Test ROM

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Test ROM

Post by GrahamN » Tue Oct 31, 2017 1:10 pm

I am currently looking at a repairing a BBC-B Issue 7 and have run into a slight problem.

The symptoms are - continuous tone, no LEDs, and sometimes what looks like a very large cursor on the screen (though this is intermittent). Changing the link (S25) makes no difference, though removing it completely gives me a screen with a black screen and white blocks or a blue screen with lines.

I have scoped all the data lines, and they seem OK, though I don't currently have a working board to compare.

I have downloaded a Test ROM from here (by 'tricky') and this flashes some colours on a badly distorted screen, but nothing recognisable and no text.

There is a test ROM listed in the manual, but I only have that as an SSD file - is there a method to convert to a ROM without using a BBC?

I should add that I've tested all socketed chips.

Help regarding the file conversion would be appreciated, or any suggestions as other checks I can make would be appreciated.

Thanks

User avatar
paulv
Posts: 3631
Joined: Tue Jan 25, 2011 6:37 pm
Location: Leicestershire
Contact:

Re: Test ROM

Post by paulv » Tue Oct 31, 2017 2:02 pm

Just a feeling based on the symptoms but I'd swap the two 6522's over to see what happens.

Paul

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 2:18 pm

Thanks Paul

One of the first things I tried :)

I also tried 2 new 6522 as well as a known working CPU, 6845, 74LS245 (IC14) and VC2069 as well as the 7002 and Serial ULA. Swapping any other chips now requires de-soldering, so I'm reluctant to do that until I get a better handle on what may be wrong.

dominicbeesley
Posts: 625
Joined: Tue Apr 30, 2013 11:16 am
Contact:

Re: Test ROM

Post by dominicbeesley » Tue Oct 31, 2017 2:35 pm

Do you want a physical ROM blowing or just the binary extracting from the SSD?

Either way post up the SSD and PM me your address if you want me to blow and EEPROM for you...

D

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 2:47 pm

Thanks Dominic

I've attached the SSD file. Hopefully it is OK, I can't remember where I downloaded it from now, but it is apparently taken from the service manual. I have a programmer, so just a binary file would be excellent.
Attachments
Test ROM.zip
(16.57 KiB) Downloaded 27 times

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 3:06 pm

Just in case it's familiar to anyone, I've attached a pic of the screen. This ONLY happens when no jumper is fitted on S25. If I put on a jumper either way, I get a blank screen with an intermittent cursor.
No Jumpers.jpg

dominicbeesley
Posts: 625
Joined: Tue Apr 30, 2013 11:16 am
Contact:

Re: Test ROM

Post by dominicbeesley » Tue Oct 31, 2017 4:44 pm

Files are attached.

I can't recall the exact details but I remember ending up with something like that when I broke the nmi pin on the wd1770 upgrade and noise would cause random NMIs...

You can read ssds using the tools here: https://swhs.home.xs4all.nl/bbc/mmbeeb/windows.html

D
Attachments
TestRomFiles.zip
(1.66 KiB) Downloaded 25 times

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 5:12 pm

Brilliant - thanks Dominic.

I'll get back to it tomorrow and see how I get on.

The board doesn't have a disk interface, but it may well have had one in the past as S9 has been cut. I've never re-linked this before, does it really need to be re-made, do you know?

User avatar
vanekp
Posts: 545
Joined: Thu Nov 30, 2000 7:09 am
Location: The Netherlands
Contact:

Re: Test ROM

Post by vanekp » Tue Oct 31, 2017 5:18 pm

Have you tried inverting the memory by putting an inverter on S25 see this post viewtopic.php?f=3&t=2190&start=30#p38006 go down to the section "Using a 74LS04N hex inverter to invert the signal from the North pin of S25"
If there is no disc interface you then need to install S9.
Peter.

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 5:51 pm

Given that both memory banks are failing, I didn't see the point of inverting the signal. I guess I can try it if the test ROM doesn't help, but I'm not sure what it would prove? In the past I've used an inverter to help work out which RAM chip to change, but I've always been able to boot with one bank or the other first.

