A Softy 2 lives

[Timbucus]

Originally Posted by SiriusHardware

That reminds me, - Tim - have you now been able to fully test out that Softy 2 you were lucky enough to acquire a little while back? I recall you were having trouble finding a co-operative display for it but you seemed to have found a way around that.

Pretty much as per post #55 I found that my Sony monitor for my MBC worked well enough to exercise it enough to burn and verify an original EPROM. It also worked very well on my little Sony Black and White TV.

I have yet to make a video about it which is on my list of tasks to do it justice and demo it fully.

For that I wanted to make up a cable to use it as a live ROM replacement for something as well so I need to find a 24 pin IDC connector in the wider EPROM shape to put on the end of the ribbon.

I have also been scouring marketplace etc for a small 12inch B&W TV to make it even more usable.

Tim

[SiriusHardware]

Ah, you mean like these:

https://www.peconnectors.com/male-id...-.100/hws5097/

(Just the first ones I found, and inconveniently located in the USA). I will keep my eyes open. I dare say an already made up cable with 2 x 24 way DIP headers on the end would also do. Not that either item is especially common nowadays.

[SiriusHardware]

Just found some, have PMed you to avoid falling foul of forum rules.

[ortek_service]

Yes, I recently-found these IDC 0.6" wide DIL-headers are now getting harder to find. And only certain sizes are still more-widely available from Farnell / RS?

I do also recently recall, from discussions on an open-source EPROM-Emulator project forum on Groups.io, that some makes of these have the connections coming-out in a different order, so pin1 isn't at the edge of the ribbon cable.
- Maybe some cheaper ones from the Far-east, that don't have very-good data on them.

[SiriusHardware]

I've seen images of another Softy 2 which I inexplicably could not find again a mere day after I had found them. One of the images was of the emulator cable which had a dual row female IDC cable mount which would fit onto the male dual-row IDC pin connector on the Softy 2 PCB, and on the other end one of those 24-pin DIL headers, the implication being that the connections were a simple one to one match. Even so I would not assume that, and I would check, with a meter if necessary, to make sure that the Softy 2 signals were landing up on the correct pins of the 24 pin header.

[ortek_service]

Dataman (UK) still sell replacement leads for the S3 & S4, which use a standard 2row 0.1" wide header.
But I think these were DIL28 and DIL32 sizes.

[ortek_service]

Just spotted these DIL24 IDC Headers, quite good value
https://www.peconnectors.com/male-id...-.100/hws5097/
Although I didn't spot any data on these, to show if the start-end of the ribbon cable is pin1 of the DIL or actually pin24, as it seems some types have unconventional swapping of the DIL's two 'columns' sides.

As well as DIL28 & DIL40 ones, plus cable-mount headers for standard DIP ones in various ways.
https://www.peconnectors.com/idc-cab...ectors-page-3/
(and do show 1+ price, when selecting each one)

Although as they are USA-based, I'm not sure about shipping to the UK, and costs.
- If they are the actual manufacturer of these, then may be a UK distributor.

[Realtime]

Just in case anyone is still looking for 24 Pin DIL ribbon cable headers (IDC) there are some good quality 3M ones available here. The original price in 2006 was about £9 each from Farnell.

[ortek_service]

Originally Posted by Timbucus

Pretty much as per post #55 I found that my Sony monitor for my MBC worked well enough to exercise it enough to burn and verify an original EPROM. It also worked very well on my little Sony Black and White TV.

>>
>>
I have also been scouring marketplace etc for a small 12inch B&W TV to make it even more usable.

Tim

For Info, ChrisOddy recently told me that whilst working on his design of a replica Softy-2, he discovered an issue with the video output he'd added to his replica Softy-1 design and it not working with some monitors.

