Determining the problem and how to fix it

Smoke coming from the BBC micro

This is almost certainly from the power supply. See my experience:

17 October 2003: Using my own BBC B (bought 31-12-1983), C2 exploded with a bang, a puff of smoke and a horrible stench...

exploded capacitor picture 1 exploded capacitor picture 2

The power supply works fine with this defective capacitor, at least for a short while. It's just a electrical noise suppressor and the failure is not problematical, won't cause shorts. This exploded capacitor C2 won't cause broken soldering connections on the board from pushing away neighbouring components, which might happen with C1 (also a noise suppressor).
I replaced C2 with another capacitor and the PS works as before. The replacement capacitor I got from a dead BBC PS, and that showed cracks all over the outer plastic casing same as the exploded one has... I hope it doesn't explode soon :) Time to order some new ones!

Collected tips and tricks from me, usenet, bbc mailing list, and the www

WHS:

- non-working/intermittent LED lights: Check the back of the keyboard PCB for broken tracks at the LED base. Resolder if broken. Check the keyboard connector on the keyboard, I've seen a few loose pins on one BBC causing weird LED behaviour. Resolder the joint if this is the case.

- non working keys: check the rear of the keyboard PCB, if you find a break in a track: scratch the tracks (to get to the copper) on the 2 sides on the break and solder them together. Also check the keyboard connector as with the LEDs above. If that's not it, it could be a bad contact inside the key switch: spray contact cleaner onto the shaft of the non-working switch, and press it a few times to let the contact cleaner get into the switch.

- Non starting beeb that gives a continuous tone and has the cursor visible in the top left hand corner of the screen: Often, this is caused by corrosion on the keyboard connectors (the one on the main PCB and on the keyboard PCB). Spray contact cleaner on the pins and into the cable ends. For more possible causes see below.

- Startup is irregular, can take a long time, and when started, pressing BREAK doesn't work most of the time: This can be caused by corrosion on the pins/socket of the NE555 timer IC in case it is socketed (see top-left corner of the PCB). Remove the IC, apply contact cleaner to the pins and socket, let it dry for an hour or so, and insert the IC. Other causes can be faulty capacitors in the reset circuit (see circuit diagram).

BBC ML: Mike Tomlinson:

Non booting beeb 1:

>Also, what happens between the machine issuing the first and second 
>beeps?

The first beeeeeeeee noise is random data on the slow data bus (hanging
off the system 6522) being interpreted and amplified by the 76489 sound
chip.  It stops once the OS has configured the system VIA.  If the OS
cannot start, or configure the hardware, the noise never stops - this is
what you are hearing.

The second beep is the control-break type reset beep (VDU 7).

>Something else I forgot to mention in my last post in that I get slightly 
>different pitched sounds each time I switch it on. Pressing the break key also 

>causes different pitches and causes different LEDs to turn on. Sometimes the
>tape relay clicks on as well.

In order:

Take out the user 6522 (this is below the 6502 and provides the user and
printer ports)

Take out the system 6522 (between the 6502 and the 8271) - the machine
won't boot properly without it, but you should see a cursor on the
screen if everything else is working.

Take out the 8271 and the DFS ROM.

Make sure you have an OS ROM in the leftmost ROM socket - this should be
marked PB04, I think - the OS and BASIC ROMS are PB04 and PB05.  I can't
remember for sure which is which at the moment (too warm here today.)

Check IC14, 74LS245 - to the right of the 6502.  If this is socketed,
try replacing it.  It's a common failure point.  Do not use a substitute
- e.g. 74HC245, 74ALS245.

Non booting beeb 2:

No tone at all, or a continuous tone?  Usually a hardware failure on a
Beeb results in the OS being able to initialise, and you will get a
continuous "beeeeee...." where the sound is the same pitch as the first
part of the normal raucous startup "beeee-BEEP!"

Do you have -5v on the purple wire from the PSU? (Not needed for the
machine to boot, but it's used by the sound circuitry, so you will get
no sound if it's not connected or working.)  It's also used by the
serial port.

