DDI-5 weirdness...

[WacKEDmaN]

so ive borrowed a friends DDI-5... its been working great over the last few days.. ran rtype 128 and the wonderboy demo fine....

well today ive gone to play...
PARADOS is showing on the ROM line on boot...(AMSDOS also works fine!)
Disk access is working perfectly fine....

but when i tryed loading up rtype 128 today it would load the "stereo" image and then lock up after it faded out....
and wonderboy demo started throwing pixels and would lock up after moving like one screen over....

so i decided to run Noels diagnostics (from dsk on DDI-5) and check the upper ram

on the main screen i see ram value set to 640KB... (it should only be 576KB right?!)
i run the upper ram test...it says total upper ram is 576kb...runs fine through the first 8 banks (00 to 07)..
but when it should be done...and just show dashes for the rest of the banks...it throws red in the first 2 segments of bank 08(white for the rest of bank 08 segments)... dashes for bank 09-0F...
it then on bank 10 it does same as at bank 08....it continues to keep doing it forever! (until i ESC and it tells me bank 00 to 06 failed (not 07?! see images)

ive tried reseating it multiple times, cleaned edge connector on 464 board and ddi-5...
it was working perfect up until today...(and the disk access is still smooth as silk)
is there any chance one of the logic chips could do this?
or the strange possibility of pin 21 on the PPI floating?! i broke a pin off days ago! so no connection to LK3 and now i have a Solavox or Amstrad depending how its feeling!! 

im just about done with the whole CPC... its driving me crazy that things work, then all of a sudden they dont without even touching the system... ill play around and get it working again and then out of the blue..boom.. screw you!

[eto]

I had something similar while developing my internal RAM expansion. In my case, the RAM signals just were not right. Amstrad diagnostics can then no longer properly detect RAM and displays all kind of nonsense like non-existing banks or banks spread across the screen.

However my screens were different so I can't say what particular problem it could be in your case. 

[eto]

Just for the sake of being 100% sure: this is a 464, nothing else is connected to the 464 and you are powering it with 5-5.1V?

[WacKEDmaN]

yup eto.. 464 with only DDI-5 connected.. 5.15v input (4.95v on LS244/LS373)..pulling 1Amp (200mA above normal without DDI-5)

i also still have solder on the fingers of the edge connector, so it makes a really tight connection.. (and yes i checked for shorts between the fingers!)

[SerErris]

I do have the same problem with mine. 

I have it connected directly to my 464 or to the MX4 board. No difference.

Also if you load a picture from video memory into the banks and replay to the screen, you will see some pixels being wrong (not a lot in my case).

I have no other idea than that my RAM chip is bad. 

I will record some screenshots how this looks like, but the RAM test is exactly the same from the DIAG rom. 

[WacKEDmaN]

yeah i tried playing Gryzor last night....
half the levels where all messed up (but still playable all the way to the end..with cheats!)

i really hope the onboard ram chip hasnt gone bad... ill be kicked up the ass so hard! 

im beginning to think all my issues stem from using cheap ass sockets on my board... ill have to replace them all with machined sockets

[eto]

seems my previous post got lost. Not sure what happened.

If there would be any issues, even randomly, with the 512KB, it should be visible. The diagnostics properly test 512KB but then goes mad, when it starts to test RAM banks above 512KB. 

Could you run the ramtest.bas on the attached DSK and share what it reports?

[WacKEDmaN]

seems my previous post got lost. Not sure what happened.

If there would be any issues, even randomly, with the 512KB, it should be visible. The diagnostics properly test 512KB but then goes mad, when it starts to test RAM banks above 512KB. 

Could you run the ramtest.bas on the attached DSK and share what it reports?

pic attached is what in seeing the one you provided

EDIT: restarted and ran it again and got differnt postition of 7E and FE ...looks like its bad :/

[eto]

first: I think I might not have the addressing of the A lines right.

But: nevertheless it should not display 7f or ef

I have a suspicion that it is not properly reacting if the A-lines are being used for RAM areas above 512KB. It should basically just ignore those and continue to deliver data from the 512KB based on D3-D5. But this looks more that something is turned off once the A lines are in use too. Maybe floating data lines? 

Did you ever test the DDI5 with Amsdiag before the current situation? Do you know that it ever behaved differently? 

[WacKEDmaN]

yeah i have seen it pass noels diag a few times... aswell as games like rtype 128 working fine...

..its not looking good..ill have to give it back to mate and have him test it... to rule out my cpc board at least...
it seems to just be the ram.. coz disk access works fine, with both amsdos and parados..otherwise 64k stuff is working fine...

only floating line the the one on the PIO which snapped off around the time the DDI-5 ram started having issues... pin 21 which goes to LK3 and pull up resistor....so it maybe that (but i wouldnt think so..that only sets name surely?)... that chips pins looked quite badly rusted on top.. ive gotta order a replacement....

[WacKEDmaN]

apparently the DDI-5 was a bit flaky for mate before he gave it to me!..so i cant rely on it...
i am thinking the rams toast tho