I have just obtained a BBC B which is not entirely working.

When you switch it on, the caps lock LED lights up... and nothing else happens. No brrr-beeep, no screen display. No keys seem to work, except for Break which has the effect of momentarily pausing the very subtle "static" noise that always comes from the speaker on BBCs. This noise sounds different to usual - it sounds repetitive, as if the processor is stuck in a loop rather than executing something.

I hope I'm not the only one who listens to the background noise from the BBC speakers...!

Seriously though, if anyone can suggest what might be wrong I'd appreciate it.

John

If you disconnect the keyboard, do you get even a flashing cursor to appear?

I'd take the keyboard out completely, simply as I had one Model B I fixed by simply cleaning the dust etc that collects under the keyboard out. Must have been something conductive in there. Plus check the jumpers on the keyboard are correct and not run with bits of foil and other random metal junk.

Anyway the next step after checking they keyboard is to reseat some of the chips and see if that fixes anything.

Righto, that's given me plenty to be going on with. Thanks very much. The machine is in good order physically but is absolutely filthy so the keyboard theory sounds good. It'll take me a while to give it a good clean then I'll report progress.

Also, it says on the back that it was made not by Acorn but by BST. I don't know if this is unusual or not, but the keytops look different to my other BBC.

If nothing else works, you might want to try changing IC32 (74LS259). I had a Beeb with similar symptoms, hardly any keys would work and the machine kept locking up until BREAK was pressed, and loud white noise came out of the speaker all the time. It was fixed by changing this IC32.

Scrub that comment about the machine being made by BST. I've just looked and its actually BSR. And my other beeb is made by them as well. So presumably this is the usual manufacturer.

Beebmaster wrote:
> hardly any keys would work and the machine kept locking up until BREAK
> was pressed, and loud white noise came out of the speaker all the time.

Mine doesn't even work that well !
No keys work and the machine is permanently locked up. The white noise is very faint, not loud.

However I haven't yet taken the keyboard out.

A good starting point is to remove all unnecessary ICs in order to get it to boot. You can remove both 6522s as well as the keyboard, plus any disc interface, the serial co-pro, 6850 and 7002. Like that it should come to the BBC Computer 32K BASIC screen with a prompt.

Mark.

I thought the BBC B wouldn't boot up unless it finds a keyboard? (Just gives a MODE 0 flashing cursor?)

Yes, it will boot to a basic prompt if you remove the system VIA.

Mark.

That's something I was not aware of.

* Tries to remember for future reference if I need to for any reason *

Ditto!

I thought the same as Ian too.

So far I have had 5 spare minutes with a screwdriver and I've removed the keyboard. Still no luck with booting but I've noticed some odd variation in what happens when I switch on.

Sometimes the screen is completely black. At other times the machine generates rolling horizontal bars on the screen. These aren't solid, they fade between white and grey.

Next I'll try pulling out some ICs.

I have removed all the ICs that Mark listed except the 6850, which is not socketed. There has been no change to the machine's condition, but I have made some more observations about what it's doing.

1) I only get the "stripy lines" effect if I leave it off for a minute or two before switching on. If I switch off then on again, the screen is just black.

2) If I switch off and on quite rapidly, I can get the cassette relay to switch. Maybe all BBC's do this if you switch them off and on again too quickly though!

3) Once I got blue lines on the screen - but only once.

4) All PSU output voltages are correct.

5) I have tried pushing down various ICs but this has had no effect.

Does anyone have any further ideas?

cheers

john

I have been examining the machine with an oscilloscope and have found that everything is working, except for the IRQ line on the CPU which is held permanently high. As I've removed both 6522's, I think the only possible source of spurious IRQs must be the 6850, the only non-essential IC that I haven't yet removed.

I've attempted to desolder it, but it's not easy! What's the best way?

john

Laserbeams and prayers...

I'm not sure from what you've said but do you realise that the interrupt request line is actually /IRQ (or NOT IRQ) and hence will only be low when an interrupt is being requested. Thus, high is it's normal state with (very) short periods of low during an interrupt request.

R85 3.3k ohm ties the IRQ high, checked its not the problem ? Long shot as resistors don't usually go wrong.

You don't need to unsolder the entire chip, just unsolder pin 7 enough that is no long makes contact with the board. Check with multimeter, then scope.

Method I'm using for chip removal is simple in theory, but annoying in reality (not recommended so I'll leave it out). Do the above and if that proves its the chip, take to it with some find side cutters. Snip the pins off flush with the chip itself, then using needle nose pliers and soldering iron to remove the rest.

As I said in my previous post, if the 6502 /IRQ line is high then there are no spurious interrupt issues to pursue. Thus, attacking the 6850 is a pointless exercise.

Now...you may say the 6850 is worth ignoring, however, I have had a board that was totally dead, and it actually turned out to be the 6850 was causing a drag on the data lines - removed it and the board popped back into life.

However, that would be a VERY (LOL)rare occurance, I've only ever seen one board that did that - but it just shows that anything "can" happen.

John, I presume you've done the 16k test and it's still dead? It may have more than one dead RAM chip, you can usually look on the I/O and see if any of them look different to the others. Otherwise, you'll need to go thru it methodically with a scope and the diagram.

Mark.

