stardot.org.uk

I finally pushed my work to GitHub: https://github.com/myelin/acorn_postbox

It's a bit messy; I'll see if I can document the connections a little better in time, but hopefully this is enough for Andy to try it out

Another update... I have now spent way too much time learning about the IOEB and trying to wrap my head around the 12MHz/8MHz IO system in the A5000. I have been unable to figure out what's wrong with the IO system, unfortunately. I've soldered test points to all the buffer lines (Wbl* and Wbe* for ...

Congratulations! It's a good feeling when that desktop appears!

Unfortunately you won't get the second part of the ROM checksum error test unless you have a POST box connected, but there's an easy way to see if any of the address lines are bad, if you have a scope or logic analyzer: When an ARM2 or ARM3 chip is in reset, it'll execute NOPs and page through its e...

The one in my A310 (which has an original MEMC) is 0276 032-01. I see two of those in Moon_Man’s first pic. I'm poking around to try to find what the 0283 041-03 chips are. No luck so far. RO2 ROMs are 0283,022-01 through 0283,025-01 Pre-IOEB A5000 PALs are: 0292,022 0292,023 0292,024 0292,025 0292,...

Updated boards #1 and #3, which both now boot into RO3 and can access a floppy disk! This means the next step is to make a dealer test disc and see if they pass the tests

Stay well, Dave! Sounds like a very sensible plan.

I'm very interested in the RISC OS ROMs and the VLSI chips. What would you consider a reasonable offer for that lot?

Thanks! I just updated the posts above for boards #2 and #3: board #2 now enters the self-test but crashes early on, and board #2 passes the test and is looking really promising.

It disappeared from there for a couple of years, and returned in late 2018 once Raspberry Pi 3 support was added: https://www.riscosopen.org/news/articles/2018/10/25/at-last-risc-os-is-back-on-noobs There's a super active thread on the RISC OS Open forum about porting to multicore CPUs: https://www....

Now where's some clever bod who can make a RISC OS multi-OS board when you need one ?? It's on the to-do list for Arcflash to support A7000/Risc PC; right now I have a working adapter board, and have managed to run RO3.5/RO3.6/RO3.7/RO4.02 off it in my Risc PC, although I haven't gotten the bootloa...

Here's the third board: serial number AV-061-9603. This one now boots into RISC OS 3, and is looking good so far! This board was the mystery of the bunch. At first I thought it had been cannibalized for parts, as so many bits were missing, but then I realized that it was actually in the early stages...

Here's the second board: serial number AV-120-6394. Working on this one still. The best I've gotten is a partial POST, crashing during the ROM checksum test. This one looked okay at first glance, and came with ROMs and the keyboard/mouse controller, but the ROM sockets fell apart when I took the ROM...

Here's the first board: serial number AV-060-2489. This one now boots into RISC OS 3 and everything is working so far! When I got it, it looked in terrible condition. The whole left side of the board is in pretty bad shape, but after replacing the ROM sockets it started up to a red POST failure scre...

I came into possession of three A3000 motherboards in varying states of repair in late 2018, thanks to Mark Haysman of Retroclinic. I did a bit of work on them back then, replacing rom sockets and NVRAM chips, but never actually got any of them fully working. Thanks to philpem’s work on the POST box...

I have a 128MB SIMM that registers as 128MB in my Risc PC but only 64MB (or sometimes nothing at all) in my A7000... I assume bad contacts somewhere!

Nice work!

Haven't touched it in a while; need to make a v2 board with a few extra resistors, or possibly just build some more v1 boards with the resistors soldered on top, given that the buffers I used have been end-of-lifed (argh) by TI due to one of their suppliers EOLing the BGA substrate they used. It wor...

Just a thought but I’m assuming fitting extra RAM off the motherboard still leaves the onboard RAM connected to the address and data bus and just disconnects the enable lines to the onboard RAM. Could the onboard RAM still be loading the address/ data bus weirdly and causing an issue? Steve This wa...

On my A5000 it was caused by four bad chips...not IC59 but the HC75 next to it, and also three of the RAM chips. Please can you advise how you established RAM chips had failed? Using a makeshift POST debug adapter that lets you read out the detailed info reported by the power-on self test. In my ca...

lcww1 wrote: ↑

Wed Feb 26, 2020 11:16 pm

That's a great photo! I think that it captures something of the spirit of this heroic endeavour!

I forwarded this on to my wife, who caught this shot -- she says thank you And that Titanium Cyborg cover is amazing!

This one? https://robotdyn.com/samd21-m0-mini.html

That has an ATSAMD21G18 on board, which will absolutely do... I've checked the pinout and it looks all good except for PA28 not being brought out, but it'll be easy enough to swap that onto another pin, so not a problem.

Here's what the POST output looks like on an A7000 with bad SRAM but no other POST faults: SELFTEST R2.05 ROM : DRAM 0: 00 MByte Data : Data-F FFFFFFFF DRAM 1: 00 MByte Data : Data-F FFFFFFFF DRAM 2: 04 MByte Data : Addrs : Byte : DRAM 3: 00 MByte Data : Data-F FFFFFFFF DRAM 04 MByte IOMD : IOMD : 5...

I got my CPLD breakout board (same as yours) wired up, found an 8MHz crystal, and got that oscillating using 330R resistors and a 10nF cap in the end. Note that you can get various clocks from the expansion ports... if you have a backplane, there's a 2MHz clock on C27 on each podule slot, and an 8M...

IC59 is on the A300 board, not the A3000. The closest equivalent functionality is in IC28 in the A3000. My guess would still be track damage rather than chip damage. Work through the circuit diagram checking all the connections are still ok. Seconding this... the RAM control line failure error can ...

Some interesting results... After replacing IC57 with a 1Mx4 chip from my 4MB SIMM, I got 0F000000 for the DRAM test, which means that D31-28 were now OK, but after replacing IC53 as well, the error changed to 80000000 (D31 bad). I looked at all the chips it could possibly be, and eventually suspect...

Interested in the RISC PC 486 card (and of course the ARM3 upgrade, but that appears long gone!)

Will PM

One chip removed! The Chip Quik allot was magical — didn’t even need hot air. Just squeezed out some flux, set my soldering iron to 400F, added Chip Quik, went around with the soldering iron again, and picked up the chip. Here we are now, with one missing DRAM chip and all my bodge wires: 56C13595-1...

Just ordered some - thanks! I’m going to practice first on a 16MB SIMM I don’t need too... I really don’t want to do any more damage to that part of the A5000 than necessary

Haven't had a chance to try this yet, but I realized I happen to have a 4MB SIMM containing eight MT4C400WDJ-6 chips with the same weird SOJ footprint as the RAM on the A5000, which are likely to be 1Mx4 60ns, i.e. suitable for the bottom memory row on the A5000. Time to give the hot air station a t...