Control/Dimming SSR failure?

[Lightguy5]

I think I'm losing SSRs in my CD80 racks at an alarming rate. The load isn't excessive (one or two 575 lamps) but the dimmer cuts in/out of full at 99%.
My most recent purchase of cubes were Tyco 240D25. They've been lasting weeks (or in one case, days) and not months or years. They're 2.4K dimmers, running on an Ion (and no, they're not profiled to non-dim).
Is there a reason that a 25A SSR isn't sufficient? or do I just have a bad batch?
Could this be something else that I'm not thinking of?

 

[venuetech]

did you use thermal paste or a thermal pad when you replaced the ssr?

 

[JD]

The Tyco 240D25, is this a direct replacement? Usually, you want a safety factor of about 300%, so using a 25 amp SSR on a 2.4kw dimmer does not make sense. SSR-480D50 is the smallest I would go. Better off with the OEM part.

 

[Lightguy5]

did you use thermal paste or a thermal pad when you replaced the ssr?

Yes, the paste.

 

[jstandfast]

You have to be really careful with heat sink compound not to over-apply it. Joel Epstein (former head of field service @ Strand) told me once that the OEM spec for the layer of heat sink compound is measured in microns. According to Joel, if the stuff is over applied,it begins to have a negative effect as far as heat sinking goes.

 

[derekleffew]

You have to be really careful with heat sink compound not to over-apply it. ...

See, I was told the opposite, for an L86 dimmer module:
1. Scrape all of the old paste off the tray.
2. Liberally butter the bottom the SSR, using the entire contents of the included tube.
3. Tighten the bolts holding the SSR in place.
4. Remove all squeeze-out.

jstandfast, as this isn't the first time this topic has arisen (see http://www.controlbooth.com/forums/...2-need-etc-l86-2400-power-cube-ssr-specs.html , posts 5-8), let's hope an expert chimes in.

In any case, it's unlikely that this is the problem. More likely the wrong SSR, or a control issue.

 

[Footer]

You want at least twice as much power as your final output. Your way undersized it and pushing them too hard. Buy these: Dimming & Control - Parts

Your breaker is what controls the final output, not the cube itself.

 

[DavidNorth]

I cannot guarantee that your continued failures are undersized SSRs, but I will say that using a 40A or 50A SSR for a 20A circuit is the common design and should be used. You could have some issues that are causing failures such as inadequate cooling (bad fans, clogged filters, blocked airflow, hot room), shorted wiring in fixtures and/or cables, shorted SO distro cables or shorted wiring in distro. Get some larger SSRs, see what your failure rate is, and if the new cubes are still popping, it's time to dig deeper into your system.

SSRs from ETC (for L86 dimmers) used to ship with a small packet of heat sink compound. I think it equated to about an ounce and yes, you were to use the entire tube. It was a very small tube.

I personally use a very large tube, squeeze out some onto the SSR, spread it over the entire SSR base, and bolt it down. I would say that you only need to use enough to equate to the thickness of 4-5 dollar bills. Any more is wasteful and any less may leave voids.

It was necessary in L86 racks to use the stuff as the SSRs are bolted down to sheet metal trays that were not perfectly flat, as you can imagine. In addition, with the mounting tabs at the far end of each cube, the pressure across the moutning plate is not even.

Remove all old grease and/or sil-pads, use an adequate amount of thermal grease, clamp firmly with the mounting hardware, wipe off the (dust/dirt grabbing) excess, verify your terminals are tight, and give it a try.

David

 

[Lightguy5]

Thank you all, for the probable solution!
Now I want to understand.
Why, with only a single 575w unit on the dimmer, is a 25A cube insufficient in this instance?

 

[dj41354]

I'd guess that the inrush current is causing the problem.

I'm not a lighting guy, so I don't have specific experience with lights on SSR's, but I've used SSR's on motors (where there are inrush issues also). On motors, 3X seems to do the trick... but this may not be enough for some incandescents. If you really want to get to the bottom of this, you need to determine the inrush event with the lights you're using, and then look carefully at the SSR data sheet. They normally give you the number of AC cycles that the SSR can deliver higher current than the continuous rating. Also, not all SSR's (with the same rating) are the same.. some have higher instantaneous current ratings than others.

