Faulty DSE D-3800 Power Supply, similar to Manson EP-925

[Raff]

Hi Everyone,

I have been given a faulty Dick Smith D-3800 power supply.

The fault is that at maximum output setting, the voltage at the output terminals is around 3.5 volts.

I have tried the following:

Replace the LM339
Replace the LM723
Isolate the shutdown by lifting one end of D12
Check the TIP31C that drives the 5 2N3055 output transistors, it is fine.
Check all emitter resistors for the 2N3055s
Check voltages on both the LM723 and LM339
The supply voltage for the LM723 is fine at 24 odd volts, the 15 volt zener is also fine hence the 15 volts to the LM339 is fine.
The output at pin 10 of the LM723 is the same as the output for the whole PSU minus about 0.6 volts across the output transistors.

The fault remains.

Oddly with a replacement UA339 IC the "overload" LED remains lit all the time, with the original LM339, it turns on at ether the maximum or minimum voltage output setting.


Its doing my head in, as for the life of me I cant work out just how its meant to work in the first place as the supply for the output transistor in the LM723 regulator is the actual output voltage of the entire PSU after the series pass transistors, not before. It seems to my tired old brain that its a chicken and egg situation. 

Can someone please cast an eye over the attached schematic?

Thanks,
Raff.

[Raff]

Sorry had a brain fart.

The output of the PSU is not the supply for the regulators output transistor, its the inverting input for the comparator, the non inverting input has the reference voltage divided down to 2 volts on it.

Damn brain 

[Armadillo]

So, you managed to get it to work?

[Raff]

No sadly it doesn't work, and I cant for the life of me work out why. 

[Armadillo]

No sadly it doesn't work, and I cant for the life of me work out why. 

While all the components are in the circuit - power it on and can you measure the voltage at the points as shown in the attached with respect to the common ground output indicated as output "-"

[Raff]

Ok will do tomorrow.

Thanks for the input.

Raff

[wilfred]

I was curious about how this PS differed to my DSE Q-1760. I found this during the search. Probably you've seen it already. But maybe not.
https://pa0fri.home.xs4all.nl/Diversen/EP925/ep925eng.htm
and I found this list of voltage measurements from here http://www.ozlogger.net/forum/viewtopic.php?t=103 (see attached pdf)

[Armadillo]

Ok will do tomorrow.

Thanks for the input.

Raff

Additionally, please also measure the voltage at these points;

[Raff]

I was curious about how this PS differed to my DSE Q-1760. I found this during the search. Probably you've seen it already. But maybe not.
https://pa0fri.home.xs4all.nl/Diversen/EP925/ep925eng.htm
and I found this list of voltage measurements from here http://www.ozlogger.net/forum/viewtopic.php?t=103 (see attached pdf)

Thanks for that, yes the Q-1760 sure looks similar.

I did find those pages when I was searching for information on my power supply, they make for interesting reading and that chart of voltages will be handy.

Regards,
Raff.

[Raff]

Ok I have had a chance to do some measurements and observations on the power supply.

Attached are the edited voltage measure points pictures with voltages as measured added, the readings that range from a low to higher value are when I adjusted the voltage adjustment pot from minimum to maximum setting.

I hope the text is legible.

Thanks for looking,

Raff.

[Raff]

I'm starting to look suspiciously at those 5 2N3055 series pass transistors...

[Raff]

Results of last testing for today.

I used an external variable PSU to supply base current to each of the 5 2N3055 series pass transistors in turn, with the D-3800 powered up and the base leads from its TIP31C driver disconnected.

All 5 2N3055s worked fine, I could wind up the base voltage to each, and see the output of the D-3800 rise with it.

Next I tried an outboard TIP31C, this resulted in the same fault as originally.

Next I tried the external TIP31C with the LM339 unplugged (I have installed sockets for this and the LM723), result, same fault.

Very odd indeed, so I tried replacing C10 a ceramic cap that's mentioned as a possible cause of this sort of fault, no go, same problem.

I even put the LM723 into a breadboard and tested it, it works fine.

That's it for today, I'm too sore, cranky, and tired to tinker until tomorrow. 

Raff.

[Armadillo]

Hi Raff;

With your measurement results, here are the possible faults;

Firstly ensure that Pin 11 of LM723 is of the correct voltage. It should be 26.3 volts. [YES / NO]

Then measure the voltage at Pin 10.  if you varies the potentiometer, the voltage should be able to go up greater than 20 volts. If you cannot achieve that, then Change LM723 for defective output transistor in the LM723.  [YES / NO]

After the above is done and if found correct; the remaining suspected faults are;

1.0 The line marked in RED from Pin 10 of LM723 to the base of the Driver Transistor is suspected. Check the Resistor value [looks like 22 ohms] along the line and check the continuity of the line also, from pin 10 to the resistor and from the resistor to the base of the TIP transistor. [YES / NO]

2.0 The capacitor C17 could be shorted or check ESR value. Take it out from the circuit. [YES / NO]

Let us know the good news.

