High efficiency table tennis lighting

[markp]

Hi All,

My TT club currently has five 500W halogen lamps per table. They are
prone to overheating and are very inefficient, but the lighting quality is
excellent.

We are now looking for more efficient lighting systems with longer lifespan.
They should have the same colour spectrum as the halogens and not have any
strobing effects.

Does anyone have any recommendations for high efficiency alternatives to
halogen lighting? From what I've read so far it looks like we will need a
colour temperature of around 3000K and a very high CRI. High frequency warm
flourescents look interesting, but is it possible to use a combination of
HPS and HID to give the desired spectrum, if so how do you reduce the
strobing effects?

Any info or links appreciated!

Mark.

 

[Victor Roberts]

Hi All,

My TT club currently has five 500W halogen lamps per table. They are
prone to overheating and are very inefficient, but the lighting quality is
excellent.

We are now looking for more efficient lighting systems with longer lifespan.
They should have the same colour spectrum as the halogens and not have any
strobing effects.

Does anyone have any recommendations for high efficiency alternatives to
halogen lighting? From what I've read so far it looks like we will need a
colour temperature of around 3000K and a very high CRI. High frequency warm
flourescents look interesting, but is it possible to use a combination of
HPS and HID to give the desired spectrum, if so how do you reduce the
strobing effects?

Any info or links appreciated!

First, I should say I am a light source and ballast designer. You may
get different and perhaps better answers from the lighting designers
that frequent this group.

To eliminate strobing you should use high frequency (electronic)
ballasts. These are common for fluorescent lamps, but now just
starting to be available at reasonable cost for HID lamps. Also, HPS
has terrible color rendition, and while metal halide is better, it is
not as good as fluorescent.

I recommend that you use high CRI tri-phosphor T8 lamps on commercial
grade electronic ballasts. Some electronic ballasts have inadequate
line frequency filtering and will produce flicker. There is no way to
identify these high ripple electronic ballasts from the ballast data
sheet. You might be able to get power line ripple data from the
ballast manufacturer, or you may have to test products yourself.

 

[Andrew Gabriel]

Does anyone have any recommendations for high efficiency alternatives to
halogen lighting? From what I've read so far it looks like we will need a
colour temperature of around 3000K and a very high CRI. High frequency warm
flourescents look interesting, but is it possible to use a combination of
HPS and HID to give the desired spectrum, if so how do you reduce the
strobing effects?

Why is very high CRI relevant to TT?
Do you put priceless artworks round the walls,
or lay out fabric samples on the table?

 

[markp]

Why is very high CRI relevant to TT?
Do you put priceless artworks round the walls,
or lay out fabric samples on the table?

--

No, but people are used to playing under halogen and incandescent lights
which do have very high CRI. The shirts players wear for example are
multicoloured, the balls can be yellow or white, bat surfaces are red and
black. It must look as much like incandescent lighting as possible otherwise
people are going to barf at it. The venue we have is purpose built for TT
and is of international standard, in fact we've had international matches
played there. We're not talking about a leisure centre ping-pong setup,
otherwise we wouldn't have 2500W of halogen lighting over each table!

Mark.

 

[Tekdyne Lighting]

I would highly recommend to use a high efficiency Compact Fluorescent Lamp.
(With a built in Ballast). Check out www.turolight.com.

 

[Victor Roberts]

Thanks Vic, this is very useful information. We had a lighting company set
up some halide lights just to see what woul happen - evaluation shows
significant flaws, namely the lamps used have too high a colour temperature
and are way too blue, and there's a hell of a lot of strobing. I've done
some basic searching around and figured out pretty quickly that these lamps
are going to do that! You're suggestion of tri-phosphor T8 lamps seems to be
very sensible, a good CRI and if we choose warm white tubes they'll be close
to what we've already got with halogens.

I'm not sure the term "warm white" is used with tri-phopshor lamps, as
it was one of the "official" colors for halophosphate lamps. It seems
you need a 3000 K lamp, or perhaps even a 2700 K if it is available.

 

[markp]

I'm not sure the term "warm white" is used with tri-phopshor lamps, as
it was one of the "official" colors for halophosphate lamps. It seems
you need a 3000 K lamp, or perhaps even a 2700 K if it is available.

As far as I can see this way of describing colour temperature still seems to
be used, e.g :
http://www.edw-uk.com/e-wholesaler/ewcats/cat8212.htm

Mark.

 

[Victor Roberts]

As far as I can see this way of describing colour temperature still seems to
be used, e.g :
http://www.edw-uk.com/e-wholesaler/ewcats/cat8212.htm

The designations used on the Web site you give seem to be made up by
the distributor. Neither the Philips North American catalog or the
Philips UK Web site use the term "warm white". In addition, Philips
uses TL-D in Europe instead of T8. The use of "warm white" could be OK
to describe a 3000 K lamps except for the fact that there is no
indication which of three Philips warm white lamps they are *really*
selling.

