specified for use with electronic ballasts only)
Fig. 7.24: Magnetic ballast for T5 lamps – here still as prototypes (with paper labels), now available from stock
Fig. 7.24: Magnetic ballast for T5 lamps – here still as prototypes (with paper labels), now available from stock

Due to the curious fact mentioned earlier that the Directive allows higher losses in an electronic ballast than in a magnetic one, e. g. a 54 W T5 lamp with a class A3 ballast may have a systems power of 63 W (Table 7.11), yielding a ballast loss share of 14.3%, while the magnetic B1 system with a 58 W lamp – formally and officially – must not exceed 64 W and is thereby limited to a loss share of 9.4% (Table 7.1). But it was also mentioned there that in practice the lamp power with a magnetic ballast is found to be only between 53.5 W and 54.5 W, and that in the end of a day the systems power is crucial and not its split across lamp and ballast. Howsoever, through the theoretical or the practical approach, the T5 lamp hits a tough challenge to match the expectation to provide a better efficiency than a good T8 magnetic system has. On top of this, the unfortunate fact that in one system the rated light output is reached more or less around the rated power intake and in the other one even far below, both catalogue data and the Directive yield unrealistic payback times. Unfortunately this will never ever be discovered, since the electricity consumption of the lighting installation is not registered separately and because during a renovation a new system will always replace an over-aged one which is insufficient in all respects. Never ever will e. g. an optimized modern magnetic system be replaced with an optimized modern electronic system. So the energy savings remain a matter of belief and trust in what the specifier specifies.

Last but not least, it is not at all true that T5 lamps could be operated with electronic ballasts only. They are meant to be, but they also work at 50 Hz with a magnetic ballast. Accordingly, one major manufacturer brought a series of magnetic ballasts to market (Figure 7.24). One company from Hungary had the matter put to the test and had a series of T5 systems tested at a university (Table 7.1).

Fig. 7.14: Repetition of Fig. 7.10, but without zero base suppression – and suddenly the incandescent lamps shows up at the very bottom of all lighting techniques
Fig. 7.14: Repetition of Fig. 7.10, but without zero base suppression – and suddenly the incandescent lamps shows up at the very bottom of all lighting techniques

Last but not least, it is not at all true that T5 lamps could be operated with electronic ballasts only. They are meant to be, but they also work at 50 Hz with a magnetic ballast. Accordingly, one major manufacturer brought a series of magnetic ballasts to market (Figure 7.24). One company from Hungary had the matter put to the test and had a series of T5 systems tested at a university (Table 7.1).

To top

Energy savings due to high frequency

It has been mentioned several times so far that the improved efficiency of an electronic system is in part due to the better lamp efficacy at high frequency. This effect does indeed exist, albeit it becomes difficult when it is attempted to quantify this effect. In their publications fabricators are hence courageous enough to rate it as high as 10% to 15%. Prof. Ron Hui from Imperial College London and Hong Kong University would not quite believe this either, procured 20 T5 sample lamps 28 W by Osram and another 20 pieces by Philips and measured. At average, an efficacy improvement of only 3.6% (against 50 Hz operation with a newly developed magnetic ballast for T5 lamps) was registered. Not even the mysterious »;Ballast lumen factor (BLF)« will help offsetting this effect, the strange factor of 0.95 which, for reasons none of the lighting experts can (or wants to?) explain, is multiplied by a lamp's light flux when evaluating a test. For electronic gear BLF = 1. So even though on paper 5% of its efficacy is taken off the magnetic system quite legally here, this still does not suffice to fully do away with its advantages. The company carrying out the market surveys under the EU's commission, who wrote up very detailed reports about this, subsequently states in there: »;Please note that for the reference ballast a normalized ballast lumen factor of 0.95 has been chosen (this suggests that manufacturers tend to under-run lamps on average on magnetic ballasts).«

