Amiga 3000 Upgraded with 32 Zip Chips

[CrashSerious]

OK, I've recently purchased a working 3000 and then bought 32 of the proper ZIP memory chips to place into the sockets on teh Motherboard.

I moved the jumper for Bank 0 to tell it there are ZIps and not Dips in Bank 0. I've removed the DIPs in Bank 0 and I have double (and triple) checked the zips appear to be inserted properly (Pin 1 to the left looking from the front of the 3000.)

When I boot into Workbench (from floppy), it shows 2Mb chip RAM (ish) and 0Mb of other memory.

What could it be? Have I missed something obvious?

 

[Nelson]

Maybe taking a look at all the Jumpers will help? http://www.l8r.net/technical/t-3000-jumpers.shtml

 

[DofD]

Also, take it step by step. Install 2MB of ZIPs at a time. The ZIP sockets may also have oxidation on them, causing bad connections, and should be cleaned. Check for bent or improperly inserted ZIPs. ZIPs are notorious for not going in properly. A couple ZIPs not seated properly has caused me issues in the past before I went with a ZIP to SIMM adapter.

 

[CrashSerious]

Thanks for the replies!

I have deduced, by chip swapping, that 100% of the RAM I bought is OK. They are also all the same chip numbers, so same type.

I can get 4Mb in both Bank 0 and Bank 1 working for a total of 8Mb. I have also swapped the other 8 into bank 1 in two different passes and each set of 8 that I was unsure about are 100% operational. Seems like when I populate Bank 2 is when this occurs. (Since Bank 2 won't populate, I haven't tried Bank 3-- I figured I would go a bank at a time to deduce where the fault was)

Since I can get 8Mb of 1Mbx4 Zips working, without teh DIP chips in their place... I assume the jumpers are right. (I've also visually inspected them after I set them to quadruple check)

The Weird thing is that when Bank 2 is populated, there's 0Mb of FAST RAM shown AND by accident on a whim I noticed that SysInfo reports a DRASTIC speed decrease. (29% of a A 3000/25, when it IS an A3000/25)

Going to have to look more closely at bank 2, I guess. I have a battery coming from AmigaKit I bought off Ebay-- Once it arrives I will take out the Motherboard and install it and reflow the solder joints on the ZIP Sockets plus inspect them.

What would you suggest doing to clean the sockets?

- - - Updated - - -

Maybe taking a look at all the Jumpers will help? http://www.l8r.net/technical/t-3000-jumpers.shtml

Good information, Thank you.

This is what I believe to be an early model, it had the Super Kickstart ROMS in it. There is no J853. But, maybe I should check the RAMSEY REV? (Checked... it is an -04, if that matters... off to research that)

 

[DofD]

That's good progress in that all the ZIPs work and banks 0 and 1 are okay. So something is certainly not right with bank 2. It is a good idea to check the solder joints on the ZIP sockets. A broken solder joint will cause issues. What revision is the motherboard? From what you describe it sounds like the ZIPs in bank 2 are being seen, but can't be initialized and added to the system. The performance drop may be due to the system trying to initialize bank 2 over and over as it sees the memory but errors when trying to add it. Off the top of my head would be a broken solder joint, or a broken ZIP socket, or maybe a broken trace. Cleaning the ZIP sockets is a pain. Generally inserting and removing a chip from the socket a few times will wear any light oxidation off. A cleaning with IPA and a tooth brush shouldn't hurt. Another option is to get hold of a set of Diag ROMs and see what they report on the memory tests. I'd also give bank 3 a try to make sure that is working okay as well.

 

[CrashSerious]

Been a bit since I could report more progress.

Thank you for the response. I actually did a bit of all that, though my tip was dirty and needing replaced so I stopped trying to reflow the joints, I did try IPA and using the chips to clean it while rocking them in and out with the IPA.

This bank does look dull so a re flow is definitely in this systems future for Bank 2. The system did have minor battery leaking (just a dime (1.5cm) sized spot on top of the board) but that was neutralized and cleaned. Nothing else but that plane (ground) was impacted and I have a hard time thinking it jumped across the board and flipped to the the bottom to the Sockets.

The sockets are in good shape themselves, I bought it from the guy who bought it new and it likely never had anything in the Sockets previously.

