Re: SIMM 32 MB :-(

From: Magnus Kollberg <Magnus.Kollberg_at_emw.ericsson.se>
Date: Wed, 25 Feb 1998 10:34:07 +0100

Hi,

> >Cool. What does Doug's mem test do?
>
> That's interesting: regardless of position of the SIMMs it shows these
> lines:
>
> ================
> ================
> ==============
>
> so it looks like last 2 MB are missing :-O But the ToolKit always sets the
> TOS to 48 MB of memory. Perhaps a bug in TK 5.07? Doug???

Doug? Foug? Frog? Oppsss.... sorry. ;-)
 
> >Says who? :-)
> >
> >I have seen figures for 128MB AB040s but I don't know of anyone testing it.
>
> I've seen the figures as well, but Douglas came with different opinion when
> he was testing and developing TK 1.x

I know that Doug said that, but on the other hand, noone has tested....
 
> I would do that if it was possible. There are only Pentiums in our office
> and all have 4x16 MB. I am missing the second 32 MB to be able to test it
> :-(

So you must have 2??
 
> >Does the errors indicate that a address line is broken? Should be easy to
> see
> >if you look at which address are broken.
>
> The TTRAMTST was writting $ff to every address and then reading it back. The
> report was like this:
>
> $28b8ff7 - read $fb
> $28b8ffb - read $fb
> $28b8fff - read $fb
>
> so it looks like one of the 16 chips is partially broken (because the error
> repeats every 4th byte address and always one bit is missing).

Yup.
 
> >Make a program which puts a "walking" 1 on the data bus. Just take a
>
> sounds like you know how a memory test should be written. Go ahead and write
> it, please :)

I think not. I think you should be able to do it your self. Anyway, you can skipp
the data bus check as it must work, so you just have to go through the address bus
and that will be quik. You just need a loop which loops 32 times. 1 time for each
address lines and write some data at that address.
 
> >If this does not work I guess you have to make a patch to "remove" the
> >faulty addresses. :-\
>
> I don't like it. Perhaps Doug could have a look at the TK - it detected some
> problems while testing the memory (see the missing '==') but forgot to
> update system variables, I think.

Maybe. Ask him.
 
> >I do think the problem is with the SIM.....
>
> I strongly hope so. Otherwise I lost another 100 DM and mainly my confidence
> in AB40 :-((

Why? Just because of some faulty addresses? ;-)

//Magnus Kollberg
Received on on. feb. 25 1998 - 11:20:00 CET