Bench power supply shenanigans.

[Spidey]

I just got my first variable power supply a week or so ago and was experimenting with charging batteries out of a power tool pack that was saying it was fully charged when it clearly wasn't.
 
Anyway long story short I got one of the cells to my goal voltage which I think was about 4.15v and I turned the output of the supply off thinking that was the safest way before disconnecting the clips and I think the power of the battery must have fed into the power supply and burned something up from the smell and noise which there was a lot of before I was able to unclip. Now my problem is after that happened I tested the power supply and it refused to go over 3 volts and it was signalling that it was being current limited. THEN the next night I tested it and it worked fine as if nothing ever happened, I even ran a laptop off of it at 19v and  3.5 to 4.2 amps for about a half hour no hiccups.

So now I'm wondering what the hell burnt up and why it started working again? I'm so confused here.

The power supply is a Hyelec HY3005B, Rated for up to 30 volts/5Amps. I took it apart to see if I can visually see what may have burnt up but no luck. Any Ideas? I was thinking of repeating the failure to see if I can see the component failing.

[Richard Head]

The PSU may not have had an anti-parallel diode across the series pass transistor. Without the diode the battery can back feed power and avalanche the base-emitter junction of the transistor, among other things.
I wonder if it was an oriental brand.

[CatalinaWOW]

With perhaps a thermal relay?  Got hot before it opened, then cooled and reset overnight.  There are lots of materials that let out some magic smells long before there is any visible damage.

[Seekonk]

I did a test one time to see what the limits were of 1/2W CF resistors.  I cooked them till they turned red for a few minutes.  Never could get one to fuse open. Resistances changed less than 5%.  I'd say don't worry about it.  No need to plan a failure.  You'll have plenty of chances for inadvertent failures in the future

[timb]

I just got my first variable power supply a week or so ago and was experimenting with charging batteries out of a power tool pack that was saying it was fully charged when it clearly wasn't.
 
Anyway long story short I got one of the cells to my goal voltage which I think was about 4.15v and I turned the output of the supply off thinking that was the safest way before disconnecting the clips and I think the power of the battery must have fed into the power supply and burned something up from the smell and noise which there was a lot of before I was able to unclip. Now my problem is after that happened I tested the power supply and it refused to go over 3 volts and it was signalling that it was being current limited. THEN the next night I tested it and it worked fine as if nothing ever happened, I even ran a laptop off of it at 19v and  3.5 to 4.2 amps for about a half hour no hiccups.

So now I'm wondering what the hell burnt up and why it started working again? I'm so confused here.

The power supply is a Hyelec HY3005B, Rated for up to 30 volts/5Amps. I took it apart to see if I can visually see what may have burnt up but no luck. Any Ideas? I was thinking of repeating the failure to see if I can see the component failing.

I've had similar things happen with older power supplies before. If it appears to be working, I wouldn't worry too much about it. Sounds like one of the power transistors was back fed by the battery and avalanches, causing it to dump all the energy back into the supply capacitor. This either over heated the transistor or a thermal fuse. Once cooled it should work normally again.

[Spidey]

I guess the consensus is - if it ain't broke, don't try to break it again to try and fix it. My concern really was that I may have severely shortened the lifespan of the unit. 

[Spidey]

I had an idea to run a diode inline with the positive wire to prevent back feed, does this sound like a good idea? Also there's like 10+ different types of diodes and i'm not sure which would be best. 

[salbayeng]

Tantalums can make an enormous amount of smoke and pong before exiting stage right.
Sometimes they explode, and there's just a few red bits left!