Lights Dim & Hot Tub

[mphillipps]

Twice I've been ask to find a solution to house lights dim when a hot tub cycles on.
In both cases the hot tub is on a separate circuit (needless to say).
Is there a way to prevent light within a house form dimming each time a hot tub goes on?

 

[charlie b]

Not easily, I think. You are getting a high amount of voltage drop during the time the large loads (associated with the hot tub) are running. Replacing the conductors to those large loads with larger size conductors might improve the situation, but as I say that is not likely to be an easy task. I would start by looking at the sizes and lengths of those conductors, and checking the load values, and doing a VD calculation.

Welcome to the forum.

 

[mphillipps]

Charlie B,
Thanks for your quick reply.
Both installation look to be professional. Runs are <50ft.
I've not work the math myself... yet. Both inquiries were of the "While you are here..." type.
Since I do stand behind my work, "maybe solutions" are hard to sell.

I'm hoping for that silver bullet or even a copper one.

 

[gar]

090306-1340 EST

mphillipps:

I assume from your description that the circuit is --- power company transformer 120-0-120 --- all wires and connections and other items from transformer to the main panel bus bars --- and the main panel . These are the common items associated with the lights and the hot tub. Thus, these items are the only common impedance relative to the tub and other loads. The remainder of the circuit from the main panel has a --- 240 breaker supplying the hot tub and no 120 loads from this 240 breaker --- from one or more separate 120 breakers the lights in question are supplied.

This arrangement would imply no neutral problem because no tub current flows thru the neutral. The entire problem is related to the impedance of the hot conductors from the transformer to the main panel, and this includes the internal impedance of the transformer, the service drop, meter, wires from meter to main panel, the bus bars in the main panel, and all joints in this path.

Solutions --- probably none relative to the circuit described above unless there are some obvious bad connections --- change the lamps to GE dimmable compact fluorescent lamps, maybe others are as good --- use a soft on-off control of the hot tub.

The GE dimmable CFL that I tested has very little intensity variation with respect to voltage change when a sine wave is applied. However, it cuts out at about 100 V. If the voltage dip resulting from turn on of the tub is not too great, then the voltage variations might not be noticeable. I have not run a transient test of an abrupt change from 120 to 100 with this bulb to see how noticeable it would be.

See my web site for plots of intensity vs voltage for several different lamp types under steady state conditions. Compare photo P14 (the GE with sine wave excitation) with P9 (a standard tungsten filament with sine wave excitation).
http://www.beta-a2.com/EE-photos.html

I have no knowledge of the existence of a soft on-off control for the tub. This would likely consist of a Triac phase controlled to gradually ramp up the current from 0 to max over maybe 20 to 30 seconds, and a similar ramp down at turn off. This would result in a gradual change in intensity that might not be noticeable.

Increasing the wire size to the tub would have no useful effect, and it would produce a minuscule worsening effect. The branch circuit to the tub is not common with the circuits to the lights. It is only common from the main panel to the transformer.

All my comments are based on the assumptions listed at the beginning.

.

 

[Dennis Alwon]

Gar are those GE CFL's truely have full range dimming. Last I saw was that Phillips had a CFL that Lutron could dim but that was the only CFL on the market that Lutron would say was truely a full range dimming CFL.

 

[gar]

090306z-1417 EST

Dennis:

The GE I referenced does have dimming capability when driven by a phase shift dimmer as shown in my plots P17 and P18. However, when the same lamp is supplied with an adjustable amplitude sine wave it has rather constant light output as shown in plot P14. Here the intensity dropped about 5% from 130 V to 100 V. The rate of change of lamp intensity is about 1/6 of the rate of voltage change.

.

 

[gar]

090306-1729 EST

Dennis:

I have just run a limitred experiment more directly related to what happens in a transient condition.

I put a 23 ohm resistor in series with the GE bulb and adjusted the supply voltage to get 120 V across the lamp. This required an input voltage several volts higher. To produce a change in voltage across the bulb I shunted the bulb with 230 ohms. This reduced the bulb voltage to about 111 V. Visually I did not detect any intensity change. Tonight I may try this with a larger voltage drop.

.

 

[Dennis Alwon]

090306-1729 EST

Dennis:

I have just run a limitred experiment more directly related to what happens in a transient condition.

