Model B PSU fuse go bang.

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
Post Reply
mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Model B PSU fuse go bang.

Post by mpk » Tue Nov 13, 2018 8:15 pm

I just acquired what turns out to be a fairly early Model B (built by Cleartone) on eBay, sold as "untested" which appears to be a euphemism for "broken". First business being "replace the caps", I took the PSU to bits, replaced C1, C2 and C9 and.. oh hey, the fuse is missing! Not just blown - not there at all. It looks as though the caps were already replaced in the past (they have little orange stickers with the values written on them in pen) and the unit has definitely been tinkered with a bit. Checked the state of the bridge rectifier as that's another possible failure point and it seems fine.

Okay, then - replace the fuse, put it back together, power it up without connecting it to a load and the fuse fails catastrophically (as in loud bang, impressive flash as the fuse wire vaporises and the fuse shatters).

Okay then indeed. Such a catastrophic failure suggests at least to my untrained brain that the failure is pretty close to the input as the fuse's extremely prejudicial failure mode suggests it's getting the mains shorted across it and nothing else seems to have blown up. The fuse being missing when I got it suggests this isn't the first time. But where to start looking?

Any suggestions much appreciated. I've taken a look around for "blown fuse" problems, but there don't seem to be any involving such persistent and.. conclusive fuse-blowing as this.

User avatar
AndyF
Posts: 1241
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: Model B PSU fuse go bang.

Post by AndyF » Tue Nov 13, 2018 8:44 pm

I found this > https://www.eevblog.com/forum/projects/ ... ows-fuses/

Although there are likely some topics on this forum here too.

Is the transformer shorted completely (primary coil) that would explain an instant popping mains fuse, also C1 if shorted would appear to do that as its connected across the supply lines. I do not have a PSU to hand, and its been about 3 years since I took one apart for the more usual cap change so I'm not sure of what resistance reading the primary lines will provide.

One thing that springs to mind here check its not been set to 110v ( ! ) or something. I recall at college years back switching a 'scope on someone had thoughtfully set to 110 for some reason, that went pop too. I'm assuming here you're supply voltage (mains supply line) is meant to be 240 or so. Not sure just seen your 'location'

EDIT... Google tells me Austria's (not Australia) supply voltage is 230V, mind you 'Aus is the same voltage apparently. I thought there's was 110 for some reason, no matter.

https://www.power-plugs-sockets.com/austria/

Attaching a pic (well two actually!), unsure if its completely accurate or not though , found it in my docs:
BBC_PSU_DIAG.zip
(736.28 KiB) Downloaded 12 times
Last edited by AndyF on Tue Nov 13, 2018 8:47 pm, edited 3 times in total.
Andy

* NEW * The Jetset Willy and Manic Miner community :)

Adventure games ported across to the BBC (in progress) as soon as I can find some time!

User avatar
1024MAK
Posts: 8860
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Model B PSU fuse go bang.

Post by 1024MAK » Tue Nov 13, 2018 10:15 pm

I produced a schematic of the mains / primary side circuitry in this thread. One day I will get around to finishing the secondary side...

For further testing of your dodgy PSU, use the light bulb tip, see this post.

Mark
Last edited by 1024MAK on Tue Nov 13, 2018 10:17 pm, edited 1 time in total.

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Wed Nov 14, 2018 10:26 pm

Well, I've had a good poke around with a multimeter and nothing seems particularly amiss, although I can't be particularly sure about the transformer as there's no reference out there for what the internal resistance should be. I should maybe remove it and poke at it further, although I know enough about electronics to know that fiddling around with anything that handles mains voltages is unwise unless you know exactly what you're doing. :)

It's tempting to just start looking for a spare PSU, although they're probably thin on the ground around here.

User avatar
1024MAK
Posts: 8860
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Model B PSU fuse go bang.

Post by 1024MAK » Wed Nov 14, 2018 11:34 pm

Fault finding a fault on the primary side of a BBC Micro PSU

To narrow down the fault, you could lift one end of resistor R12. You may have to remove capacitor C2 to get to it.