To top

Energy savings with dimmable ballasts

Fig. 7.25: Light outputs of different systems employing T5 and T8 fluorescent lamps, plotted against the absolute electrical systems power input
Fig. 7.25: Light outputs of different systems employing T5 and T8 fluorescent lamps, plotted against the absolute electrical systems power input

So if you want to save energy you will try to reduce the lighting level automatically, dependent on the level of available daylight. As you have learned in Section 7.4, the reduction of the voltage fed into magnetic ballasts, although it does save energy, does not reach far enough to call it a »;dimming technique«, so you will try with dimmable electronic ballasts. But again, the question was how far the savings potential would go. Measurements were commissioned with an independent certified lighting laboratory by the German Copper Institute DKI and the company M&R Multitronik to complement the measurements on magnetic ballasts described in Section 7.4. In order to obtain objective, comparable results compliant with the existing measurements reported in Section 7.4, a twin electronic ballast together with two commonplace, readily available T5 lamps (triphosphor, colour rendering index 840) were used, since it has turned out in Section 7.3 that a twin electronic ballast usually has lower losses than two single-lamp ones. As for the lamps, the lowest wattage of the biggest available size (1449 mm) was chosen because the greatest efficiency could be expected from these. This led to a rating of 2*35 W.

The T8 lamps had been tested before with an ambient temperature of 25°C according to the standard where they usually perform their best efficacy. The T5 lamps were additionally measured with an ambient temperature of 35°C, deviating from the standard, since for some good reasons they are optimized to this ambient temperature.

Fig. 7.26: Light efficacies of different systems employing T5 and T8 fluorescent lamps, plotted against the relative electrical systems power input
Fig. 7.26: Light efficacies of different systems employing T5 and T8 fluorescent lamps, plotted against the relative electrical systems power input

The results were summarized in Fig. 7.25, where the systems' light outputs were plotted against the respective electrical power intake. Further, a line was included in the plot, representing a constant efficacy of η = 80 lm/W, which should represent a guideline for the efficiency in today's lighting installations. In this way the following becomes evident:

  • The efficacy of any T8 system increases during input power reduction. Generally speaking, the values in the lower segment lie above the 80 lm/W »;guideline«, while in the upper half they lie below, and especially in the overload range they strongly tend to flatten out.
  • The T5 lamps exhibit the inverse behaviour: Efficiency decreases during dimming. Values in the upper range tend to lie above the »;guideline«, while values in the lower range will rather lie below.
  • The improved efficiencies of the T5 lamps at 35°C against the values measured at 25°C become quite obvious.
  • But unfortunately this type of plot is not very adequate for a direct comparison of either system against the other one because there are not any two lamps T5 and T8 with equal electrical power ratings available.

It was therefore successfully tried to find a different method to compare both of the systems to each other by plotting the light efficacy against the relative system power (Fig. 7.26). In this type of graph, a direct comparison of different systems should be possible when keeping the following remarks in mind:

  • For the T8 systems, what is meant by relative systems power is the ratio of the measured systems power at the respective voltage divided by the systems power measured at rated voltage of the same system (for instance, with the old magnetic ballast class EEI=C the reference point representing 100% is 69 W, that of an improved magnetic ballast class EEI=B1 is 61.4 W, which represent the respective systems values measured at 230 V).
  • For the T5 system, what is meant by relative systems power is the ratio of the measured systems power at the respective dimming level divided by the systems power measured when set to full light output (100%, i. e. same system with dimmer set to full power).
  • For ease of orientation, the minimum requirements for class A1 are plotted in the chart in stroke-dotted lines once for a reference ambient temperature of 25°C and once for 35°C.
  • The non-dimmable electronic ballast also included in the measurements could not reasonably be displayed in this format, since its power intake, along with the light output, is invariable and would have yielded only a dot.

