A5000 repairs

discuss the archimedes & risc pc, peripherals and risc os/risc os on pi
nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 4:34 am

Hi everyone!

First time poster here, so please be gentle :)

Just got into the Acorn machines after reading about them for quite a while, been into Amiga's and Atari's, so thought it was time to branch out and got myself an A5000 as a starting point.

At the moment, it only boots into Supervisor mode. I have tried doing a CMOS reset, but it doesn't work, that is, i don't get a red border that i keep reading about. So had a look at the board, this is what i found:
IMG_2653.JPG
Pins 5 thru 8 of the RTC chip have lifted from the board, its held on only by the first 4 pins, which would explain why i can't perform a CMOS reset(correct me if i'm wrong)

Now i have to order a few parts, and have had a look at the service manual, but your help with the parts numbers, just to make sure i am getting suitable parts.

The RTC chip, this one:
https://au.mouser.com/ProductDetail/NXP ... qWdA%3D%3D

Diode @ D15, this one:
https://au.mouser.com/ProductDetail/Vis ... sV2Q%3D%3D

The tantalum caps, looking at these ones, because i found 1 with a crack right thru, and another one with a crack on the side (internally exploded??)
https://au.mouser.com/ProductDetail/KEM ... 1zJw%3D%3D

If there are any suggestions for better, more appropriate parts, please let me know.

Obviously, there will be a lot of broken traces, which i will need to go through and repair, and replace quite a few SMD resistors and caps.

Also, is it common in these machines, that the SMD caps leak? All the 47uF ones on mine have, so full cap replacement is required too :(

Sorry for the long post, but just want to make sure i get things right the first time around :D

Cheers,
Nick

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 7:25 am

There's not too much evidence of all the SMD caps going on these - but it's worth noting yours seem to have. It'd be good if you could put up a list of replacements for your motherboard version if you do.

Alas the battery damage is all too common- if you search through here you'll find a myriad of threads about fixing these up. You might well find the IDE and floppy have gone as they're up in that area. The fact it boots to anything is good! I'm assuming you've found the schematics?

Myelin did an excellent talk describing his epic repair which you can find on the abug site:
http://abug.org.uk/index.php/2020/08/02 ... l-pearson/

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 7:51 am

danielj wrote:
Wed Sep 16, 2020 7:25 am
There's not too much evidence of all the SMD caps going on these - but it's worth noting yours seem to have. It'd be good if you could put up a list of replacements for your motherboard version if you do.

Alas the battery damage is all too common- if you search through here you'll find a myriad of threads about fixing these up. You might well find the IDE and floppy have gone as they're up in that area. The fact it boots to anything is good! I'm assuming you've found the schematics?

Myelin did an excellent talk describing his epic repair which you can find on the abug site:
http://abug.org.uk/index.php/2020/08/02 ... l-pearson/
The SMD caps are reminiscent of the Amiga 600/1200 caps, not a question of if, but when they'll leak. I guess the A5000 is of the same era, so its possible they might suffer the same fate. To add to my post earlier, there are also a couple of the smaller 10uF SMD caps that have leaked too.

I'll post some more pics of the bad caps at some stage.

And i will definitely post my parts list once i get confirmation from the gurus out there that the parts mentioned earlier are correct. I just want to see it boot up first and get a desktop up :D

Once thats done, i'll focus on the PSU, because that fan!!! So loud!!!

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 9:04 am

The RTC is right, make sure the tant is rated over what the original was voltage-wise, and for the diode, if the part number matches you'll be fine?

Caps - it's more bad batch than era. The A600/1200 are notorious, but other machines notsomuch. Consequently there's not been much evidence of them going in the A30x0/A4000/A5000/RiscPC machines other than a few anecdotal examples so more evidence is a good thing!

Fan-wise, I fitted one of these and it's basically silent: https://www.amazon.co.uk/gp/product/B002IG5GYE

d.

