semi dead speccy 48k who can help

techy · October 2012

Hi forum ,

i am busy recovering a 48 k speccy issue 3 board unit.

With the help of a more experienced spectrum user from the uk , i managed to get it limping.

so ;

Somewere in the past ( maybe 20 years ago ) someone tried to get an eprom in it instead of the rom ( which was gone ) .

So i finished the " 74ls32" eprom modification , arranged a replacement erpom and the unit boots.

Then i decided to do the reset-button-upgrade as well , and the unit boots.

After that i got the keyboard reconnected , speccy boots -> and the screen remains with the copyright logo , so no response from the keyboard.

Again with the help of " the more experienced user " i tried to get response from speccy by trying to connect the connector pins of kb1 and kb2 ( to simulate a key-click ) , NO reaction from speccy.

i the verified all connections to ula and cpu via diodes d1-d8 and resistors r65-r69 etc,etc , and all looks fine to me compared to the schematics of issue 3 ,

i now know the keyboard needs a new membrane ( cracked and very brittle as you touch it , it breaks ) .

BUT , what else might be defective ?? ( i assume ula but how to prove that ?? )

Ula is ofcourse the "6C001E-6 " , so nobody changed it and maybe it has gone 20 years ago............

parts are rather hard to find over here , so who can shine a light on this or maybe has a working ula for sale ??

techy

Spirantho · October 2012

Did you check the keyboard ribbon socket connections were actually connected to the ULA correctly? They can break....

RetroTechie · October 2012

Welcome here, techy! :smile:

techy wrote: »

So i finished the " 74ls32" eprom modification , (..)

I'm not sure what mod you are referring to. :-? An EPROM can take the place of a ROM directly, as long as connections on the board (socket) are correct for the EPROM (27(C)128?). There's even some wire links on the board to help with that. So where this LS32 comes into play I don't know, but it's not for using an EPROM I think.

EPROM contents will 'leak' after a long time (think decades), so if you used an old one that wasn't programmed in recent years, that could be an issue here & reprogramming a fresh EPROM might solve the problem.

BUT , what else might be defective ?? ( i assume ula but how to prove that ?? )

Ula is ofcourse the "6C001E-6 " , so nobody changed it and maybe it has gone 20 years ago............

Usually "image shows" directly translates as "ULA is okay", although for example tape or keyboard input pins could be damaged. I have my doubts that keyboard isn't working at all if you use the wire trick on keyboard connectors, as ULA keyboard pins surely won't all be dead... So plz. get a keyboard layout in terms of rows & columns and how they relate to keyboard connector signals, try all 40 combinations, and double-verify a wire you use for this purpose actually makes the connection. Worry about the busted keyboard foil later...

parts are rather hard to find over here

Which is where? (add to your profile?). For replacement ULA's you could try RWAP or a German shop (Sintech).

But I doubt you need a replacement ULA in this case. Most other Spectrum problems are the keyboard foil (replacements available in same places as mentioned above), and power supply / memory chip related (mostly RAM, lower 16K in particular).

1024MAK · October 2012

RetroTechie wrote: »

I'm not sure what mod you are referring to. :-? An EPROM can take the place of a ROM directly, as long as connections on the board (socket) are correct for the EPROM (27(C)128?). There's even some wire links on the board to help with that. So where this LS32 comes into play I don't know, but it's not for using an EPROM I think.

The ROM used in the Spectrum has a /CE input, a /CS input and a /OE input.
Sinclair made use of this to make the decoding circuits for the ROM simple.

EPROMs only have one /CE or /CS input and a /OE input, so the /CE and /CS inputs have to be ORed and fed to the EPROM /CE or /CS input. Hence the use of a 74LS32 chip.

Some other computers use different ROMs, or use only one /CE or /CS input and ground the other input.

Mark

RetroTechie · October 2012

I would be very surprised if ULA's /ROMCS signal didn't include /MREQ. For example the schematics for Chris Smith's Harlequin project suggests that it does.

You can use a wire bridge (the left "H" in schematic) to put that /ROMCS signal on pin 20. Next you have Z80 /RD signal wired to pin 22 - perfect (just going by the schematic here).

Which would make the EPROM enable its output if (A15 low, A14 low, /MREQ active and /RD active), which is what you need. Leaves pin 27, modify on socket to connect to +5V and nowhere else. And pin 1, probably already wired to +5 but I don't see that in schematic so needs checking.

