LED's & Microwave

[GoldDigger]

I have told him this is probably what he will have to do, that I will try whatever means is easily available.

We really need a better description of "flicker".
If the LEDs continually fluctuate in light output while the microwave is running continuously, then there is a driver problem which can be attacked, probably with filtering. If the LED output only changes when the magnetron in the microwave cycles on and off, then it may just be a voltage drop issue which can either be addressed with wiring changes or ignored. If the LED output change as the magnetron cycles is greater than for an incandescent in the same fixture, then the voltage drop would have to be excessive or the driver in the fixture very badly designed. Measuring the voltage change between microwave on and off is a good starting point. And if the LEDs get brighter when the microwave starts up, there is a high resistance neutral that should be looked at.

Now that we have had fun speculating, the correct answer will be found on site, and I would appreciate hearing from you after you get a look at it yourself.

 

[gar]

130509-2324 EDT

Today I bought a new Cree 6 W bulb. A quick test of just changing sine wave voltage excitation showed power input of 5.4 W at 130 V and 4.9 W at 100 V. Voltage ratio is 1.3 and the power ratio is 1.1 . An eyeball intensity test did not show much light change in comparison to an incandescent. A noticeable time lag was observed in power input changing from 5.4 to 4.9 W when voltage was reduced.

jellmc you need to run some controlled tests on your LED and other LEDs and see what will change intensity. I have not done any phase shift dimmer tests on the above bulb yet. The bulb is identified as being dimmable.

.

 

[gar]

130710-1029 EDT

Some more quick tests.

Same LED as previous post, and a standard 60 W incandescent tungsten bulb. Tests with and without a dimmer. The dimmer is a 3 terminal type. This means the dimmer can provide phase shift control with and without a load. Full on is almost a full sine wave, minimum is somewhat more than a 90 degree delay in turn on.

Line voltage 122.7 V.

No dimmer ---
60 W incandescent 59 W.
6 W LED ............. 5.5 W.

With dimmer ---
60 W incandescent 6.5 W to 56.4 W.
6 W LED ............. 1.3 W to 6.3 W.

The 6.3 W was unexpected. Have not done any light intensity comparisons. The LED seemed to have a smooth characteristic with dimmer control.

To make an LED dim effectively with a phase shift dimmer it is likely that the circuitry senses the turn-on phase angle to control current to the LEDs.

With just a Variac control of voltage to the LED, not much dimming takes place until 60 V, and dropout is about 15 V and 0.5 W.

Generally this bulb performs much better for dimming than a GE dimmable CFL.

.

 

[jmellc]

Gold Digger, the flicker I saw was the more or less normal flicker seen in many lights when a heavy load comes in & out. Lights flickered a second when MW turned on or off. Flickered a second when magnetron cycled in or out. Customer had seen lights strobing at times too. I never saw this.

New info. Customer has now identified that the strobing is caused by an expresso machine that had not been mentioned before. It also causes strobing in another room. I am looking now for a filter to try out.

 

[kwired]

Gold Digger, the flicker I saw was the more or less normal flicker seen in many lights when a heavy load comes in & out. Lights flickered a second when MW turned on or off. Flickered a second when magnetron cycled in or out. Customer had seen lights strobing at times too. I never saw this.

New info. Customer has now identified that the strobing is caused by an expresso machine that had not been mentioned before. It also causes strobing in another room. I am looking now for a filter to try out.

Still sounds like a voltage drop problem more so than any interference problem to me.

If these items are on different circuits than the lamp in question then you need to look at the service or applicable feeder as well as service drop or lateral size and length and even the size of the POCO transformer can have some impact if it is not very high capacity.

Have you monitored voltage at this lamp when those loads are switched on/off?

 

[gar]

130714-1259 EDT

New experiment. This time relative to intensity change from motor startup.

