UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Reply
 
Thread Tools
Old 5th Jun 2023, 7:38 pm   #21
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

In line with our usual policy of being unfailingly helpful, a sketch (attached) for a 74S287 PROM reader adaptor.

I did notice, though, that the 74S287 is soldered in. I have to admit I would be crippled by indecision, on the one hand not wanting to disturb pristine original 40+ year old solder, but on the other hand knowing that if that PROM ever fails and there's no independent backup of the code in it, the unit will be toast.

I've been trying to work out what the PROM is used for, as it is very much part of the video generation circuit my guess would have to be that it is the character generator.
Attached Thumbnails
Click image for larger version

Name:	Read_74S287_as_2716_frm.jpg
Views:	54
Size:	76.3 KB
ID:	279237  
SiriusHardware is online now   Reply With Quote
Old 5th Jun 2023, 8:17 pm   #22
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: A Softy 2 lives

Yes I believe it is the character generator and I assume the marking is to show it is blown maybe?

I know that it was postulated in the previous thread that it was probably the same code as the Softy 1 as the circuit is essentially the same and it only can generate a very limited character set.

I am am torn as if I fail to remove it then the read may be moot - I wonder if I can do it in circuit with the analyser clipped on top of it.
Timbucus is offline   Reply With Quote
Old 5th Jun 2023, 8:20 pm   #23
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: A Softy 2 lives

Thanks for the sketch by the way! I am glad we have a Softy2 to pick apart but, as you say very aware of the scarcity.

Imagine if I ever find a SCRUMPI 3...
Timbucus is offline   Reply With Quote
Old 5th Jun 2023, 8:50 pm   #24
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

Quote:
I assume the marking is to show it is blown maybe?
Blown as in 'has been programmed', I would hope, yes. Shame they didn't put it in a socket but I don't suppose they anticipated that code ever having to change, unlike the main firmware where there is a good chance that improvements and bug fixes may be incorporated over time. From what you say it does sound as though the PROM contents could be the same as in the PROM in Softy 1 but the only way we'll ever know that for sure is to compare it with the corresponding PROM code in SOFTY 1, which presumably Chris O has.

AnalogueMan may be quite bemused by all this talk of 'rescuing' the code from the programmed devices but ancient programmed devices are the bane of antique microprocessor based units like this because if they fail the unit is dead from that moment on unless the code to programme replacements with can be found somewhere - hence the need to read the code out of both of the programmed devices while they still work.
SiriusHardware is online now   Reply With Quote
Old 5th Jun 2023, 8:58 pm   #25
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

I wonder if you could connect a very fast Arduino to the A0-A7, Q0-Q3 lines and the 'G' line and have it capture the code coming out of the PROM for every possible value on A0-A7? Maybe that's something your analyser is clever enough to do.
SiriusHardware is online now   Reply With Quote
Old 5th Jun 2023, 9:22 pm   #26
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: A Softy 2 lives

Hm the little one is only 8 channels + clock so would only have four address lines - not sure about the HP maybe.
Timbucus is offline   Reply With Quote
Old 5th Jun 2023, 11:09 pm   #27
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,252
Default Re: A Softy 2 lives

It might be possible to wire a battery backed ram onto the top of the prom, so every read from the prom is written to the ram. Then disconnect the battery backed ram and read the contents on a prom programmer.
Mark1960 is online now   Reply With Quote
Old 6th Jun 2023, 12:43 am   #28
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

Brilliant idea, I would never have thought of that. I think probably the best way to pick the signals up from the PROM without having to solder anything to anywhere would be to use an IC test clip, if Tim has such a thing.

Capture the whole 256 bytes like that, read it in an EPROM reader and then ask Chris to verify those 256 bytes against the SOFTY 1 character generator PROM code he has.

The only problem I can forsee is, because this is a video generator circuit it may be running too fast for writes to the battery backed SRAM to succeed (PROMs were often used in this sort of application because they were much faster than other contemporary memory devices).