Yes of course, anything is always possible but to clarify, jms said he'd already scoped everything and all is well apart from the 6502 IRQ line which I think he then mistakenly assumed to be a problem because it was held high. I meant therefore that attacking the 6850 for an IRQ problem is not sensible

Thanks for all the advice. I hadn't realised that the IRQ line is normally high rather than low, but I was following the advice in the Acorn Service Manual (pages 39 & 40) which says that the IRQ line should not be constantly high or low. When I saw it flatlining at 5v I assumed this must be an indicator of a problem.

The full list of Service Manual test results is as follows:

i) The 6502 reset line was high but goes low when pressing Break. [PASS]
ii) The IRQ line was permanently high [FAIL]
iii) Clock input and output on 6502 was 2MHz, and the 1, 2, 4, and 8 MHz signals were present on VIDPROC [PASS]
iv) Pin 7 of the 6502 (SYNC) was not permanently high or low, indicating that the CPU hadn't stalled. [PASS]
v) The read/write line of the 6502 was a square wave, which I assume means "working normally" (the Service Manual doesn't define what working normally means in this case!) [PASS]
vi) There are no horizontal / vertical sync signals on pins 39 & 40 of the CRTC, which the Service Manual says indicates that the CRTC has not been successfully programmed [FAIL]

Based on the above I thought the 6850 must be the culprit, but I haven't tried the 16k RAM test, but perhaps I should do that first.

john

It might also be that the 6845 CRTC has failed as that could also lead to the symptoms that you describe. Do you have a spare you can try swapping in ?

Or another machine with a socketed 6845 that you could try the one from your Beeb in ?

Cheers.

Phill.

No, I don't have a spare. However the 6845 is socketed, so if I build up enough evidence that its duff, it'll be easy to change.

I was just reflecting on the fact that I got a bit over-enthusiastic with my original diagnosis of the fault - just goes to show that a little knowledge is a dangerous thing!

On the positive side, nobody seems to object to me borrowing the company's oscilloscope and checking my dead machine in the workshop, in fact they are fascinated by seeing a BBC micro, so if this fault finding exercise takes a long time it doesn't matter too much!

Doh! What am I saying?

I've got another BBC here (a working one) which I can rob the 6845 out of if necessary.

I have finally got round to completing all the basic tests which people here have suggested. The latest one being swapping the 6845 with that from a known good machine. Again no luck. So thus far I've tried:

1) Switching RAM banks in and out
2) Swapping CRTCs
3) Removing both 6522s, the 8271, the ADC and serial processor.
4) Checking clock pulses and 6502 reset lines, all OK
5) Various other basic checks as described on pages 39 and 40 of the Service Manual.

The only definite fault I've detected is the lack of horizontal and vertical sync signals on pins 39 & 40 of the CRTC, which I assume is due to it not having been programmed correctly rather than a hard fault with the IC.

The only non-essential IC that I've not removed is the 6850.

I'm still quite keen to experiment further with the company's oscilloscope, or to swap components. Ian's suggested changing IC32, has anyone got any alternative or additional thoughts?

john

Some more ideas...

Have you done a full link survey as per the service manual or by comparison with your good Beeb? You didn't say where you got it from but eBay Beebs often have expansion cards ripped out by unwitting sellers for separate sale and in doing so they can easily leave the Beeb in a very sad configuration.

Try swapping the OS and Basic roms and whilst doing so check that there are no damaged pins or socket ways. Again, previously fitted expansion cards can wreak havoc with chips and sockets.

Visually check all the PCB front expansion edge connectors for bent (and hence shorted) pins.

Check all socketed IC's for correct notch-north orientation.

When you say you've scoped the 6502 for normal signal operation, are you sure that each and every address, data and control line is swinging from close to 0v to close to 5v - i.e. none are being adversely loaded and struggling to exceed 2.5v.

Thanks Martin, that lot will keep me going for a good while!

I don't think there have been any expansion boards fitted, because the machine looked very much untouched internally and externally.

With regards to Ian's suggestion of changing IC32, I see from the circuit diagram that its closely associated with the system VIA and is something to do with reading the keyboard. That suggests to me that it probably isn't at fault, because I've removed the system VIA and unplugged the keyboard without any change to the machine's symptoms.

john

OK, I've tried out Martin's ideas and got a couple of possible leads to follow up. No problems with all the simple stuff, including swapping out the OS ROM (I've removed all the others now to eliminate them from the investigation).

Regarding scoping the 6502, I found 5V, 0V and 2MHz signals on the pins which should obviously carry those signals. But I found the following things which I thought might be a bit suspicious:

1) Pin 4, IRQ, was constantly high (which a previous poster has explained is normal) but Pin 6, NMI, was constantly low. Is this normal?

2) All the address and data lines were 4V peak to peak, not 5V.

3) On the address lines A1-A8, some (not all) of the 4V peaks also had a little extra square bit on top, bit part way along, taking the peak up to 5V.

4) Data lines D5 and D6 (pins 27 & 28) showed a wobbly and spiky signal, not like a square wave at all. Other than these two, all the address and data lines were bouncing up and down in a more-or-less square wave type way, but not at any fixed frequency.

Do any of these point to anything further? I'd be delighted to hear any hints on how to interpret this.

No, it isn't, and it will quite obviously suffocate the machine completely. So make that one your first priority ...