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

Status
Not open for further replies.
Yeah, I need to get back to that next. It does go all the way on when the PSU is run at 47 amps, I want to be sure it does go full on UNTIL 47 amps. I have to take the night off the rest of tonight though.
 
Let me know if you think this is a problem...see the video in the link (sorry it's sideways). The volatge is fluxuating a little bit. Aftet I recorded the below, I attached the same PSU to a lipo charger and loaded it up to about 30 amps and the voltage did not fluxuate at all, even with the little voltmeter, so I think it is something to do with the dummy load itself that's causing it. Maybe the little bit of fluxuation is OK though? Also note, when I load it up on the dummy load, I noticed that the TVS was getting warm...not hot, but warm.

https://www.youtube.com/watch?v=Tlf1JVfkfdg
 
Hmm. Might be some noise or something.

Lets try disconnecting the wire between R4 and R6 and see what happens. Keep in mind when you do this there is no over temperature protection.

https://www.electro-tech-online.com...fan-load-bank-pwm-circuit.128753/#post1074167

Does your amp meter go as high as where the voltage starts to fluctuate? If so you could stick it in series with the + lead and see if the current also varies. The load box should try to keep the current constant no matter what happens to the voltage.

One other thing.... Try measuring the 24 volts with the meter on AC - fairly low scale.
 
The voltage is a little high to start with. I wonder why... What TVS do you have in there?
 
I need to go out for a bit, but one other thing to try.
Tie pin 7 (+sense) of the PSU to the +12 output of the same supply and pin 9 to ground on its own supply.
I think this will get you better regulation and closer to 12 volts nominal.
 
I am using the original TVS we selected at Digikey, I ordered more. I already tried connecting the current sense pins and the voltage still jumped around. I am doing some work on my plane right now but will try your other suggestions soon. So I take it by your suggestions that you so think it is a serious problem that needs to be fixed, or can I ignore it?
 
The voltage is a little high to start with. I wonder why...

The ATX PSU specification allows for a ±5% deviation from nameplate voltage on the +12 lines. I have NO idea if server PSU's follow the ATX specs or not though. Could easily be tighter tolerances considering they are intended for more mission critical systems. Your call ronv.

Fluctuating Voltages.
Do we/should we have decoupling caps on the OP-Amps power pins? Without I could easily see getting oscillation that would cause these symptoms. Definitely need to do the current and AC voltage test's and see if it's a voltage meter artifact or genuine fluctuations. Then we can go from there.

Broken PSU
This turned out to be the problem after all eh? Good to know, I guess I can't be right all the time . Anyway, if you want jocanon you can send the dead unit to me and I can diagnose/repair it. I fixed most of the ones I broke Let me know and I will PM you my address.

This load bank can certainly be a PSU breaking tool, but only because it holds the DUT to it's nameplate ratings. IMHO PSU's are almost never put out the power they claim to be able to. But I am talking about computer PSU's, which is heavily tuned toward consumer marketing and not actual facts.
 
I don't know as it is a problem. The thing that bothers me the most is the warm TVS. It sees zero current unless the voltage is over 26 volts. That makes me wonder if the ripple (AC) is pushing it over. Maybe we should have gotten 28 volt TVS's. Anyway I thought I read in the RC forum that using the sense pins made the voltage a little lower. That was really what I was looking for.
 
Well, I guess its a good thing that the temp sensor isn't turning on and off from noise. I can see from the video that the current is just +/- the last digit so that seems good.
I'm beginning to think the nominal voltage is just to close to the TVS voltage.

Posts are passing in the night.
 

Well, now I am not so sure it was a broken PSU. I just hooked up the "broken" one and it is working just the same as all the others. The volt meter that was giving me the high readings is still giving strange readings, on the "broken" PSU it says 28 or so volts under full load, on a not broken PSU it says 21 or 22 volts but in both cases the other two volt meters always agree and are what you saw in the YouTube video. My gut is telling me its some noise in the dummy load and one of my volt meters is just not tuned correctly, but what do I know .

I will probably not test each PSU at its max for very long because I don't want it to be a " PSU breaking tool" . Probably test at max for a min or two then turn it down to 35 to 40 amps.

I don't think the voltage is too high, it should be 12.5 per PSU or 25 when the current sense pins are open.
 

I agree, it doesn't seem like the TVS should be getting warm. I can order some 28 volt TVSs. Connecting the current sense pins did lower the voltage a little, but it still fluctuated.
 
See if the TVS is still getting warm. Did the voltage go down to like 24.5?
Last measurement. This may not be a real good one, but the AC voltage as you ramp it up.
The fluctuation is not such a worry, especially since the current is stable - The load box is a current device the power supply is the voltage device. So the box is holding the current steady no matter what the voltage.
 
AC voltage is 112.5 under no load, then at 47amps it's 106.5vac.


I don't have the current sense pins hooked up anymore if that is why you're asking about the TVS, but yes, it is still getting warm. I only hooked up the current sense pins on one PSU and ran the load on 12v, so the TVS didn't get warm then, but it was at 12v and voltage still fluctuated, but I am not worried about it if you are not, I will just get a higher rated TVS.
 
Last edited:
I guess we fooled that meter.
I think I would just jumper the sense pins and give it a go.
You have a few decoupling caps that () was asking about?
 
Hm, I really don't want to jumper the sense pins, because I need a solution that will work for a quick plug in of each PSU I build to test each one, and I will not be jumpering the sense pins on the ones I sell. The guys over at RC Groups have mentioned they don't think jumpering the sense pins would add any tangible benefit for our purposes so nobody does that. If I did, it would be an extra step (take longer to build) and the voltage would be lower. Can I just get a 28 volt TVS and call it good?

We put these on the power inputs to the op-amps:

https://www.mouser.com/Search/Produ...rtualkey59420000virtualkey594-A104M15X7RF5TAA
 
I think you are ok. I think I would try the 28 volt TVS if the 26 volt one is hotter than the leads it is soldered to.

I think power supplies are rated in watts. ()blivion probably knows for sure. So if you have 12 volts at 50 amps you should only expect 46 amps at 13 volts. The added voltage may actually be making things worse depending on voltage of the battery you are charging.
 
I see what you are saying, but for example, the Power Lab 8 (PL8) charger will take up to 32 volts, so 24 (or 25 in our case) volts is actually a limitting factor. In other words, the PL8 won't even use the full 47 amps of the PSU because of the voltage only being 25.
 
You want to know something funny...I just realized tonight that the reason the thermistor wasn't doing anything on the 24v fan control the first time I tried it is that I ordered a 10 ohm thermistor instead of 10K. I thought it was a bad thermistor and switched it out for another one that actually was 10K so it worked, but it wasn't a bad thermistor, just a bad order .

btw - ronv, it says you exceeded your stored PM limit and won't let me send you any more PMs.
 
Status
Not open for further replies.
Cookies are required to use this site. You must accept them to continue using the site. Learn more…