Help with Water Pump

[salty joe]

Thanks Alec.
At Supply+1 I got 20.9, at Supply+2 I got 14.9 and at +12.6 I got 14.2.

The PS connected to Supply+2 apparantly failed and is good only for 15V. The V adjustment no longer works. Maybe that's where the faint crackle sound and puff of smoke came from when I accidently shorted the battery for a split second.

I show 20.6 at +Vext. Is that a good number? Seems awful high.

 

[KeepItSimpleStupid]

@Joe
Take an AC voltage measurement of supply #2.

 

[salty joe]

Hi KISS. Glad to hear from you.

I disconnected the old power supplies from my circuitry and now they both show the normal range of voltage adjustment. If I'm going to trust my filter with these power supplies, that's very good news. Both of the old pumps I have been working with seem to be just fine, so that's really good too. I tested both power supplies with the AC mode of my DMM and they both showed around 42. That would be volts I guess.

Evidently, my circuitry is wrong. I started off on this last leg thinking I had plenty of room so I spread things out on the boards in a nice leisurely way. Then we added a couple things, made some changes and pretty soon I had stuff on one board that would have been better on another board. I am leaning hard towards a do over. None of these last circuits are as hard as the PDMs. It's just that I've gotten them into a jumble. Plus I am almost certain that the little bitty crackle I heard and puff of smoke I saw came from somewhere on the Filter Pump Power Backup circuit. That makes me kind of itchy.
Anyway, I really do appreciate all the help you guys have given me. Thanks.

 

[ronv]

Sounds like a short somewhere or the 317 melted and the supply is trying to charge the battery with really high current. Check in to out on the 317 to see if it is shorted. be really careful now It could be the only thing keeping the +12 from going to +20 is the battery.

 

[KeepItSimpleStupid]

Yup. Sounds like the supply is current limiting. The 42 should me millivolts. Volts would be no good. The test was to ATTEMPT to look at the output ripple of the power supply. Very high ripple meaning bad sully. For switching supplies ripple is a lot higher, 100-300 mV p-p and usually a high frequency which your meter would not respond to.

 

[alec_t]

I show 20.6 at +Vext. Is that a good number? Seems awful high.

That's fine. Spot on.

I disconnected the old power supplies from my circuitry and now they both show the normal range of voltage adjustment.

Good news. So PS2 shut itself down because of overload and has now recovered. I agree with ronv that the '317 probably died when the battery was shorted. It was likely the cause of the overload. I'd be inclined to swap it for one of your spares rather than test it in situ and risk further damage.

 

[salty joe]

Hi ronv, I'm glad you're still here, thanks.

OK, before I scrap it I'll replace the LM317 and see where that puts us.

I thought 20.6V at +Vext going to pin 4 of the LM324 was way high, thanks for letting me know it's good.

Thanks all.

 

[alec_t]

OK, before I scrap it I'll replace the LM317 and see where that puts us.

Of course if the 317 died because of a short which is still present somwhere then your replacement 317 will probably die too . So perhaps it would be better to start over if you've got all the ingredients.

I thought 20.6V at +Vext going to pin 4 of the LM324 was way high

Nah, the little LM324 can take a supply voltage up to 32V.

 

[salty joe]

I corrected an error at LM324 and replaced LM317. I compared resistance between legs in the old and new LM317 and they were not the same. I think it’s safe to say the old LM317 was damaged.

I ran the pumps on the battery for 40 minutes and plugged the PSs back in. After about 5 minutes, the charging indicator LED started to flicker. LM317 was barely warm. After a few more minutes, the LED became more solid. At that point there was 12.69V at the battery terminals. 10-15 minutes later the LED was on solid and there was 12.76V at the terminals. LM317 was pretty warm. It’s been about 30-35 minutes and there is 12.97V at the terminals. 15 minutes later and we are up to 13.2 at the terminals. The heat sink at LM317 seems to be doing the job. If this is normal, I think we might be good to go. I’m shutting everything down for now because I don’t want to let this run without a babysitter quite yet.

We still seem to have an issue with one of the old PSs.
Thanks one and all.

 

[KeepItSimpleStupid]

General warning:

You can''t trust in-ircuit resistance readings. Close to zero in both directions - probably.

For the LM317T, the key troubleshooting points are:

~ 1.2 V from Vout to Vadj and around 3V from Vout to Vin.

Vin ~ 3 V greater than Vout.

 

[alec_t]

I compared resistance between legs in the old and new LM317 and they were not the same.