Although he'd previously worked-out the different resistor values required to use the original UM1111 or the different-input levels but more-common UM1233 modulators (plus had AC-coupled that with a 10uF non-polarised electrolytic), he found the added video output was DC-coupled at the wrong level so it only tended to work with monitors that had an AC-coupled input.
But he's changed some component values on his Softy-1 replica design, to fix this,and updated the schematics etc. on his website for those who've built one of those.

And he'll make sure his Softy-2 replica design will also have a correct dc-levels video output (IIRC, they removed some of the buffer stages that were on the Softy-1 design were eliminated on the Softy-2- to save cost? - that makes having a correct DC-levels video output more-involved)

[ortek_service]

Originally Posted by Timbucus

Well thanks to Sirius, AnalogueMan and Phil_G I am now the proud owner of a Softy 2 EPROM programmer.
>>

and despite some ominous markings on a chip:

Attachment 279100

Well it seems Dataman were quite fond of using 'XXX' to mark PROM etc. IC's that have been programmed on these, as they also did this (but on a label) on another (Black-cased version) that was co-incidentally next to Chris Oddy's replica at RCF2023.


Although there was an odd '77' logo printed label on the SC/MP firmware EPROM's window, whereas it seems Tim's one didn't have anything covering that (unless it had come off) and relied on the case to prevent daylight etc. eventually erasing it.



- Plus see attached photo of main board / underside of PSU, and ZIP-file of all photos I got without any auto-compression of image by forum.

[ortek_service]

For Info, ChrisOddy got his Softy-2 replica working at the RCF
- After disassembling the contents of the firmware (which Tim had posted a copy of, on post #12:
https://www.vintage-radio.net/forum/...4&postcount=12
of this thread, after Tim had read-out the MM2716Q EPROM from his working Softy-2)
With Chris discovering that the first instruction byte was bizarrely 00h=Halt!?
And, comparing the contents with the firmware-EPROM from the black-cased one, Chris found that first byte should actually have been 08h
(plus bit3 was also incorrectly 0 on most of the other bytes in Tim's readout copy).

Chris also found that when he re-read the original EPROM from Tim's Softy-2, the (USB-powered, so maybe +5Vcc could be a bit-low) TL866 Programmer Chris had brought along also read incorrect contents from this. But his Acorn System programmer did read it OK (same as one from black-cased unit?)
I also tried using a Dataman 48Pro+ (Elnec re-badge), and that did seem to read Tim's original EPROM OK (1st Byte was 08h, and I noticed a block of FFh's towards end of contents, that is mainly 'D7h' on previous readout - so indicating bit5 may also be often stuck-low on that copy).

Therefore the version originally attached on this thread (as well as copy of this Chris had put on his Softy-2 webpage) are Faulty.
And so need to get these updated, before anyone else can use them to create a replica etc.

It is strange that this EPROM is giving different output data in different equipment. But maybe (especially with no label over window?) it's started to have the onset of 'bit-rot' (although you'd expect bits to go to 1, when charge is lost, on most EPROM's). And erasing plus re-programming it may help in preserving its contents.

[Mark1960]

Originally Posted by ortek_service

With Chris discovering that the first instruction byte was bizarrely 00h=Halt!?
And, comparing the contents with the firmware-EPROM from the black-cased one, Chris found that first byte should actually have been 08h
(plus bit3 was also incorrectly 0 on most of the other bytes in Tim's readout copy).

Does Softy 2 have hardware to latch the halt flag? If not then 00 will have very little diference from 08. There will be an extra fetch of the op code, but without the extra hardware to drive CONT the scmp will not halt and the flag will be cleared in the following instruction fetch.

Bit 3 different on most other bytes would be more serious.

[Mark1960]

Remember also that the instruction at 0 is not executed after reset, only if the program counter wraps round from 0FFF.

[ortek_service]

Well in those cases, it's possible that the first Byte wasn't actually the issue.
- But it was an indication that something probably wasn't quite right, and finding that bit3 (and also bit 5?) was stuck low on most other locations would definitely have resulted in it not working!

