I Messed Up - Torroidal Power Supply

[metrologist]

I have a moderately large DC power supply, 80V 20A. It can be wired for either 120V or 240V. The instructions were rather simplistic and I blindly followed them...

There are two reds and two blacks on the primary. The secondary is already wired to an unregulated dual output rectifier circuit with large filtering caps.

I wanted to wire for 240V and the instructions said to take any red and a black and connect them together, then take the remaining red and black and connect them to the two hots. I did not quite get that "a" black had to be "the" specific black, and the way they came out of the torroid was not obvious. Naturally, we know why I'm posting...

OK, after taking a deep breath, I threw the switch, the lights went dim, and my torroid started making rice krispy sounds. I saw a slight puff of smoke, and quickly shut the mess off. Then it dawns on me that there are two individual primaries that needed to be wired in series, and all I did was wire them in parallel reverse polarity 

I undid the mess and wired it up correctly and it seems to power up fine. None of the components on the output seemed to be blown up, and the DC voltage with moderate load seems just fine.

But, I saw smoke and I'm sure something bad has happened. I suspect some of the winding insulation has burned, and I am sitting on a ticking time bomb before it melts down an destroys the stuff on the output. Does this seem about right? Or is this just a bit Chicken Little?

[T3sl4co1l]

You plugged it in unfused?

[Red Squirrel]

I was testing a UPS to see if I can repair it, I narrowed it down to possibly the transformer but it was somewhat uneducated guess.  I did the mistake of hooking up what I assumed with the primary side directly to 120v to see if I get voltage on the other end, but I got nothing. I then tested the ohms and got nothing - was open... I should have tested the ohms FIRST.  I don't know if I blew it, or if it was indeed already blown, and indeed problem.  I'm going to go with it was already blown.   

UPS transformers make great door stops though, for when you open windows in summer and don't want doors to randomly slam.

[metrologist]

You plugged it in unfused?

There's a dual 30A circuit breaker on the panel. Does that count? I better install that 10A switch breaker, huh?

[rdl]

Circuit breakers are to protect the wiring in the walls. They're too slow to protect devices plugged into the wall outlets.

[T3sl4co1l]

You need a 8-10A fuse to protect a transformer that size, and actually have it protected.

Breakers may operate too slowly to save a transformer, certainly one triple the rating!  Which explains your observation, more or less.

Tim

[metrologist]

I bought a two pole, red rocker style 10A circuit breaker switch. Is that good enough. (like this one: http://www.digikey.com/product-detail/en/te-connectivity-potter-brumfield-relays/W33-S4B1Q-15/PB262-ND/254480)

And should I just buy a new power supply? I know, it's an impossible question to answer. Who knows what could have happened. Does connecting the windings in reverse polarity make the load even worse, or provide some cancellation of EMF induced current?

[T3sl4co1l]

Hmm, maybe. I still don't like the idea of relying on a fragile mechanical breaker for protection.  Good old cartridge fuses are cheap and reliable, and you should never have to replace them (or flip a breaker) in properly designed equipment.

And the faster clearing speed of a cartridge fuse means less damage to the transformer -- no sizzle to worry about.

Tim

[David Hess]

And should I just buy a new power supply? I know, it's an impossible question to answer. Who knows what could have happened.

I have had power transformers degrade over time and fail later leading to fires after abuse like that so I would definitely consider it a safety hazard.

Does connecting the windings in reverse polarity make the load even worse, or provide some cancellation of EMF induced current?

That is what happened; the 120 volt primary windings connected anti-parallel so the inductance mostly cancelled out and it looked like a short to the power line.  I am amazed that the 30 amp circuit breakers did not trip.

What is the resistance of one of the primary windings?  Maybe that plus the wire run resistance was enough to limit the current so that the circuit breakers did not immediately trip.

[metrologist]

You's are not instilling any piece of mind

But I guess there is a silver lining in here...

[Zero999]

That is what happened; the 120 volt primary windings connected anti-parallel so the inductance mostly cancelled out and it looked like a short to the power line.

Yes, that sounds like what's happened.

A similar thing would have happened, if the two windings were in series and one was backwards but the current would have been lower as the resistance would be double.

[David Hess]

You's are not instilling any piece of mind

Good, I almost had a house burn down because of this.

You are relatively safe because the transformer is inside of what is effectively a fire shielded enclosure.  Add some appropriate circuit protection on the line side like a fuse or fast circuit breaker.  Then run a supervised long term test over a couple of weeks where you check the temperature of the transformer at the end of each run.  I suppose you could do a high potential test but I am not sure if it would reveal anything useful.

