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

[jocanon]

How difficult would it be to add an on/off switch to the PSU (the DUT). I imagine it would be as easy as just splicing it in on the mains wires and then finding somewhere to mount it...am I missing something?

Also, if I plugged in the charger and got that spark we were talking about earlier because it is charging up the capacitors, would it potentially damage my charger, or just the PSU, or niether? I am thinking if it would probably not damage either, but if anything could potentially damage the PSU only, because it is seen as shorting the PSU leads while it is charging up the capacitors(although the PSU protections, i.e. over current protection, should kick in), but the charger should be fine, is that right?

 

[()blivion]

First question:
I don't know about server PSU's but with standard ATX PSU's you turn them on by shorting the green wire to ground. If your PSU has the normal connectors on it like that then you could do this, but I didn't see them in your image. No matter what you can install a switch in the main line if that's you want.

Second question:
The sparking is nothing more than high current running through the point the metal makes a connection causing a microscopic piece of metal to ablate (explode). It is more harmful to your connectors than anything. After a time enough metal will have melted off to mi-shape your connectors enough that they will get poor contact. And by "after a time" I mean a few hundred years or so more than likely.

Edit: Disregard the statements that were here.

Bottom line:
I would say, You don't really need to worry about sparks when first connecting devices. Worry more if there is no filtering capacitor or TVS.

 

[jocanon]

OK, thanks ()blivion, I can take pictures of the internals of the PSU when I get home, for the power switch question.

 

[jocanon]

Here are a bunch of pictures of the mains power.

 

[jocanon]

I am trying to think of ways to make just one switch for the 24v PSU (2 run in series)...I am thinking maybe the best solution might be to just rig an in-line on/off switch on the power cord itself where I rig a Y connector so I plug them both in with one plug and put the in-line switch on the single cable part of the Y connector. Seems like that will work, then I won't have to do any modding on the mains inside the PSU.

 

[()blivion]

I am trying to think of ways to make just one switch for the 24v PSU (2 run in series)...I am thinking maybe the best solution might be to just rig an in-line on/off switch on the power cord itself where I rig a Y connector so I plug them both in with one plug and put the in-line switch on the single cable part of the Y connector. Seems like that will work, then I won't have to do any modding on the mains inside the PSU.

Yeah. With mains power I normally use a DPDT switch. This disconnects both the hot and neutral. Both will shock the crap out of you in reality, so I'd just as soon unplug them both.

 

[jocanon]

Hoping to FINALLY get some time this Labor Day weekend to build the dummy load...I've been working 12 to 13 hour days lately! I will post up pictures as soon as I get it done.

 

[()blivion]

Yeah, I have been busy lately also. No worry's.

 

[jocanon]

I got two of the Op-Amps soldered up on the PCB and I wanted to test it before I finished the rest and it works, yeah! I will finish the rest later and then post some pictures.

 

[jocanon]

I got all the Op-Amps soldered and noticed that there is continuity between the positive and negative feedback loops of all the Op-Amps except one. There is no reason that one should be any different than the others so I am assuming it is faulty and will have to take that one off and put in a new Op-Amp. Is it normal to have this happen or is it something I may have done to cause it like getting it too hot when soldering?

 

[()blivion]

Is it normal to have this happen or is it something I may have done to cause it like getting it too hot when soldering?

Hard to say. Depends on how you tested it. It could be other components, or a slightly different configuration.

 

[jocanon]

Yeah, OK, so I am an idiot, haha. I re-did the entire Op-Amp and it still didn't work. Come to find out I just didn't have a good enough connection on the .13 ohm resistor since I only laid it on the ground wire but didn't solder it yet. I did the same with the other .13 ohm resistors and it made a good enough connection in all cases but one, now that they are soldered all of them work... duh!

Edit:
At least it was good practice!

 

[()blivion]

These sort of things happen. Doesn't make you an idiot. Just remember, if you break something... get it on film so we can laugh at you.

 

[jocanon]

I just realized a mistake in my schematic, luckily I haven't built that part yet, I was showing R25 connecting to pin 4 an 3 but I think it should be 5 and 3. I think everything else is right.

 

[jocanon]

I am thinking I will put the temperature sensor somewhere between the 10 MOSFETs where it gets the hottests, or should input it further away from them?

 

[ronv]

Sounds like you are moving right along. I would put the sensor at the hottest place since it is there to protect the MOSFETs. The hottest one would probably be the one to blow under full power and we don't have a lot of margin.

 

[jocanon]

Thanks ronv, what about ()blivion's comment a few posts back on when he put the schematic in a simulator, he said the over temperature alarm was going off continously, are the ohm ratings of the resistors right? If I do have this problem, which resistor(s) would I have to tweek? I am thinking probably R12 or R13.

Status update:
I got everything soldered onto the PCB late last night. I still need to connect all the components that are not on the PCB to the PCB, the MOSFETs, fans, alarm, LEDs, then paint the whole thing with a conformal coating (after I make sure it all works of course), thinking of using polyurethane for this which I have laying around from a wood working project.

 

[ronv]

I think the alarm will be ok. The base needs to be about .65 volts lower than the emitter for the transistor to turn on. It is only about .45 lower.

Your question about R25 is also correct.

 

[()blivion]

I just realized a mistake in my schematic, luckily I haven't built that part yet, I was showing R25 connecting to pin 4 an 3 but I think it should be 5 and 3. I think everything else is right.

Oh... yup... the schematic definitely has it wired wrong. You can move R25 to pin 5 if you want, or R24 to pin 11, it doesn't really matter as either works. We just want to give the unused gates something to do. Though as I said, with comparators I don't think it actually matters as much. Unlike other logic gates/IC's comparators are made for analog input. They may oscillate without tie ups/downs though. So it's sort of best practice to do it the way we have it right now.

 

[jocanon]

AMAZING!!! It works as advertised !!! I even tested out the temperature shut off and that worked exactly as we planned. I couldn't even get the MOSFETs hot enough to activate it though, even with 1200 watts so I put it on the current sense resistors (because those get HOT) to test it. The second I turned the water off the temperature on the MOSFETs would climb rapidly, but with the water running they stayed nice and cool I think the hottest they got was about 60C. The only thing that worries me, and maybe it is no problem, is how hot the current sense resistors get. Those things cook! They would get up to over 80C at about 650 watts and trip the over temp circuit when I put the temp sensor on one. How hot can those safely get?

Btw, the DUT over current protection did kick in at 55 amps, it just shut down and then I had I turn off power, but then it started right back up just fine.

One thing the dummy load did unexpected was when powering down the alarm beeps a couple times, not a big deal as long as it is not hurting anything, I doubt it is.

Anyway, in summary, awesome job guys! This thing is great!!! I will post up some pictures tomorrow, it is too late tonight and I have to go to work in the morning. Thanks!!!!!