Toastrack memory paging problem

[beanz]

I have a toastrack that seemed to work fine and loads 48k games no problem but doesn't run 128k games.

Using the diagnostic rom on the retroleum smartcard V2 I've run the 48k memory test which passed ok and then the 128k memory tests...which I "think" passes ok but then fails on memory bank switching (that might be a memory chip fail too, I don't know hence why I'm asking here).

Here's the error I'm getting, ran the 128k memory test multiple times with the same result...I want to say it failed when on "Bank 8" but again I might be wrong about that.

Any clues on what I should be looking at to fix this as far as ICs etc go, thanks.

[cj7hawk]

PAL
issue?

[beanz]

PAL
issue?

?

[1024MAK]

No, he means that there may be a problem with IC29, which is a Programmable Array Logic (or PAL) chip. It’s known as type “10H8”.

The bad news is that these are no longer made
The worse news, is that their replacements from most manufacturers, GAL (Generic Array Logic) are also obsolete
But there are still some supplies of these on trading sites. You do need a suitable programmer though. I suggest you look for someone to “blow” (program) one for you.

I don’t think this is a DRAM chip fault. In most cases (but not all), when a DRAM chip fails, it normally affects most, if not all, of the address range. As these are 64k bit chips, the toaster would not be able to function as a 48K if this type of failure occurred.

Your fault may also be caused by a failure of diode D32 (1N4148), resistor R138 (470Ω), IC28 (74LS04), or IC31 (74LS174). All of which are much easier to obtain, and these should be cheaper to get. If you can’t get a 74LS04 or 74LS174, a 74HCT04 or 74HCT174 respectively should also work.

If you have some test gear (equipment), that could narrow it down further…

In the meantime, you may want to read this topic…

Mark

[cj7hawk]

Mark, every so often I gain a glimpse into just how much experience you have with the Spectrum hardware...

Me: PAL.

You: Your fault may also be caused by a failure of diode D32 (1N4148), resistor R138 (470Ω), IC28 (74LS04), or IC31 (74LS174). All of which are much easier to obtain, and these should be cheaper to get. If you can’t get a 74LS04 or 74LS174, a 74HCT04 or 74HCT174 respectively should also work.

Mark, you never cease to impress me with your instant knowledge of everything spectrum related. Thank you also for the many times you've helped me in the past.

Beanz -Anyway, in case it is the PAL, you can get a replacement that fixes some issues here: https://www.bytedelight.com/?product=un ... grey-model

Or if you can find a burner and a GAL, you can get instructions to DIY here:
https://spectrumforeveryone.com/technic ... -machines/
and there's a copy of the JEDEC file ( programming file ) here:
velesoft.speccy.cz/zx/umbrella/gal16v8-new.zip

PALs tend to die more than other chips, due to the very high speed of operation, ability to source or sink a LOT of current, and significant heat buildup while doing so, so that they can run with propagation delays less than 10ns despite a huge number of internal gates. So there's a reasonable chance it's the PAL, but given the clean list above, it's worth checking those items first since it will take time for a GAL or Programmer or assistance to make one happen.

David.

[beanz]

Thanks guys

[lister_of_smeg]

No, he means that there may be a problem with IC29, which is a Programmable Array Logic (or PAL) chip. It’s known as type “10H8”.

The bad news is that these are no longer made
The worse news, is that their replacements from most manufacturers, GAL (Generic Array Logic) are also obsolete

Can an Amtel ATF16V8C PLD be used as a replacement? These can emulate simple PALs (including 10H8) and still appear to be in production.

[1024MAK]

Can an Amtel ATF16V8C PLD be used as a replacement? These can emulate simple PALs (including 10H8) and still appear to be in production.

Yes. However, some programmers designed for Lattice GAL chips don’t work with some Amtel chips. Sorry, I can’t remember the details.

Mark

[cj7hawk]

All Gal16v8 models are interchangeable, but as mentioned, getting the right programmer for the brand you have is important. A gal can replace an equivalent Pal. The main difference is that GAL have a way to set outputs as inputs or even switch between input and output. Pal outputs are output only.

Though I've seen ee PAL chips that are newer that worked just like gals

If you have convenient access to a GAL programmer then it's easy enough to replace and set if it improves things. Otherwise you might want to check the other things mark mentioned first as well.

[beanz]

Just got around to ordering a new multiplexor chip off ebay, (https://www.ebay.com/itm/164532401824) I also got the unrainer mod from bytedelight which I'll install at the same time so I'll let you guys know if it works.

[beanz]

So the PAL/unrainer didn't fix the problem, I installed it last night and it's still doing the same thing. The PCF chip is still on the way (I saw on a Noel's retro lab video this was causing a similar problem to mine on one of his 128 toastracks). I'll gather the other parts mentioned too.

[zx64]

Can you write some simple test in assembly to see if the paging actually fails?

[zx64]

Yes. However, some programmers designed for Lattice GAL chips don’t work with some Amtel chips. Sorry, I can’t remember the details.

Mark

There is open source PAL programmer (afterburner) - I used it myself and can confirm it can burn ATF16V8C.
Another option is buy TL866 programmer

[beanz]

Last night I tested all the diodes on the board after CJ mentioning one that could be bad...I found a diode that seems bad next to the ULA that produced a reading BOTH ways on my multimeter (don't know what board designation it is as it's probably under the diode itself..see photo), it was late so didn't have time to change it, I'll do that tonight...not sure if it's related or not but guess it needs changing regardless.

[beanz]

Diode didn't fix the problem.

Waiting on the The PCF chip I guess.

[beanz]

Well this is a real pain in the ass...the new PCF chip didn't fix it, the suggested diode D32 (1N4148), IC28 (74LS04), or IC31 (74LS174) didn't fix the problem...I couldn't find resistor R138 (470Ω)...anyone point me to where it is on the board (I think the numbers are all below the resistors themselves for a lot of them).

I've ordered a new ULA just for the hell of it (no idea if it could be related)...is it worth getting a new rom chip for it?? again if it was a 48k unit it be fine and passes the upper and lower ram tests...runs 48k software...it only fails on paging and running 128k stuff.