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

[jocanon]

I'm not not dead yet, as they say in Monte Python. I got the copper pipe soldered last night. Just doing this one small step at a time. Tax season is killing me!

 

[jocanon]

One interesting thing to note, the stainless steel scouring pad I put inside the copper pipe can catch fire if it gets too hot. It started smoldering and set off my fire alarm when I was soldering the coper pipe fittings. Looking back, it might have been best to put the stainless steel in the pipe after it was soldered.

 

[()blivion]

Looking back, it might have been best to put the stainless steel in the pipe after it was soldered.

LoL... Any crash you can walk away from... was not a crash my friend.... It was a successful landing.

 

[jocanon]

LoL... Any crash you can walk away from... was not a crash my friend.... It was a successful landing.

True, very true.

 

[jocanon]

OK, wheels up in my head turning again, I was just thinking about the grounding issue once again...since I have to float the DC supply on one of the PSUs when running them in series to get 24v (I have to isolate one of the PSUs from ground or it will short when I connect them in series), if I ground the case, will I need to be sure that none of the circuitry that is connected to the leads of the DUT in anyway touches the case when the 24v DUT is connected or it will short out, i.e. the second PSU in the 24v series set up would no longer be isolated from ground if anything touches the now grounded case? I do not forsee that being an issue, I will just have to be aware of it and be sure that nothing can physically come in contact with the case.

Edit:
Please let me know if any part of my question above doesn't make sense. I had a hard time formulating it into words.

 

[jocanon]

Okay, so here is a progress report...I was able to get the mosfets soldered onto the copper pipe. I tested each mosfet on my test circuit after soldering and they all still work, the heat didn't destroy them. I ended up using the torch to solder them, the soldering gun just wasn't cutting it on that big pipe. So I pre-tinned the copper pipe with A LOT of solder using the torch. I also pre-tinned the mosfets using my temperature controlled soldering station. I never pre-tinned the back of a mosfet before, so I didn't know what to expect. I thought the solder would flow onto it and lay flat as it usually does, but it didn't, it balled up on the back of the mosfet. I think that worked out alright though, because then I layed them all on the copper pipe and they didn't lay flat since there were balls of solder on the back of them. I heated the copper pipe from the back with the torch (from the back as to not apply heat directly to the mosfets). As soon as the temperature got hot enough and the solder started to flow, the mosfets would fall down and lay level in the pool of solder so the balls of solder on the back served a purpose to let me know when the solder was flowing. Once the solder started to flow on one set of mosfets, I would quickly move down the line to the next mosfets as not to leave the heat on them too long because I was worried about damaging the mosfets. It worked like a charm, when I was all done I tested each mosfet individually and they all still work. I will do some more work on it tonight, but for now I have to go to my day job for a while.

 

[()blivion]

OK, wheels up in my head turning again

If that was intentional airplane jargon, and that was the best you could come up with, then you should be PUN-ished. (Kidding, also failing at trying to be funny.)

Will I need to be sure that none of the circuitry that is connected to the leads of the DUT in anyway touches the case?

In general, you should make the circuit board as covered with insulation as you can get it. It would be really bad for anything conductive to touch any part of the circuit board at all.

Speaking specifically, if our load circuit has it's ground attached to the case ground, then anything from our circuit that touches the case will be grounded. This could be as simple as screwing with the feedback causing things to flake out. Or of course it could just let a whole lot of current through one single part, instantly exploding it. And I really do mean exploding, like, with a loud bang and smoke and such.

Things probably will be hooked up this way. So short answer is yes, keep anything electronic from touching the case unless specifically told to do otherwise.

[Would] the second PSU in the 24v series set up no longer be isolated from ground if anything touches the now grounded case?

No, that supply is still "isolated" in the way it needs to be. It is the high side, the most positive supply. And it's positive is the only wire coming out of it to the rest of the world really. The rest is isolated. The thing that needed to be accounted for so that things were safe was the other supply's positive. Which is why you needed to isolate the other one's ground, because that ground and supply were attached directly to each other.

Basically, The statement...

I have to isolate one of the PSUs from ground or it will short when I connect them in series

Should be corrected to...

I have to isolate one of the PSU's Grounds from it's case otherwise it will short the other supply when I connect the two together

 

[()blivion]

I also pre-tinned the mosfets using my temperature controlled soldering station. I never pre-tinned the back of a mosfet before, so I didn't know what to expect. I thought the solder would flow onto it and lay flat as it usually does, but it didn't, it balled up on the back of the mosfet.

It should have flowed onto it and flattened out (wet) the backs well.

That means...
(1) You didn't clean them well enough.
(2) You didn't get them hot enough.

(2) is the most likely culprit. Though having a look at your second picture, I can see the the solder is tented in the right direction right against the MOSFET's to suggest that when you put them on the pipe, they were fully wet with solder before it eventually solidified. So no harm no foul I guess.

The best test would be to try and pull or pry one off, it should be hard enough that you would destroy one physically before the solder let's go.

IF THEY WERE NOT SOLDERED TO THE PIPE WELL ENOUGH THEY WILL OVERHEAT AND DIE!!!

