Variance of Lamp Wattages

derekleffew

Resident Curmudgeon
Senior Team
Premium Member
In replying to a thread on LED fixtures, I did an experiment, and was somewhat surprised at the results. For an actual comparison, I used my new Kill-A-Watt, a "trueRMS" device.


A random 500W GE EHD lamp (of undetermined age):
Unplugged:
Volt: 124.9, Amp: 0.00, Watt: 000, VA: 000, PF: 1.00

Plugged into GPO (Australian term for General Purpose Outlet):
Volt: 121.1, Amp: 4.91, Watt: 596, VA: 595, PF: 1.00

A different 500W GE EHD lamp (of undetermined age):
Unplugged:
Volt: 124.9, Amp: 0.00, Watt: 000, VA: 000, PF: 1.00

Plugged into GPO (Australian term for General Purpose Outlet):
Volt: 121.4, Amp: 4.31, Watt: 524, VA: 521, PF: 1.00

For all examples above, VA=Vrms*Arms, and PF=W/(Vrms*Arms).

The question is:
A.) Do Stage/Studio Lamps vary that much in wattage right out of the box, i.e. Are manufacturing tolerances really that lax?
AND/OR
B.) Does the age of the lamp matter as to how much wattage it consumes, i.e. Does a lamp draw more (or less?) power as it ages?


I realize both are compound questions, so be specific in your answers. You may use charts and/or diagrams in your responses if you wish.
 
As a lamp ages its resistance goes up and therefore its wattage goes down, this is most important in 240 volt countries running series pairs of 120 lamps as the older lamp will take more of the power and decrease its life quickly{usually par 64's}
 
Anyone else want to weigh in with his/her opinions? [Remember there are no wrong answers (only incorrect responses!)? Has "that lamp guy" scared everyone off? Who else knew that a 500W lamp is not really 500W?
 
What is missing is information like what is the nominal voltage of your lamps, I would guess 110 volt as they are running in excess of their rated wattage and their is no rational explanation for that to occur.
Also pretty dodgy wiring if your dropping 3.8v at 4.3 amps
And what is the accuracy of the meter?
 
Last edited:
The EHD is a 120V lamp. This is the "meter" in question. I disagree with the "dodgy wiring" statement, although most of the 50' 14/3 extension cord was coiled inside a "Cord-o-matic" at the time of measurement, could that matter? I suspect one should expect a drop in voltage when any significant load is applied to a circuit, correct?
 
Best price I have found, and local too, although mine was a gift. Wait, I'm wrong. Less expensive on the internet, but isn't everything?
 
Would you be kind enough to check this device against a Fluke or something known, it looks flash but the claims for it may be a little exaggerated, when things like this don't add up first check all the data.
I would tend to distrust the meter more than the lamp
 
Good idea! As soon as I receive my Fluke 336A I'll do that. However, I have repeated the same measurements on TWO different EHD lamps, and get the same results each time. I even exchanged the lamps between the two fixtures, and the same lamps read the same measurements, +/- 0.5%.
 
Last edited:
On the subject of the original question:
Back when I was using a lot of ACL's, I used to meter and rate lamps so that I could use "balanced" sets. With ACL's (28v), you use 4 in series off of a dimmer, so matching them insures that one will not receive more stress that it's three friends. What I found was pretty consistent. (Jig was a 30 amp variable transformer adjusted to output 28 volts under load, using a clamp meter, lamps were 4559.)

1) Lamps from the same manufacturer and of the same date code batch were very close. (less than 1%, hard to see on meters of that day.) It should also be noted that all of the new lamps I tested had a draw in excess of 21.5 amps, leaving them above their 600 watt rating.

2) Lamps from different batches or different manufacturers varied up to 10% out of the box.

3) Lamps of different ages varied up to 20% from new. (always lower on the current)

At the time, I also tested a few 1k par bulbs at 120 volts, and found the same type of variance.

