Poorly BBC model B

[SiriusHardware]

This machine has never worked since I've had it.

OK, that's daunting as we could be dealing with literally anything up to and including a lightning strike victim. These are the sort of challenges we like, though.

With regard to your test, 17 flashes is a successful full RAM test I believe - it only goes on to tests 18 and 19 (VIA tests, which it would obviously fail at the moment due to the missing VIA), if you put the test chip in the BASIC ROM socket so it looks like that 245 buffer may actually have been at fault although I agree it's better to replace like for like. It's weird that the Chinese programmer - which I have to tell you I have very little faith in - declared the chip OK but I assume you consider the European-designed retro chip tester to be the ultimate authority - I'm not too familiar with the latter item, can it test large peripheral ICs? Like 6522s?

So that's what's happening with the test ROM, what now happens if you start the machine with the OS ROM at IC51 and the not-quite-right 245 buffer fitted?

[SiriusHardware]

Regarding the screws missing from my two machines, I have found that there are complete kits of replacement screws available so I've ordered one complete kit which I will 'share' between the two machines. I have to say these machines seem amazingly well supported by third party parts and accessory suppliers.

[Syrinx1]

Are you getting a continuous tone from the speaker?

IC32 74LS259 is connected to IC3, various keyboard signals including caps, shift lock and the sound generator so might be worth investigating if IC3 has been ruled out.

[Syrinx1]

... the IC32 latch is also involved in screen memory for the various screen modes.

[SiriusHardware]

As regards a possible 'proper' try-out replacement for the 74LS245 data bus buffer it looks to me as though the 1MHz bus connector could manage without its 74LS245 data bus buffer (IC71 on my diagram) for the time being, but I would suggest testing that chip in the retro chip tester first before trying it as the data bus buffer (IC14 on my diagram) - we really have no idea what is broken and what isn't at the moment, so we can't assume that any chip borrowed from some non-essential position in the machine is definitely working.

[SiriusHardware]

And - sorry to have to ask but which IC are we calling the 'System' VIA and which are we calling the 'User' VIA?

I make them

"System" VIA = IC3

"User" VIA = IC69

(Again, these are reference numbers on my version of the diagram and I don't know if they change from one mainboard to the next).

I raise this question because since one of the VIAs was removed and left out, we seem to have lost all activity on the keyboard LEDs which are driven (via IC32, as Syrinx1 says) from the System VIA.

[Syrinx1]

As far as I know, the IC numbering should be the same across BBC issues ... yes the Sytem vs User via terminology is confusing. My understanding is that IC3 is the system via (interfacing to internal systems e.g. keyboard, sound etc.) IC69 is the user via for interfacing to external peripherals - for example the user port.

[SiriusHardware]

I should actually have stated IC reference numbers for the VIAs from the start, if there has been any confusion arising, that will be my fault entirely.

You made an interesting point about IC32 also being involved with the screen memory in some way, so the absence of IC3 would explain a lot that's now going on.

Thanks for confirming that the IC reference numbers have remained the same across all PCB versions.

[SiriusHardware]

My MMC drive has arrived! Eprom version MMFS 1.53. Mine came with an SD card and a whole lot of other stuff on a writable CD - good job I still have a few PCs which do have optical drives - looks like great value for money.

The SD card PCB itself is tiny, really fit-and forget. I could (minor grump) wish it did not occupy the user port because the 'physical computing' side of the BBC B - the ability to connect it to things and control them via the user port, is one of my favourite things about it, so that is essentially lost while the interface is fitted.

[Syrinx1]

Re: IC32 and screen memory, just to clarify as it's not immediately apparent from the circuit diagram, it's the CO and C1 signals that go to IC27,IC40 and IC36.

[Syrinx1]

@SirusH There's an MMC version for the printer port you could try.
The hardware side of things for both user and printer ports is nothing more than a level shifting spi micro sd card adaptor as used in arduino projects - cost about £3.

Just needs wiring correctly to an appropriate connector.

[Syrinx1]

... and the correct ROM burning.

[SiriusHardware]

Sounds interesting, I may look into that.

I think if I could wave a magic wand I would have had it attached to the 1MHz bus connector so it remained neatly hidden away and protected as it is now, but then the interface-side circuitry would need to be a bit more complicated, with address decoding and buffering on board. And I suppose it would upset those who do actually use the 1MHz bus connection.

[Syrinx1]

... it's also easy to install a modern 28pin sram in one of the ROM banks - just need to pick up the write enable line and additional address line fir a 32K chip... here's one such guide:

http://www.retroclinic.com/acorn/swr/swr.htm

You can then experiment with different sd card filing system ROMS without the need for eprom programming.

If you do install multiple fs ROMS you might also want something like the advanced rom manager ROM to make it easy to selectively disable fs ROMS and avoid problems where they both use the same commands.

[SiriusHardware]

Thanks for the info, although that is starting to sound rather advanced.

I have access to a pile of 28-pin Winbond 27E512 EEPROMs at work which I could either program 4 different images into with the images selected by a couple of links on the highest address lines, or simply duplicate the same image four times through the chips and plug one or another in directly as the mood takes me. I could just plug an extra conventional 28-pin socket onto the mainboard socket to prevent undue wear and tear on that socket.

[Syrinx1]

... I'll leave you to play with your beeb now SiriusH

[wave solder]

IC32 74259 passes its test in the Retro Chip Tester

Since replacing IC14 (I have now fitted the correct TTL LS one) the VIA IC that appeared to be doing something does not affect how the board operates in either sockets. The caps lock LED just stays permanently lit.
I have ordered 2 VIA ICs and will see what happens once I try those.

The Retro Chip Tester cannot test those, it also only has a 32 pin ZIF socket.

[SiriusHardware]

OK, it will be interesting to hear how you get on. It's sounding as if there may be more problems to track down, but you can cross those bridges when you come to them.

[SiriusHardware]

I'm just thinking that in theory it should be possible to make an Arduino-based 6522 / 6520 tester with loop-back connections onboard so that every function of every pin could be exercised and tested. It could even identify whether the IRQ output is open-drain or totem. Might be a useful tool for Beeb and / or PET fanatics.

[Syrinx1]

That sounds like an interesting and useful project, and why stop at the 6522! Is this something you are considering developing yourself, or are you just putting the idea out there?