Philips has 32-watt T8 lamps with 3000 K rare earth phosphor with
three different CRI's: 78, 86, and 95, using the designation TL730,
TL830 and TL930. From the site you give I can't figure out which of
these 3000 K Philips lamps is being sold as "warm white".

 

[Victor Roberts]

As far as I can see this way of describing colour temperature still seems to
be used, e.g :
http://www.edw-uk.com/e-wholesaler/ewcats/cat8212.htm

Also note that this distributor gives a rated *voltage* for each of
the T8 lamps. Fluorescent lamps are rated for an operating *current*,
not voltage. While they do have a typical operating voltage when
operated at rated current and rated temperature, the operating voltage
of a 600mm long T8 lamp is less than a 1200mm long T8 lamp, which in
turn is less than a 1500mm long T8 lamp which in turn is less than a
1800mm long T8 lamp. Yet, this distributor lists the voltage of each
of these four lamps as 240 volts!

 

[JM]

quoting:

My TT club currently has five 500W halogen lamps per table. They are
prone to overheating and are very inefficient, but the lighting quality is
excellent.

We are now looking for more efficient lighting systems with longer lifespan.
They should have the same colour spectrum as the halogens and not have any
strobing effects.

Does anyone have any recommendations for high efficiency alternatives to
halogen lighting? From what I've read so far it looks like we will need a
colour temperature of around 3000K and a very high CRI. High frequency warm
flourescents look interesting, but is it possible to use a combination of
HPS and HID to give the desired spectrum, if so how do you reduce the
strobing effects?

500w hologen lamps are aprox. 10,500 lumens each.

5 x 10,500 = 52,500 lumens


For standard T8 fluorescent lamps, you would need aprox. 19 32w lamps to do
the job. The resulting wall of lamps would probably run too hot and
inefficient.

T5 HO 80w lamps are approx. 6,500 lumens each. You would need only 8 lamps
to do the job. Downside is ballasts will operate only one lamp, so 8
ballasts will be needed. Input power 91w per ballast will yield approx
52,000 lumens for all 8 ballasts, with 72 LPW system efficiency.

ballast: Sylvania QTP1x80T5HO120PSNE 120V 20/CS 1/SKU
lamp: Sylvania FP80/830/HO


T8 HO 86w is approx. 7,600 lumens initial each. You would need 7 lamps to do
the job at 1 BF. Each ballast can operate either 1 or 2 lamps each, so you
would need only 4 ballasts. Input power for the two-lamp ballasts is 185w,
and the one-lamp ballast is 99w, total 654w input power. This will yield
53,200 lumens @ 81 lpw initial system efficiency.

ballast: Advance RCN-2S86
lamp: Philips F96T8/HO/TL830 PLUS ALTO

 

[markp]

quoting:


500w hologen lamps are aprox. 10,500 lumens each.

5 x 10,500 = 52,500 lumens


For standard T8 fluorescent lamps, you would need aprox. 19 32w lamps to do
the job. The resulting wall of lamps would probably run too hot and
inefficient.

T5 HO 80w lamps are approx. 6,500 lumens each. You would need only 8 lamps
to do the job. Downside is ballasts will operate only one lamp, so 8
ballasts will be needed. Input power 91w per ballast will yield approx
52,000 lumens for all 8 ballasts, with 72 LPW system efficiency.

ballast: Sylvania QTP1x80T5HO120PSNE 120V 20/CS 1/SKU
lamp: Sylvania FP80/830/HO


T8 HO 86w is approx. 7,600 lumens initial each. You would need 7 lamps to do
the job at 1 BF. Each ballast can operate either 1 or 2 lamps each, so you
would need only 4 ballasts. Input power for the two-lamp ballasts is 185w,
and the one-lamp ballast is 99w, total 654w input power. This will yield
53,200 lumens @ 81 lpw initial system efficiency.

ballast: Advance RCN-2S86
lamp: Philips F96T8/HO/TL830 PLUS ALTO

Thank you very much for providing this information!

The lights at the moment are standard halogen floodlight types mounted to
point downwards. They actually had to drill holes in the top to stop
overheating, but they also had to mount quite high above the table to spread
the light out evenly. The general feeling is they can be brought down to 2/3
of this height with suitable reflectors which I guess would just about halve
the amount of light needed (assuming square law). This has to be
experimented with, but you have given me an excellent starting point to do
this.

Thanks again!

Mark.

 

[Victor Roberts]

500w hologen lamps are aprox. 10,500 lumens each.

5 x 10,500 = 52,500 lumens


For standard T8 fluorescent lamps, you would need aprox. 19 32w lamps to do
the job. The resulting wall of lamps would probably run too hot and
inefficient.