Hence, the above description facilitates the following observations:

  • The T5 system under test by far exceeds the minimum requirements.
  • It becomes even clearer now that the efficacy of the T8 system increases due to power re-duction (and accordingly drops inadequately in the overload range), while the efficacy of the T5 system is best at full power and drops during dimming.
  • At full load and 25°C ambient temperature the T5 system is about equally efficient as the best T8 magnetic system (EEI=B1).
  • At full load and 35°C ambient temperature the T5 system is ≈10% more efficient than the best T8 magnetic system is at 25°C.
  • At ≈75% of their respective electrical power input measured at 230 V or, respectively, of the undimmed lamp, the efficacy of the best T8 magnetic system is about equal to that of the T5 system at 35°C.
  • When reducing, respectively dimming, the systems power to ≈60%, the efficacy of the T5 system even drops below that of a T8 system with an ancient class D magnetic ballast which was rescued from a scrap metal container back around 1986.
  • When reducing to ≈50% input power the possible range of application for the voltage re-duction technique ends. Otherwise the lamps will go out completely. A greater dimming range can be implemented with dimmable electronic ballasts only.

This facilitates the following conclusions:

  • Dimmable ballasts provide only a rather limited energy savings potential. Who wants to save energy should reasonably employ a combination of voltage reduction and subsequent grouped automatic switching (e. g. from the aisle side to the window side in an office) after exploiting the (limited) »;dimming« potential of voltage reduction – optionally, wherever possible, applying a technique which comes without any need for stand-by consumption and using electronic starters7, which spare on the lamp life as well as on the employees' nerves wherever switching occurs more frequently than once a day.
  • The voltage reduction technique is no replacement for dimming. Who wants to dim has to use dimmable electronic ballasts. On the background of today's knowledge all techniques for dimming magnetic ballasts that have ever been around are makeshift solutions and do not satisfy modern needs. They should therefore not be considered any longer.

Still, these considerations do not yet include the following circumstance:

Dimmed operation of fluorescent lamps represents permanent cathode heating operation. The position »;Lights off« is usually identical with the position »;Dimmed down to 0«, where no more light is generated, but the filaments continue to be heated (Fig. 7.27), albeit these states are often confused with each other. After all the lights are off in any case. As far as only the daylight sensor exerts control over the electronic ballasts in this way but over the rest of the time it is made sure that e. g. in an office the light is turned off properly after work and on weekends, this is not yet a tragedy. It has not proven a good practice, however, to completely turn off the supply voltage to the lighting e. g. by switch timers. After all, the cleaners are yet to come somewhat later on, and nobody knows exactly how much later, and at some point in time there will be someone working right through into the night or on a Saturday – usually the boss himself. When the lights go out this will rouse trouble, and this trouble will happen only once. The facility manager, being the target of the corresponding complaints, but not in charge of coming up for the electricity bill, will immediately sabotage the timer switch. So the lights will always be »;dimmed down to 0« whenever the light is »;off«. This also sabotages the underlying energy savings efforts carried out during the planning stage and also the payback calculations which these were based upon. In cases of favourable exceptions, this is taken into regard, and the installation is configured adequately, so that users do not lose their daily savings at night. This is the case, for instance, when a daylight sensor and a presence indicator work together on the »;project« of dimming but only the daylight sensor really dims, while the motion indicator releases a signal equal to »;turn lights off completely« (including cathode heating). Subsequently solely the stand-by consumption of the ballast is left, which at present is limited to 1 W, from 2012 onwards only more 0.5 W, which limits the loss.

As a refurbishment, each sensor and each actuator in such a control system will actually require a power supply of its own to draw power from the mains. The net DC power requirement of any one such device may be as low as a few milliwatts, but a power supply unit including a transformer will be needed in every case. The smallest commercially available transformers, however, have a power rating of ≈1 VA and a no-load loss rate of ≈1 W. While the load loss is irrelevant at such a low degree of loading, it is the no-load losses in the great number of such devices adding up to the bulk of the stand-by power demand in such an installation. Modern control systems, which can be refurbished fairly easily if the commensurate cables or at least empty installation tubes have been fitted already during the building's construction stage, use one single centrally located power unit and supply both DC power distribution and the signals via the same cable. This technique incurs the potential to reduce the stand-by consumption down to a fraction, although, by principle, a permanent demand for power also during periods without demand for light remains for all of the time. Hence, it remains to be considered in each individual case whether the implementation of a less sophisticated savings technique together with latest magnetic ballasts, which shuts off the entire lighting completely or, for instance, groups of it, could not be the cheaper and at the same time less power consuming approach.