>I am an electronics engineer and I have lots of test equipment, but I
>don't 'do' computers, so I'm a bit in the dark in trying to diagnose the
>fault. I have the Beeb circuit diagram. I suppose I should look at the
>power-up reset first, around IC16.

If the machine is totally dead, (no sound at all) and you are sure the
power supplies are present and correct, first thing to check is the
16MHz crystal (easier if you have an oscilloscope.)  ISTR it's somewhere
near the 6502.  This signal is divided down by the Video ULA to provide
8MHz, 4MHz, 2MHz and 1MHz signals around the board and is essential for
the machine to work.  It seems to fail more readily if the machine has
been stored in a damp location.

Normally a reset line fault (there are both RESET and notRESET, where
notRESET is logically inverted) will result in a continuous tone on
power up.

The machine will happily boot with the following ICs missing:  user 6522
VIA (ic3, below the 6502), 6850 ACIA (serial), 68B54 ADLC (econet), all
user ROMS including basic (but the system ROM in the LHS most ROM
socket, marked PB04, must be present, of course),  and the 8271/1770
FDC.  It will sort-of boot but will not be usable with the system 6522
VIA missing (to the left of the 6502) - you'll get a screen with a
flashing cursor but will not be able to type anything.  So a valid
diagnostic technique with a faulty Beeb is to remove all the above
chips, see if the machine boots (if not, the fault is severe and needs
professional attention) - if it does, the chips can be replaced one at a
time, powering up in between until the faulty component is isolated.

Finally, if ic14 (74LS245, to the right of the 6502) is in a socket,
this is a high-failure component and replacement is worth a try.  but
the replacement must be a 74LS245 and not an ALS245, HC245 or other
alternative.

Dead BBC PSU:

WRT BBC PSUs - the most common fault on those is capacitor C9.  If you
can find someone able and prepared to replace it for you, this fixes 70%
of dead BBC PSUs.  Another common problem is that the mains cable breaks
at the strain relief grommet at the back of the PSU, usually due to
people winding the cable around the machine to store it.

Don't work on PSUs if you have no prior experience - they can have a
very nasty bite even if left unplugged for a long time.

Usenet:

(I've replaced the @s by Ys in addresses)

Dead BBC PSU:

From - Mon Feb  4 02:23:54 2002
Path: news.euronet.nl!news.euro.net!image.surnet.ru!surnet.ru!newsfeed.rt.ru!news.tele.dk!small.news.tele.dk!195.238.2.15!skynet.be!skynet.be!bnewspeer01.bru.ops.eu.uu.net!bnewsifeed01.bru.ops.eu.uu.net!lnewspost00.lnd.ops.eu.uu.net!emea.uu.net!not-for-mail
From: Mr S C Winsor <stuartwinsor YYY argonet.co.uk>
Reply-To: Mr S C Winsor <stuartwinsor YYY argonet.co.uk>
Subject: Re: PSU for BBC B
Newsgroups: comp.sys.acorn.hardware
Date: Sun, 03 Feb 2002 23:19:46 BST
Message-ID: <na.0635194b03.a80820stuartwinsor YYY argonet.co.uk>
References: <a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk> <1012746550.18199.0 YYY eos.uk.clara.net>
X-Newsreader: NewsAgent 0.85 for RISC OS
Organization: ArgoNet, but does not reflect its views
Distribution: world
Lines: 24
NNTP-Posting-Host: userac73.uk.uudial.com
X-Trace: 1012779629 news.dial.pipex.com 8513 62.188.131.17
X-Complaints-To: abuse YYY uk.uu.net
Xref: news.euronet.nl comp.sys.acorn.hardware:97543

In article <1012746550.18199.0 YYY eos.uk.clara.net>, "Ian Stocks"
<ian YYY frogstar.demon.co.uk> wrote:
> Are there any common faults on the BBC power supply, are they easy to
> fix?

Assuming the later switch mode (I believe there were some very early ones
with linear PSUs). If I remember correctly, there's an electrolytic
capacitor which is close to a semiconductor on a heat sink which runs quite
warm. This accelerates the "drying out" and loss of capacity.

However, by now, replace all electrolytics. They are the commonest cause of
lack of output from PSUs.