I can easily remake the link, but I've seen many boards in the past where the link has been cut, but the disk interface removed without issue, so I'm a bit skeptical that it would be that easy a fix - still an easy thing to do, so I will try tomorrow when I get back to the board.

Any other suggestions are very welcome!

User avatar
vanekp
Posts: 545
Joined: Thu Nov 30, 2000 7:09 am
Location: The Netherlands
Contact:

Re: Test ROM

Post by vanekp » Tue Oct 31, 2017 6:02 pm

As far as I know moving the jumper always uses the same bank of ram for page zero only way to switch it is to use an inverter on S25.
Peter.

User avatar
tricky
Posts: 2704
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Test ROM

Post by tricky » Tue Oct 31, 2017 6:16 pm

I don't know if you found the thread or videos of what is supposed to happen with the SelfTestOS ROM, but here are the video links.


URL:http://youtu.be/CAR3MrxAXlQ

The keyboard LEDs and relay should turn on and off in time with the sound effects, this should partially test the VIA and sound via the slow data bus.


URL:http://youtu.be/ookuXJdF_AA

The first screen should be the MODE 7 Frogger splash screen and then various bit patterns in mode 0 showing the first 20K, changing the first 16K, then the last 20K, changing the second 16K.

This should test the teletext chip, 6845 and ULA, but knowing which isn't working may take a little thought.
The test patterns are there to help identify dodgy RAM/control chips, but again more investigation may be necessary.

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 7:29 pm

Thanks for that.

I shall be able to get back to the computer tomorrow late afternoon, so I'll report back once I've tried the ROM and inverter / link.

dominicbeesley
Posts: 625
Joined: Tue Apr 30, 2013 11:16 am
Contact:

Re: Test ROM

Post by dominicbeesley » Tue Oct 31, 2017 7:45 pm

The ROM that Graham posted up doesn't do anything that spectacular, are you talking about the same thing?

D

GrahamN
Posts: 26
Joined: Wed Nov 18, 2015 11:49 am
Location: Burnham on Crouch, Essex
Contact:

Re: Test ROM

Post by GrahamN » Tue Oct 31, 2017 8:10 pm

Tricky was talking about a test ROM he had created and which I shamelessly downloaded from here earlier.

I've just had a few minutes to try it again (I haven't had time to burn the one Dominic created for me yet), as well as the inverter and S9 link

The S9 link made no difference at all that I could see.

The inverter showed the same image that I posted up earlier with no links on S25 at all.

The Tricky Test ROM switched the cassette LED on straight away and the relay clicked. The cassette LED then switched off, but none of the other LEDs came on at all. I did create a video though...

https://www.radiophile.co.uk/images/bbc_test.mp4

User avatar
tricky
Posts: 2704
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Test ROM

Post by tricky » Wed Nov 01, 2017 6:37 am

I guess that the 6502, OS ROM and whatever drives the cassette relay are working at least partly and some different values look like they are being picked up by the display circuitry, but not much else.
I wild guess would be whatever circuitry detects the the memory mapped hardware is being selected, but I'm a software guy :(

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 10:24 am

Hi, new to the forum here! I used Trickys test rom (thanks btw - great work) to just sanity check my system. It mostly behaves but just occasionally I can get a freeze or something (break usually sorts it) - I put that down to lose sockets perhaps. But having run the test ROM I am thinking I have intermittant RAM issues.

Does the out of place pixels there indicate a problem? See attached!

BTW - the patterns do vary a bit - it's not consistant.
IMG_4544.JPG
Last edited by GadgetUK164 on Sun Jun 24, 2018 10:31 am, edited 1 time in total.
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 11:18 am

The strange thing is the dot patterns change there, and I can see more than just the 2 or 3 dots as shown in the photo. Same happens when switched to 16K mode too. Yet all the games I load work and look fine?!?! I just don't get it! Is there a RAM test program for the BBC? Something I can write in basic perhaps?
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 11:32 am

