Help with PSU (Temp control fan, load bank, & PWM circuit)

Wow! I got some pretty wild readings! I measured the DC voltage. When running the single PSU (12vdc) It was a stable 12.7 volts no matter what the current level, from 0 all the way to 50. When I ran the 2 in series for 24vdc, totally different result...It was about 24.3v at 0, it decreased quite steadily to about 30 amps, at 30 amps it was only 13.5vdc, then it started spiking, it got all the way to 170vdc at 47 amps and seemed to be climbing, but I turned it off! What's going on here???

OK, those numbers were just a little bit too crazy...I took another reading with a different cheapo voltmeter from Harbor Frieght and it is not 170vdc...It was about 25v but then around 40amps the voltage becomes unstable and starts to jump around going as low as 20vdc at times, but never above 25. Now I don't know what to trust though, I am starting to loose confidence in my cheap $5 voltmeters from Harbor Frieght. I probably need some sort of scope or something but I don't have access to one. Do you think the voltage becoming unstable is most likely related to the PSU or the dummy load? Maybe tomorrow I can load it up with chargers and measure the voltage.

edit:
I also noticed the current level starts to becomes unstable at around 30 amps (which makes sense if voltage is jumping around).

Hmmm. Not good.
Check to see if it is the supply with the isolated ground. In other words measure each one on it's own as a 12 volt supply then with the two together see what the voltage is across each one. They both should be solid at 12 volts. You don't need to go up so high as to have it go wild.

The fact that the TVS are blowing suggests the PSU.

I am starting to loose confidence in my cheap $5 voltmeters from Harbor Frieght.

Click to expand...

LOL

Just a thought but judging from the fact that this only happens when the two supply's are connected, we may be getting some oscillating in the current sharing path and it's making things wacky.

Maybe we can put a RC circuit on the current sharing pin? Give it something like hysteresis?

Even when run in serie, each unit alone looks to be a pretty steady 12.5 volts, it drops slightly, but when I measure from ground of one unit to +24 of the other it starts to drop a lot around 40 amps.

Edit:

I take that back, the one that's not isolated is steady, but the isolated one drops above 40 amps even by itself.

I have to boggie now, but maybe a link to the isolation process??

There are only two steps involved in the isolation, it in the picture with words on it in the first post of this thread...I need to get going too, it's getting late.

Isolation Check

(1) Disconnect everything you can from the "isolated" supply
(2) Put a multimeter to the ohms/continuity setting.
(3) Hold the probes together, note what the meter does.
(4) Hold the black probe to the ground of the "isolated" supply.
(5) Scratch the red probe tip into the metal case a bit. (To get good contact to fresh metal below it.)

*IF* the meter reads the same as when the probes are connected to each other, then the PCB is not properly isolated from the case. Otherwise, you should be good.

Here are some things to check if test shows problems. These things are a mix of long shots you would never think of and common sense stuff that's easy to overlook. But the basic idea is simple, we are looking for ANY connection that exists from the main PCB ground to the case still.

(1) Did you put washers on BOTH sides of the PCB?
(2) Is the screw holding down the PCB making contact with any traces/copper on the PCB? (The inside of the PCB hole sometimes has metal.)
(3) Is the divider under the PCB missing any pieces or misaligned?
(4) Are there any protrusions from the case that could be touching the PCB?
(5) Is the green ground wire from the mains power socket screwed to the case, or the PCB? Is the screw screwed into the case?
(6) Is the new fan circuit wiring to anything other than it's own power? If so, is it case ground?

I did make sure there was no continuity between DC ground and the case. I just double checked again and it is properly isolated. I used a washer in both sides. There is more info on the process here:
https://www.rcgroups.com/forums/showthread.php?t=1581061

I did notice just now that all the PSUs have 1169 ohms between +24vdc and DC ground, but the one I am testing the op-amp temp control circuit on only has 1055. Don't know if this matters. Also, if I disconnect the isolated PSU from the non-isolated one there is no continuity between DC ground and case, but when I connect them in series there is 1055 ohms between DC ground on the isolated PSU and case, which would make sense because I just connected the positive lead of the non-isolated PSU to the DC ground of the isolated PSU and the DC ground of the non isolated PSU is connected to the case of the isolated PSU through the AC earth ground pin.

FWIW, here is a link to the video I shot earlier today.

https://www.youtube.com/watch?v=t5rglrH168E&sns=em

I supose there is a possibility the supply has gone bad since running by itself it still has a problem. Is it possible to restore it to normal and try it by itself?

I get a little lost in the RC Group thread. Is there still a capacitor attached somewhere to frame ground? If there is thats not a problem--just curious. It sounds like some other people have had similar problems?

Yeah, it is a long thread with a lot of information scatter around...I did not add a capacitor to attach to frame ground, are you talking about something that is already part of the unit or a mod?

Already a part of the unit.

Did they talk about that in the RC groups thread? I must have missed it, do you know what post number?

It might have been a different supply. Anyway it is not important. Do you think it is a bad supply or something with the modification?

hm, I don't know. I guess it could be a bad supply. When I get home today I want to to try a different supply and see if I get the same result. I can also try undoing the mod on the isolated supply and see what happens.

Do you think it would help at all if I were to connect the current sense pins to the load leads? Do you know what I am referring to?

No, I don't think that would help. If we are talking about the same pin it is just to make the supply compensate for the voltage drop in the leads.
Might be a good idea to test one before and after the modification. How many TVS's do you have?
My bet is that it is just a bad supply, but if it is the ground mod it might be noise as sometimes the line filters used in switchers use frame ground.

I am all out of TVS's, blew all three

Would you say it would generally be a bad idea to test it with the TVS detached, especially if it is sending high voltage through the dummy load, high enough to bust the TVS? I could probably find another one at Radio Shack on my way home from work before I test them tonight if needed.