BK Precision 1670A power supply repair help

[Robomeds]

Hey all, I was hoping the members here might have some suggestions and/or advice for me. 

I'm attempting to repair a BK Precision 1670A power supply that was in the recycle bin at work.  Here is the BK site with the model:
http://www.bkprecision.com/products/power-supplies/1670A-triple-output-30v-3a-digital-display-dc-power-supply.html

The problem is the variable voltage output is basically pegged at 52 volts (this is ~1V less than the voltage across the bridge rectifier). 

I've been unable to find a schematic for this supply but this forum has already come through with a schematic for the BK 1670 (no A) supply.  The circuits are very similar. 
https://www.eevblog.com/forum/chat/bk-precision-good-customer-service/msg122755/#msg122755
Thanks to MikeK who posted the schematic in his post last year!

Looking at the schematic here is what I've been able to tell.  Please note I will refer to the parts on the 1670 schematic because they are similar to mine. 

1.  The output electrolytic (C5) on the supply is bad.  This isn't suprising as it was subject to 52V and was rated at 35.  I assume it's just a buffer and for test purposes I should be able to run the supply without it (it has been removed).

2.  The op amps are getting ~18V across their power rails (IC1A.B)

3.  The voltage control pot (VR2) seems to have failed open.  As a test I have bypassed it with a 1k resistor.  Without the resistor I was seeing the full rail voltage (53V) across VR2.  With it I was seeing 16V.  I don't think there is anything too significant about that given the current path from the + output through VR2, R14 and D9 to the negative output. 

4.  Unlike the schematic, this circuit has only one 2N3055 transistor (TR1, no TR2 or associated R2). 

Basically what I can tell is TR1 is being left at, for lack of a better term, full open.  Neither of the op amps are shutting the thing down.  With all that said, does anyone see any obvious thing I should check (physically everything looks good).  What might be the common suspects?  TR3?  TR1? 

Thanks!

[Robomeds]

OK, I have a bit more information.  In the schematic there is an NPN transistor, a TIP32C that appears to be working with the main current control transistor in a Darlington type arrangement.  I pulled it and found that it seems to have been fried.  It fails a diode check with basically a short between all the pins.  It has a slight bluing to the metal tab suggesting it got hot.  Any idea why the transistor that controls the power transistor might have cocked itself?  Is this a common thing?

[wraper]

If it isn't mounted on heatsink it can be very hot because it dissipates significant amount of power. I have seen some another PSUs with same failure. Some of those died when output was shorted with voltage set near to max. Before schematic starts limiting output current, through this transistor may flow current high enough to damage it. You should check that current limiting works properly. That also can be the cause.

[amspire]

Wrapper is right about the heat in the TIP32C. If it is not on a good heatsink, it is probably the most stressed component in the supply. It probably has to dissipate something like 3W maximum  and that is pushing it for a small clip-on heatsink. What kind of heatsink does it have?

Whatever it has, it would be a good idea to find a way to give it a bigger heatsink. It sounds like at 3A out, this transistor has been running very hot.

If the relay contacts have stuck or the relay control circuit (Tr6 and TR7) have failed, then that would almost double the power to the TIP32C and the output transistors. After replacing the TIP32C, before I loaded the supply, I would check that the relay is working correctly. I do not know the switching point, but when the output is below a certain voltage - such as 12V, it should switch to the lower voltage winding and so the voltage on C31 will drop to possibly half.

If the relay contacts have stuck, then a good tap on the relay might free the contacts. If the problem has been caused by a relay issue, then I still think a better heatsink for the TIP32C is a good idea as it would mean that even if the same problem happens again, the TIP32C will not overheat. With a stuck relay, the output transistors will become very hot at full current and a low voltage output, but this will take some time, and hopefully next time you will notice the heatsink getting very hot.

[Robomeds]

The tip32c had no heatsink as delivered.  I think I could get one and add it to the setup.  amspire, the TIP32C was used in a Darlington like setup so it shouldn't ever see 3 amps.  That said it does see the current needed to switch the main transistor as it handles 3 amps.  I hope a thermal failure was the original problem.  I do wonder if perhaps the 50k variable resistor was the original failure.  At 50k even with a 50V potential across it, it shouldn't have an overheating issue.  That works out to be .05W on a part rated for .15W.  I pulled the thing apart and the resistive track failed very near the terminal connection.  I believe 50K ohm corresponded to 30V so an open circuit (well 470kohm thanks to the resistor in parallel) might have the power supply trying to drive rail voltage to the output. 

