New to board A3000 question

[mray]

So I thought I would update the ram problem to those that viewed and replied to my posts. I Have included a couple of photos To show the results or the ramtest program I ran. I might mention this a great little program, (highly recommend). The photos
reveal I have several problems with the most notable the address lines on ram bank 0 bad. With possible actual bad ram at other locations. I believe this was mostly caused by myself by firmly seating the existing ram (OCD). So I will be diving into this more deeply.. And thoughts are welcomed.
Thank you

[cdh]

It seems pretty consistent from your address test output that all eight ZIPs in your Bank 0 are having issues with A0, A1, A2, A7, A8, A9. This suggests to me that either there is a short or open on the motherboard, or one of the ZIPs is shorting some of the address lines. You might try removing half of the ZIPs in bank 0 to see if you can make the problem go away. If that doesn't work, try removing the other half and just insert one ZIP at a time. I use a magnifier and overhead light when inserting ZIP chips, so I can verify all pins are going in to the proper socket before I push. It helps to line up one row (ZIP chip at an angle), then rock it slightly to line up the second row of pins.

As for testing ZIPs, I'd always start with the data line test to verify the ZIP chips are working at a basic level. This test is the simplest and can usually point out chips with data line or RAS/CAS issues.

[mray]

I thought I might finish this thread. I finally got my system back up and seemingly
populated with approximately 16MB of ram. It turned out to be a real problem
that would not have been solved without the program ziptest. So if I would have a recommendation for upgrading ram. Download the Ziptest first, run the program check if there are any problems, sort them out then start with the upgrade . install the upgrade ram one bank at a time rerun the program. If no problems, continue upgrading in steps.
Considering the lightweight,cable length, and small form factor of the 3000 how would one physically do this in steps. I am including a picture how I made a cheap testbed that would make this practical. I purchased a cheap wooden crate at a local home improvement store, cut it down, and screwed the powersupply and drives to it. Allowing it to bridge the chassis.

[mray]

Cdh, thank you for the reply, I have at some level completed the upgrade I do wish to agree with your accessment of the data and I may get back into the motherboard question
at some point . While trying to sort this out The system went hard down, very disappointing. The system would not boot up with a could not find ram error. Being frustrated at that point I pulled all zips off the board . A couple of days afterward I pulled
256X4 DIP ram off a 2091 pcb and installed getting the system back up with one MB ram.
I ordered from Hong Kong 1MBx4 DIPs allowing the system to boot with 4MB ram
Started reinstalling slowly the zips while running ZIPTEST correcting problems during the process I ended up with banks 1-3 full and 16MB ram total, including the installed DIPS. I reran the Ziptest it still shows address problems on bank 0. So I am left with
partial answers to what happened and why, If this was preexisting, how did It ever work, How does The Ziptest work in respect to the DIP ram? Before I started reinstalling the Zips I attempted to OHM check and scope all the lines ( could have missed something due to the vast number of lines to check also did not check for line to line short just start to stop continuity ). I am still looking at the prints and thinking about
the issues .
Thank you again for the reply
I am open to any further thoughts or recommendations
My email is this: mraysommers@comcast.net

[mray]

I thought I might add to the results of what should have been an easy ram upgrade
Since I am still very new to the 3000 It did not realize that as a 1.4 alpha boot rom
unit could effect the ziptest progam The alpha 3000 uses approximately 512K ram
to load the KS os so that would explain why It would show less than 16 meg and
faults during the ziptest on bank zero

[cdh]

When I wrote ZIPTest, I never thought of the fast memory being used to hold the kickstart ROM, such as is done with the MMU in the early Amiga 3000 units. If I ever come across someone locally (California SF Bay Area) who has one of these, I'll ask to borrow the Amiga and adapt ZIPTest to be able to detect and handle testing the remapped memory.

If anyone has the inclination to do this, check out the amask_to_address() function -- it'll need to be changed to detect the correct address (relative to the ROM address or FASTMEM_TOP).