Stuart. 

-- 
 __  __  __  __      __ ___   _____________________________________________
|__||__)/ __/  \|\ ||_   |   /
|  ||  \\__/\__/| \||__  |  /...Internet access for all Acorn RISC machines
___________________________/ stuartwinsor YYY argonet.co.uk

101 uses for a Pentium: No1 - A slow cooker.



-------------------------------------------------------------------------------
From - Thu Feb  7 01:09:47 2002
Path: news.euronet.nl!news7.euro.net!newspeer.clara.net!news.clara.net!dispose.news.demon.net!demon!news-hog.berkeley.edu!ucberkeley!enews.sgi.com!news.xtra.co.nz!xtra.co.nz!kendallj
From: Kendall Julian <kendallj YYY xtra.co.nz>
Subject: Re: PSU for BBC B
Newsgroups: comp.sys.acorn.hardware
Message-ID: <4b049cc78dkendallj YYY xtra.co.nz>
References: <a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk> <1012746550.18199.0 YYY eos.uk.clara.net> <na.0635194b03.a80820stuartwinsor YYY argonet.co.uk> <a3lm57$u85$1 YYY news6.svr.pol.co.uk>
User-Agent: Pluto/1.14i (RISC-OS/3.70)
Organization: TFCC
Lines: 23
Date: Wed, 06 Feb 2002 16:48:27 +0100
NNTP-Posting-Host: 210.55.47.171
X-Complaints-To: newsadmin YYY xtra.co.nz
X-Trace: news.xtra.co.nz 1012967305 210.55.47.171 (Wed, 06 Feb 2002 16:48:25 NZDT)
NNTP-Posting-Date: Wed, 06 Feb 2002 16:48:25 NZDT
Xref: news.euronet.nl comp.sys.acorn.hardware:97646


> It's some years ago that it failed.  At the time I thought maybe dry joints
> around the power components.  This fixed the problem for a while. But then
> went faulty again.  Again I did the joints and it was ok for a short while.

> It's a couple of years since I looked at it.  Do you know what the power
> components are?   I'd guess mosfets but can't remember.  Is there a circuit
> diagram somewhere?

> AndyW

Dug out my service notes (cough cough from the dust) - the only two types of
fault I came across were failure of C9 (220uf 25v 105C) on the primary side,
and C18 (100uF 25v 105C) on the secondary side. As always with switch-mode
power supplies, make sure the high voltage filter caps (C6,C8) are discharged
first.

Kendall

-- 
Kendall Julian
Napier, NEW ZEALAND
email: kendallj YYY xtra.co.nz


-------------------------------------------------------------------------------
From - Thu Feb  7 01:10:31 2002
Path: news.euronet.nl!news2.euro.net!uunet!ash.uu.net!lore.csc.com!nntp.abs.net!feeder.qis.net!sn-xit-02!supernews.com!postnews1.google.com!not-for-mail
From: news YYY sprow.co.uk (Sprow)
Newsgroups: comp.sys.acorn.hardware
Subject: Re: PSU for BBC B
Date: 4 Feb 2002 05:04:25 -0800
Organization: http://groups.google.com/
Lines: 39
Message-ID: <2ec0aa9a.0202040504.76e7e3aa YYY posting.google.com>
References: <a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk> <1012746550.18199.0 YYY eos.uk.clara.net>
NNTP-Posting-Host: 136.170.128.116
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 1012827865 24710 127.0.0.1 (4 Feb 2002 13:04:25 GMT)
X-Complaints-To: groups-abuse YYY google.com
NNTP-Posting-Date: 4 Feb 2002 13:04:25 GMT
Xref: news.euronet.nl comp.sys.acorn.hardware:97560

"Ian Stocks" <ian YYY frogstar.demon.co.uk> wrote in message news:<1012746550.18199.0 YYY eos.uk.clara.net>...
> "AndyW" <andy YYY spamtrap.homeeng.swinternet.co.uk> wrote in message
> news:a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk...
> > I'm looking for a PSU for a BBC B  (though I guess BBC A  PSU is the
>  same?)
> > Or maybe a complete BBC B
> >
> > Hopefully someone is about to throw one out?
> > Prefer to collect.  I'm in NW London.
> 
> I could do with a couple of BBC power supplies if anyone has a few to
> sell/donate (obviously after Andy has got one first). I currently have
> mine hooked up to an old AT power supply, makes putting the lid back
> on difficult.
> 
> Are there any common faults on the BBC power supply, are they easy to
> fix?

