Question about linear PSUs

classic non-acorn hardware, software & languages
Post Reply
User avatar
jonb
Posts: 2686
Joined: Sat May 21, 2011 1:42 pm
Location: South Coast of England
Contact:

Question about linear PSUs

Post by jonb » Wed Oct 03, 2018 10:09 am

Hello

I have a PSU here that needs to be tested and fixed. It's a great beast of a thing with several outputs having a similar circuit, all separated by running off individual windings from a huuuuuge transformer. Here is an example, it's the 5v 30A rail:

5v30A PSU Scematic.JPG
5v 30A PSU

The PSU also has 10v, -10v, 12v, -12v, -15v 24v rails at varying amperage capability (0.5 - 2.5A, so only the 5v rail is really juicy at 30A).

So, to the question. Do I need dummy loads to test this thing, given it is linear and uses voltage regulators?

(Before anyone asks, yes it's for a particularly rare computer that I will be restoring, but as I'm still working on one of the Superbrains I don't want to start a discussion about it just yet).

cmorley
Posts: 1317
Joined: Sat Jul 30, 2016 8:11 pm
Location: Oxford
Contact:

Re: Question about linear PSUs

Post by cmorley » Wed Oct 03, 2018 10:19 am

I'm sure it will sit there unloaded just fine (unless faulty!) but to be sure none of the pass transistors are ill then you'll need to load it up.

A string of 6v motorcycle bulbs would work I suppose. You'll want to test the SCR crowbar too because if it does fault later on in service and that isn't working then your undisclosed rare computer will not be very happy. :S


User avatar
jonb
Posts: 2686
Joined: Sat May 21, 2011 1:42 pm
Location: South Coast of England
Contact:

Re: Question about linear PSUs

Post by jonb » Wed Oct 03, 2018 1:31 pm

WOPR? What's that then?

(Edit: Oh the computer in WarGames?)
Last edited by jonb on Wed Oct 03, 2018 1:34 pm, edited 1 time in total.

User avatar
jonb
Posts: 2686
Joined: Sat May 21, 2011 1:42 pm
Location: South Coast of England
Contact:

Re: Question about linear PSUs

Post by jonb » Wed Oct 03, 2018 1:37 pm

cmorley wrote:
Wed Oct 03, 2018 10:19 am
I'm sure it will sit there unloaded just fine (unless faulty!) but to be sure none of the pass transistors are ill then you'll need to load it up.
All lines are good apart from 15v and 24v (run the monitor & 8" drives) so I should be OK to power the main board. The bad lines occupy a separate PCB too, so there's no chance of contagion.
A string of 6v motorcycle bulbs would work I suppose. You'll want to test the SCR crowbar too because if it does fault later on in service and that isn't working then your undisclosed rare computer will not be very happy. :S
How is this tested? Short the output?

cmorley
Posts: 1317
Joined: Sat Jul 30, 2016 8:11 pm
Location: Oxford
Contact:

Re: Question about linear PSUs

Post by cmorley » Wed Oct 03, 2018 2:09 pm

jonb wrote:
Wed Oct 03, 2018 1:37 pm
A string of 6v motorcycle bulbs would work I suppose. You'll want to test the SCR crowbar too because if it does fault later on in service and that isn't working then your undisclosed rare computer will not be very happy. :S
How is this tested? Short the output?
Nooo... it looks like an overvoltage protection. I would probably look at disconnecting pin 5 of IC2 and feed in a voltage from external source/pot. Then I'd raise the voltage until the SCR fires & take note of the threshold and see if it is sensible (or some of the passives might be off/cooked around IC2 gvinig the wrong threshold or IC2 itself).

The transformer is likely to make some loud buzzing and get quite hot doing this so I would limit how long I tested with the SCR shorting the output.

Other's may have other/better suggestions!

User avatar
Elminster
Posts: 4247
Joined: Wed Jun 20, 2012 9:09 am
Location: Essex, UK
Contact:

Re: Question about linear PSUs

Post by Elminster » Wed Oct 03, 2018 2:32 pm

On the subject of dummy loads I have one of these on order, probably on a container ship somewhere.

Had reasonable reviews for cost. Cant say if any good as not arrived yet, but was cheast I could find, without actually building one myself.

https://www.ebay.co.uk/itm/283016302782

User avatar
jonb
Posts: 2686
Joined: Sat May 21, 2011 1:42 pm
Location: South Coast of England
Contact:

Re: Question about linear PSUs

Post by jonb » Wed Oct 03, 2018 2:39 pm

cmorley wrote:
Wed Oct 03, 2018 2:09 pm
The transformer is likely to make some loud buzzing and get quite hot doing this so I would limit how long I tested with the SCR shorting the output.
Hmm, right.

Well the PSU has voltage adjusters for every rail so perhaps I could stick a load on each and up the voltage with the adjuster until the crowbar shorts it.

TBH I'm not entirely comfortable doing this. I actually have two of these PSUs and one of them had a shorted out regulator diode. When I applied power the transformer started crackling and smoking! I hit the power switch pronto, and then disconnected the offending regulator, but now when I turn it on it plays dead (even the fan stays off). Probably the fuse has gone.