I've noticed that one minute dots appear on the black bar, other times the white bar, and they dont follow the same pattern either. One minute no dots at all, power of and on, 2 or 3 rows of random dots in white, or black columns. Thinking it could be something related to RAS or CAS, and perhaps only affecting video or something. What i just do not understand is how the system is so robust if these patterns are indicating a fault. I've played hundreds of games on this and it does behave normally?!?!
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

User avatar
1024MAK
Posts: 7787
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Test ROM

Post by 1024MAK » Sun Jun 24, 2018 12:27 pm

There are various BASIC memory test programs in the Fault finding thread ;-)

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

User avatar
1024MAK
Posts: 7787
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Test ROM

Post by 1024MAK » Sun Jun 24, 2018 12:38 pm

Hello GadgetUK164

Why not introduce yourself in the introduce yourself section :wink:

Your machine may have a fault in the video circuitry side. Hence a running program would not suffer from corruption of the RAM contents.

Is the problem present when the machine is cold (after being off for two or more hours)?

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 2:55 pm

Thanks! I am not convinced there's an actual fault tbh - unless there's something specific about the way it tests vs normal operation (ie. perhaps its testing in some way that normal software would not typically run?) I see no graphical issues at all, and the machine does behave normally. I questioned it because it would seem to suggest a RAM fault, but in usage I dont see any problems whatsoever (even in display). I wonder if there's a bug in the test code there that assumes a specific video ULA revision or something?

I've spent the last 2 hours scoping everything around the RAM section and in particular where it connects to the 6845 and Video ULA and everything looks fine, and games look and work perfectly. If it was an accurate indication of a RAM fault I am sure I would have at least seen a pixel out of place somewhere (maybe in higher res screens), but I haven't?!?!?

I will continue working my way through software and games and report back if I discover anything.
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

cmorley
Posts: 616
Joined: Sat Jul 30, 2016 7:11 pm
Location: Oxford
Contact:

Re: Test ROM

Post by cmorley » Sun Jun 24, 2018 3:15 pm

My dev beeb has the notorious overheat RAM fault. It will go wrong if the lid is on in a warm room. You can see it in graphical corruption and games which really work the machine out will crash - e.g Whitelight
The game "Microball" (37B on beeb.MMB menu) doesn't refresh the screen so it picks up random pixels. Lid off it works perfectly 24/7.

So I'd expect memory faults to show graphical errors and random crashes.

The RAM is interleved/shared between the CPU and video. So it could be possible I suppose for the video memory reads to be a glitching I suppose or the data into the ULA (which clocks out the pixels) without main RAM being faulty. I've never seen this but I don't repair the things that often...
Last edited by cmorley on Sun Jun 24, 2018 3:16 pm, edited 1 time in total.

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 3:46 pm

Thanks! It just doesn't make sense to me lol! Your suggestion there about the serialisation is the best bet I think, but I've ruled out the Video ULA, and I've swapped the 6845 and SAS5050 too to rule them out. But again - I see no issues either! Those dots appearing in the test there (in all modes) make me think it cannot be anything but the serialisation. But then surely I would see something in games, but I don't! I just dont get it at all!
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

cmorley
Posts: 616
Joined: Sat Jul 30, 2016 7:11 pm
Location: Oxford
Contact:

Re: Test ROM

Post by cmorley » Sun Jun 24, 2018 4:09 pm

Only the RAS/CAS generation from the 6845. IC8, 9, 10 &11. Two work in mode 7 and two work in graphics modes. (IC 10 & 11 are for teletext mode IIRC)

Those chips are fabled for being out of spec and I think they are the ones which it is rumoured only 1 manufacturer's chips actually worked. I suppose you could get wrong bytes on occasion being fed into a working ULA. But the computer runs fine because IC12 & 13 do the RAS/CAS for the CPU.

But if this were the case then I'd expect you'd see it in games.

I think you can run Tricky's test ROM in an emulator & see what happens on a working (virtual) machine. I've not tried it.

Edit: and IC45... the RAS/CAS selector pins 9-12
Last edited by cmorley on Sun Jun 24, 2018 4:15 pm, edited 1 time in total.