Last edited by SiriusHardware; 6th Jun 2023 at 1:05 am.
SiriusHardware is online now   Reply With Quote
Old 6th Jun 2023, 2:49 am   #29
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,252
Default Re: A Softy 2 lives

Looks like a 4MHz dot clock, with a 4 bit shift register. It might need fast access to read the data into the shift register at the correct time but hopefully the data is stable for a microsecond. Maybe the 4MHz clock can be used as write enable, but probably need to check it with a scope to see when data and address are stable.
Mark1960 is online now   Reply With Quote
Old 6th Jun 2023, 8:20 am   #30
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

If speed was an issue which prevented the available non-volatile RAM from capturing the output from the PROM you could slow the whole system right down temporarily by fitting a much lower frequency crystal, 1MHz or less, so that the 74S287 would be clocked / accessed more slowly. Of course it would not be possible to view the video output from the unit while running at 1/4 speed but the assumption would be that it would still be trying to render the contents of memory to the screen as hex characters, so hopefully every location in the character generator PROM would be 'exercised' anyway, slowly enough for the RAM to be able to catch it all.
SiriusHardware is online now   Reply With Quote
Old 6th Jun 2023, 9:26 am   #31
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

Looking at the images Tim has posted so far it seems that the character generator only ever generates hexadecimal characters 0-9, A-F, is that right? I haven't seen other alphabetic characters in any images posted so far.
SiriusHardware is online now   Reply With Quote
Old 6th Jun 2023, 9:43 am   #32
TonyDuell
Dekatron
 
Join Date: Jun 2015
Location: Biggin Hill, London, UK.
Posts: 5,151
Default Re: A Softy 2 lives

If it's anything like the oriignal Softy (which I have), then yes the character generator only displays 0-F.

A 'location' on the video display appears to contain 2 characters. The high 4 bits of the memory data are fed to the character generator, the output bits of that are serialised and sent to the video output. Then a mulitplexer switches over, the low 4 bits of the same memory data are sent to the character generator and the process repeated.
TonyDuell is offline   Reply With Quote
Old 6th Jun 2023, 9:55 am   #33
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

Thanks for that - so it seems likely that the character generator would run through its whole character set within a very short time after power-on without any need for user intervention, which is what we would need in order to capture all possible address inputs vs. all resulting 4-bit data outputs.

Last edited by SiriusHardware; 6th Jun 2023 at 10:13 am.
SiriusHardware is online now   Reply With Quote
Old 6th Jun 2023, 10:00 am   #34
TonyDuell
Dekatron
 
Join Date: Jun 2015
Location: Biggin Hill, London, UK.
Posts: 5,151
Default Re: A Softy 2 lives

FWIW the character generator n a Softy 1 is also a 74S287 and may well be the same programming. I can't remember if it's socketed n my Softy or not, but you could probably convince me to remove it and read it out.
TonyDuell is offline   Reply With Quote
Old 6th Jun 2023, 10:40 am   #35
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

Tim has speculated that the Softy1 / Softy2 char gen may contain one and the same code but Chris Oddy of these parts has already read the code from a Softy 1 char gen so I think yours can remain unmolested for the time being - thanks for the offer though.

Just looking at available options for physical devices to back the code up into, obviously a historically correct DM74S287N would be preferable but according to the PROM equivalence list on Mike's Arcade there are several pin compatible devices:-

https://www.mikesarcade.com/cgi-bin/...ge=PromRef.txt

256x4 (Tristate)

N82S129N (Philips / Signetics)
AM27S21 (AMD)
6301 (MMI)
DM74S287 (National Semiconductor),
24S10 (Ti)
7611 (Harris),
93427 (Fairchild?)

Not mentioned specifically is the MH74S287 (Tesla) which presents the same difficulty to get programmed as its big brother the MH74S571.

I have one or two NOS tubes of blank N82S129Ns since they were used for channel programming in Pye MX290 series radios which I had an interest in at one time. I'd be happy to donate a couple to a good cause once in posession of the correct code to program into them.
SiriusHardware is online now   Reply With Quote
Old 6th Jun 2023, 8:37 pm   #36
ChrisOddy
Pentode
 
Join Date: Jan 2021
Location: Northampton, Northamptonshire, UK
Posts: 131
Default Re: A Softy 2 lives

The SOFTY 1 character PROM dump can be found here: https://theoddys.com/acorn/proms_and...1/softy_1.html together with my Python emulator. I would be surprised if the SOFTY 2 is different as the character set is very simple ?

