Page 1 of 1

ZX diag cart and 128+ misbehavior

Posted: Sun Jan 09, 2022 7:10 pm
by bola_dor
Hi, i have this Diagnostic Cartridge with Brendom 's ROM,
zxdiag-main.png
that works flawlessly with many ZX 48 and clones..
But gives me this result with a perfectly working ZX spectrum 128 6U board..
20220108_190834.jpg
I think you can imagine my cold sweat running at that moment... I just wanted to try the color bars test after fixing the yellow/blue ghosting...

I could load the tap version using a divmmc with an all OK result...
20220108_192323.jpg
Much better image than before but still some green ghost that I don't know if has a cure...

Anyway my concern now is why is the diagnostic cart frightening people like this.. ?? I tried it on another +48 and works perfect ( I don have a second 128 but a black +2 and I don't know if it's compatible)
I've tried anther PSU and checked voltages under load as I thought that could be an issue but 5v are firmly 5v with multimeter.. I don't know were is the 128 using the 12v from the coil???? There are no 4116 nor 1889 IC (fortunately) ...
Any thoughts?
Thanks!!

Re: ZX diag cart and 128+ misbehavior

Posted: Thu Feb 03, 2022 5:48 pm
by zedexspecky
I'm having a similar problem but with Smart Card V2 ROM Diagnostics v1.40.
The RAM tests seem to be reporting bad ram for the 48K model and shows the IC numbers at fault for a 48K spectrum (IC6,IC9,IC18).
I have a 128K +2 grey model similar to toastrack board and these have 16 RAM chips starting from IC17 to IC32.
Could be the machine model is not getting detected correctly for some unkown reason?

Re: ZX diag cart and 128+ misbehavior

Posted: Thu Feb 03, 2022 11:49 pm
by 1024MAK
bola_dor wrote: Sun Jan 09, 2022 7:10 pm Hi, i have this Diagnostic Cartridge with Brendom 's ROM, that works flawlessly with many ZX 48 and clones..
But gives me this result with a perfectly working ZX spectrum 128 6U board..
The +2A / +2B / +3 / +3B have different ROM chip control signals on the edge-connector/expansion port, so hardware designed for 16K/48K/+/128K (toastrack)/+2 (grey) is not compatible.

All the 128K models use bank switching.

I can’t think what is causing this issue. Obviously the test code can run. But why it thinks the RAM is faulty is beyond me.
zedexspecky wrote: Thu Feb 03, 2022 5:48 pm I'm having a similar problem but with Smart Card V2 ROM Diagnostics v1.40.
The RAM tests seem to be reporting bad ram for the 48K model and shows the IC numbers at fault for a 48K spectrum (IC6,IC9,IC18).
I have a 128K +2 grey model similar to toastrack board and these have 16 RAM chips starting from IC17 to IC32.
Could be the machine model is not getting detected correctly for some unkown reason?
If the ROM is switched out (which it must be as the expansion card ROM is in use), it’s hard for code to be fully certain which machine it is running on if that machine has a hardware fault…

Mark

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 12:42 am
by Moggy
My EPROM programmer was made for the ZX81 and 16/48k Spectrum yet works perfectly well on my 2A. :?

Is it a case of the ROM chip control signals not being required for a device like this, genuine question?

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 12:53 am
by bola_dor
1024MAK wrote: Thu Feb 03, 2022 11:49 pm
bola_dor wrote: Sun Jan 09, 2022 7:10 pm Hi, i have this Diagnostic Cartridge with Brendom 's ROM, that works flawlessly with many ZX 48 and clones..
But gives me this result with a perfectly working ZX spectrum 128 6U board..
The +2A / +2B / +3 / +3B have different ROM chip control signals on the edge-connector/expansion port, so hardware designed for 16K/48K/+/128K (toastrack)/+2 (grey) is not compatible.

All the 128K models use bank switching.

I can’t think what is causing this issue. Obviously the test code can run. But why it thinks the RAM is faulty is beyond me.


Mark
Mine is a Toastrack, I don't remember using the diagnostic cart on my black +2...
My first thought was a dirty connector.. I ve cleaned it.. and DivMMC is working fine on it.. the Toastrack is working fine.. so I guess the problem is at the cartridge it self...

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 12:57 am
by bola_dor
zedexspecky wrote: Thu Feb 03, 2022 5:48 pm I'm having a similar problem but with Smart Card V2 ROM Diagnostics v1.40.
The RAM tests seem to be reporting bad ram for the 48K model and shows the IC numbers at fault for a 48K spectrum (IC6,IC9,IC18).
I have a 128K +2 grey model similar to toastrack board and these have 16 RAM chips starting from IC17 to IC32.
Could be the machine model is not getting detected correctly for some unkown reason?
Have you ran the RAM version of the test?
It doesn't recognice your ROM as a 128 model then bank switching may not being performed properly..
Is your cart the same as mine? And are you using Brendom Test?

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 1:07 am
by zedexspecky
I'm using this slotted diag rom: http://blog.retroleum.co.uk/20170918/smart-card-v2/
I cannot access any menu test functions as the initial boot the diag rom auto tests the lower/upper ram.
If any ram errors are detected then I cannot proceed any further.
The ram faults are always 16K/48K spectrum (IC6,IC9,IC18) bits 0 and 3 even though I have a 128K +2.
I have socketed all of the 16 ram chips and installed new ram, swapping them all around just results with
the same ram errors in the same location.

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 1:58 am
by bola_dor
zedexspecky wrote: Fri Feb 04, 2022 1:07 am I'm using this slotted diag rom: http://blog.retroleum.co.uk/20170918/smart-card-v2/
I cannot access any menu test functions as the initial boot the diag rom auto tests the lower/upper ram.
If any ram errors are detected then I cannot proceed any further.
The ram faults are always 16K/48K spectrum (IC6,IC9,IC18) bits 0 and 3 even though I have a 128K +2.
I have socketed all of the 16 ram chips and installed new ram, swapping them all around just results with
the same ram errors in the same location.
Ok. I have a different diag cart.. mine is not compatible with Amstrad models..
If your machine is running well and the ram loaded tests are running ok (look for the testram.tap file on the release package here https://github.com/brendanalford/zx-dia ... i/Firmware ) then we may have some similar issues with the diagnostic rom cartridges. But if you actually have a hardware problem then you should focus on solving it..

Re: ZX diag cart and 128+ misbehavior

Posted: Fri Feb 04, 2022 2:48 am
by 1024MAK
Moggy wrote: Fri Feb 04, 2022 12:42 am My EPROM programmer was made for the ZX81 and 16/48k Spectrum yet works perfectly well on my 2A. :?

Is it a case of the ROM chip control signals not being required for a device like this, genuine question?
Yes.

Mark