@Elminster - I looked at the item but it appears to be for testing battery charge / discharge. Not sure it'd cope with 5v 30A? Bit of the old magic smoke?

User avatar
Elminster
Posts: 4247
Joined: Wed Jun 20, 2012 9:09 am
Location: Essex, UK
Contact:

Re: Question about linear PSUs

Post by Elminster » Wed Oct 03, 2018 2:54 pm

There are several models, I believe there is one for 30A. If it is DC source they will work. (only a few of the models charge as well)

Edit: ZKETech do seem to do a 5v 40A version. (Much more expensive though)
Last edited by Elminster on Wed Oct 03, 2018 3:04 pm, edited 4 times in total.

cmorley
Posts: 1317
Joined: Sat Jul 30, 2016 8:11 pm
Location: Oxford
Contact:

Re: Question about linear PSUs

Post by cmorley » Wed Oct 03, 2018 3:03 pm

jonb wrote:
Wed Oct 03, 2018 2:39 pm
TBH I'm not entirely comfortable doing this. I actually have two of these PSUs and one of them had a shorted out regulator diode. When I applied power the transformer started crackling and smoking! I hit the power switch pronto, and then disconnected the offending regulator, but now when I turn it on it plays dead (even the fan stays off). Probably the fuse has gone.
I'd probably creep up on it by disconnecting the SCR first or testing the crowbar in isolation. Maybe someone has a suggestion of how to test that without being quite so brutal?

philb
Posts: 206
Joined: Sat Aug 05, 2017 7:05 pm
Location: Cambridge, GB
Contact:

Re: Question about linear PSUs

Post by philb » Wed Oct 03, 2018 4:58 pm

Agreed, I don't think you want to do a full "live" test of the SCR crowbar circuit. If you do fire the crowbar then at best it will almost certainly blow a fuse, which might be irritating to replace, at worst you might end up overstressing or destroying some of the other components in the process, and in any case you can never be 100% certain that the SCR itself will survive the experience.

One option would be to temporarily disconnect the SCR anode and put a light bulb in series with it. Then (without any valuable load connected obviously) you should be able to disturb the regulator to the point where the crowbar would fire and check that the light comes on. The series resistance of the bulb ought to prevent anything else very bad from happening.

User avatar
jonb
Posts: 2686
Joined: Sat May 21, 2011 1:42 pm
Location: South Coast of England
Contact:

Re: Question about linear PSUs

Post by jonb » Wed Oct 03, 2018 5:23 pm

Well, I have had a go anyway.

PSU 1 started making a crackling sound and the transformer started smoking. I unplugged it immedately!
PSU 2 had nothing on 15v / 24v lines but everything else looked good.

I went back to PSU 1 and discovered the blown regulator diode. Tested it and it was shorted out (both ways) so disconnected it and fired her up again. Dead. Even the fan, so swapped the fuse from PSU 2 into it as it had popped. Now on testing I am getting all voltages despite missing a rectifier diode on the 24v circuit. I imagine it is running with 1/2 wave rectification.

As I don't need the 24v line just yet (it provides power to the 8" drive) and it is separate from the other lines I am in the position of being able to power the beast up for the first time. Question is, dare I? It's capable of talking via one of its serial ports rather than its built in monitor / keyboard, so I thought that would be the way forward. Hmm, perhaps I should introduce it more formally.

User avatar
1024MAK
Posts: 10292
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Question about conventional PSUs

Post by 1024MAK » Wed Oct 03, 2018 6:52 pm

For load testing, I strongly recommend filament lamp bulbs. No heatsinks needed and you can see if the output is stable (no flicker) and you get a positive indication that the PSU is working. And if you already have a bunch of suitable lamps, no extra cost :D

So multiple 6V motorcycle and 12V car filament lamps (especially in SBC/B15 cap types) with a selection of different wattages is what I suggest. And of course, you can under run high wattage 12V lamps on a 5V supply :wink:

For crowbar circuit testing, it’s normal practice for me to put an appropriately rated filament light bulb in series with the circuit, normally where the DC fuse would be. Or if the design does not have a fuse (which would be a poor design IMHO), between the transformer secondary winding and the PCB connections, or if off board diodes/bridge rectifier, between these/this and the PCB.

The lamp then limits the ‘fault’ current and provides a positive indication that the crowbar circuit has operated.

In the schematic in your first post, you also need to measure the voltages across ALL the 0.05R resisters in the emitter circuits of transistors Q3, Q4 and Q5. As well as the base-emitter and collector-emitter voltages of these transistors. This should be done with a load drawing about 25 to 50% of the full load current. This is the best way of checking that all three transistors are sharing the load current.

As well as testing the voltage of each output both with no load, and with a 50% to 100% load, you should also use a ‘scope to check for ripple and noise on the ouputs (again with a 50% to 100% load).

Also, I don’t see ANY fuses in the circuit in your schematic. I strongly suggest that a fuse be added between the cathodes of CR1/CR2 (40HF10) and the PCB. Also between the cathodes of CR1/CR2 (40HF10) and the centre tap of the top transformer secondary.

Mark
Last edited by 1024MAK on Wed Oct 03, 2018 7:11 pm, edited 4 times in total.

Post Reply

Return to “other vintage hardware, software and languages”