This will disconnect transformer T2 and transistor Q2 from the primary circuit.

Mark

aotta
Posts: 168
Joined: Fri May 26, 2017 8:57 am
Location: Italy
Contact:

Re: Model B PSU fuse go bang.

Post by aotta » Thu Nov 15, 2018 6:34 am

I exploded 2 fuse while recapping micro psu in the last weeks, and in both cases it was because residual fluxing paste take a short circuit between 2 tracks (the gap between tracks near the 2 paper caps are very thin, and ac 220v is not like low voltage dc).
I clean pcb with IPA, mounted new fuse and solved.
Since you found your psu without fuse, your case is probably not related to recap working, but doing a close inspection between tracks, looking for brown signs in the pcb, may be useful
Last edited by aotta on Thu Nov 15, 2018 6:40 am, edited 1 time in total.

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Thu Nov 15, 2018 10:05 pm

aotta wrote:
Thu Nov 15, 2018 6:34 am
I exploded 2 fuse while recapping micro psu in the last weeks, and in both cases it was because residual fluxing paste take a short circuit between 2 tracks (the gap between tracks near the 2 paper caps are very thin, and ac 220v is not like low voltage dc).
I clean pcb with IPA, mounted new fuse and solved.
Since you found your psu without fuse, your case is probably not related to recap working, but doing a close inspection between tracks, looking for brown signs in the pcb, may be useful
I think it's not impossible that the current state of the PSU is the resuilt of the recapping work - I can see a scenario where the caps blew, someone replaced them, but then the fuse blew so they assumed the whole PSU must be bad and gave up. There is some flux residue on the PCB - I'll dig out the alcohol and clean it off. The worst that can happen is that I blow another fuse..
1024MAK wrote:
Wed Nov 14, 2018 11:34 pm
Fault finding a fault on the primary side of a BBC Micro PSU

To narrow down the fault, you could lift one end of resistor R12. You may have to remove capacitor C2 to get to it.

This will disconnect transformer T2 and transistor Q2 from the primary circuit.
Good tip - I'll give that a try. Am I reading the value on that resistor correctly? Looks like 0.47 ohms..

If all else fails, I've also asked my dad if he can dig out the old BBC B from the attic and send it over to me. That machine was well used over many years and abused horribly by me in my attempts to do experiments, but it should still have a good PSU (modulo the need to re-cap it).

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Fri Nov 16, 2018 8:33 am

Cleaned the flux residues off the track side, lifted the leg of R12, powered it.. and the fuse went pop again. The investigation continues - I think I'll desoider the replacement caps I'd already installed to get a better look at any damage the original C2 might have caused when it exploded.

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Fri Nov 16, 2018 9:45 am

So, a bit more desoldering and debugging later - all the caps before the BR seem to be fine, but even after desoldering C1, C2, C13 and C14 there was a short between the AC lines. Eventually tracked this to the BR - confusing, as I'd checked it out before and it seemed fine. Turns out it has a short in one direction only, which is why I hadn't noticed it last time I checked - just dumb bad luck that I'd had the probes the other way round.

The offending component's a KBP10 which judging from the way the legs have had to be bent out to fit the PCB is probably a retrofit, but there's what seems to be a date code (8338) which might suggest otherwise.

Does anyone know off the top of their head what the usual BR in these supplies is? I figure it's a good idea to check before I go ordering a couple of KBP10s..

User avatar
1024MAK
Posts: 8860
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Model B PSU fuse go bang.

Post by 1024MAK » Fri Nov 16, 2018 11:03 am

Yes, the bridge rectifiers are normally KBP10. At least on the PSU’s that I have worked on (about five so far).
Have a look at the last photo on this page :wink:

Mark
Last edited by 1024MAK on Fri Nov 16, 2018 11:08 am, edited 1 time in total.

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Fri Nov 16, 2018 8:11 pm

Aha! Yes indeed. I guess the legs are staggered like that to help with separation of the AC and DC sides, then.

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Tue Nov 20, 2018 2:43 pm

Okay - after waiting a few days for some replacement bridge rectifiers (I ordered 5 on the assumption I'd blow up a couple more :) ) I now have a working Model B!

