BBC Model B faults

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
Post Reply
Hinotori
Posts: 33
Joined: Fri Jun 02, 2017 11:38 am
Contact:

BBC Model B faults

Post by Hinotori » Mon Dec 17, 2018 2:24 pm

Hi,

I was wondering if anyone has any insight or advice they could offer in troubleshooting a BBC Model B Micro I am trying to fix.

It was working perfectly until recently but now displays random characters diagonally across the screen. This image is of it running 10 PRINT "TEST" 20 GOTO 10

Image

I have been looking through the "troubleshooting the BBC Micro" book here:
http://primrosebank.net/computers/bbc/d ... 0Micro.pdf

and have swapped out IC2 (6845) and IC6 (ULA) for working chips from another machine and it hasn't rectified the issue.

It seems like the machine is running fine, in terms of being able to code in a line of program and run it, with the LEDs displaying as expected and the speaker beeping as it should, but everything that is displayed is corrupt. The form of visual corruption changes depending on which MODE it is in.

I have checked the Vcc on IC2, IC5 and IC6. Although the Troubleshooting book says that IC6 should receive +5V on pin 18 it is a mistake and it should be pin 16. So, all three chips seem to be receiving +5V fine.

It is an issue 7 motherboard.

Any advice or shared experience with this kind of issue would be really greatly received.

Many thanks all!
Last edited by Hinotori on Mon Dec 17, 2018 3:07 pm, edited 4 times in total.

User avatar
DutchAcorn
Posts: 1973
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: BBC Model B faults

Post by DutchAcorn » Mon Dec 17, 2018 5:26 pm

So it does boot up?

What happens if you remove the jumper on S25 (top right of the PCB)?

You will get a 16K BBC, so not all the modes will work.
Paul

Image

Hinotori
Posts: 33
Joined: Fri Jun 02, 2017 11:38 am
Contact:

Re: BBC Model B faults

Post by Hinotori » Wed Dec 19, 2018 11:09 am

Hi,
Many thanks for the reply.

It does happily boot up and functions as normal (beeps, caps lock light etc) apart from the display.

When I remove S25 jumper the bootup screen looks like this:
Image

When I replace the S25 jumper it looks like this:
Image

Do you think it might be a RAM issue?

Thanks again

User avatar
KenLowe
Posts: 434
Joined: Mon Oct 18, 2004 4:35 pm
Location: Scotland
Contact:

Re: BBC Model B faults

Post by KenLowe » Wed Dec 19, 2018 12:12 pm

Hinotori wrote:
Mon Dec 17, 2018 2:24 pm
The form of visual corruption changes depending on which MODE it is in.
What actually happens if you switch to a different mode (eg MODE 1) and run your test again?
Last edited by KenLowe on Wed Dec 19, 2018 12:15 pm, edited 1 time in total.

User avatar
AndyF
Posts: 1150
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: BBC Model B faults

Post by AndyF » Wed Dec 19, 2018 12:25 pm

Does the ULA have a heatsink on it ? They can get a bit warm sometimes.
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
tricky
Posts: 3098
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: BBC Model B faults

Post by tricky » Wed Dec 19, 2018 12:40 pm

The mode 7 corruption looks like it is every 128 bytes, which might suggest an address line error or the chip that drives it. As it seems OK in the bottom 16k (boots OK), it could just be whichever chip drives that a7.

Try:
MODE 0
COLOUR 129
CLS
And see what you get on screen; it should be white.

Hinotori
Posts: 33
Joined: Fri Jun 02, 2017 11:38 am
Contact:

Re: BBC Model B faults

Post by Hinotori » Wed Dec 19, 2018 2:09 pm

Hi All,

Many thanks for the help!

My ULA has a heatsink on it, they do get very warm.

Here are a couple of photos of it in different modes.
Image
Image

And this is a photo of the screen following inputing
MODE 0
COLOUR 129
CLS
Image
Last edited by Hinotori on Wed Dec 19, 2018 2:11 pm, edited 1 time in total.

Kazzie
Posts: 396
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: BBC Model B faults

Post by Kazzie » Wed Dec 19, 2018 4:43 pm

You've tried booting the machine with S25 removed (enable upper 16k only) and replaced (presumably North, enabling both banks for 32k). Have you tried S25 in the South position? This will enable only the lower bank of RAM.

We know the machine boots and executes programs, which indicates that the lower 16k is fine. If the video corruption goes with S25 South, then that indicates the fault is with the upper bank.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

Kazzie
Posts: 396
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: BBC Model B faults

Post by Kazzie » Wed Dec 19, 2018 5:17 pm

On closer examination of where your mode 0 screen starts to display a white screen properly (7th row, 32 characters in), this is from memory location &4000 onwards, i.e. the upper bank of RAM. That's unexpected. as your screen is showing corruption in the lower bank of RAM, but faults in the lower bank will usually stop the beeb from booting correctly. :?