I can hear the relay click when the system is turned on so I'm not worried about it.  That doesn't mean TR6 and TR7 haven't failed.  I'll do some diode checks on those guys tomorrow.  In the mean time I ordered a new transistor and pot.  I can see the failure of the pot being perhaps a root cause.  I'm less sure about the transistor.  I would assume BK (or the ODM) was smart enough to size it correctly on this rather common supply.  Then again I don't know what happened to this supply when it went bad in the first place. 

Thanks for the thoughts!

[MikeK]

I got that schematic by contacting BK.  You might try the same, they were very helpful.  Tell them how much you like their products.

[amspire]

The tip32c had no heatsink as delivered.  I think I could get one and add it to the setup.  amspire, the TIP32C was used in a Darlington like setup so it shouldn't ever see 3 amps.  That said it does see the current needed to switch the main transistor as it handles 3 amps.  I hope a thermal failure was the original problem.  I do wonder if perhaps the 50k variable resistor was the original failure.  At 50k even with a 50V potential across it, it shouldn't have an overheating issue.  That works out to be .05W on a part rated for .15W.  I pulled the thing apart and the resistive track failed very near the terminal connection.  I believe 50K ohm corresponded to 30V so an open circuit (well 470kohm thanks to the resistor in parallel) might have the power supply trying to drive rail voltage to the output. 

The TIP32C will not get 3A, but the 2N3055 transistors are very low gain, so the current in the TIP32C can be over 100mA. with a 25V drop, that can mean over 2W dissipation, and even 1W is pushing it without a heatsink.

I doubt the pot problem caused the TIP32C to fail. In general power supply driver transistors need a heatsink. I would be surprised if originally, it didn't have a clip-on heatsink.

I can hear the relay click when the system is turned on so I'm not worried about it.  That doesn't mean TR6 and TR7 haven't failed.

Sounds like the relay works, but once you have fixed the other problems, you just need to check it switches as you go from low to high voltages on the output.

[wraper]

When this transistor is already hot it can easily go beyond safe operating when something happens. Max power dissipation without heatsink is 2W at 25 C ambient temp. and it derates when ambient temperature is higher. There is no need to exceed 3A (5A peak) max current to burn it. Shorted output together with hot transistor can make the local hot spots on silicon crystal, and it can make puff. I attached pic with safe operating area of TIP32C. As you can see at DC it is only 1A max at 40V and 2A 5ms pulse, 600mA 50V DC and 1A 5ms pulse. So when capacitor before transistor is charged for example to 50V (no load), and output is shorted, transistor can easily go beyond SOA for a short moment before current limiting kicks in (depends of it's speed). And absence of heatsink just makes it much worse. I highly recommend you to put some heatsink on it.

[SeanB]

I would also recommend replacing the 2N3055 with something slightly better, they are not really rated for 3A at high dissipation, and in a power supply often they will have a gain approaching 1 under certain conditions. I replace them with 2N3773's normally, better SOA and somewhat more robust. As well place 1 33R 0.5W resistor between base and emitter lead when replacing it, it helps with reducing leakage current when hot, and is good on many power supply discrete darlingtons to help turn off the slow output device and reduce leakage current in it.

[Robomeds]

All,
Thanks for the great information!  I can say with confidence the tip32c did not have a heat sink from the factory.  I'm the only person who ever pulled this supply apart and I know it didn't have a heat sink (didn't and shouldn't of course are two different things). 
BTW, the supply described in this thread
https://www.eevblog.com/forum/reviews/%28au%29-get-em-whileif-you-can-0-30v-2-5a-psupply-$49-dicksmith-last-stock!/
is the same as the 1670 in my reference schematic.  One of the pictures shows the TR3 transistor next to the relay (same basic setup in mine) and like mine it has no heat sink.  Again that doesn't mean it shouldn't.

BTW, does anyone know a good resource for reading up on power supply basics?  This is just a trip down curiosity lane here, nothing else. 

Thanks again!

[quantumvolt]

Hi.

Your supply is fairly simple and should not be difficult to fix. That however requires that you do exactly what I suggest so that I can get information that lets me deduce instead of guess. Also you must be a fairly precise in your answers to my questions. If you want to give it a try, I will  explain every step so that you can judge for yourself whether it is logical sound and non destructive to your unit.

