LED Lights in Kitchen Are Flickering -

[J R Saunders]

I have a customer that experienced flickering lights daily for 11 MONTHS and spent 4 figures with various electricians with negligible results. We then tightened every wire-nut junction and device termination in the house, re-torqued all panel connections and added a good SPD. Monitored and recorded the service entry and added second ground rod. Improved, but did not totally clear the problem. As a last resort, we zeroed in on the AC Unit, adding a new Service Disconnect and ran AWG 8 stranded copper, THHN, 2 conductors in rain-tight conduit almost 90 feet from the unit to the panel, and kept the existing10/2 Romex in parallel.

Full-load voltage drop came in at under 1% and the problem has never shown up again (so far - 3 months).
I am giving him a 5-year workmanship warranty with confidence.

Your thoughts, anyone?

 

[tortuga]

As a last resort, we zeroed in on the AC Unit, adding a new Service Disconnect and ran AWG 8 stranded copper, THHN, 2 conductors in rain-tight conduit almost 90 feet from the unit to the panel, and kept the existing10/2 Romex in parallel.
.../...

Your thoughts, anyone?

What do you mean by "kept the existing10/2 Romex in parallel"?

 

[LarryFine]

Welcome to the forum.

What do you mean by "kept the existing10/2 Romex in parallel"?

Agreed. That's not good or legal.

 

[J R Saunders]

What do you mean by "kept the existing10/2 Romex in parallel"?

Exactly as stated; the existing 10/2 Romex was left as found, and the new conductors are added. Both sets are measured to be the same length.

We can of course unterminate the 10/2 if needed, but we are thinking the voltage drop will be higher?
Right now, the flickering problem appears to be resolved with the increased current handling capacity.

 

[LarryFine]

It's against code to parallel conductors smaller than 1/0 (other than EGCs).

Yes, you should abandon the #10.

 

[J R Saunders]

Welcome to the forum.


Agreed. That's not good or legal.

Exactly as stated; the existing 10/2 Romex was left as found, and the new conductors are added. Both sets are measured to be the same length.

We can of course unterminate the 10/2 if needed, but we are thinking the voltage drop will be higher?
Right now, the flickering problem appears to be resolved with the increased current handling capacity.

Can you suggest the NEC section to clarify "not legal" please. And help me understand your reasoning about "not good".

Respectfully;
JRS

 

[J R Saunders]

It's against code to parallel conductors smaller than 1/0 (other than EGCs).

Yes, you should abandon the #10.

Excellent.
Appreciated.

 

[LarryFine]

Can you suggest the NEC section to clarify "not legal" please.

310.10 (H)(1)

And help me understand your reasoning about "not good".

To difficult to keep currents equally shared, especially with mismatched sizes.

 

[readydave8]

Bigger wire to AC unit solved flicker problem

I have several times tried to solve light flicker by running bigger wire to AC unit, since it never worked I quit doing

Can someone explain why it worked this time?

 

[ptonsparky]

I agree the 10/2 and 8 should not be paralleled by NEC but the intent was not for increasing the breaker to 70. It would help to reduce the VD to the unit. It may also cause additional VD in total to the house.

 

[Fred B]

Bigger wire to AC unit solved flicker problem

I have several times tried to solve light flicker by running bigger wire to AC unit, since it never worked I quit doing

Can someone explain why it worked this time?

Not sure of the size you tried to upsize to in past but this "non compliant" parallel conductor had created a circ mil equivalent of between a #6 and #4, per table 8. Pretty large for a normally not much more than a 25A load.

Not sure if that alone would make all the difference, but, I would also lean toward looking the service capacity and voltage drops all the way back to that point. If from that point the service can't "keep up" with the usual surge demand from a large load that will usually cause flicker, and equipment level VD elimination would have limited success.

Perhaps, the allowed practice of allowing for 80% wiring of service wire in residential wiring, should be a consideration when adding more of the high demand loads and the use of more sensitive light loads such as LED. These very transient temporary high startup loads that are not overloading but will cause a VD that can be more of a nuisance issue that code really mostly ignores and leaves it to the designer (electrician in most cases in residential installations) to consider.
Code minimum is not always the best for this.

Another source of flicker on the LED is transients from the new VFD motor loads like on AC and other motors, and not sure if larger wire would help this or not.

