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

[()blivion]

Did we add over current protection? I thought we just added protection against over temperature. We added a feature to read the current, but I don't think it shuts down if it goes over a certain limit...I could be wrong. In post 141 ronv talks about the over temperature cutoff.

*Looks at schematic, this time with glasses on*.... OH! your right.... That must be why I had forgot

Looking into a way to do the MOV's other than 10 of them. If you get a common 63 Amp circuit breaker, and have the MOV parallel to the dummy load, but all that on top of the circuit breaker. Then when the MOV trips FROM OVER VOLTAGE, it will short the large current into the breaker, opening the total circuit and protecting it from the over voltage. This will also not require such a large MOV.

Edit: Just a theory right now BTW. It would have to be a fast breaker.

 

[()blivion]

Something like this....
https://www.digikey.com/product-detail/en/ERZ-V20D470/P7319-ND/227579

Combined with this....
**broken link removed**

Like this...
View attachment 66504

So, normally the MOV is open, and does not conduct. When an over voltage is present, the MOV trips, becoming a short circuit, bypassing the dummy load. This dumps all the powersupply's current straight into the breaker, causing it to pop (hopefully before the MOV does) opening the circuit and removing any voltage, over voltage or otherwise.

Edit: I'd like to stay, but I gota go do stuff. I'll see you tomorrow or something.

 

[jocanon]

Hm, it's getting a little up there in price, not too high, but high enough to make me wonder how likely it is that I will need it. I wonder what the probability is, obviously that would be impossible to know for sure.

 

[ronv]

Just my 2 cents.

I think there is already a fuse in the form of the .13 ohm resistor. If the FET shorts either the supply will shut down or more likely the 325 watts through the .13 ohm 5 watt resistor will force it open.
In normal operation the adjustment will only allow it to go to 6 amps max so no problem there. (current limit)
Only a few things that I would worry about.
1- You drop something on the pipe that shorts it to ground. In theory the DUT should shut down. If you had a fuse it probably wouldn't blow because the suppy couldn't supply enough current to blow it.
2- If you ever forget it is on and just pull the plug on the supply the arc will probably blow stuff up. A single MOV or TVS on the load box from +24 to ground might help.
3- Could be the FET shorts to the gate. A 750 ohm resistor between the op amp and the gate might keep the op amp from blowing. I don't think I would worry about this one.
Debug is probably the greatest risk.

On another note. It might not be to hard to disable the decimal point on the little volt meters if we knew where that pin was.

 

[()blivion]

On another note. It might not be to hard to disable the decimal point on the little volt meters if we knew where that pin was.

Agreed.

 

[jocanon]

If you ever forget it is on and just pull the plug on the supply the arc will probably blow stuff up. A single MOV or TVS on the load box from +24 to ground might help.

I don't know enough about electronics to know what is going on, but there are two things I think I should let you all know about that could effect things:

1. I don't know if this is related to what you are talking about above, but I know from experience if I plug my battery charger into the PSU after the PSU is already turned on I get sparks, so I have to make sure that I connect everything first, then power up.

2. Yesterday the power to my whole house flickered off for a fraction of a second. It was very quick, but enough to shut the computer down. I wonder what problems, if any, this would create for the dummy load if everything was hooked up and powered on at 1200 watts. Maybe nothing, would it just power down and then kick back on without problems (#1 above wouldn't be an issue because everything is still connected). Or do I need to get some sort of battery back up system to protect the dummy load...note: I do not want to do that so I would like to avoid it at all cost unless it is really necessary.

Edit:
Called the power company, the short black out is something they are aware of and they are trying to fix it right now. It happened twice yesterday and once this morning. We've lived here for a couple years and this is the first time it's happened, so I don't think it is that frequent.

 

[jocanon]

1. I don't know if this is related to what you are talking about above, but I know from experience if I plug my battery charger into the PSU after the PSU is already turned on I get sparks, so I have to make sure that I connect everything first, then power up.

I don't know how feasible it would be to make it so that it doesn't spark when you connect a load after turning it on. If this could be done it would be one less warning I would have to attach with each PSU.

 

[ronv]

We could probably fix the sparks when you plug it in, but I'm not sure how to prevent them if you unplug it with the load on. The transorb or transient supressor may protect it. You would still get the sparks but nothing would blow. (at least in the load box)
Have you ever unplugged the charger while it was charging?

 

[jocanon]

Have you ever unplugged the charger while it was charging?

No, I have not ever done that.

 

[jocanon]

We could probably fix the sparks when you plug it in

That would be really cool if we could do that with a relatively simple cost effective solution that I could replicate for each PSU. There are a few guys selling these in the RC community, but I do not believe anyone has found a solution to that problem yet, at least not in versions that are up for sale.

 

[jocanon]

I found this short discussion on it:

https://www.helifreak.com/archive/index.php/t-346137.html

Looks like it's the capacitors charging up when you first connect the charger that creates a surge/spark and the PSU is designed to shutdown if the surge is too large.