That's a valid comparison if both 317's were out of circuit when you measured them.

The heat sink at LM317 seems to be doing the job. If this is normal, I think we might be good to go.

Your measurements of the charging progress are a good sign that the charging circuit is now healthier. I would have expected the green LED to have been on right from the start, though. Perhaps the trimmer needs a tweak for that 'Pin13 - Pin12 = 50mV' setting. "Pretty warm" sounds good for the heatsink: "very hot" would have justified extending it. So the old PS is still sulking?

 

[salty joe]

Thanks KISS, I'll check the new LM317 later.
The comparison I made between the old and new LM317s was with both disconnected on the bench.
I'll jump R1, disco the battery and adjust the trimmer. Thanks Alec.
Sulk is a good way to put it for that PS. It seems fine until placed under a load.

 

[salty joe]

Perhaps the trimmer needs a tweak for that 'Pin13 - Pin12 = 50mV' setting.

Good call. The green LED came on solid from the getgo. There was 12.76V at the battery terminals, after 5-10 minutes it rose to 12.91V.

From my previous notes on LM317, I have ADJ pin 1, OUT pin 2 and IN pin 3.
With battery being charged and pumps running;
1.24V between pins 1 and 2.
7.23V between pins 2 and 3.
13.01V at pin 2.
20.3V at pin 3.

Not all the numbers you were looking for KISS, but hopefully the battery being charged changes things.
Everything is shut back down.
Thanks a million!

 

[alec_t]

Those numbers look good. You're getting there, Joe!

With battery being charged and pumps running

So PS2 was working ok again? The volts at Pump+1 and Pump+2 should be very similar if it is (assuming both PSs are putting out the same volts). If not similar, switch off, disconnect the battery and check that all four Schottky diodes (D2 x 2, D3, D4) feeding the Pump+1 and Pump+2 terminals are good. Even in circuit their forward and reverse readings with your DMM should show if one has failed. A failed-short diode could result in PS2 trying to charge the battery directly via the short circuit, which would pull down the PS2 volts.

 

[KeepItSimpleStupid]

Aside:
I killed one supply of mine by backfeeding. That;s wht I suggested the reversed biased diode across the supply, Not sure you have one there. Now with batteries involved, it's possible here.

 

[salty joe]

I checked all 4 diodes on the diode setting with 3 showing ~ 258 and 1 showing 175. I was going to scavange one from my junkpile, but it also showed 175. It turns I have 2 different brands. They were all sold as 7.5A Schotkey diodes. I hope that's what they are.

I set the shifty PS to 20.6V and connected a pump directly to it (after disconnecting all else) and the volts dropped to 16 or so with virtually no adjustment. Got the same result with only the second pump hooked up. With nothing hooked up, the lowest voltage is a little better than 19. There is a power indicator LED on the PSs and when the problem PS is under a load, it flickers.
I guess it's time to shop for another PS. As far as I know, there is not a reverse biased diode.

 

[KeepItSimpleStupid]

258 mV means Schotkey.

You can fit the power supply witth a reverse boased diode at the terminals. Anode to - cathode to +.

If you want a little more protection on the power supply used for charging, fit a high current diode (The rating of the supply) in series with the (+) output and put the + sense lead to the anode of the diode. So, it's the cathode to the power supply, + sense to the anode and draw power from the anode/sense junction.

This modification is commonly suggested by big supply manufactures when they can be subjected to reverse voltage.

 

[alec_t]

Looks like the diodes are all ok then.

I set the shifty PS to 20.6V and connected a pump directly to it (after disconnecting all else) and the volts dropped to 16 or so with virtually no adjustment. Got the same result with only the second pump hooked up.

A good test. That's pretty conclusive; the PS indeed seems poorly . Is it openable to attempt surgery?

@KISS
There's already a reverse-biased Schottky (D7) across the '317 (following your suggestion ages ago), and a diode in series with the + output.

 

[salty joe]

But if I open the PS, the warranty is void.
Looks like a piece of cake to open. Sounds like fun, thanks. I'll break the seal on it tonight. The problem might be getting my computer to accept an image. If we need to get a pic up, should not be too hard to find a way.
Well, it's nice your good idea is in there KISS. I'm just trying to follow the plan.
Thanks guys.

 

[alec_t]

But if I open the PS, the warranty is void.

Yeah, right. Seriously, if it's still under warranty can't you claim? It should have had overload protection.

The problem might be getting my computer to accept an image.

You've sent pics before (BTW gif or png is better than jpg). Do you have some new problem with the PC?