 

[readydave8]

I can understand why voltage drop upstream of lights can cause flicker, but not how downstream voltage drop can (or is it the other way around? Having trouble visualizing stream)

 

[readydave8]

Also mention of driving 2nd ground rod in attempt to solve problem has me wondering what's really going on

 

[Elect117]

I can understand why voltage drop upstream of lights can cause flicker, but not how downstream voltage drop can (or is it the other way around? Having trouble visualizing stream)

Inductance of motor loads can create a momentary dip in voltage as current increases. This is visible as mainly dimming but can also be a flicker. By reducing the voltage drop of the circuit with the lights or the load you can alleviate the effects because it means reducing the circuit's impedance which helps push more "milkshake through the straw" without needing to swing the voltage as much. The voltage sag is proportional to the inrush current and the impedance. So as impedance goes down and the maximum inrush is the same, the voltage won't experience as severe of a dip.

The inrush can be so high that it effects loads in panels and sometimes, even on the utility.

You can also try balancing the legs better or moving the breaker so that the lights are not at the bottom of the panel.

Also mention of driving 2nd ground rod in attempt to solve problem has me wondering what's really going on

I don't think this did anything. It was probably a step they took and are attributing it to it helping. It would only help if they are seeing current coming back through earth while the unit was running which is unlikely the case. The second ground rod can help with noise and reduce the impedance to earth but I don't see that helping with the AC unit starting current in a meaningful way.

 

[hbiss]

Only thing I can think of is the larger conductor made it easier for the compressor to start due to the reduced voltage drop. The compressor draws less starting current which comes from all the way back at the service. However, it all would depend on the AC system. This might work with old school condensing units, but with modern VFD systems it probably would be a waste of time because they are soft start to begin with.

Ground rod did absolutely nothing.

-Hal

 

[Jraef]

It's against code to parallel conductors smaller than 1/0 (other than EGCs).

Yes, you should abandon the #10.

Not to mention this:

...and ran AWG 8 stranded copper, THHN, 2 conductors in rain-tight conduit almost 90 feet from the unit to the panel, and kept the existing10/2 Romex in parallel.

Assuming, because the new conductors were in "rain-tight conduit" that this was outdoors, that Romex itself was likely a Code violation (assuming it was not "UF" cable).

 

[kwired]

Exactly as stated; the existing 10/2 Romex was left as found, and the new conductors are added. Both sets are measured to be the same length.

We can of course unterminate the 10/2 if needed, but we are thinking the voltage drop will be higher?
Right now, the flickering problem appears to be resolved with the increased current handling capacity.

Code issues aside all that will do is reduce voltage drop to the AC unit. If anything will allow the unit to draw even more amps which can draw voltage down even more at the service/feeder levels.

Not sure exactly how/when this light flickering was going on, but suspect maybe during starting of the AC unit? What you possibly did was increase amount of current the AC can draw during startup which might have shortened the starting time period which maybe helped lessen the flicker effects?

 

[kwired]

Inductance of motor loads can create a momentary dip in voltage as current increases. This is visible as mainly dimming but can also be a flicker. By reducing the voltage drop of the circuit with the lights or the load you can alleviate the effects because it means reducing the circuit's impedance which helps push more "milkshake through the straw" without needing to swing the voltage as much. The voltage sag is proportional to the inrush current and the impedance. So as impedance goes down and the maximum inrush is the same, the voltage won't experience as severe of a dip.

The inrush can be so high that it effects loads in panels and sometimes, even on the utility.

You can also try balancing the legs better or moving the breaker so that the lights are not at the bottom of the panel.



I don't think this did anything. It was probably a step they took and are attributing it to it helping. It would only help if they are seeing current coming back through earth while the unit was running which is unlikely the case. The second ground rod can help with noise and reduce the impedance to earth but I don't see that helping with the AC unit starting current in a meaningful way.

But at same time some voltage drop in the supply conductor to a induction motor is essentially a crude "soft starter" and lessens the peak current level but at same time causes it to take longer to accelerate the motor. You may still get light dimming out of the situation but will have less of a drop in voltage but the event has a longer duration as well.

 

[don_resqcapt19]

It is my opinion that voltage drop acts as a poor man's soft start. The reduction in voltage to the unit reduces the inrush current and decreases the depth of the voltage sag that the high inrush current. When you put in larger wire the inrush gets higher and the depth of the voltage sag increases. There will be a time trade off, and the sag will be of shorter duration with the larger wire, but may be more noticeable because the amount of the sag is greater.

 

[ptonsparky]

It is my opinion that voltage drop acts as a poor man's soft start. The reduction in voltage to the unit reduces the inrush current and decreases the depth of the voltage sag that the high inrush current. When you put in larger wire the inrush gets higher and the depth of the voltage sag increases. There will be a time trade off, and the sag will be of shorter duration with the larger wire, but may be more noticeable because the amount of the sag is greater.

This process has been used for submersible wells for some time. They can exert quite a bit of stress on the drop pipe at startup so VD is beneficial.