Here's a couple of technical references discussing inrush current on incandescent lamps:

“When an incandescent lamp is initially turned on, the cold filament is at minimum resistance and will normally allow a 10X to 12X peak current. Within 3 to 5ms the current falls to approximately 2X the hot current”
(source):
http://www.allegromicro.com/en/Products/Design/an/an295012.pdf

“The electrical resistance of the tungsten filament at room temperature is initially quite low. When electrical power is first applied to the lamp, a large in-rush current causes rapid heating of the filament. The resistance of the filament rises to a value five to ten times the cold resistance, which causes the amount of current drawn by the lamp to stabilize and the lamp to emit a stable light output. Depending on the size of the filament, the in-rush period can be from tens of milliseconds to hundreds of milliseconds. This in-rush current requirement should be taken into account in the selection of the power source for a specific lamp application.”
(source):
Tungsten Filament Lamps — International Light Technologies

“The reduced resistance of a cold bulb results in larger current draw when it is first powered; called "in-rush current". The filament color temperature (Kelvin) at which the lamp is designed to operate determines the inrush current. A 300'K lamp will have an inrush current about the same as its operating current, but a 2,500'K lamp will draw about 12 times the current when cold. Most lamps will heat to normal current ratings in approximately 30-100 milliseconds.”
(source):
Incandescent Lighting

 

[DavidNorth]

Correct. In-rush on a cold tungsten filament can be 10-14 times its operating current rating for up to 3 cycles. It is this reason that you need to have the appropriate cube and also why magnetic breakers with medium to high delay trip curves can withstand this high-amp onslaught. Our experience with lamp and dimmer testing is that only takes about 3 minutes for a tungsten lamp to get cold enough to draw this amount of current at turn on. Think about that next time there is a dimmer check/preheat disussion......

So here are some other probable reasons for going with a higher current rated cube.

- The larger the current rating, the large the die and the better ability it has to withstand dead shorts and apparrent shorts. Note that a cold filament looks like a dead short for a very brief amount of time.
- You never know when you will put more that 575W on that dimmer. So if it holds on 575W but later you put 3 - 575Ws on there, it may pop. Put the right part in.
- The dimmer design may be using some of that curent overhead and cheating on cooling to take advantage of the cube design. I'm not saying that is the case on any specific manufacturer's design, but best not to take chances. A combination of heatsink surface area and forced air cooling are where engineers can rebudget energy.

That all make sense?

David

 

[Lightguy5]

So here are some other probable reasons for going with a higher current rated cube.

- You never know when you will put more that 575W on that dimmer. So if it holds on 575W but later you put 3 - 575Ws on there, it may pop. Put the right part in.

That all make sense?

David

Oh, I'm all over the whole long-term issue, and having the under-rated cube is bad, but one of the cubes was installed and failed THIS WEEK, with no change in the load in the meantime. That's what I got curious about.

Anyway, new parts on their way and I'm going to try the heat pads that Crydom also sells for them.
That's also very enlightening about the power surge through cold filaments (no pun intended, but I'll use it anyway). Thanks, guys!

 

[JD]

Oh, I'm all over the whole long-term issue, and having the under-rated cube is bad, but one of the cubes was installed and failed THIS WEEK, with no change in the load in the meantime. That's what I got curious about.

Anyway, new parts on their way and I'm going to try the heat pads that Crydom also sells for them.
That's also very enlightening about the power surge through cold filaments (no pun intended, but I'll use it anyway). Thanks, guys!

Watch your voltages too! The 240v may appear reasonable, but most vendors use much higher voltages. I would not consider using one under 480 volts. Most SCR and Triac dimmers use devices with a PIV of 600 volts. Without getting into a long explanation, a 240 volt SSR can fail on a 120 volt circuit. Remember, 120 is an "RMS" value, the peak is much higher. Add in some inductance flyback and surge variations and you can exceed the device ratings and blow the device even if there is very little load attached.

It is hard to determine true PIV on an "engineered" device such as a SSR, but as there is no external snubber cap on this dimmer, the OEM may either contain a snubber, or be of such PIV that flyback from the circuit will not exceed it. I would buy at least one OEM and try it out. If that one does not fail while others do, we can safely conclude that the choice of SSRs is the problem.