But I guess there is a silver lining in here...

What silver lining is that?

[metrologist]

I learned something about transformers.

[David Hess]

I learned something about transformers.

True wisdom would be learning from someone else making this mistake but the opportunity does not always present itself.  At least it turned out well except for the transformer and your wallet and who knows?  The transformer may have survived to operate for its expected operating life or longer now that we know it is not weak.

[johansen]

I wouldn't worry about the transformer, unless the insulation changed color.

the crackling sound is the windings expanding due to thermal expansion causing the mylar and other plastics to creak as they shift a bit..

[VK5RC]

If it will be powered up and unsupervised for long-ish periods I would ditch it,  have a look at the impedances of the primaries,  comparing one to the other,  also fire each side up individualy to see if you get the same voltage out on the secondaries,  if out by much,  I would ditch it.  I nearly went insane trying to parallel some secondaries in a big toroid. No smoke but I was thinking of starting a fuse factory.  HiHi.

[Kilrah]

I wouldn't worry about the transformer, unless the insulation changed color.

How can you see insulation color deep down in the coil, which is where most of the heat buildup would happen? 

[SeanB]

Use an insulation tester to see if there is increased leakage current between the primaries. If there is ( and on a good transformer this will be over 1G of resistance at 1kV DC test voltage) then you have damaged carbonised insulation, and the transformer is now a candidate to rewind or scrap.

[Assafl]

I have a moderately large DC power supply, 80V 20A. It can be wired for either 120V or 240V. The instructions were rather simplistic and I blindly followed them...

There are two reds and two blacks on the primary. The secondary is already wired to an unregulated dual output rectifier circuit with large filtering caps.

I wanted to wire for 240V and the instructions said to take any red and a black and connect them together, then take the remaining red and black and connect them to the two hots. I did not quite get that "a" black had to be "the" specific black, and the way they came out of the torroid was not obvious. Naturally, we know why I'm posting...

Something doesn't seem right. It seems each red and black form a winding - where either the red (or the black) always has the "Transformer Dot" (the symbol used in a schematic to assign the winding direction). In that case it wouldn't matter if Coil1Black is Connected to Coil2Red or Coil2Black to Coil1Red. In both of these - the Dots are in series (properly).

Perhaps the trafo is wired incorrectly?

[metrologist]

...it wouldn't matter if Coil1Black is Connected to Coil2Red or Coil2Black to Coil1Red. In both of these - the Dots are in series (properly).

Perhaps the trafo is wired incorrectly?

Oh, you're right. I'm not sure why I was saying parallel. For 240V, it would need to be wired series.  So what it really means is that I connected one primary together (shorted) and blasted the other primary with 240V. The lights dim anyway when I first click it on because of the large filter caps.

[electr_peter]

...it wouldn't matter if Coil1Black is Connected to Coil2Red or Coil2Black to Coil1Red. In both of these - the Dots are in series (properly).

Perhaps the trafo is wired incorrectly?

Oh, you're right. I'm not sure why I was saying parallel. For 240V, it would need to be wired series.  So what it really means is that I connected one primary together (shorted) and blasted the other primary with 240V. The lights dim anyway when I first click it on because of the large filter caps.

Thanks for clarification. I also thought that you may have connected 120 winding to 240V and shorted the other winding, but your "parallel connection" explanation threw me off.

[Assafl]

...it wouldn't matter if Coil1Black is Connected to Coil2Red or Coil2Black to Coil1Red. In both of these - the Dots are in series (properly).

Perhaps the trafo is wired incorrectly?

Oh, you're right. I'm not sure why I was saying parallel. For 240V, it would need to be wired series.  So what it really means is that I connected one primary together (shorted) and blasted the other primary with 240V. The lights dim anyway when I first click it on because of the large filter caps.

LOL. If you shorted the second primary then the filter caps are not the reason your lights dimmed.

Hint - it is the shorted primary, acting as a shorted secondary.... In fact, I'd gamble that your secondaries are perfectly fine. I'd test the primaries with a Megger, though...

[metrologist]

I'll have to see if there is a place that will test it. It powers a machine that is not used that often, and is somewhat supervised when running.

[qno]

Next time you do something like this put an light bulb in series with the transformer and do not load the power supply.
The light bulb will light up and slowly dim due to the rush-in current of the power supply if all is ok.

Else the light bulb will act as an current limiter and indicator that "something is wrong"