Chris
ChrisOddy is offline   Reply With Quote
Old 6th Jun 2023, 11:58 pm   #37
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

If it was me I would be supremely confident of removing the PROM to read it and refit it without leaving any trace due to a combination of decent desoldering equipment and decades of daily practice as a repair technician. I have to appreciate that it's not such a routine process for most other people, and this would be a fairly high-stakes operation which must not be messed up.

Given the apparently high likelihood that the character PROM is actually the same as the one in the Softy 1, I think if I were Tim I would ride my luck and deal with PROM failure if it ever happens.

We could resolve this another way - Tim gives you all the info you need to make a working Softy 2 replica and you then put a Softy 1 PROM in it and see if it works. If it does, that will answer the question without inflicting any unnecessary disturbance to the historically original fabric of Tim's Softy 2.

If it doesn't work, we will be back to needing to see the actual code from a Softy 2 PROM.
SiriusHardware is online now   Reply With Quote
Old 7th Jun 2023, 11:58 am   #38
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: A Softy 2 lives

I am reasonably confident with desoldering and have the tools to achieve it so I may pull the PROM as the simplest way to do this especially if Sirius can program a substitute. I will examine if there is room for a socket in which case we could also prove the the substitute worked...

Anyway here is a closeup of the chars 0-F as a sequence in the front of RAM buffer.

Click image for larger version

Name:	IMG_5201.jpg
Views:	41
Size:	122.0 KB
ID:	279293

On a more prosaic note I think I need other repairs anyway. I got hold of a BNC cable so now have a slightly better way to view the screen - although I still lose a small portion of the bottom line and the digital scan controls on this monitor seem to crash out when I try them (arghhh).

Click image for larger version

Name:	IMG_5188.jpg
Views:	40
Size:	93.8 KB
ID:	279292

Anyway if you look carefully you can see some corruption of the display towards page 2 of RAM buffer (the grey background graduations are used to judge them).

I filled the boundary areas with repeating 0123456789ABCDEF so you can see the effect better.

Click image for larger version

Name:	IMG_5202.jpg
Views:	41
Size:	161.9 KB
ID:	279294

This video shows that this is quite catastrophic as the 'cursor' jumps around the columns as it comes up one line at a time so there is a frame shift happening as well at this point which perseveres all the way down,

https://youtube.com/shorts/FHxVTTVNRJU?feature=share

I filled the boundary areas with repeating 0123456789ABCDEF so you can see the effect better.

Do not be confused by the first and last few columns they I believe are artefacts of the video generation that would normally be in the overscan area which are duplicates of the main data - this leads to a duplicate cursor anyway - hopefully Chris can confirm he can see that effect - it is clear here in the upper part of the buffer where the cursor triggers twice perhaps...

Click image for larger version

Name:	IMG_5203.jpg
Views:	29
Size:	161.6 KB
ID:	279295

Last edited by Timbucus; 7th Jun 2023 at 12:05 pm. Reason: Make clearer
Timbucus is offline   Reply With Quote
Old 7th Jun 2023, 1:46 pm   #39
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

As a double check, do you also see that corruption via the modulator / RF connected TV?

If the monitor you are currently using is specifically meant for CCTV use, does it have a 75R termination with an on / off switch next to the BNC input? What happens if you move it to the position it isn't in now?

If you do consider putting the PROM in a socket which will obviously raise it up higher off the PCB, remember how @~! hot those things get - if you put it too close to the 'lid' it could end up heat-browning the enclosure.
SiriusHardware is online now   Reply With Quote
Old 7th Jun 2023, 1:57 pm   #40
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,411
Default Re: A Softy 2 lives

In your picture in post #15 shot from the CRT TV screen the characters throughout the second pale band down look perfectly OK.

I'm not sure if what you are getting on the flatscreen isn't just the result of the flatscreen monitor trying and failing to put what it may see as a highly unorthodox video signal on the screen as best it can.
SiriusHardware is online now   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 12:29 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.