diagROM expected behaviour with bad chipram

project23 · 20 May 2019, 16:03

Hey all, I have an AMRAM-16 512k trapdoor expansion i'm trying to get working in my 500.

The board seems okay, but the memory doesn't show at all in workbench, so I moved over to diagrom...

The thing is not even diagrom see's the memory. Now is this the expected behaviour? I know that with fastram during a memory check diagrom will just flag bad bits as it progresses, and if the ram is missing altogether it'll all be marked as 'bad', but is it the same process with chipram?

On the A500 itself _EXRAM is pulled to ground correctly, and there is Vcc to all 16 of the ram chips. I see activity on all address/data lines, but as I say there just isn't anything in either Workbench or diagrom.

Anyway, if this is to be expected and diagrom treats chipram a bit differently then that's fine, i'll assume the card is somewhat dead. If not, then I suppose this particular ram 'invisibility' might point to a specific problem.

Cheers,

John

Daedalus · 20 May 2019, 18:40

DiagROM checks for present chipRAM on first boot in a similar fashion to Kickstart - by writing at certain intervals and checking for a correct reading or garbage. If it's garbage, the chip RAM size is calculated based on that and no further checks are done. Depending on the architecture of the expansion, all it might take is a single RAM chip to be faulty to present bad bits across a large range of addresses and thus stop the entire expansion from being detected.

project23 · 20 May 2019, 19:06

Aha! Thanks Daedalus - that's what I was thinking. It's a single chip per bit, this card, so it really could just be one chip not working and the whoooole thing is screwed.

I'll double check by trying it in another 500 and if I get the same result i think its safe to say its screwed.

I don't wanna go out and buy another as i'm building the A314 pretty soon - so I think i might do the 1MB motheboard mod using the RAM from the second 500, and then on the second 500 do the mod to use only the expansion ram so that when the A314 is done that 500 utilises the whole 1MB of the expansion.

Cheers dude,

John

20 May 2019, 16:03	#1
project23 Used Register Join Date: Jul 2018 Location: Liverpool Age: 40 Posts: 437	diagROM expected behaviour with bad chipram Hey all, I have an AMRAM-16 512k trapdoor expansion i'm trying to get working in my 500. The board seems okay, but the memory doesn't show at all in workbench, so I moved over to diagrom... The thing is not even diagrom see's the memory. Now is this the expected behaviour? I know that with fastram during a memory check diagrom will just flag bad bits as it progresses, and if the ram is missing altogether it'll all be marked as 'bad', but is it the same process with chipram? On the A500 itself _EXRAM is pulled to ground correctly, and there is Vcc to all 16 of the ram chips. I see activity on all address/data lines, but as I say there just isn't anything in either Workbench or diagrom. Anyway, if this is to be expected and diagrom treats chipram a bit differently then that's fine, i'll assume the card is somewhat dead. If not, then I suppose this particular ram 'invisibility' might point to a specific problem. Cheers, John

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Some downtime expected due to server updates	FrodeSolheim	support.FS-UAE	15	27 April 2019 01:06
Diagrom?	rincewinda	support.Hardware	4	13 February 2019 15:45
DiagROM V0.9 Out	Chucky	support.Hardware	14	31 August 2017 15:35
Address register expected	Nightfox	Coders. Asm / Hardware	4	12 August 2016 11:51
Which games you didnt expected to be so good??	JudasEZT	Retrogaming General Discussion	17	13 April 2009 23:38

20 May 2019, 18:40	#2
Daedalus Registered User Join Date: Jun 2009 Location: Dublin, then Glasgow Posts: 6,084	DiagROM checks for present chipRAM on first boot in a similar fashion to Kickstart - by writing at certain intervals and checking for a correct reading or garbage. If it's garbage, the chip RAM size is calculated based on that and no further checks are done. Depending on the architecture of the expansion, all it might take is a single RAM chip to be faulty to present bad bits across a large range of addresses and thus stop the entire expansion from being detected.

20 May 2019, 19:06	#3
project23 Used Register Join Date: Jul 2018 Location: Liverpool Age: 40 Posts: 437	Aha! Thanks Daedalus - that's what I was thinking. It's a single chip per bit, this card, so it really could just be one chip not working and the whoooole thing is screwed. I'll double check by trying it in another 500 and if I get the same result i think its safe to say its screwed. I don't wanna go out and buy another as i'm building the A314 pretty soon - so I think i might do the 1MB motheboard mod using the RAM from the second 500, and then on the second 500 do the mod to use only the expansion ram so that when the A314 is done that 500 utilises the whole 1MB of the expansion. Cheers dude, John

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)