iRAM/640 - Ram Expansion - public repository

[Blank_Reg]

Showing 640 (briefly)

[Blank_Reg]

Upper test

[Blank_Reg]

On the board

[eto]

Based on that I would expect that one of the ICs is malfunctioning or that there is a short somewhere on the PCB. Unfortunately Amstrad Diagnostics is not very useful to identify if the address logic of an expansion misbehaves. Depending on the error it report all kind of weird values. And what I see here is definitely weird.

If the seller has tested the module before shipping then it might just be a loose IC. You can try and remove all ICs from their sockets and insert them back again.

After having that done you can try again with Amsdiag. If you can change the ROM I would recommend to use the more recent 1.3 or 1.4 version.

https://www.cpcwiki.eu/index.php/ROM_List#Background_Utility_ROMs
(1.4 is available directly on that page, last column. 1.3 will be available when you follow the link to the Github repository)

Just recently I have also written a minimal RAM Test in BASIC. Unfortunately I can only find the screenshot at the moment. It's just a few lines though. Output should be like in the screenshot attached (but only until 10, the rest should be --). This might give a bit more clarity what is going on.

However, if the reseating does not resolve it, you should open a Ebay case and ask for a replacement or refund. I could imagine that one of the ATF chips is not fully working. They are definitely not from the same batch and that kind of difference could be a sign that they are recycled ICs coming form a cheap source like AliExpress. While that is not necessarily bad, if he did not test the expansion it might have slipped through that a pin is malfunctioning.


btw: where are you located?

[Blank_Reg]

That's very useful information indeed - many thanks!

I'm going to send the module back to the vendor so they can test, but I suspect one of the two ICs or the RAM itself (tho I normally find a RAM chip is either working or is completely dead, not just semi working). One of the ICs under a magnifying glass looks like it might have a funny 'leg'... I've done the prise and reseat on the chips - this is what got me from nothing to semi working before, though after a power off /on, that went again.

I had another one on order from a vendor in Poland, though this seems to be a non starter as it's been at least 4 months now and nothing has materialised...

I'm located in Southport. I've spent the last 2 years or so building up a collection and have got far too much kit now tbh! I recondition it and restore it usually. I'm a systems architect and HPE engineer by trade - used to deploy blade system, simplivity, alletra etc etc

I don't have a ROM programmer/ kit for the latest build unfortunately. Maybe my next thing to get and learn (flashing ROMs is so easy on modern equipment I'm spoilt tbh).

[Rabs]

I'm located in Southport.

I don't have a ROM programmer/ kit for the latest build unfortunately. Maybe my next thing to get and learn (flashing ROMs is so easy on modern equipment I'm spoilt tbh).

I'm down on the South Coast. Think I have spare PCBs and parts if you want a go at building your own. The programmers are not too expensive. Message me if you need help.

Rabs

[Blank_Reg]

My feeling is that the vendor can sort out or refund the i640. If I can't get hold of one after that then build your own could be a viable option, thanks.

As for the dual Amsdos / Parados ROM, I'm all ears if you can burn an up to date replacement, obviously name your costs.

Cheers.

[eto]

but I suspect one of the two ICs or the RAM itself

Yes, the RAM is less likely as at least one bank can be tested and if the SRAM was just malfunctioning you should see bank 8 being available in AmsDiag but never bank 7.

However, it might also be related to the 74hct174 IC. If you do have a multimeter you can test the connections between SRAM and 74HCT174 that select the 64K page. Maybe there's an issue. 

If there's no connection issue, you can check if A16-A18 are set correctly if selected:
In BASIC just type
 
out &7f00,&c4 => A16-A18 should be 0
out &7f00,&cc => A16 should be 1, all others 0
out &7f00,&d4 => A17 should be 1, all others 0
out &7f00,&e4 => A18 should be 1, all others 0
out &7f00,&fc => A16-A18 should be 1

[Blank_Reg]

On PAL2, 3rd / 4th from right to left bottom row looks iffy; bent. I'll test with multimeter on reverse pin.

I'll test A16 - A18 on Friday when I have a chance with multimeter.