I don't understand your message here. See below.

T5 HO 80w lamps are approx. 6,500 lumens each. You would need only 8 lamps
to do the job. Downside is ballasts will operate only one lamp, so 8
ballasts will be needed. Input power 91w per ballast will yield approx
52,000 lumens for all 8 ballasts, with 72 LPW system efficiency.

ballast: Sylvania QTP1x80T5HO120PSNE 120V 20/CS 1/SKU
lamp: Sylvania FP80/830/HO


T8 HO 86w is approx. 7,600 lumens initial each. You would need 7 lamps to do
the job at 1 BF. Each ballast can operate either 1 or 2 lamps each, so you
would need only 4 ballasts. Input power for the two-lamp ballasts is 185w,
and the one-lamp ballast is 99w, total 654w input power. This will yield
53,200 lumens @ 81 lpw initial system efficiency.

ballast: Advance RCN-2S86
lamp: Philips F96T8/HO/TL830 PLUS ALTO

Four-foot, 32 watt T8 lamps are more efficient than either T5/HO or
T8/HO lamps. For the same amount of light they would produce less
heat, not more, in spite of the fact that you would need to use more
lamps. T5/HO or T8/HO might be the best lamps for this application,
but not because they would be more efficient or cooler.

 

[markp]

Four-foot, 32 watt T8 lamps are more efficient than either T5/HO or
T8/HO lamps. For the same amount of light they would produce less
heat, not more, in spite of the fact that you would need to use more
lamps. T5/HO or T8/HO might be the best lamps for this application,
but not because they would be more efficient or cooler.

Would that not require more ballasts? Wouldn't the inefficiencies that this
introduces negate the extra efficiency of using 32 Watt T8 lamps over T8 HO
lamps?

Another quickie if I may, presumably these electronic ballasts have high
power factor inputs as standard. Is that correct? This is of concern for
lighting many tables.

Thanks!

Mark.

 

[markp]

No, 32-watt, four-foot T8 lamp and ballast SYSTEMS are more efficient
than T8/HO and T5/HO lamp and ballast SYSTEMS. The ballasts for both
types of lamps operate at comparable efficiencies. I do not understand
why fewer ballasts would be assumed to have lower power loss.

OK, For each electronic ballast I assume there's an active PF correction
circuit producing DC which is then used to drive some kind of inverter (?).
Having many active PF correctors and inverters to me is less efficient than
having one of each, but having thought about it more there's going to be
higher current in a larger one and lower current in the smaller ones and
some power loss is proportional to the square of the current, so I'm not
sure now whether that logic stands up!

For example are you saying that four small electronic ballasts driving one
lamp each would have the same efficiency as a single large ballast driving
four lamps?

Of course, since different types of ballasts from different
manufacturers do have different efficiencies, you could put together
systems that contradict my assertion. However, that would be an
aberration. If you used products of equal quality, the 32-watt
four-foot T8 beats the T8/HO and T5/HO every time. The lower efficacy
of HO lamps is based on a fundamental aspect of fluorescent lamp
operation, it is not an issue of poor design or any such thing.


Commercial grade electronic ballasts all have high input PF, but the
same cannot be said for residential grade units. You need to check the
ballast data sheet to be sure. Another problem is the tradeoff between
THD and inrush current. Electronic ballasts with very low THD, listed
as < 10%, often have very high inrush currents, while electronic
ballasts listed as < 20% usually have much lower inrush currents. I
have not seen ballast inrush current data on any ballast data sheet
from the major manufacturers, so you will have to call their tech
support staff if this is important. Or, just use ballasts rated for
< 20% THD.

OK, I'll check. Why high PF input stages need nowadays to have high inrush
currents is a little confusing, I thought they would use active power factor
correction circuit with controlled smooth inrush, but I'll look out for it
too!

Thanks!

Mark.

 

[AC/DCdude17]

X-No-Archive: Yes

A 500W halogen lamp can put out 10,000 lumens from a lamp no larger than a
pencil.

To get the same amount of light out of F32T8 fluorescent lamps, assuming each
puts out 2600 lumens on a ballast factor 0.88 ballast(considered normal output),
you'll need four lamps and each one of these lamps is 4' long by 1" diameter and
light can not be focused the same way a halogen light can.

Since you have five halogen lights per table, you're looking at installing
TWENTY 4' lamps.

Once you figure out how you're going to arrange those twenty lamps, it's not
hard at all.


What kind of halogen lamp do you have? If you have a standard RSC T3 (similar
diameter and slightly shorter than a pencil), you can try a GE HIR lamps. They
will put out the same amount of light as 500W halogen while using 350W. While
it's not the best result, it's an improvement.

