exxos's Atari, Amiga & retro forum

Hi.

I just recap an ASP34-4 power supply for my Atari 1040 STF (also found on some STEs).
I followed the Exxos page and Exxos values for improvment

https://exxosforum.co.uk/atari/last/psu/#ASP34

C5, C4 : 47uF 350V (33uF 350v replacement)
C15, C14, C7 : 680uF 35V (330uF 35v replacement)
C13 : 4700uF 16V
C12 : 4700uF 16V (2200uF 16V replacement)

I also changed transistors Q2 (BU508) and Q1 (2SC1384).

Here is the schematic of the ASP34-4
After reassembly I tested the output power supplies using a 2.2 Ohm (20W) load for the 5v output and a 10 Ohm (20W) load for the 12V output rail.


Now when reading the output voltages, I obtain:

6.4V instead of the expected 5V
14v instead of the expected 12v

Not being an electronics specialist, could you tell me on which components acting to try to bring the voltage rails back to 5v and 12v?
Thank you for your help.

This is odd and a difficult one. It might be because you changed the transistors. Slight variations in part number may have different current gains.

Or it may be you 5v load isn't connecting properly causing the PSU to overvolt.

exxos wrote: 13 Sep 2023 17:09 This is odd and a difficult one. It might be because you changed the transistors. Slight variations in part number may have different current gains.

Or it may be you 5v load isn't connecting properly causing the PSU to overvolt.

I usually use this load board to test PSU and it is the first time I have these ouputs.
I have to check the ref of the original transistors but I think they were a bit different

What about the R9 resistor ? Would it be implied in the case ?

Do you think I can connect this PSU to the Atari motherboard ? Any risks ?

exxos wrote: 13 Sep 2023 17:09 Slight variations in part number may have different current gains.

In fact I replaced
original Q2 C3040 by BU508A (ref given on the schematic) and
original Q1 C2236 by 2SC1384 (ref given on the schematic)

Here are the datasheets. Can you get some differencies that would make higher output voltages ?

doc wrote: 13 Sep 2023 17:32 What about the R9 resistor ? Would it be implied in the case ?

I would suspect the voltage issue started when you changed parts ? Looking to change other parts to fix the problem isn't the answer.

Do you think I can connect this PSU to the Atari motherboard ? Any risks ?

I doubt connecting a malfunctioning PSU to a ST is a good idea... In particular as the voltages are a lot higher than they are supposed to be.

doc wrote: 13 Sep 2023 18:15 Here are the datasheets. Can you get some differencies that would make higher output voltages ?

While it is nothing I have put to the test, I can't advise much.

Though the differences are clear on the BU508A & 508D..


I don't know why one is in milliamps and the other is amps.. But anyway..

508A = HFE ~ 9 Gain.
508D = HFE - 15 Gain.

For instance, if a transistor has an HFE of 100 and you apply a 1 mA change in base current, it would result in a 100 mA change in collector current under appropriate operating conditions.

Theoretically a larger current swing in the transistor would yield a larger flux in the core and could possibly cause voltage overshoot as you are seeing. Transistors are incredibly slow devices in reality and they may not be able to regulate the voltage correctly if it is having to deal with current which is double what is supposed to be.

Of course it is only speculation to your problem, But it is not something I have specifically put to the test on these power supplies. Changing the capacitors alone should not yield to such behaviour.

Very interesting, but still stuck ...

I tested all resistors R10, R11, R12, R13, R14 et R15 and they are fine.
Q3 is also fine.

Output voltages are still 6.28V and 14.23V :(

So the only differences are Q1 and Q2 ...

What if I try to change values R13 and R14 to bring 5v back ? (newbee idea)

I finally found the culprit : the 1uF tantal cap was dead

:bravo:

Tantalum caps suck. Replace them with a ceramic or similar .