The advice you have been given over are imo fairly poor - you are not about to re-spec or improve a standard variable series regulator PSU with current limiter that has been produced like this at least since the op-amp was established as an electronics 'Lego-brick' (at least 40 years ago). You might "improve the design" later if you like - but for now I suppose you want  to repair the unit that is.

---

1.  The output electrolytic (C5) on the supply is bad.  This isn't suprising as it was subject to 52V and was rated at 35.  I assume it's just a buffer and for test purposes I should be able to run the supply without it (it has been removed).

2.  The op amps are getting ~18V across their power rails (IC1A.B))

#This is correct - the voltage is the sum voltage over ZD1 and ZD3#

3.  The voltage control pot (VR2) seems to have failed open.  As a test I have bypassed it with a 1k resistor.  Without the resistor I was seeing the full rail voltage (53V) across VR2.  With it I was seeing 16V.  I don't think there is anything too significant about that given the current path from the + output through VR2, R14 and D9 to the negative output.

#Do you mean D10?#

4.  Unlike the schematic, this circuit has only one 2N3055 transistor (TR1, no TR2 or associated R2).

#This is OK - it only means that your unit is specified for a lower current or has less over-dimensioning of the main series transistor TR1#

Basically what I can tell is TR1 is being left at, for lack of a better term, full open.  Neither of the op amps are shutting the thing down.  With all that said, does anyone see any obvious thing I should check (physically everything looks good).  What might be the common suspects?  TR3?  TR1?

OK, I have a bit more information.  In the schematic there is an NPN transistor, a TIP32C that appears to be working with the main current control transistor in a Darlington type arrangement.  I pulled it and found that it seems to have been fried.  It fails a diode check with basically a short between all the pins.  It has a slight bluing to the metal tab suggesting it got hot.  Any idea why the transistor that controls the power transistor might have cocked itself?  Is this a common thing?

----

Well, if you agree: Replace C5 with any 100 microF or bigger for testing. Replace TR3 with any NPN at hand for testing. Your worst scenario is that TR3 burns. Without C5 for possible stability issues and TR3 for possible working regulation, it is imo no point in measuring anything. And before I know that TR3 burns again, I see no point in searching further. If you go for this, put in 10 kOhms for VR2 and put the Current Limiter knob in the OFF-most position. And no load on the output...

Reason for testing this first: The regulating feedback loop in your PSU is IC1A, TR3 and TR1. In one single step we will know if the loop works at all, or TR3 burns again. If TR3 does not burn, measure the output voltage of the PSU. Then change the value of VR2 by adding another 10 kOhms in parallel with the 10 kOhms that should be there by now (you can put in whatever you have between 5k and 50k), and measure output voltage again. If they are different, report latest Base voltage / Emitter voltage for TR3 and TR1, and voltage on the + and - input and output of IC1A ...

Good luck

[quantumvolt]

BTW, does anyone know a good resource for reading up on power supply basics?  This is just a trip down curiosity lane here, nothing else.

#They are all over the web. Go look! Your PSU is in principle identical to this circuit:



Source: http://www.radio-electronics.com/info/circuits/transistor_current_limiter/transistor_current_limiter.php

I'll leave it as an exercise to identify:

1. The series regulator element

2. The current sense resistor(s)

3. The current limiter device

4. The output voltage sensing / difference amplifier / adjustment loop

5. The reference voltage element


In addition your PSU have:

An automatic High/Low unregulated voltage control circuit with a relay

Voltmeter and Amperemeter with separate PSU's

Separate 5 V and 12 V PSU's

Separate PSU for a Led, IC power, Voltage referecse and power to relay control circuit


That's it ... and I forgot - it is 

[Robomeds]

Thanks quantumvolt!

I've ordered a few replacement parts from Digikey.  By the end of the week I should have the new pot, TR3 and output caps.  I'm curious why TR3 would have failed in the first place.  I don't know if a failure of the voltage control pot would have run the thing too hot.  Since I have no idea what load the amp was running when it failed, it's also possible the thing was simply overloaded in some way. 

Either way, I agree it seems like a really simple circuit and shouldn't be hard to fix. 

PS, yes I mean D10, not D9!

Thanks again!

[CheddarAlan]

I know this is antique stuff, but it is a useful P/S.  It had 60+ volts applied to the O/P.  I found a weirdo as well as the expected Txs and OP amp, which led me a dance.  The current limit red led was leaky in reverse which gave the same original external symptoms of 40+ volts at the terminals.

It might help someone.

Alan D.