 

[Victor Roberts]

X-No-Archive: Yes

A 500W halogen lamp can put out 10,000 lumens from a lamp no larger than a
pencil.

To get the same amount of light out of F32T8 fluorescent lamps, assuming each
puts out 2600 lumens on a ballast factor 0.88 ballast(considered normal output),
you'll need four lamps and each one of these lamps is 4' long by 1" diameter and
light can not be focused the same way a halogen light can.

Since you have five halogen lights per table, you're looking at installing
TWENTY 4' lamps.

As far as I remember, no one in this discussion has claimed that
fluorescent lamps are as small as an incandescent lamp that produces
the same amount of light. And, until we know more about THIS
APPLICATION, it is not clear that five concentrated sources would be
better than 20 larger sources.

 

[Douglas G. Cummins]

X-No-Archive: Yes

A 500W halogen lamp can put out 10,000 lumens from a lamp no larger than a
pencil.

To get the same amount of light

From his earlier post, you'll find he's currently using floodlamps
pointing downward. You can't assume that he's using a 500W T-3 bulb in
a fixture. He also says they're mounted quite high off the table to get
a uniform illuminance around the table. He also says that he will be
bringing the lights 2/3rds closer to the tables, so matching fluxes
won't work in this case.

 

[Victor Roberts]

OK, For each electronic ballast I assume there's an active PF correction
circuit producing DC which is then used to drive some kind of inverter (?).
Yes.

Having many active PF correctors and inverters to me is less efficient than
having one of each, but having thought about it more there's going to be
higher current in a larger one and lower current in the smaller ones and
some power loss is proportional to the square of the current, so I'm not
sure now whether that logic stands up!

If the same size wire and power semiconductors were used, it might,
but higher power ballasts used larger wire and ferrite cores in their
transformers and inductors, and larger power semiconductors. The goal
is to keep the ballast efficiency in a "reasonable" region while not
making it too expensive.

For example are you saying that four small electronic ballasts driving one
lamp each would have the same efficiency as a single large ballast driving
four lamps?

Well, we started discussing "normal" power linear lamps vs. fewer HO
lamps. But the situation is basically the same for two lamp vs. four
lamp ballasts. I will have to check the manufacturers data to give you
a real answer, but my experience is that most electronic ballasts
operate with an efficiency of about 90% or slightly better. I will
check some real data and get back to you.

 

[Victor Roberts]

For example are you saying that four small electronic ballasts driving one
lamp each would have the same efficiency as a single large ballast driving
four lamps?

I did say that, but, here is one comparison that indicates I am wrong.

Consider two rather normal and similar electronic ballasts: the
Advance REL-2P-32-SC and the Advance REL-4P-32-SC. The first runs two
32 watt T8 lamps with a ballast factor of 0.88 while consuming 58
watts. The second runs four of the same lamps at the same 0.88 ballast
factor while consuming 112 watts. This means that one REL-4P-32-SC and
four 32-watt T8 lamps will operate with 3.6% higher efficacy than two
REL-2P-32-SC ballasts and the same four 32-watt T8 lamps. Note that
the input wattage numbers are rounded to the nearest watt, so the
actual difference in efficacy could be different than the number I
calculated.

The question for HO lamps is different in. In this case the HO lamp is
less efficient than the normal power lamp, so any improvement in
ballast efficiency would have to more than offset the loss in lamp
efficacy. Let me see if I can find some data.

 

[Victor Roberts]

OK, here is an example of HO vs. normal power lamps. I decided to use
T5 because the T8 situation is more complicated. Most regular power T8
lamps run on instant start electronic ballasts while all HO lamps are
rapid start or one of its variants, such as programmed rapid start. On
the other hand, all T5 lamps run in programmed start or programmed
rapid start mode. I also decided to compare two-lamp systems, because
as shown in my previous note, there is a small difference between
two-lamp and four-lamp ballasts. It is also important to compare lamps
with the same length, as, all else being equal, longer lamps will have
higher efficacy.

The first system uses an Advance ICN-2S28@120 programmed start ballast
operating two Philips 46-inch F28T5/830 lamps. The lamps are rated for
2900 lumens at 28 watts, the ballast has a ballast factor of 1.03 with
two F28T5 lamps and draws 64 watts. This gives a system efficacy of
93.34 lm/W.

The second system uses an Advance ICN-2S54-90C@120 programmed start
ballast operating two Philips 46-inch F54T5/830/HO lamps. The lamps
are rated for 5000 lumens at 54 watts, the ballast has a ballast
factor of 1.0 with two F54T5 lamps and draws 120 watts. This gives a
system efficacy of 83.33 lm/W.

The normal power T5 system has higher efficacy.

BTW - premium T8 lamps on premium instant start electronic ballasts
can have system efficacies as high as 98 lm/W. So they beat both T5
systems.