Cpc 6128+ - memory failing?

[keith56]

Bah! it looks like my CPC 6128 Plus'es extended ram is failing... at least it explains why my game suddenly became unloadable...

Please see the screenshots, I attached a screenshot from
http://www.cpcwiki.eu/forum/programming/memory-check/10/
and from the |drive command... it seems two of the banks of extended memory have 'vanished'.. demoting my 6128 to a lousy 6096!

If anyone has any ideas of what may have gone wrong, please let me know! - fortunately it's in the extended memory, and  I have an X-mem - but I could do with the machine as a regular 6128 plus for game testing.

[gerald]

You may double check with the cpr in this message. The test method is different from SyX one.
http://www.cpcwiki.eu/forum/technical-support/testing-the-ram/msg127288/#msg127288
If the extension ram is indeed failing, you will have to change it.

[keith56]

Well that doesn't look very good!

I take it this is what I need?
http://www.mutant-caterpillar.co.uk/shop/product_info.php?products_id=2646

Is there any possibility that it's another part? (such as the bank switcher chip?) I don't quite understand why 32k seems to still be working in the first test.

[gerald]

Well that doesn't look very good!

I take it this is what I need?
http://www.mutant-caterpillar.co.uk/shop/product_info.php?products_id=2646

Yes ! a pair of socket would not hurt 

Is there any possibility that it's another part? (such as the bank switcher chip?)

All banking is done within the ASIC.

I don't quite understand why 32k seems to still be working in the first test.

SyX test is using a basic method (fill with 0x00/0xff/0x55/0xaa) which may be not see all kind of defect.
My test does 2 pass, one with a method similar to SyX one, the second using 32 bit LFSR that make it really unlikely to have false positive.