And presto: EPROM working fine, no extra LS32 needed like I said.

1024MAK · October 2012

So what we need to know is does the ULA include /MREQ in the /ROMCS logic.
If yes, then as RetroTechie says above, an extra OR gate (74LS32) is not needed.

However, if the ULA circuit that generate the /ROMCS does not include /MREQ then the OR gate is needed.

As Sinclair included arrangements to feed /MREQ to the ROM, why would they include it in the ULA? If /MREQ is included in the logic for /ROMCS, why feed /MREQ to the ROM?

Maybe Chris Smith can reveal the answers...

Mark

guesser · October 2012

One signal is the Chip Enable and the other is the Chip Select. The links allow you to change which pin on the socket corresponds to those signals

The edge connector is connected to the !ROMCS pin of the ULA (34) via a resistor. This means that the ROMCS on the edge connector overrides whatever the ULA is outputting and prevents the rom chip being selected.

guesser · October 2012

Actually, I just realised that was probably all what you already worked out and not actually helping.

I've been up for over 24 hours so I think I'll just try to sleep for a bit.

1024MAK · October 2012

:eek: :eek: :eek: :roll: :roll: :roll: :razz: :razz: :razz:

roko · October 2012

1024MAK wrote: »

If /MREQ is included in the logic for /ROMCS, why feed /MREQ to the ROM? Maybe Chris Smith can reveal the answer...

He gave no reason, still presented the facts on page 181 of his book: /ROMCS=A14+A15. The diagram is on page 183.
You really got me doubting, Retrotechie!

techy · October 2012

Hi all ,

I got the spectrum long time ago becouse it was not working.

When i opened it , it had an eprom , pin 27/28 connected and track to pin 27 cut , and indeed it did not work that way.

Thanks to mark :smile: , i found a starting point an moved on to verify and modify the unit .

There is no connection between pin 1 ( 5v ) and pin 28 on the issue 3 pcb .

I tried the obvious to verify the change in rom//eprom , and with the addition of a 74ls32 it worked without a problem.

There are several ways to do this modification , but all i can say is that a straight swop of rom vs eprom cannot work becouse of signals needed by the 2 different components.

You can use passive components to achieve the same ( 2 diodes , 1 or 2 resistors ) all it does is get the appropiate signal to pin 20 but that is more difficult to arrange than simply piggback and 1 or-gate of an ls32.

All i could find about having both possibilities ( rom and eprom ) is to modify pin 27 ( wire it to pin 14 ) on the rom , and then plug it into the socket .

Unfortunaltely i cannot test that as i don't have a rom.

I also verified the eprom , arranged a fresh one , and powered up the spectrum .

It displays the copyright message , but does not respond to the keyboard.

i verified all kb1 and kb2 pins towards ula and z80 , all seems to be ok.

Both shops don't have the correct ula , so the search continues.

thanks for all replies , that is what we are here for.

techy

RetroTechie · October 2012

roko wrote: »

He gave no reason, still presented the facts on page 181 of his book: /ROMCS=A14+A15. The diagram is on page 183.
You really got me doubting, Retrotechie!

I got myself doubting now... :lol:

/ROMCS including /MREQ makes sense as it would allow use of ROMs with 2 (and not 3) enable inputs. Would seem like a silly mistake on Sinclair's part not to support that as an option. Also Chris Smith's schematics for /ROMCS generation include /MREQ as an enable signal. But perhaps that's just wishful thinking, and original ULA decoding is simpler...

Should be easy to determine though:

Wire up a 27(C)128 as proposed earlier by me, and see if that works
Put an ULA in a breadboard, set all 8 combinations for /MREQ, A14/A15 inputs, see what /ROMCS output does, and draw conclusion (preferably repeated for 5Cxx version ULA's)

roko · October 2012

Indeed Retrotechie, the thought as such is valid. And tempting....
Your final test would come down to proving Chris Smith right or wrong. Alas his work on the real ULA leaves no room for that, it is not only impressive but also rather conclusive....
In the concerning Harlequin diagram the problem is solved indeed, probably because :-) the use of a standard EPROM is anticipated there.

semi dead speccy 48k who can help

Comments