User avatar
flaxcottage
Posts: 4197
Joined: Thu Dec 13, 2012 8:46 pm
Location: Derbyshire
Contact:

Re: A5000 repairs

Post by flaxcottage » Wed Sep 16, 2020 9:06 am

I can help with the fan.

I used this one on mine - https://www.amazon.co.uk/gp/product/B00 ... UTF8&psc=1

It is so quiet I thought it was not working! The hard drives also make a huge noise. I replaced mine with dual CF cards and now the machine is virtually silent.

Good luck with the soldering. If you are methodical these boards can be resurrected.
- John

Image

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 11:36 am

It's perhaps worth noting that the machine will (should!) boot to desktop with DEL held down (hold it down during power-on and keep held until desktop/beep) even without a battery and even without the RTC/CMOS chip present at all. Whilst working on the machine, before replacing lots of parts in this area, get to this state.

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 2:10 pm

@danielj - Thanks for confirming parts, and your right about upping the voltage, might be why they have exploded and cracked.
I'll definitely get photos up in regards to the caps though and also the board number/revision incase others out there might come across the same leaky caps.

And these psu's REALLY need a quiet fan!! Surprised its still spinning though, fair effort for a near 30 year old fan. I recently replaced the fan in my Amiga 2000 psu with a Noctua, and very quiet :)
IanJeffray wrote:
Wed Sep 16, 2020 11:36 am
It's perhaps worth noting that the machine will (should!) boot to desktop with DEL held down (hold it down during power-on and keep held until desktop/beep) even without a battery and even without the RTC/CMOS chip present at all. Whilst working on the machine, before replacing lots of parts in this area, get to this state.
Interesting about the reset. I have tried that, kept it pressed all the way through the flashy floppy lights, then it does a beep, and it just goes to supervisor. So there could be a more fundamental fault?

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 2:31 pm

nikoskon74 wrote:
Wed Sep 16, 2020 2:10 pm
Interesting about the reset. I have tried that, kept it pressed all the way through the flashy floppy lights, then it does a beep, and it just goes to supervisor. So there could be a more fundamental fault?
Does the keyboard work? At all? Does DEL work? Since you have supervisor prompt, you should also be able to try *unplug to see which modules are unplugged - after a DEL-reset, most of them should be alive (specifically including desktop). Try issuing *desktop anyway No RISC OS machine should remain at the supervisor prompt after a DEL-reset - what you've described isn't normal/expected operation.

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 2:38 pm

@ian that's seen quite a bit if the cmos is fubar. No amount of Del pushing will sort it out if the IC can't be read/written to.

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 2:45 pm

IanJeffray wrote:
Wed Sep 16, 2020 2:31 pm
nikoskon74 wrote:
Wed Sep 16, 2020 2:10 pm
Interesting about the reset. I have tried that, kept it pressed all the way through the flashy floppy lights, then it does a beep, and it just goes to supervisor. So there could be a more fundamental fault?
Does the keyboard work? At all? Does DEL work? Since you have supervisor prompt, you should also be able to try *unplug to see which modules are unplugged - after a DEL-reset, most of them should be alive (specifically including desktop). Try issuing *desktop anyway No RISC OS machine should remain at the supervisor prompt after a DEL-reset - what you've described isn't normal/expected operation.
Yes, the keyboard does work, including the DEL key, because i can type and delete, so the keyboard works (when i get to the supervisor prompt anyway), not sure if it is actually registering a keypress while its booting though.

Tried using the UNPLUG command, and there is a long list of unplugged modules, but cannot see the Desktop module because there are so many unplugged, it might be at the top of the list and it goes off screen.

When i try typing desktop, i get "File not found" error number &D6

I have also tried the RMReInint Desktop command. Then when i type Desktop, i says: Error: SWI &400F2 not known (Error Number &1E6)

Could this be a ROM set issue? Maybe one of the 4 roms is faulty? I could burn a new set, but not sure what EEPROM's i should use or where to get the 3.11 rom bin files.