Items: 9 ft long #16 extension cord. Bench outlet. 60 W tungsten filament 60 W 120 V bulb. GE 26 W 120 V dimmable CFL. Cree 6 W 120 LED bulb, same as previous posts. Unloaded 1/3 HP 120 V 60 Hz Montgomery Ward induction motor used in many different experiments.

One at a time the bulbs are tested. These are plugged into the 3 socket end of the extension cord. The motor is plugged into the same end of the extension to provide a momentary drop in voltage.

Results:

1/3 HP motor transient on inrush.
Incandescent --- the normal expected flicker, appears greater than the LED.
GE dimmable CFL --- no noticeable flicker.
Cree LED --- an extremely short flicker, less noticeable compared to the incandescent.


1500 W resistive heater.
Incandescent --- very slight change in intensity.
GE dimmable CFL --- no noticeable change.
Cree LED --- an extremely short and negligible flicker, no obvious steady state change.

.

 

[gar]

130714-2115 EDT

When I use the 3 terminal phase shift dimmer to adjust the Cree intensity the dimming action starts immediately as I move the adjustment from maximum to minimum. This implies that the phase angle delay of the dimmer signal is being used to control dimming. There is a possible analog way to do this, but unlikely the method. Also dimming action is fairly uniform.

So, if the Cree bulb is excited with a full sine wave of full voltage, and a signal is superimposed on the sine wave at some point in the cycle, then it is possible the bulb might go into a dimming state.

To see if this was possible from a simple type of parallel load I created the following experiment:

The Cree was plugged into the extension cord. The extension cord and supply provide some common source impedance.
The phase shift dimmer was also plugged into the extension cord.
An 800 W heater was connected to the dimmer output.

At no position of the dimmer was there abnormal operation of the Cree. This experiment causes a time controlled (synchronized) drop in voltage within the excitation cycle at the dimmer turn on time in the cycle.

At this point I know enough about the Cree bulb to suggest its use as a test at your problem location. If your customer location produced the same results as with whatever LED you are having to troubleshoot, then it might be possible to simulate a signal to try for the same result.

What is important to know is whether the problem LED is uniquely at fault or if there may be a general type of problem.

.

 

[weressl]

130714-2115 EDT


An 800 W heater was connected to the dimmer output.

Is the dimmer designed to control tungsten-like loads, simultaneously with performing its designed control operation of LED's?

 

[GoldDigger]

Is the dimmer designed to control tungsten-like loads, simultaneously with performing its designed control operation of LED's?

The dimmer is not feeding the LED. It is just there to create a pulsed voltage drop at the LED, based on the phase-shift-switched current drain of the heater. In other words it is a worst case noise source (at least if you leave out RF noise.) The LED is a dimmable LED which is not connected to a (different?) dimmer for this test.

 

[gar]

130714-2330 EDT

GoldDigger:

Your response prompted the addition of 2 ohms to the source impedance. At 90 deg phase shift for the very fast change in voltage the step magnitude is about 17 V. Again at no point in the dimmer range was there any abnormal result. Under minimum dimming (close to 0 phase shift delay), maximum average current thru the source impedance, there was some reduction in intensity as would be expected. But there was no flickering anywhere in the adjustment range.

weressl:

By a three terminal dimmer description I am referring to a dimmer that has at least one hot input, one hot output, and a neutral connection. This kind of dimmer requires no output load to operate. Power for the electronics is derived from the input hot and neutral and is totally independent of whatever the load might be. This type of dimmer is required for all sorts of different loads that might not provide enough energy to operate the electronics in a two wire dimmer.

This dimmer will work with incandescent, CFLs and LEDs designed for phase shift dimmer operation, and regular fluorescent fixtures designed for dimming. Probably could also be used for DC motor operation, but has limitations compared to a control designed for DC motor control.

.

 

[hbiss]

Your problem is undoubtedly caused by voltage drop but whether it is normal or not is up to you to determine. The situation with LED lamps is that intensity changes caused by even a minor voltage drop looks significant because the response is instantaneous. There is no persistance like with incandescent and fluorescent that tend to smooth such disturbances out.

-Hal