Chris said he'll look into adding facility to generate a checksum on his Acorn System EPROM-programmer version, as it could be useful.
IIRC, Data-IO programmers do generally self-check all their firmware EPROM's against a stored C/S, rather than just print it on a display and have it on IC's label.
I'm not sure there's an easy way to do it on the Softy (1 or 2) as screen is just a display of bytes in its RAM. Although there is a top status line, with seemingly a few spare unused '00' Bytes. So it might be possible to put something in those / maybe even a F/W version number from within some spare Bytes in the EPROM (seems to be quite a few FF's near the end, that are maybe between the main code and a table of data?).

It may also be possible to output its own checksum / status over its TRANSWIFT Serial / cassette or the parallel printer interfaces - but might not be enough space in the standard 2KB EPROM to implement much more.

[Phil__G]

It seems to be convention to start any SC/MP assembler program with either a 00 halt or a 08 nop, for an O/S or monitor that starts after a reset the contents of location zero are immaterial since as Mark says the first instruction fetch is from 0001. For example, National Semis 'Kitbug' starts with a nop, SCIOS starts with a halt, NIBL starts with a nop - the contents of 0000 doesnt in itself indicate a problem
For the Softy rom it was just a bad read, possibly a TL866 speed problem since Chris also got a bad image using his own TL866 - but a good read from his Acorn programmer, and with a good eprom image the Softy burst into life

[SiriusHardware]

With apologies to Tim, it seems like we need the attachment to post #12 deleted and replaced with Chris O's resurrected code - Chris or someone else with a 'good' copy of the code would need to liaise with a friendly moderator to get that done. If that is too complicated then maybe we can just get the text of the post amended to say that the code attached to it has been found to be faulty.

The value of the first byte is a probable red herring as others have pointed out, but Tim's original EPROM may be very borderline already and it may just be sheer luck that the Softy 2 itself and a 'real' programmer are still able to read the code from it correctly. It could really do with being over-programmed with good code (with a 'proper' programmer) to set the bit-rot clock back to zero.

Another thing which was discussed earlier was the need, or not, to read the character PROM from Tim's original Softy 2, but I take it the PROM code from a Softy 1 has been tried and found to work in a Softy 2 replica?

[SiriusHardware]

Does the Softy-2 have the ability to view / copy from its own EPROM space? If so, you could use it to make a new EPROM for itself.

[Timbucus]

Originally Posted by SiriusHardware

Does the Softy-2 have the ability to view / copy from its own EPROM space? If so, you could use it to make a new EPROM for itself.

Actually that is what I did to get a working copy I think - I never thought to check that the one read on the TL had no errors - this is the second works sometimes EPROM we have seen in as many months...

I agree on the note - Chris now has my original machine to copy the case so I cannot upload the fixed EPROM as I don't have it - hopefully he will do so here in the thread later at least for preservation.

To answer the character PROM yes they are conformed the same on 1 and 2.

[SiriusHardware]

Your Softy 2 is ironically probably the best 2716 programmer you will ever own, unless you manage to land yourself something like a DataMan S3 at some point.

I have absolutely no faith in the current generation of modestly priced USB powered programmers, they mostly just can't produce the +25V needed to programme one reliably and I have seen some weird behaviours from them in various repair threads over on VCFED. They are probably perfectly OK with anything from 2764s onwards.

I would, however, have expected just about anything to be able to read a 2716 OK since there are no outlandish voltages involved during the read process. If it's just a matter of speed then it is hard to imagine why one particular bit would be more affected than others.

Witjh regard to the 'bad' firmware file attached to #12 of this thread, maybe you, Tim, (as the OP of that post) could flag that up to one of the mods and ask them to 'quarantine' or remove the attachment and add a note to the effect that the file has been withdrawn because it was found to be bad?

[Mark1960]

I was a bit surprised no one took up electronwilly’s offer of 2516s, they’d probably be good for softy programming. Maybe nobody noticed his offer.

https://www.vintage-radio.net/forum/...d.php?t=204865