Of all the PSUs I've ever repaired I attribute 90% of them to a dead
chopping transistor (the vertically mounted one in TO3 case).A BU208A
is a suitable replacement.
The way to tell if it's this is to unplug all the power leads from the
motherboard and switch on - you should see 5v as usual.With the power
leads reconnected you'll struggle to get 2v out of it.

The other 10% are:
- dried out startup cap.It's the tiny electrolytic on its own just
near
  loads of hot components in the middle of the board.If the PSU
squeaks a
  little when turned on then sighs when turned off it's prolly that
- shorted bulk decoupler on the 12v output.This was attributed to a
knackered
  disk drive,not really the PSUs fault
- catastrophic failure: things splitting their sides and oozing goo
and bits
  of molten metal everywhere.That's when I give up!

Sprow.



-------------------------------------------------------------------------------
From - Fri Feb  8 20:55:03 2002
Path: news.euronet.nl!news.euro.net!news7.euro.net!newsfeed.freenet.de!fu-berlin.de!uni-berlin.de!213.122.75.112!not-for-mail
From: "Philip Pemberton" <philpem YYY bigfoot.com>
Newsgroups: comp.sys.acorn.hardware
Subject: Re: PSU for BBC B
Date: Thu, 7 Feb 2002 08:42:52 -0000
Lines: 43
Message-ID: <a3tef4$1aadti$1 YYY ID-122086.news.dfncis.de>
References: <a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk> <1012746550.18199.0 YYY eos.uk.clara.net> <a3mfft$lh9$1 YYY helle.btinternet.com>
NNTP-Posting-Host: 213.122.75.112
X-Trace: fu-berlin.de 1013071141 44382130 213.122.75.112 (16 [122086])
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Xref: news.euronet.nl comp.sys.acorn.hardware:97703

"Dom Wright" <DomWright YYY btinternet.com> wrote in message
news:a3mfft$lh9$1 YYY helle.btinternet.com...
> "Ian Stocks" <ian YYY frogstar.demon.co.uk> wrote in message
> news:1012746550.18199.0 YYY eos.uk.clara.net...
> >
> > "AndyW" <andy YYY spamtrap.homeeng.swinternet.co.uk> wrote in message
> > news:a3jd2l$nfp$1 YYY newsg2.svr.pol.co.uk...
> > > I'm looking for a PSU for a BBC B  (though I guess BBC A  PSU is the
> > same?)
> > > Or maybe a complete BBC B
> > >
> > > Hopefully someone is about to throw one out?
> > > Prefer to collect.  I'm in NW London.
> >
> > I could do with a couple of BBC power supplies if anyone has a few to
> > sell/donate (obviously after Andy has got one first). I currently have
> > mine hooked up to an old AT power supply, makes putting the lid back
> > on difficult.
> >
> > Are there any common faults on the BBC power supply, are they easy to
> > fix?
> >
>
> I have had a couple fail on C2 (on the Mains end of the board).  One of
> these took little diagnosis - case cracked, sparks and smoke coming out,
> large Neon arrow saying "The Fault Is HERE!";-)
>
> I replaced them with similar ones from Maplin (Part No. SC93B: 0.1uF,
> 275Vac) for 29p.
You would have done better ordering some 105deg. C (high temperature), "Low
ESR" capacitors from www.farnell.com/uk and fitting them instead. 105deg
caps are generally regarded as the best to use in switchmode power supplies.
I've got enough equipment down here to recap a power supply and I wouldn't
mind having a go sooner or later. I'm not taking apart my M128 again,
though!

Later.
--
Phil.
philpem YYY bigfoot.com
http://www.philpem.btinternet.co.uk/

To email me, go to this page