I put a 23 ohm resistor in series with the GE bulb and adjusted the supply voltage to get 120 V across the lamp. This required an input voltage several volts higher. To produce a change in voltage across the bulb I shunted the bulb with 230 ohms. This reduced the bulb voltage to about 111 V. Visually I did not detect any intensity change. Tonight I may try this with a larger voltage drop.

.

It's Friday night. Relax go for a swim, go dancing, anything-- just put the tools down.

Gar- you amaze. I sometimes feel like you speak another language.

 

[gar]

090306-2239 EST

Dennis:

We did go to the movies. So experiment is delayed.

I do not recommend that anyone try to solve a flickering light problem with the GE type of CFL until they are absolutely certain that this flicker is not caused by a bad connection somewhere. If one can determine that flickering is just the result of the normal circuit impedance, then I think this CFL might be a useful solution.

It would not be a solution for my wife because she has a great dislike of fluorescent light.

Dennis if you have a customer with the flickering problem and since you have apparently used the GE CLFs it would be an interesting experiment if you would try one of these bulbs in one of these problem situations. If voltage dips to 100 V or thereabouts, then this may not be a solution.

Some interesting background. The psychology class that I took was taught by Wilson P. Tanner (Spike). He was working on a military contract at that time and for years afterwards on signal detectability. This involved experiments with human subjects. In class he indicated he was looking for subjects and the pay was I believe about $1.25/hr, but it may have been less. There were times when there were bonuses based on performance, this was part of the experiment.

At this time the experiment consisted of four subjects, four desks each with a four button key pad, a dark room with controlled ambient light, and a white translucent screen that we viewed. Uniform backlighting of the screen was adjustable. In addition there was a special projector behind the screen that could light one of four known locations with a controlled light intensity of adjustable duration. Besides this the signal was made know in time by an audio signal. This is described as a signal known exactly. So in both position and time the light spot was known except for which of the four positions. In each test one of the four locations was illuminated and we were to answer which position by pressing the appropriate button. Feedback was provided after each test on our correctness.

The position of a spot was determined from numbers from a random number table. Therefore, there was no pattern. On a pure guessing basis or simply by pressing the same button all the time you would be correct 25% of the time. Any score above this meant that you were getting some information from the flashed light.

This whole subject area is concerned with low level signal detection in a background of noise. Generally these experiments are based on the assumption of gaussian white noise.

To start training new subjects the spot light levels were sufficiently bright that you could answer with 100% success. Then the light levels were reduced and different durations used.

Once one had been in the experiment for a while it was possible to detect with some probably above 25% a flash of light that you had no idea that you really saw it.

.

 

[gar]

090307-1259 EST

Dennis:

I have redone the experiment with different resistors. Now the source resistance is about 10.4 ohms and the shunt resistor 57.5 ohms. Still using the GE dimmable CFL, only a sample of one, the Variac input voltage was adjusted so that with the shunt across the bulb the lamp voltage was close to 102V, and with the shunt removed the voltage was about 122 V. I did not see any change in light intensity between the two voltages. Definitely showed lamp cutout at about 100 V.

Different bulbs of the same model number may have somewhat different cutoff voltages.

Now I would like to see some real world experiments to see if the flicker problem still exists or not.

.

 

[gar]

090309-1840 EST

I took a quick look at the current waveforms to the two types of CFLs.

I need some information on how to identify a Lutron dimmer that has three wires (hot in and out, and neutral) and that provides phase shift control from 0 to 180 deg. The two wire unit I have is totally unsuited to looking at how the GE CFL responds to dropout and restart. I could find no useful technical information at the Lutron web site.

Dimming of the GE does not start until the phase shift is near 90 deg. Logically this is to be expected.

.

 

[gar]

090311-1103 EST

Dennis:

You said in previous post that the GE CFL was fully dimmable. What model dimmer would you use to achieve this level of dimming? Also what would be your definition of fully dimmable?

I called the Lutron technical support line and I could not determine that they make a dimmer that can phase shift from 0 to 180 deg. In my opinion this would require a neutral wire.

With the cheap Lutron dimmer with which I have been experimenting it has the typical two wire dimmer problem of not being able to phase shift to 180 deg and if set to near minimum dimming it won't restart on return of power after loss of power.

A dimmer requiring a neutral should not exhibit this dropout problem and could be dimmed to a lower level.

.