 

[()blivion]

The input filter capacitors in your Li-Po charger are empty when it's not connected to anything. When you connect them, the empty capacitor needs to be filled, the PSU sees this "filling up" process as a short circuit momentarily. However, PSU's usually have an output capacitor, so your basically hooking one charged capacitor to another empty one. Most all of the PSU's that I have seen won't care about this current wise because they don't have current sense after the output filter capacitor. So I would not suspect the over current to get tripped from something like that.

I do however see a lot of computer PSU's with comparator sensed under/over voltage shutoffs that are attached to these output caps. So it's more likely that the under voltage is getting tripped when the caps charge sags. In order to be tripping over current protection, your chargers capacitors would have to first drain away the PSU's output capacitor, then draw enough current from the inductor for a large enough current to be seen up by the over current sense. Server PSU's are completely different animals though, as well as the so called "multi-rail" PSU's. Both these could easily put current sense as the first thing right after the wires.

Also note that those sparks will not happen when you attach our load circuit because it does not have an input capacitance like your chargers do.

I think what ronv is talking about is the fact that when you remove the leads to the load while it's running the high current going through the stray inductance of the rest of the circuit (cables mostly) will have no where to go and will thus cause a high voltage spike from "inductive kick". And as he said "The transorb or transient supressor may protect it." Or we could devise a snubber circuit that should work well for such things.

 

[jocanon]

Just when you think you are done...haha, no, in all seriousness, I would rather catch this now then after I find out the hard way.

Edit:
Btw - I hope my above comment does sound ungrateful...you guys, ()blivion & ronv, are awesome!

 

[()blivion]

Btw - I hope my above comment does sound ungratefule...you guys, ()blivion & ronv, are awesome!

No problems, Not at all ungrateful sounding to my eyes. Even if it was, one unwritten rule of Internet forms is "that which is typed is not always what it appears to be." A lot of sentences can turn quite nasty with just the flip of a single word, or even a letter. And also the beat or flow of a sentence can be totally altered with things like punctuation.

For example.

Q: Is this plan of mine good?

A: "Your idea is, sucky ideas should be obvious."
*OR*
A: "Your idea is sucky, ideas should be obvious."

Two totally different messages, But the smallest of difference in them. When you get used to these kinds of mistake the internet doesn't seem like such a disrespectful place.

But it's certainly good to assume the worst impression and apologize right away. Makes you look like an awesome person.

 

[()blivion]

So how far are you on this project? Can we expect some pictures of what looks like a circuit board soon? It's OK if not, taking your time is the mark of a true expert.

 

[jocanon]

taking your time is the mark of a true expert.

Then I must be a true expert I was hoping to make a lot more progress on it over the weekend, but I ended up having to work on tax returns all day Saturday and I didn't have much time to do anything on Sunday. I hope it will be done soon. I want to draw out, or at least have a really good idea in my head, of how I am going to put everything on the PCB before I start soldering in order to minimize mistakes/rework. So, I see that as the next step, maybe I will find some time to do that tonight.

The other thing I need to figure out is wire gauge...I think I may already have this figured out, but just wanted to draw it all out and be sure. I think I will use some 10 gauge silcone wire I got from **broken link removed**, claims it can handle 150 amps continuous current, for the wire from the PSU leads to the dummy load. I, and many others, have already tested this wire under high loads and it holds up. I am thinking some 18 gauge wire I took out of a CPU power cable will work for the connection to and from the MOSFETs that will be carrying up to 6 amps. Then for all the other connections/jumpers, I think I could use 22 gauge, or something like that, very small wire since there is a very small amount of current running through the logic circuits. One part I wonder a little about is the wire from the Op-Amp output to the MOSFET gate. I don't think this is carrying anything over .5 amps either (probably much less), but not sure. Can I use small 22 gauge wire there as well?

 

[()blivion]

One part I wonder a little about is the wire from the Op-Amp output to the MOSFET gate. I don't think this is carrying anything over .5 amps either, but not sure. Can I use small 22 gauge wire there as well?

If this were any other project involving FET's, it could be drawing a decent amount of current. With this circuit though.... 22 gauge is perfectly fine.

We are actually going to be drawing something like less than 0.00001 Amps per FET gate.

 

[jocanon]

Cool, thanks.

 

[ronv]

I should have read your link. I thought you were talking about the load box.
Anyway, something like this would probably work for the supply tripping from the undervoltage.
Still a big spark but not quite as big and the supply should stay up.
Maybe we should probably add a power on LED to the box?

https://www.ametherm.com/datasheets/ms350r550.html

 

[jocanon]

So that part you linked to, would that go in the server PSU between the positive and negative terminals to prevent it from shutting down and reduce spark when connecting a charger? And the LED are you saying to put that in the dummy load, or were you referring to the PSU? If the dummy load, I was going to try to use a blue light that is already installed in the PC case I bought for this project. I think I can just plug into that for a power on light.