I have not repeated the experiment in recent years. My opinion on the subject is that there are not only manufacturer tolerances at play, but that the manufacturers intentionally design the lamps to be a little "heavy" to start with. As the lamp ages, the filament becomes thinner, and therefore draws less current. Resistance is a measure of amps passing at a certain voltage (V/A=R), therefore, a lower current at a given voltage means the resistance of the lamp increased with age. As wattage is amperage times voltage, a lower amperage at the given voltage would indicate the lamp has dropped in wattage with age. Results were very consistent. The filament of a 4559 is very short and thick, so I never ran into coil winding shorts. I would suspect that is the only time an old lamp would draw more.
 
(Attempting a table on CB for the very first time):

Readings of two different 500W, 120V GE EHD lamps, both of undetermined age and usage:
Lamp|KillAWattAmps|Fluke336Amps|Fluke30Amps|KillAWattVolts|Fluke336Volts|Fluke30Volts
Lamp "A"|4.91|4.7|4.2|121.1|120.7|120.1
Lamp "B"|4.33|4.1|3.7|121.6|121.1|120.5

I guess the primary lessons learned from this are:

1) The Kill-A-Watt is accurate enough for home use, and tends to over-estimate.
2) No two meters will ever read the exact same, unless same make & model, and even then, doubtful.
3) Voltage, Amperage, and Wattage fluctuate slightly constantly, so an averaging or weighted feature is a good thing to have on a meter, as well as a peak hold.
4) The table function on CB is pretty cool, but I still prefer MS Excel.
 
There may be a reason for Kill-A-Watt always "rounding up". The device is marketed as something you can use to calculate your energy usage in order to further reduce the amount of energy you use. If it estimates higher, it might spur more people into action. No problem with me. I guess I just shouldn't buy one of these after all, though.
 
the other thing you learn is that the advertising as .2 % accuracy is at best misleading.Ref. the Kill-a-watt ad.
The differences in current readings between all three are way out of tolerance and should be rechecked, were all three in series at the same time to get an accurate comparison ?
Happy New Year.
 
Found, belonging to my roommate, another meter, a Micronta (Radio Shack) #22-160 Analog ammeter, 25+ years old; and repeated the tests, with all meters in line simultaneously.


All readings are in Amps, for a 500W GE EHD, 120V lamp.
Lamp|KillAWatt|Fluke30|Fluke336|Micronta
Lamp"A"|4.92|4.3|4.7|4.7
The Micronta was only an AC Ammeter, so could not measure voltage, but others read voltage same as in the table in post above. Interesting that the Micronta and the 336 were the two closest to one another, on both lamps. Perhaps this illustrates why many electronics techs/electricians prefer the Simpson analog meter over all others. Accuracy, and that they're old.;)
 
There was mention made of an extension cord being used rolled up and of voltage drop... Now given we are talking about AC power, yeah there will be a straight resistance, but you will also get an inductance because of the coil.

Now someone made comment about 3.8V seeming like a lot of voltage drop for 4.3A. It is only 3% of supply and that voltage drop and that current draw give a nuisance impedance of about 0.9R. But it's not that simple... I don't know where derek did these measurements, for all I know he could have been at home i the middle of the day with net to nothing on. That being the case, the 4.3A could have significantly increased his home's current draw at the time. Suppose for a moment there was no load. When the volt meter was applied, you would have been looking at the open circuit voltage; because there is no load, there is not complete circuit ad so there is no current flow. Hence all resistance along the way is negated. But still supposing nothing else was turned on, when the test lamps were energised, you start pulling current all the way out to he street. So you are looking at the impedance through the incoming mains, through the meter, a service fuse, a circuit breaker or fuse, potentially an RCD - with a coil and hence an inductance, and then maybe 50 metres of smallish cable from switchboard to lamp, so it's not really just a voltage drop across the extension cable in use, that measurement would have to be made with a meter between the active at one end and the active at the other and then doubling it to add in the neutral as well...

Just some food for thought... What was the question again?
 

Users who are viewing this thread

Back