I have Diagrom downloaded and a couple 27C400's waiting to receive it as well, but I'm going on vacation for a week and with it back together 95%, I'm going to do that after I get back-- so I don't forget where I left the screws/etc.

As for Bank 3, I placed chips in banks 0, 1, and 3 and just saw 8 MB... but I wasn't sure they would be registered or not with Bank 2 empty.

The MotherBoard is a revision 8.9, IIRC. I'll verify once I take it apart again.

That's good progress in that all the ZIPs work and banks 0 and 1 are okay. So something is certainly not right with bank 2. It is a good idea to check the solder joints on the ZIP sockets. A broken solder joint will cause issues. What revision is the motherboard? From what you describe it sounds like the ZIPs in bank 2 are being seen, but can't be initialized and added to the system. The performance drop may be due to the system trying to initialize bank 2 over and over as it sees the memory but errors when trying to add it. Off the top of my head would be a broken solder joint, or a broken ZIP socket, or maybe a broken trace. Cleaning the ZIP sockets is a pain. Generally inserting and removing a chip from the socket a few times will wear any light oxidation off. A cleaning with IPA and a tooth brush shouldn't hurt. Another option is to get hold of a set of Diag ROMs and see what they report on the memory tests. I'd also give bank 3 a try to make sure that is working okay as well.

 

[DofD]

Have a good vacation there!

When you get back, take a good look at the solder points on the ZIP sockets. I was reading some reports where people have found that not all the connections on the ZIP sockets were actually soldered.

Been a bit since I could report more progress.

Thank you for the response. I actually did a bit of all that, though my tip was dirty and needing replaced so I stopped trying to reflow the joints, I did try IPA and using the chips to clean it while rocking them in and out with the IPA.

This bank does look dull so a re flow is definitely in this systems future for Bank 2. The system did have minor battery leaking (just a dime (1.5cm) sized spot on top of the board) but that was neutralized and cleaned. Nothing else but that plane (ground) was impacted and I have a hard time thinking it jumped across the board and flipped to the the bottom to the Sockets.

The sockets are in good shape themselves, I bought it from the guy who bought it new and it likely never had anything in the Sockets previously.

I have Diagrom downloaded and a couple 27C400's waiting to receive it as well, but I'm going on vacation for a week and with it back together 95%, I'm going to do that after I get back-- so I don't forget where I left the screws/etc.

As for Bank 3, I placed chips in banks 0, 1, and 3 and just saw 8 MB... but I wasn't sure they would be registered or not with Bank 2 empty.

The MotherBoard is a revision 8.9, IIRC. I'll verify once I take it apart again.

 

[CrashSerious]

So I thought I would post a resolution to my own thread for future generations.

I was able to get my 3000 working 100% today with 16Mb of Fast Ram (Zips), 2Mb of chip RAM, and able to install OS 3.1.4 with my Kickstart 3.1.4 purchase from July.

What was happening:

- I couldn't get more than 8Mb of FAST RAM working in the system. The moment I put in bank 2 (or 3) the system wouldn't show more RAM or lock up.
- I couldn't install OS 3.1.4 to teh SCSI (or IDE with a AT BUS 2008 using a (several actually) CF card). I would get a 80000004 error at somewhat random places in the install.
- The install was attempted on several SCSI Drives as well as a SCSI2SD I bought because SCSI Drives were about the same price.
- New SCSI Cables were purchased and nothing seemed to work
- Occasionally I would get a red screen upon start up, but it woudl go away if I powered off and back on.
- I read a million posts it seems, some pointed to RAM issues. This made sense since I had that "RAM" problem, but moving chips from bank to bank always yeilded teh same results... 8Mb recognized.
- I could copy from df0: to Ram: 100% everytime. When I wrote large chunks to teh HDD, things went south.

What Fixed it?

One or both of two things....
- I erased my 27C400s to replace 2.04 on them to see if I still had teh issue under 2.04. But I noticed that one of them didn't erase while the other did (they were in there together). I erased it again, same thing. I did this a couple times. Finally I was able to get it erased, but not before I just grabbed another 27C400 to burn the 3.1.4 ROM to. Threw away the suspect ROM since the error was always in the same place. Bit Rot?
- I enabled SCSI Reselection with the HDTools that came with 3.1.4--- I don't think this was the fix though.

When I replaced the ROMs--- EVERYTHING WORKED. 16Mb of RAM, installation, everything!

So... double check your writes.