Edit: The second from the left and the third from the left both look the most suspect. Judging by the eyelets the others look to be well bonded.

 

[jocanon]

Yeah, I thought I probably didn't get them clean enough or hot enough when I pre-tinned the mosfets. But they got A LOT hotter when I soldered them with the torch onto the pipe. I tried pulling them and they are definitely on there tight enough that they will not come off without causing physical damage, so I think they are good to go.

 

[jocanon]

The one you have is the right one. Sorry. You don't need to worry about it reading amps. For it to do that it needs a shunt (sense resistor). We can just convert amps to volts and measure it that way. The only drawback is that a reading of 5.1 will mean 51 amps. The same way with temperature. 0.6 volts will represent 60C. We should maybe amplify this one to make it 6 volts = 60C since the meter only has 1 digit after the decimal point.
Will that be okay or should we work on making the readings complete?

Here is a link to a small power supply for the logic.

**broken link removed**

How difficult would it be to amplify the readings so that 51 amps = 51volts and 60C = 60 volts for proper readings on the volt meters? If too much trouble then I won't worry about changing it at this point, but if it is simple and I can just pick up a couple more parts to modify it I might do that.

 

[()blivion]

All things being equal, you did a decent job. I think it's serviceable. The only thing I would have done differently with the same equipment is maybe have all the FET's in the PCB before soldering them to the pipe. You may run in to alignment issues when it comes time to board the FET's now, which may lead to having to modify the prototype holes to fit the MOSFET's better. Starting with them in the board makes them both look better since they are all lined up already, and makes it easier to wire the circuit.

Unless you plan on running wires to each of the FET leads one by one? I don't recommend this, though it can be done.

 

[()blivion]

How difficult would it be to amplify the readings so that 51 amps = 51volts and 60C = 60 volts for proper readings on the volt meters? If too much trouble then I won't worry about changing it at this point, but if it is simple and I can just pick up a couple more parts to modify it I might do that.

Difficult. For one, the meters need to survive 60+ Volts, and for two, we would need a supply to generate 60+ volts. And that's not even mentioning the fact that an amp that runs on 60+ volts is going to be hard to make/find.

 

[jocanon]

Difficult. For one, the meters need to survive 60+ Volts, and for two, we would need a supply to generate 60+ volts. And that's not even mentioning the fact that an amp that runs on 60+ volts is going to be hard to make/find.

Okay, nevermind that idea then

 

[jocanon]

All things being equal, you did a decent job. I think it's serviceable. The only thing I would have done differently with the same equipment is maybe have all the FET's in the PCB before soldering them to the pipe. You may run in to alignment issues when it comes time to board the FET's now, which may lead to having to modify the prototype holes to fit the MOSFET's better. Starting with them in the board makes them both look better since they are all lined up already, and makes it easier to wire the circuit.

Unless you plan on running wires to each of the FET leads one by one? I don't recommend this, though it can be done.

I was planning on running wires to each of the FET leads. Why would this be a bad idea?

Edit:
Is it just that it wouldn’t be as clean or could it adversely affect to the overall performance of the load bank?

 

[jocanon]

Sounds like it's not a good idea to have the pins of the mosfets not secured to something because they could break off too easily that way. I am going to have to think of the best way to secure them, it may be that the best way to do that is attach the PCB to the copper pipe and then I can solder the mosfets into the PCB as well. Question is, how to attach the PCB to the copper pipe, plastic zip ties, or will they melt with the heat???

 

[jocanon]

Ok, so just trying to plan out my layout and I want to verify that I am thinking of a couple things correctly. First off, the wire gauge that attaches to the 3rd pin of the FET or source only really has to be able to handle up to 6 amps, right? I mean, we are running 10 of these in parallel but each one individually is only going to be going up to 6 amps. Then this wire out of the source pin in the FET goes to the negative feedback loop on the op amp and the .13ohm current sense resistor. Now the current sense resistor on the other side connects to the negative terminal on the DUT and this wire needs to be able to handle the sum of all the currents from the paralled FETs which in my case is up to 60 amps. Am I getting that all right?

 

[ronv]

First off, the wire gauge that attaches to the 3rd pin of the FET or source only really has to be able to handle up to 6 amps, right?
I mean, we are running 10 of these in parallel but each one individually is only going to be going up to 6 amps.

Right you are.

Then this wire out of the source pin in the FET goes to the negative feedback loop on the op amp and the .13ohm current sense resistor. Now the current sense resistor on the other side connects to the negative terminal on the DUT and this wire needs to be able to handle the sum of all the currents from the paralled FETs which in my case is up to 60 amps. Am I getting that all right?

Right again.

As they say in the old matchbook ads "You may have a promising career in electronics"

 

[jocanon]

Lol, thanks . Actually, right now I wish I did go into electronics. It's much more interesting than accounting!

 

[ronv]

This is a good project. Some different "stuff". It will be interesting to see just how well the water cooling works.

 

[jocanon]

I cut open a computer power cord and got some 18awg wire out of it. That should be enough to handle the 6 amps shouldn't it?