Replaced the BR, cleaned any remaining flux residues off, put PSU back together after replacing the original power cord with a Schuko plug so I can use it here without needing an adaptor. Not only did the PSU not blow up, it had almost bang on +5V and slightly less than -5V on the output. Good enough..

Reassembled the machine. Got constant tone and no video output, bah. Started removing chips to see if I could isolate the issue and after about half an hour got it down to the 8271. Pulled out the 8271 and we have liftoff!

Of course, now it doesn't have a disc controller. It'd be nice to troubleshoot the disc interface a bit further to see if the 8271 problem can be solved assuming it's not just completely dead.. but hey, right now it's good to have a working machine, right?

User avatar
1024MAK
Posts: 8860
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Model B PSU fuse go bang.

Post by 1024MAK » Tue Nov 20, 2018 4:13 pm

Well done =D>

You could get a 1770 upgrade module kit instead. For example, there are currently two sellers on eBay:-
https://www.ebay.co.uk/itm/253965595291 and https://www.ebay.co.uk/itm/223236465202.

Both are well known for helping Stardotters fill their houses with Acorn gear... :mrgreen: while emptying their wallets of unwanted bits of paper printed with the Queens head... (or any other respectable currency) :mrgreen:

Mark
Last edited by 1024MAK on Tue Nov 20, 2018 4:14 pm, edited 1 time in total.

User avatar
AndyF
Posts: 1241
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: Model B PSU fuse go bang.

Post by AndyF » Wed Nov 21, 2018 12:57 pm

I can say I've had a kit or three over the years from both of those suppliers and both have been top-notch.

In one case I asked if I could have a different ROM instead and they were happy to do that instead. A+ from me!

The only issue I actually had with the 1770 upgrades were the tiny links needed, I used to have trouble getting them to stay in place (not the fault of the suppliers) , in one case I just got annoyed and did a small wire link between the pins instead! :D
Last edited by AndyF on Wed Nov 21, 2018 12:58 pm, edited 2 times in total.
Andy

* NEW * The Jetset Willy and Manic Miner community :)

Adventure games ported across to the BBC (in progress) as soon as I can find some time!

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Wed Nov 21, 2018 1:28 pm

I have to admit I've ordered stuff from both of them already. Money? Oh...

A package arrived from my dad today with our old, much-used Issue 3. It's fitted with a Solidisk DDFS interface which I may attempt to transplant if I'm feeling brave, but it's.. not exactly renowned for its excellence as a disk interface so I probably won't bother. Ordering a replacement 8271 from eBay seems like the way to keep things period-accurate. :)

Couple of photos attached showing the "old" machine - my goodness, it's even more knackered than I remember it being. Note the mangled OS ROM socket which was the result of my very clever attempt to replace it with.. god knows what? By the look of it, a hammer and chisel or something.

Still, it survived all my abuse over about 12 years or so, and I'd half-bet that if I plugged it in right now it'd boot.
IMG_2904.jpg
IMG_2903.jpg
I think the "new" machine (which I'm now certain is an upgraded Model A - it's a Cleartone Issue 3 with half the RAM socketed etc.) will be the working one, while the old one can probably enjoy a graceful retirement as a parts donor.

User avatar
AndyF
Posts: 1241
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: Model B PSU fuse go bang.

Post by AndyF » Wed Nov 21, 2018 5:25 pm

:) Without looking at the instructions (as I don't have them to hand) I think to fit a disc interface on < issue four , you need a few extra minor tweaks. I might be thinking of Econet install actually as I had an issue 3 board at one time with 'net' (I know you don't from the pics and its not relevant!) but that appeared to have some kind of clock circuitry on it if I remember rightly, going back about 15+ years though!

