Screw in CFL's and actual load

[Ragin Cajun]

Been wondering just how "accurate" the label on these things is.

13W CFL, 120V, 0.200A says the label.

I have a neat device that indicates volts, amps, kw, va kwh, hz, pf, and hours. It has proved quite accurate where ever I've used it.

Plugged in the above CFL lamp.

121.7V
0.360 A
59.9 Hz
25 W
44 va
0.57 pf.

What a POS!!

I knew the pf on these things was bad, but 0.57, WHEW!!

RC

 

[KentAT]

I wonder if those numbers changed as the bulb warmed up from throwing the switch to full light output.

 

[SG-1]

What are you using to measure the CFL voltage & current ? Kills-A-Watt ?

Upon what do you base the prior accuracy ?

 

[earlejohnson]

I have a neat device that indicates volts, amps, kw, va kwh, hz, pf, and hours. It has proved quite accurate where ever I've used it

What is the device

 

[texie]

watt meter

watt meter

Given the crude power supply typical in a compact flour. lamp with all it's harmonics and distortion it would take a pretty sophistcated watt meter to measure accurately.

 

[SG-1]

The wave forms would look like this:

 

[broadgage]

IME the plug in energy and power monitors are far from accurate, especialy at low loads such as a single lamp.
Presuming that the CFL is of reputable make, I would expect the watts used to be within about 10% of the manufacturers claims.
Such lamps are sold on a huge scale by many reputable suppliers and I dont think that they would get away with selling lamps that use about twice the power claimed.

Budget lamps may be rated in chinese watts, with the actual power used and light given bearing little relation to the claims on the packet, but unlikely with reputable ones.

 

[gar]

101124-0914 EST

Ragin:

Using my Kill-A-Watt

GE 13 W 120 V 0.18 A after a few minutes warm-up
122.3 V, 0.15 A, 11 W, 19 VA, 0.62 PF initially just after turn on PF was 0.59
Good correlation with the label.

NuVision 27 W 120 V 0.45 A after a few minutes warm-up
122.2 V, 0.34 A, 23 W, 41 VA, 0.55 PF
Again good correlation with the label.

A fluorescent bulb is a gaseous discharge device. This is quite sensitive to temperature and pressure, and pressure is a function of temperature. In its operating region it may be in its negative resistance region or have reached the constant voltage area of its V-I curve.

Power factor is expected to be low because the input is a simple rectifier followed by a capacitor input filter.

The quality of design and manufacture of Kill-A-Watt meters is not particularly good, but over various test ranges I have found the measurements to be satisfactory for an inexpensive meter. My particular complaint is with problems when voltage drops to less than a value at which I expect the instrument should work. One unit started failing at about 105 V, another around 100 V. An instrument of this type I would expect to work down to about 85 V.

On a good capacitor the Kill-A-Watt will read 0 for power and current. It's accuracy on voltage isn't to bad. Appears to be monotonic, and thus is great for small voltage difference tests. Therefore useful in measuring voltage drop on a circuit as a load is switched on and off. Useful to measure the length of a wire, or to troubleshoot for high resistance connections, etc.

.

 

[Ragin Cajun]

Sorry for the delay in responding, I am out for the holiday.

Forgot the name of the device I used.

Values posted were well after warm up.

Wattage essentially matched nameplate but the current draw was much higher.

RC

 

[gar]

101124-1853 EST

Ragin:

A rerun of my above test. A different GE 12 W, but same model.

Now using my Kill-A-Watt EZ. This works to a lower input voltage.

Shortly after power on:

GE 13 W 120 V 0.18 A after a few minutes warm-up
120.1 V, 0.20 A, 15 W, 24 VA, 0.62 PF.
Good correlation with the label.

GE 13 W 120 V 0.18 A about an hour later
120.1 V, 0.18 A, 14 W, 22 VA, 0.62 PF.
Good correlation with the label.

Lowered voltage:
GE 13 W 120 V 0.18 A
100.1 V, 0.20 A, 11 W, 18 VA, 0.63 PF.

Plain 15 W 120 V Incandescent
120.2 V, 0.15 A, 16 W, 18 VA, 0.92 PF.

Plain 15 W 120 V Incandescent
100.1 V, 0.13 A, 12 W, 13 VA, 0.92 PF.


