• Please review our updated Terms and Rules here

Acorn A3000 Gamble

mbliss11 said:
I see what you are saying though. That would be on the fifth sequence and not the sixth. That would be a VIDC related error and actually per the documentation would explain why it is able to post. I am going to triple confirm it is the 6th series of flashes and not the fifth when I get home this evening.
Click to expand...
Yes, just to start with the basics. If the post code is really related to RAM errors and you've already sorted out the 1MB expansion as a possible culprit, that would mean that something is bad with the motherboard RAM subsystem; but not enough to avoid booting to OS. Sadly, there's no RiscOS version of "memtest86"...

Have you replaced the four ROM sockets?. If not, I'm still thinking that you really need to do it ASAP, and once cleared, check all data and address lines running all over the place. I can clearly see in your pics that there's a lot of corrosion in the area. A broken trace (or a bridged one) would create havoc; but maybe it's a more subtle problem, these leaks are conductive so who knows...
 
jltursan said:
Yes, just to start with the basics. If the post code is really related to RAM errors and you've already sorted out the 1MB expansion as a possible culprit, that would mean that something is bad with the motherboard RAM subsystem; but not enough to avoid booting to OS. Sadly, there's no RiscOS version of "memtest86"...

Have you replaced the four ROM sockets?. If not, I'm still thinking that you really need to do it ASAP, and once cleared, check all data and address lines running all over the place. I can clearly see in your pics that there's a lot of corrosion in the area. A broken trace (or a bridged one) would create havoc; but maybe it's a more subtle problem, these leaks are conductive so who knows...
Click to expand...
Yea I confirmed last night it was on the 6th sequence and not the 5th.

Yep all sockets in that area have been removed and replaced and I toned out all of the data lines there were a few traces I needed to repair on the board by the roms but none of the rom lines themselves those were all ok.

I have a thread on stardot as well they are helping with this too. Suggestions are looking at ic28 the ram subsystem as well. Someone had written a memory test program but unfortunately it only works on systems with 4 Megs or more.

 
4meg ram upgrade came in the mail. I threw it into the A3000 and am still presented with the same error code. I probed around IC28 and even removed it to check if it was bad I had ordered a few spares just in case. Tested fine and I toned everything out back to the onboard ram, expansion ram pins and to the MEMC and ARM chip. I can run the acorn memory scanner now mem14 and I just have that on a loop it does a quick physical test and then logical. So far it has done several passes. I am just going to let it run for an hour or so and see what happens.
 
Weird. If I'm not wrong, the 4MB upgrade disables the onboard RAM so it means that the whole RAM must be good and the "problem" is in the RAM glue logic; but where?, RAM is testing good tho, why RiscOS is returning this warning?, what kind of test is implemented?
 
I actually removed the motherboard ram as well as it was suggested it could have been bad and messing with the bus even with the 4 Meg upgrade. Still persists. Was also suggested it could have something to do with the ROM latches. I removed them off the board as well and one did test bad on my t84 after a series of tests 10 each. It’s possible it could be a failing chip so I ordered replacements as well.

I ran star fighter 3000 on it for hours over the weekend ran without a hitch
 

Attachments

  • 6ED09DA3-D9CF-4A43-AEFB-F47CDA0E51E7.jpeg
    6ED09DA3-D9CF-4A43-AEFB-F47CDA0E51E7.jpeg
    2.3 MB · Views: 6
  • 32E56A54-E3FA-4414-AA10-25250A4BADC3.jpeg
    32E56A54-E3FA-4414-AA10-25250A4BADC3.jpeg
    2 MB · Views: 6
With the original latches. I am waiting for new ones to come in. I did socket them so I can swap them in easily. I did change the order they were in originally on the board after I threw them in sockets but no change.
 

Attachments

  • 4E7BA191-2C6D-464B-9C90-89E81D58F53F.jpeg
    4E7BA191-2C6D-464B-9C90-89E81D58F53F.jpeg
    2.3 MB · Views: 6
And the game run for hours but the RiscOS test spots a difference every single boot?. All in all, seems amazing to me, how can an IC have so marginal failure?. Or...maybe, games are not usually using the feature that makes the test fail...no idea tho.

As I've run out of ideas, replacing the latches (moreover when one proved to be unstable) is the best option. Let's see...
 
jltursan said:
And the game run for hours but the RiscOS test spots a difference every single boot?. All in all, seems amazing to me, how can an IC have so marginal failure?. Or...maybe, games are not usually using the feature that makes the test fail...no idea tho.

As I've run out of ideas, replacing the latches (moreover when one proved to be unstable) is the best option. Let's see...
Click to expand...
Yea I am at a loss too. I had tried to dig up some more info on what those errors mean and what the tests are actually doing but didn’t come up with anything useful. It could be that the MEMC is just failing? I hope not otherwise I would need to source another one which would mean getting another a3000.

Yea I am going to wait for the new latches to come in and put them in there. If it still shows the error I think I may just close it up and just enjoy using it while it is still functional and hope for the best. The only other thing I could think is that I have bad roms but unlikely. I suppose I could test those why not have tested everything else haha
 
Latches came in along with a tru-mouse adapter. Unfortunately switching those latches out still results in same result. General consensus on other forum is the MMC is at fault. Good news is I found a replacement online at a reasonable price. Will have to put a 68pin socket in its place . Not a fan of surface mount components but oh well. I will also only do this if the machine fails. I’m going to continue using it in it current state since I can still run software on it and appears to work just fine.

Tru mouse adapter didn’t work originally. I had not toned out the mouse port as I was waiting to have something to test it with. Needless to say had to do some more wire runs. The pins on the mouse port are tiny and it looks a mess but it’s stable and works. I can use a real mouse now :)

For now I am closing it up once more and will use it and enjoy it until I can’t. Should that MMC come out I’ll definitely update the thread.

Any game suggestions??
 

Attachments

  • D6B58671-DF0C-418B-95D3-4B5D632AEA6C.jpeg
    D6B58671-DF0C-418B-95D3-4B5D632AEA6C.jpeg
    2 MB · Views: 10
Being the masochist I am I opened her up once again to re-repair a bodge from the mouse port only to discover keyboard membrane issues. I had ordered the keyboard repair kit Tim O had suggested earlier in the thread. Turned out great and I just desoldered the original connectors in place of the pin headers that came with the kit. Hate these membrane keyboard connectors.
 

Attachments

  • 5AFD55E7-B7D2-460B-94B7-9892BF98D7E4.jpeg
    5AFD55E7-B7D2-460B-94B7-9892BF98D7E4.jpeg
    1.7 MB · Views: 10
  • 9D9C86A7-56A8-4BCD-A7F5-EF00E275A1DD.jpeg
    9D9C86A7-56A8-4BCD-A7F5-EF00E275A1DD.jpeg
    1.6 MB · Views: 9
The original connectors did have corrosion on them originally they cleaned up decently after soaking in vinegar and scrubbing with a tooth brush and IPA. The problem with the connectors is it is friction based and the more you remove the keyboard the contact starts scrapping off. Mine was in decent shape originally but several of the tracks started getting pulled away after reseating it several times.

I have carbon conductive paint but that has a higher resistance than the silver paint. The silver paint came in a pen like applicator but the end got gunked up so I just use a little tooth pick and dip it in the paint to apply
 
I see. In that case, maybe the old ZX Spectrum trick of cutting some of the end of the connector could work...it there's enough cable length of course; can't remember if the A3000 cables were long enough.

Anyhow, your replacement will last forever probably :)
 
You must log in or register to reply here.
Back
Top