I'd still suggest trying S25 in the South position. If you're able to use a 7404 inverter to swap the banks over as in this post, that could be useful too.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
DutchAcorn
Posts: 1973
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: BBC Model B faults

Post by DutchAcorn » Wed Dec 19, 2018 5:24 pm

Kazzie wrote:
Wed Dec 19, 2018 5:17 pm
On closer examination of where your mode 0 screen starts to display a white screen properly (7th row, 32 characters in), this is from memory location &4000 onwards, i.e. the upper bank of RAM. That's unexpected. as your screen is showing corruption in the lower bank of RAM, but faults in the lower bank will usually stop the beeb from booting correctly. :?

I'd still suggest trying S25 in the South position. If you're able to use a 7404 inverter to swap the banks over as in this post, that could be useful too.
Yes, looking at the screen S25 south could make a difference! In 32K mode CAS0 is the lower 16K memory and CAS1 is the higher 16K memory. With S25 set south you will be using CAS1 only (and thus as the lower memory). CAS1 seems to be working ok...

Actually I assumed the higher 16K was causing issues, hence my request to try without S25 because that makes it run on just CAS0.
Paul

Image

User avatar
tricky
Posts: 3098
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: BBC Model B faults

Post by tricky » Wed Dec 19, 2018 7:54 pm

There are some simple BASIC programs in some of the threads for testing memory, they usually write a byte and try and read it back.
I'm wondering if this is just the display.
On the MODE 7 screen, what happens if you use the cursor keys and copy on the the corrupted characters, do you get the same character, or do you get a space/blank?

Hinotori
Posts: 33
Joined: Fri Jun 02, 2017 11:38 am
Contact:

Re: BBC Model B faults

Post by Hinotori » Mon Jan 14, 2019 1:24 pm

Happy new year Stardotters! I am back to trying to repair my faulty BBC B.

I tried with the S25 jumper in the lower position and in the upper and have the same issues.

@tricky - When I copy I just get a blank space, it doesn't copy the corrupt part of the image.

Many thanks again all for the assistance :)

Is it time to get out a scope?
Last edited by Hinotori on Mon Jan 14, 2019 1:24 pm, edited 1 time in total.

User avatar
tricky
Posts: 3098
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: BBC Model B faults

Post by tricky » Mon Jan 14, 2019 1:59 pm

That would suggest that the sa5050 isn't reading the same value as the CPU, at least if the character is copyable.
I'm not sure which lines go where, but hopefully it might give someone else an idea.

Kazzie
Posts: 396
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: BBC Model B faults

Post by Kazzie » Mon Jan 14, 2019 2:19 pm

Happy new year to you.

It could be time to start scoping things. But I'd like to double-check what you said about S25. Was it that there's no difference between just North (top) and South (bottom), or no difference between North, South, and removed?

The copying test Tricky advised seems to show that the corrupted display values aren't being written to RAM (at least not where they'd be read again). And the fact that the machine is booting fine in whatever memory configuration suggests that it's not a RAM issue.

As the screen corruption shows up in both mode 7 and non-teletext modes, and the corruption only shows up at particular addresses (below &4000 in mode 0) I'd suggest pointing your oscilloscope at the 6845 CRT Address Generator (IC2). I'd take a look at the memory address lines first, particularly MA0 (pin 4) to MA9 (pin 13) as these are used in both teletext and non-teletext modes. Another possibility is a poor connection to the data bus (pins 26-33) leading to misconfiguration of the internal registers, or even an internal fault.

The above is just an educated guess, but I can take some measurements of my own beeb for a benchmark if you'd like. Other suggestions may also be on their way from other members...
Last edited by Kazzie on Mon Jan 14, 2019 2:20 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 (mid- restoration)
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 8218
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: BBC Model B faults

Post by 1024MAK » Mon Jan 14, 2019 6:46 pm

It does look like a display system fault.

Do you have anything (such as a PC with RS232 serial port or RS232 serial port to USB adapter) that can receive serial data?

As you can get the BBC to redirect output text to the serial port. You can also redirect text input from the serial port. Then you can use a PC as a terminal. This may enable you to run some more tests and see the results a bit easier.

The five pin domino / dice DIN plug is hard to find. But CPC stock them (some other suppliers also stock them).

Of you can use the DIY technique as shown by Colin using Blu Tack (shown somewhere in the forum).

Mark

Hinotori
Posts: 33
Joined: Fri Jun 02, 2017 11:38 am
Contact:

Re: BBC Model B faults

Post by Hinotori » Wed Jan 23, 2019 1:49 pm

Thanks everyone for the advice.

At this months ABUG meeting, RobC and I were able to find and fix the issue, with very kind support from a selection of other Stardot members too. Many thanks! :)

The problem was that IC45 had failed and after running some scope checks had to be replaced as it was always high on the output of pin 10. Which connects to pin 19 of IC9 and IC11 so was causing problems in both teletext and non-teletext modes.

I hope this information helps with anyone experiencing similar issues in the future.

Thanks!
Last edited by Hinotori on Wed Jan 23, 2019 2:02 pm, edited 2 times in total.

Kazzie
Posts: 396
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: BBC Model B faults

Post by Kazzie » Wed Jan 23, 2019 2:08 pm

Thanks for reporting back to us on the cause. It makes perfect sense in retrospect, though I hadn't considered IC45 as a possible explanation at the time.

You may wish to give a short summary of the fault and solution on the Fault-finding thread, for easy reference in the future.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

Post Reply