Irrespectively of this, advertisements by the lighting industry continue devising a scenario promising an energy savings potential of, say, 80%.2 The mere transition from a magnetic to an electronic ballast on the same lamp is supposed to bring about 25% of energy savings. But, first of all, the worst and oldest magnetic system you can find in existing installations is always plotted against the best and most modern electronic solution. The existence of »;improved« energy efficient magnetic gear is ignored and the differences between these blurred. Further, it remains totally unclear which system is compared to which one, but assuming the base case was a 58 W lamp being converted here to use the best electronic ballast classified A2, this would yield a maximum permissible systems power intake of 55 W. Then the old magnetic system, against which a savings potential of 25% was tapped, must have had a power intake of over 73 W. This matches the stone old magnetic ballast mentioned earlier which was recovered from a scrap metal container already back in 1986. It yielded a light flux of 5300 lm, while the same lamp performed only 4700 lm with an electronic ballast, but this difference is dropped unmentioned.

The transition to a T5 lamp even claims for a savings potential of 50%. So this can only refer to a 35 W HE lamp, with a maximum permissible power intake of 39 W with an electronic A2 ballast. However, this system has a light output of only 3300 lm – which again does not appear worth any mention. Of course you can save energy when you replace a Greyhound bus with a Volkswagen Rabbit and keep silent about the number of seats available. Additionally, the »;cut-off« feature is mentioned as a further advantage here, which comes as an inevitable inherent constituent of the magnetic system, but there it finds no mention.

Finally, the introduction of a dimming technique is supposed to raise the savings to 80% – without any explanation of the circumstances and assumptions under which this was calculated or measured.

The comparison to the measurements described in the previous sections, however, does not show any similarity of any kind to these claims of the advertisements (Fig. 7.28).

Fig. 7.28: Light outputs of different systems employing T5 and T8 fluorescent lamps, plotted against the absolute electrical systems power input
Fig. 7.28: Light outputs of different systems employing T5 and T8 fluorescent lamps, plotted against the absolute electrical systems power input

On top of all, the value measured at 230 V had to be used for the system with the ancient ballast still rated 220 V – which, after all, is only realistic, for the line voltage rating today is 230 V. But it makes the system's power intake rise over-proportionally to 80 W, dragging the efficacy further down. Only in this way could the efficacy of the old system, which was already lousy, be deteriorated once more so as to leave a tiny little margin for improvement towards the T5HO system, or else the »;savings potential« of the T5HO lamp versus a system dating back to ≈1965 would have turned out negative, and we wanted to spare the T5 lamp such an intrusion.

The individual measurement results and calculations derived from these have been compiled in Table 7.13.

An office room was considered here where

  • for 1000 h/a the lights are on at full power,
  • for 1000 h/a half of the light suffices, for which purpose the magnetic system is stepped down to 64%, because a lower reduction is not possible, so that it actually produces more light than necessary,
  • for 1000 h/a no artificial light is needed, so that the automatic dimming function dims down to the minimum value of 2%, but where, on account of permanent cathode heating, the power intake still lies at 14%,
  • for the rest of the year absence is detected by the presence indicator, and the stand-by loss of the electronic ballasts drops down to the future maximum value of 0.5 W.
Fig. 7.15: In order to reduce the lamp input power with a given ballast the lamp impedance has to be reduced
Fig. 7.15: In order to reduce the lamp input power with a given ballast the lamp impedance has to be reduced

For the magnetic system the according states »;dimmed down to zero« and »;lights off« are identical and both match a full separation from the supply. Since also the control gear of the »;EnOcean« brand does not draw any power from the mains the power intake is equal zero here.

