Page 1 of 1
Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 10:19 am
by Lardo Boffin
My troubled Toastrack strikes again! I decided to run a diagnostic soak test and got the attached.
Interestingly different chips are shown each time. Does this likely mean:
1) The diagnostic software isn’t sure?
2) They are all faulty?
3) Something else is making them look faulty?
I have no major issue with replacing the RAM but if it is option 3 that presumably won’t help.
Aside from presumably the ULA what else is involved in controlling the RAM?
Work done so far:
Recapped
New voltage regulator
SCART blanking voltage mod
Composite mod and composite improvement (
https://spectrumforeveryone.com/technic ... eo-output/)
Many thanks!
Lardo
Re: Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 10:48 am
by Tiger
IC 29 (a PAL-IC which must be programed) controls memory accesses in conjunction with Z80 and ULA. In the case you didn't swap any RAM's between the tests, there could be a mistake in this area. RAM-test should also be possible without the adapter at the expansion bus. Do you have the same result without the adapter connected? If not the adapter is faulty...
Re: Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 10:51 am
by Lardo Boffin
Thanks. I didn’t swap any RAM between checks.
I’m using a Retroleum Smartcard which boots to the diagnostic software shown. I can try it on another Spectrum to make sure it is still ok but only have a 48K rubber key so not sure it would be a worthwhile test?
The Smartcard was fine a week ago when I checked a Spectrum +3.
Re: Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 11:14 am
by Tiger
It is also possible, that there is a problem with the contact between expansion bus and smartcard, so the test on an other Speccy doesn't help very much...
Re: Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 2:38 pm
by 1024MAK
The other chip that is very heavily involved with control of the RAM is IC27, a another custom IC. It has some of the controls for the non-contended DRAM plus does all the address line multiplexing for the Z80 for all the DRAM chips.
As Tiger says, the first thing to do is to ensure that the edge-connector is clean and spotless, and is not damaged.
Have you tried the supplied
Retroleum diagnostic software as well as the
ZX Spectrum Diagnostics software to compare the results?
To be honest, my current thoughts are there may be a poor connection somewhere. Poor chip/socket contact, dry/cracked solder joint, or cracked PCB track.
Mark
Re: Spectrum+ 128K ToastracK RAM faults?
Posted: Sat Jan 30, 2021 4:12 pm
by Lardo Boffin
The built in Retroleum checks didn’t report any errors.
Neither did the Brendan Alford tests this time...
I will let it run in soak test mode for a while to see if needs to warm up a bit before it reports errors.
Edit: Update - 30 iterations without error. So most likely a dirty connection.