I guess all I'm saying here is if you're doing the 8271 to 177x upgrade a few more extra steps *might* be needed on a < Iss 4 board. Later ones are almost all plug/play, unless the factory (thankfully not that common!) "thoughtfully" forgot a socket for the existing 8271! :oops: I've only ever seen one 'factory soldered' in though so I'm not sure how common it is, the tiny chip near the cassette socket I've changed a few of those the LM iirc, that seems about 50/50 if its socketed or not lol.
Last edited by AndyF on Wed Nov 21, 2018 5:26 pm, edited 1 time in total.
Andy

* NEW * The Jetset Willy and Manic Miner community :)

Adventure games ported across to the BBC (in progress) as soon as I can find some time!

mpk
Posts: 23
Joined: Tue Nov 06, 2018 11:58 am
Location: Vienna, Austria
Contact:

Re: Model B PSU fuse go bang.

Post by mpk » Wed Nov 21, 2018 9:16 pm

AndyF wrote:
Wed Nov 21, 2018 5:25 pm
I guess all I'm saying here is if you're doing the 8271 to 177x upgrade a few more extra steps *might* be needed on a < Iss 4 board. Later ones are almost all plug/play, unless the factory (thankfully not that common!) "thoughtfully" forgot a socket for the existing 8271! :oops: I've only ever seen one 'factory soldered' in though so I'm not sure how common it is, the tiny chip near the cassette socket I've changed a few of those the LM iirc, that seems about 50/50 if its socketed or not lol.
Yes, indeed. I backed out the entire Acorn disk interface upgrade and installed the 1770 from Solidisk's instructions exactly as it had been installed in the other Issue 3 (easy when you can compare the two machines to each other and have the service manuals) but still got solid tone at power-up. I'm trying to work out if it has anything to do with the one bit of the 8271 upgrade I didn't back out - the Issue 3-specific instruction to cut pin 9 of IC27 and solder it to the RHS of S9, which isn't required by the STL interface and seems to have the effect of leaving pin 11 of the 8271 socket floating rather than connected to the gate input. Looking at the STL carrier board pin 11 isn't connected to anything, so it should make no difference whatsoever.

Kazzie
Posts: 728
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Model B PSU fuse go bang.

Post by Kazzie » Wed Nov 21, 2018 10:44 pm

IC27 generates a non-maskable interrupt (NMI) for the CPU on behalf of the 8271 FDC or the 68B54 Econet chips when required. When those respective chips aren't installed, S9 and S2 ensure the respective portions of IC27 remain high, so no NMI (low pulse) is generated by the disconnected chips. S9 and S2 need to be broken to allow the 8271 and 68B54 NMIs to reach the CPU.

From the relevant portion of the Service Manual:
vi) On issue 1, 2 or 3 circuit boards only, cut the leg of IC27 pin 9 as close to the PCB as possible and the track connected to it on the component side of the circuit board between IC27 and IC89, then reconnect the cut IC leg to the East pad of link S9 with a short Length of insulated wire.
vii) On issue 4 boards onwards, cut the TCW link at position S9.
In later circuit diagrams, IC27 pin 9 is connected to S9 East and IC78 (8271) pin 11. However, from reading the above, it seems that up to issue 3 it was also connected to IC89 (the 68B54) by mistake. The modification described should, if done correctly, disconnect IC27 pin 9 from IC89, but leave it connected to IC78 pin 11 (via S9 east). Pin 11 shouldn't be floating.

(It may be the case that the Solidisk FDC doesn't make use of the NMI, so has no need of the above modification, but I've never encountered anything other than the 8271, so can't advise you there.)

If you've removed the 8271, and there's nothing else in that socket taking control of IC78 pin 11, you need to make sure that IC27 pin 9 is connected to ground. Otherwise, leaving pin 9 of IC27 unconnected will mean it constantly generates an NMI for the CPU, and the CPU never gets the opportunity to boot the OS. On issue 4+ boards, that means refitting S9. I'm not too sure what role S9 plays on an issue 3 board, but you could resolder the wire to connect pin 9 to pin 7 of IC27 instead, which will ground it and disable the NMI generation.
Last edited by Kazzie on Wed Nov 21, 2018 10:45 pm, edited 1 time in total.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM, ZIDEFS with 512MB CF card
Acorn System 1 home-made replica

Post Reply