Strange PSU Problem

[Danno]

Hi All,

I am repairing a couple of 30 year old PSU's but have a strange problem with one of them. 3 of them I have repaired boot and run the board fine but one won't boot the board. On the bench all the voltages are within spec with the PSU not under load the logic supply to the board is 5Vdc but as soon as the PSU is hooked up to the board and I check the voltage on the board it is only reading 3.75Vdc. If I swap it for another PSU it boots and runs fine with 5V present on the board so I know it is a PSU related problem. I have fitted new 1723 regulators, fitted new trim pots and new TIP3055's as well as the LM311.

There is circuitry on the board that creates a interrupt if voltage falls below 4.75V hence why it won't boot.

Attached is the setup and schematics for PSU, I ma going around in circles with this and would appreciate a fresh pair of eyes that might spot the obvious, thanks in advance for your help.

The PSU is of Linear design BTW.

 

[Tha fios agaibh]

Have you checked for any cracks on tracers, or bad solder joints?

 

[Martaine2005]

Hi Danno,
On anything that old, the caps could be the problem.
Tack another one to the under side of the board to see if it solves the voltage problem.

Martin

 

[Tha fios agaibh]

Hi Danno,
On anything that old, the caps could be the problem.
Tack another one to the under side of the board to see if it solves the voltage problem.

Martin

Good idea for open caps but not if their shorted.
I'm also wondering if by paralleling another cap to the underside of a good one, the double capacitance could result in blowing up something downstream. ?

 

[Martaine2005]

Judging by the OP's comment about having 5v until the unit is under load (connected to the board),
you could assume a short is unlikely.
More than likely the caps (or one) are failing.
As soon as it's connected to the main board the voltage drops and the under voltage circuit inhibits start up.
Most faulty laptop supplies have the specified voltage at the connector and the LED glows bright. As soon as it's connected to the laptop the LED (sometimes) dims and voltage present can be significantly lower. 98% of the time it's capacitors. I have a bag full of them waiting for new caps (all opened and caps bulging/leaking).
Putting another cap in parallel wont harm anything, they're before the regulators. It's only a quick check to see if voltage holds up.
I should say that I have not had a problem doing this! If it's a concern replace the caps or use an ESR meter.

Martin

 

[Danno]

Thanks all, I have ordered new caps and will replace.

To me it seems like a regulator problem, would I be right in saying that the 723 regulator samples the output voltage and adjusts itself accordingly by I am assuming adjusting the voltage to the 3055 base. I guess this is used like a tap to control voltage. For some reason the regulator isn't regulating the output voltage.

 

[Martaine2005]

Hi Danno,
If you think it could be the MC1723 as it does control the voltage and current, have you tested the voltage before the regulator and plugged the board in?

Martin

 

[Danno]

Hi Martin

No not yet, I've changed the 723 but will try what you suggest, could you confirm my theory of how the 723 works is right?

Thanks

 

[Martaine2005]

The datasheet does say that the current can be increased by several amps using pass transistors.

You said you have changed the regulator, so why do you still suspect it?

Martin

 

[Danno]

Sorry maybe not being clear, I am suspecting the regulator circuit in general not the chip itself, my understanding is that if the output voltage drops the 723 senses it and opens the tap a bit more (3055) to correct voltage to 5v, something is stopping the tap from opening and not necessarily the 723, of course this theory could be wrong hence why I'm after guidance from you guys, thanks

 

[Martaine2005]

Hi Danno,
from my reading, the LM311 is operating as the voltage comparator.
This is the chip.

Martin

 

[Tha fios agaibh]

The good news is, you have other power supplies to use as a reference.
I would put a light load (of the same resistance) on both 5v outputs, and compare voltage points from the output backwards.

 

[Danno]

Solved it, thanks for everybody's input.
When I had the PSU apart I found the fault would come and go if the board was flexed slightly, dodgy solder joint I thought. The main smoothing cap for the 5v is remote to the board and connected via a soldered wire, it was the solder joint on the wire to the board had broke away leaving the wire floating in the hole on the board. It makes sense now why the meter read 5v not under load, it was un smoothed DC, when the board was plugged in it didn't like it.

You can see the broken solder joint in the pic which is arrowed.