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Sun Jun 24, 2018 5:06 pm

Thanks! I've thought about this and I am wondering if its happening is it is a RAM fault, but sticking bits rather than just not working at all. ie. some bits when set are not being unset. Therefore, if RAM is used to hold graphical assets etc, they display OK, but only if that RAM was re-used would there be a noticeable issue. This is where I guess I need to probe with basic to do a write read test and change the values to see if any bits stick, the problem there is I am not sure how to test the area that basic uses without killing basic. 4000h to 7FFFh is the upper 16K I think (optional for model A), and I think 0000-3FFF is the based 16K? So I am thinking if I loop through 4000 to 7FFF and write / read back values - I can check that way? Then maybe invert the CAS line with a NOT gate to swap the blocks around and repeat on the lower block (which will now be the upper block)?
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

User avatar
MartinB
Posts: 4905
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity
Contact:

Re: Test ROM

Post by MartinB » Sun Jun 24, 2018 5:21 pm

Have a looky here .... :)

(If you think it's what you're after and you go for it, just answer N (North) and N (No) for a standard 32k Beeb.)
Last edited by MartinB on Sun Jun 24, 2018 5:22 pm, edited 1 time in total.

cmorley
Posts: 616
Joined: Sat Jul 30, 2016 7:11 pm
Location: Oxford
Contact:

Re: Test ROM

Post by cmorley » Sun Jun 24, 2018 5:32 pm

If it were a RAM fault I'd expect games to crash randomly. All the code is stored in RAM. One bad read and the CPU executes the wrong instruction. Or you'd see artefacts in the graphics - glitches or corrupt sprites....

Still MartinB's RAM test won't hurt!

User avatar
tricky
Posts: 2704
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Test ROM

Post by tricky » Sun Jun 24, 2018 8:42 pm

Those dots look like they are in columns, which would suggest a single ROM/Data line, but it is hard to be sure, do they also look wrong in the other patterns?
It looks like a RAM issue (or wrong writes), but as I don't really do hardware, other than resistors and a few select 74 chips, I don't know.
The source code is with the ROM, but I don't think there are any opportunities for it to write randomly.
Interrupts are disabled and only the "OS" ROM is read, so there isn't much chance for external influences.

PS Love the videos
Last edited by tricky on Sun Jun 24, 2018 8:42 pm, edited 1 time in total.

User avatar
GadgetUK164
Posts: 57
Joined: Sat Jun 23, 2018 3:59 pm
Contact:

Re: Test ROM

Post by GadgetUK164 » Mon Jun 25, 2018 5:54 pm

MartinB wrote:
Sun Jun 24, 2018 5:21 pm
Have a looky here .... :)

(If you think it's what you're after and you go for it, just answer N (North) and N (No) for a standard 32k Beeb.)
Thanks to the link to Mem Test v3 there - I ran that against modes 0 to 7 - no errors, display correct with straight lines. So there is something crazy going on with Trickys ROM that I am using. I just don't get it lol.
tricky wrote:
Sun Jun 24, 2018 8:42 pm
Those dots look like they are in columns, which would suggest a single ROM/Data line, but it is hard to be sure, do they also look wrong in the other patterns?
It looks like a RAM issue (or wrong writes), but as I don't really do hardware, other than resistors and a few select 74 chips, I don't know.
The source code is with the ROM, but I don't think there are any opportunities for it to write randomly.
Interrupts are disabled and only the "OS" ROM is read, so there isn't much chance for external influences.

PS Love the videos
Thanks! Yes they vary - never consistant, sometimes a few dots in different rows, then somewhere else. Sometimes in the black band, other times the white one. And the same variety between the modes there - but in all modes. I cannot understand how that other mem test program shows no errors and displays straight lines. The only thing I can think is either its because the code is running from ROM vs RAM perhaps - but I dont see how? Or there's a bug that I've stumbled across, but I think that's less likely. Or perhaps your code is testing in a different way to that mem test v3 prog?!? And thanks for the comment on the videos btw - much appreciated =D
Last edited by GadgetUK164 on Mon Jun 25, 2018 5:55 pm, edited 1 time in total.
Retro Gaming Mods & Repairs - My YouTube Channel - https://www.youtube.com/GadgetUK164

Post Reply