Back to the regular Kill-A-Watt
119.6 V, 0.11 A, 13 W, 13 VA, 0.99 PF.


Shows some of the disparities between the Kill-A-Watts.

.

 

[kwired]

Is a CFL with 13 watts marked on it supposed to be total watts or just the lamp portion and ballast may have other consumption?

I believe a tube with no internal ballast the wattage marked on the lamp is for the lamp only and there is other energy consumed in the ballast same for HID lamps.

 

[gar]

101124-2331 EST

I would expect it is the total package power input because it is one integrated unit. My various measurements tend to indicate this.

I believe you will find more variability in CFL power input from one bulb to another than for incandescents. I have found incandescents fairly close to their ratings.

.

 

[wireguru]

I was just looking at a CFL from a large manufacturer, its an 11w lamp, the specs says 140ma. BUT the instructions on the package say not to put more than 25 of these CFLs on a 20 amp circuit. I wonder whats going on here.....

 

[gar]

101126-1733 EST

A quick check on my GE 13 W CFL looks like I can see at least a 20 A or greater initial charging current to the capacitor in the CFL. This an eyeball test without any phase angle control of turn on phase and no camera to record the pulse. Extremely short pulse compared to 2 MS/CM or 20 MS full trace. Probably less than 10 microseconds. The peak I saw might be greater than 20 A.

Note: 2 ohms at 170 V = 85 A, but there is some inductance present in the system circuit and additional resistance within the CFL. With zero initial charge a capacitor looks like a short circuit when a voltage is first applied.

To do this without a controlled setup to trigger at 90 deg one just tries a number of random turn ons.

If 20 of these are on one circuit and all are switched on at one time, then the pulse duration might be 20 times longer at a peak of 20 A, or a shorter time and larger peak. Will depend upon the circuit impedance. If the capacitance was 50 MFD and resistance was 2 ohms, then the time constant would be 100 microseconds. I doubt the duration was that long and 50 MFD may be too large for the package size.

It could be possible that one could trip a breaker with something not too much greater than 20 of these devices.

Maybe someone else will run a more controlled experiment. Probably best to run the test with 20 bulbs.

.

 

[ELA]

Here is a scope trace I posted in another thread showing the current demand each half cycle of 6 CLFs on the same circuit.
Each CFL draws a peak of approx 1 amp and their firing phase angle is close enough to be additive.
Thus you can see that 20 lamps would result in large current "bumps" each half cycle.

 

[wireguru]

Here is a scope trace I posted in another thread showing the current demand each half cycle of 6 CLFs on the same circuit.
Each CFL draws a peak of approx 1 amp and their firing phase angle is close enough to be additive.
Thus you can see that 20 lamps would result in large current "bumps" each half cycle.

am I correct in that is showing a terrible power factor?

 

[gar]

101126-2335 EST

ELA:

Can you try catching some turn on transients with your 6 bulbs, and see if you can trap one near a voltage peak?


wireguru:

Yes power factor is poor. In the 0.5 to 0.6 range. My AC power waveforms here are somewhat flat-topped sine waves. More so than ELA's. Probably because of the 10s of thousands of computers with their capacitor input filters in the area. 30 to 40 years ago the AC line was a nice sine wave. There are over 60,000 college students within about 3 to 6 miles of me. Also lots of home and business computers as well. Many of these are almost always on, at least during the day and evening.

.

 

[ELA]

101126-2335 EST

ELA:

Can you try catching some turn on transients with your 6 bulbs, and see if you can trap one near a voltage peak?
.

No ... can you

We know there is substantial turn-on inrush associated with any capacitive input device.

Unless all 20 lamps were on a "single switch" who cares?

What is more significant is the continuous "bump" every 1/2 cycle.

... the 6 lamp test setup has been long since disassembled.

 

[iwire]

We know there is substantial turn-on inrush associated with any capacitive input device.

Unless all 20 lamps were on a "single switch" who cares?

Lighting after a power failure comes to mind ...

 

[wireguru]

its interesting the utilities would be pushing a product with such poor power factor so hard (especially to residential users that dont get billed for kvar).

why dont they make the CFLs with a better ballast circuit with PFC and have the tube replaceable?