Faulty RAM Decoder rev 0.3 - improvements

[Popopo]

Hello everyone.

I have change the way that Faulty Ram Decoder works and take datas to decode and point to a failing RAM IC.

Update: 13/04/25 all info and code available on my repository. Now with a lot of documented historical information and the final code.

Purpose
I have made changes in the ram_fail php code in order that now, it may detect a faulty RAM IC in expanded (modded) Sinclair QL boards that are increased till 512KiB or 640KiB of RAM.
So from now on, it is not for only stock QLs with 128K.

Previously

I am playing around with some test for a RAM program check.
I had stored in my hard disk many screenshots from my nowadays fixed QLs that before had some IC Ram faulty and some other users with we were playing around to repair their QLs.

I would appreciate for my final tests if any of you that got screenshots from your Minerva faulty ram could post it here.
In special and the most important... in QL withs mods that have not the stock IC RAMs but the expanded.
For example,
A stock QL has installed TMS4164, that is fine and I would like to have some. But some special QL change them for NEC D41256C or others by putting in piggyback a MT 1257 (or similar)
Those are what I am the most interested about.

The last one usually has placed in each IC slot 2 ICs in piggyback. In a way that expand the internal RAM till 640KiB, instead the first one that increase it till 512KiB.

More info and the code to download

Some demo images:

Thank you

[Popopo]

Well,
with the few material I have... it seems that now the system can tell which IC is faulty with some different versions.
I will publish the code later for free testing who want to check it out.
But with an internal RAM expansion of 512K, it tells us which IC is the faulty one, even over 128K. Telling the exactly place (ICn) that is falling. This mod for QL was a typical one from a Danish distributor.

I want to add some extra makeup before publish it later.
I cannot test faulty systems with internal 640K mod.

In a future revision (I hope not too late) it will tell also which multiplexers IC could be involved (because sometimes it is not only the IC but also could be a multiplexer).
The code also will be able in my Github for free use (of course).

And again... I really would apreciate any data from those who own an internal expanded RAM mod that keep the faulty codes (screenshots from Minerva) and (when possible) remember which IC was the faulty one (to compare the new version).

Cheers