[WaveyDipole]

Since you have already swapped the 723 and you have indicated that you are getting 15.6v at the Zener Z01, then my suspicion would be an open R8 or a break in the track between pin 10 of the 723 and R8 or R8 and the TIP31C (basically something along the red line indicated). It is only necessary to test C25 for a short and it does not need to be removed for that, but please follow all the steps outlined by Armadillo. It is conceivable that a shorted or badly leaky (electrically) C17 could have caused the diode at J1 [sic] to fail open although I would expect such a failure to be evident and possibly for the 2A fuse at F1 to blow.

[Raff]

Thanks for the input and suggestions so far guys / gals.

This thing is haunted, I'm almost sure of it. 

I have checked and measured all of the voltages and resistances as suggested.

All are fine, except the pin 10 output from the LM723 and pins 2 and 3 which monitor the voltage across the current sense resistor.

The tracks to the TIP31c driver etc are fine too.

Tested C25, no problem there so replaced anyway, J1 diode also OK, replaced anyway as I have a pretty good stock of bits acquired from many hamfests 

What I have noticed is that the fault varies, in that the output voltage will occasionally be able to be adjusted up into maybe the 9 volt range, and that its also unstable, which makes me think something is breaking down under load.

I decided to measure across the output terminals, don't ask me why I hadn't thought of this before.

It reads 245 ohms either polarity, hmm, that doesn't look right to me.

Looking at the circuit, there is nothing that accounts for that resistance, I even have the LM723 out of its socket when I measured.

I'm off to the shed later so I will report on my findings later.

I'm retired due to Ill health so can only tinker when things are going my way a bit 

[Raff]

Results of last tests attached, many thanks.

[Raff]

Gees, just noticed my bad spelling, sorry 'bout that, brain not fully functional today 

[Armadillo]

Measure and confirm Pin 5 of Lm723 is 1.99Volts and stable. If you cannot get it, check R14 and R15.

If you can get it, Let's double confirm pin 2 and pin 3 is not the problem. Bend pin 2 leg out of the circuit and plug LM723 back into the socket. Confirm Pin 10 can varies voltage upwards and stable. If cannot, replace LM723 and put back pin 2 into the socket.

Let us know your findings.
 

[Raff]

Results from latest tests and tinkering.

Ive got it to put out the correct regulated voltage so far with the LM339 out of circuit, if I put it back in, the protection circuit cuts in and drops the output to near zero.

I found that the TIP31C whilst testing out of circuit tested fine, was faulty, possibly breaking down or going low gain under load.

The terminals from the bridge rectifier to the filter cap were loose, so I soldered them on.

The LM723 I used as a replacement for the one in the power supply was also faulty, another new can style one works.

So I'm off to Jaycar tomorrow for a couple of LM723s and LM339s, handy things to have as spares anyway 

Tired and sore again so more maybe tomorrow.

Raff.

[Raff]

Update.

Its working!

Thanks to everybody who took the time to help.

Went to Jaycar, got a new LM723 and LM339 as well as some other stuff, got home and found that the bloke who grabbed the LM723s from the drawers behind the counter at Jaycar is dyslexic or blind, I have 2 4011 CMOS ICs instead, oh well.

Put the can style LM723 I have lying around back in and also a new LM339 from Jaycar, its branded as a TI part.

Power supply now works. All I can say is that a uA339 isn't quite the same as an LM2339.

Last job was to sort out the front panel meters, both of which didn't work, they had really tight mechanisms in them so I dissolved the gunk on the set screw and loosened it slightly and now they both work.

Raff

[Armadillo]

  Good to hear it;

So the receipt is a 4011 or 723? LOL

[Raff]

Good to hear it;

So the receipt is a 4011 or 723? LOL

It says LM723 on the receipt, oh well, Ill go back, lucky its only about 4 kilometres from home. 

[Raff]

Went to Jaycar, got a new LM723 and LM339 as well as some other stuff, got home and found that the bloke who grabbed the LM723s from the drawers behind the counter at Jaycar is dyslexic or blind, I have 2 4011 CMOS ICs instead, oh well.

The fault is yours for not checking at the counter that you have the correct parts. Well that's what I tell myself in similar circumstances.

The longer the drive home the greater the chance you have incorrect parts.

I normally do have a quick check of the parts I buy as its usually me that forgets that all important part I really went to get, not the other stuff I usually walk out with. 

[WaveyDipole]

So it seems that multiple problems were at play which always makes any troubleshooting interesting to say the least. Glad you got it sorted in the end.

[Steamer61]

Hi Folks - thanks for a really useful conversation. I'd like to ask what may be a dumb question - how do I get the board out of the D-3800? Looks like the meters are inaccessibly screwed to the front panel, so I need to unsolder the meters in order to remove the board. Is that right? I need to get at the LM723.