1983 Oltronix Power Supply - Died despite current limiting

[baltersice]

Hi!
I thought I could safely short this psu but I only got some magic smoke 
I have attached the schematic. I found R2 and R15 burnt and D12 shorted.
I replaced those parts and and the double op-amp for good measure, now I have a big DC offset: approx 12V Output when 0V is set.

Now for troubleshooting I am having difficulties understanding the TL431 (D11) Voltage reference. In its configuration it acts like a 2.5V zener diode but since they connected the positive output voltage to the anode ?? nothing makes sense. Wont the output voltage (>70V) be handed all the way down to the 7805 output?

[timb]

The TL741 is biasing the inverting input of that LM358 with 2.5V. It compares that with a divided down portion of the output voltage for CV operation. The TL741 itself is biased via 5V from the 7805. The 7805 is powered from a separate tap of the transformer, so it's completely floating with regards to the other two taps that deliver the 0-70V output.

What this means is that the TL431 floats on the positive output of the supply, meaning the inverting input of that opamp will be -2.5V with respect to the negative output of the supply, which is what the opamp's non-inverting input is sensing. Hopefully that makes sense?

[Kleinstein]

Very interesting type of circuit - I like it: rather low parts count and still good performance.

Its a supply with a floating regulator. So the OPs and regulating part is powered from the upper smaller transformer and 5 V regulator. So for much of the circuit the positive output (here left of the shunt) is the reference point. This is a little more difficult to understand, but a good type of design (HP used is in a lot of supplies).

So normally the diode D12,D13 will see essentially nor current. I would guess there is something else brunt, not just only the 2 resistors. So I would check the power transistors (Q3) and the TL431 first. It's also possible P6 is damaged.

The feedback path with R15,R53 and P6 is rather low impedance and thus quite some current can flow here if a small voltage is called for and more is present. The divider regulator for the voltage uses the upper side of the TL431 as a reference point.

[EPTech]

Hi There,

The TL431 is a precision shunt reference and indeed, in this configuration it acts as a precision 2.5V reference.
The 5V supply is referenced to the positive output through R1. If the other resistors have failed, this one may have failed as well. The 5V supply is superimposed on the positive main supply, otherwise you will not be able to regulate down to 0V. The voltage regulator (top right opamp) will regulate the positive output of your supply in such a way that the voltage on it's negative terminal (= positive out + 5Vfixed -2.5Vref) is exactly the same as the voltage on it's positive terminal (=voltage set by potentiometer in voltage divider between positive out + 5Vfixed and negative out).

I do not know why D12 has shorted though since it is in a high impedance path, unless you shorted the supply over a high inductance load. In that case I would recommend adding a snubber or MOV on the output once you get it repaired. Warning: such a protection device may cause the supply to respond slower.

[baltersice]

Hey, thanks a lot, guys!
I missed that floating regulator part and already thought who'd put 70V into an op-amp v+

By pure chance I found that the CC shunt was gone as well:


Replaced everything, tried CV worked perfectly, but in CC everything burned again
Gonna report when I find whats causing that ...

[EPTech]

Sorry to hear about that. My best guess is that D7 is defective or that there is something wrong with the current pot.
 Try loading the supply with a power resistor instead of dead shorting it.

[baltersice]

Got it: There are 2 ranges on the 70V Channel: 0-30V 1.2A and 0-65V 0.6A.
Q1 is used to offset the output of the CC potentiometer P5 to accommodate the 2 ranges (if I understand it correctly).
And after a replacement of that part I don't get any more magic smoke even with a direct short.

Thanks for your help

[EPTech]

Yes, that is exactly what it does and that would have been my last guess. But anyway, congrats with the fix.