4 of such magnetic systems, as described in section 7.5, generate the same quantity of light as 3 of such electronic systems do, as described in section 7.7, i. e. with 6 lamps rated 35 W each. This allows for a direct comparison of both systems. It only requires the quantification of the total light energy in lumen hours. While electric power is measured in kilowatts and light power in lumens, electric energy is measured in kilowatt hours. So accordingly the light energy generated by this number of kilowatt hours has to be measured in »;kilo lumen hours« or »;mega lumen hours«. This is exactly what was done in Table 7.13.

The sobering resume is that there is not much left of the energy savings benefit of dimming, but rather the opposite: Along with an automatic, daylight and presence dependent control the magnetic systems generate more light energy at the same peak lighting power and even use a bit less electric energy for this than the electronic systems do. Their mean annual efficiency is thereby better. This is so thanks to the following circumstanc

  • The efficiency advantage even of T5HE lamps is only marginal.
  • The efficiency advantage of electronic ballasts is similarly marginal, other than generally assumed and propagated.
  • The efficacy of electronic dimmed systems deteriorates during dimming, whereas
  • also at the minimum level of the dimming range with practically no light output a substantial »;no-load« consumption of 14 W remains flowing and
  • even when the lighting is switched »;off« dimmable ballasts need to remain in operation, which, though representing only a minor fraction of the power intake, still stands for a fraction of the overall energy consumption that is no longer negligible.

After all it would also provide quite a lot of comfort and convenience to park one's car with the engine running in the evening, since a modern diesel engine uses barely any more than 0.3 litres of fuel per hour when idling – apparently not worth the talk, hardly any more than 1% of the top speed value, and in the morning the car would already be warm and not even frozen after a cold winter's night. But nobody does that, because at the filling station it would soon become obvious that a night of idling is equivalent to an 80 km trip the next day and would hence no less than double the average consumption for such a distance!

Electricity, however, you purchase blind, paying once a year without a clue when you have consumed how much for what. This situation leaves you dependent on half-way correct indications by fabricators and installation planners about their ratings. If this is not the case nobody will realize – until the day someone takes the trouble and undergoes the expenses to measure.

In this context it needs to be pointed out once more that the measurements presented here would absolutely not compare the best available magnetic system against the poorest electronic solution, as is always the case vice versa – and has to be done in this way to elaborate any advantage for the electronic option at all. Rather, we did not only select the most efficient electronic system according to catalogue data but a rounded figure was used for the annual usage time of an office – in the according studies carried out by the EU less than 3000 hours per year are assumed for an average European office measured values of an excellent electronic ballast were used which consumed »;only« 35% of its rated power at 25% of the maximum light output, instead of just calculating with the permissible maximum of 50% input power at 25% brightness »;for reasons of simplicity« the measurement on the T5 lamps was carried out at 35°C ambient temperature because for some good reasons these lamps are optimized to operate at such temperature, rather than measuring at 25°C, as the standard requires it for historical reasons.

So if someone should intend to run town the electronic ballast – or at least its dimmable version – these »;slip holes« could and should still be used. In a description of the facts which only intends to shed the correct light upon some lightly made statements in lighting advertisements this is not necessary, for it goes without saying that electronic gear performs some excellent features the traditional magnetic technique cannot offer. As mentioned earlier, all dimming techniques that have ever existed before the invention of the electronic ballast behaved very much like makeshift solutions, while sometimes the option of a wide range of brightness regulation is just required, such as in conference rooms. Of course it also provides more of a feeling of luxury and comfort when the light dims down continuously and mostly unnoticed as daylight arises than when it is switched off abruptly. But who believes to be able to buy in more energy efficiency than a magnetic ballast with an electronic starter can offer along with the increased luxury is moving on thin ice.

To top