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 2:46 pm

If the ROMs were faulty you'd never hit the supervisor prompt, it'd be protesting CRC failures all over the place. Get the CMOS circuit sorted and you should see things recover.

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 2:52 pm

This then must be the way i need to progress, i'll have to at least install a new CMOS chip, then proceed from there.

Time to start making that order list. I have my work cut out for me :D

Thanks to all who have provided their input here, its great when you can so much help so quickly, especially for a newbie to the Acorn scene :)

I'll post the results once i get the CMOS chip delivered and installed.

Cheers

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 2:59 pm

Good luck - worth having a look here too:
https://www.retro-kit.co.uk/page.cfm/co ... -an-A5000/

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 2:59 pm

danielj wrote:
Wed Sep 16, 2020 2:38 pm
@ian that's seen quite a bit if the cmos is fubar. No amount of Del pushing will sort it out if the IC can't be read/written to.
The CMOS chip doesn't even need to be present when holding DEL though. (Been there, done that, proven it).

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 3:04 pm

danielj wrote:
Wed Sep 16, 2020 2:59 pm
Good luck - worth having a look here too:
https://www.retro-kit.co.uk/page.cfm/co ... -an-A5000/
Good site, good info too, especially with the flashing lights, helped me work out my error code from there :D
Although i have a feeling mine will be messier than his :lol:

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 3:04 pm

nikoskon74 wrote:
Wed Sep 16, 2020 2:45 pm
Could this be a ROM set issue? Maybe one of the 4 roms is faulty?
Nope, othewise you'd get POST failures or worse. RMREinit desktop on its own isn't likely to help, as it'll rely on other modules, as you saw.

If DEL really isn't working, then it could be some failure that's causing the keyboard controller to have issues starting up.

I still suggest NOT putting the CMOS chip / battery in place until you get a clean DEL-based start-up. Moreover, because you'll still nominally have to have a working DEL-based startup to set up that new CMOS chip anyway (Yes technically since you have a prompt you can do a whole-lotta typing to reinsert and re-configure as required, but there's some bits of CMOS you simply can't set that way, and a DEL-startup is the only real way to get a clean setup).

User avatar
danielj
Posts: 8334
Joined: Thu Oct 02, 2008 5:51 pm
Location: Manchester
Contact:

Re: A5000 repairs

Post by danielj » Wed Sep 16, 2020 3:06 pm

Hmm, there are certainly mixed reports on that front (e.g. viewtopic.php?f=16&t=19829&p=277070) but if the keyboard is working, a Del from instant of turn-on until something happens doesn't sort it, and the area around the cmos is toast (which it is) then I'd certainly start with repairing that...


d.

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Wed Sep 16, 2020 3:14 pm

IanJeffray wrote:
Wed Sep 16, 2020 3:04 pm
nikoskon74 wrote:
Wed Sep 16, 2020 2:45 pm
Could this be a ROM set issue? Maybe one of the 4 roms is faulty?
Nope, othewise you'd get POST failures or worse. RMREinit desktop on its own isn't likely to help, as it'll rely on other modules, as you saw.

If DEL really isn't working, then it could be some failure that's causing the keyboard controller to have issues starting up.

I still suggest NOT putting the CMOS chip / battery in place until you get a clean DEL-based start-up. Moreover, because you'll still nominally have to have a working DEL-based startup to set up that new CMOS chip anyway (Yes technically since you have a prompt you can do a whole-lotta typing to reinsert and re-configure as required, but there's some bits of CMOS you simply can't set that way, and a DEL-startup is the only real way to get a clean setup).
Is there a list of dependant modules required for Desktop, for a bare minimum boot? Would it be a good test to try and RMReinit all the unplugged modules and see which ones fail to load/plug in?

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 3:29 pm

nikoskon74 wrote:
Wed Sep 16, 2020 3:14 pm
Is there a list of dependant modules required for Desktop, for a bare minimum boot? Would it be a good test to try and RMReinit all the unplugged modules and see which ones fail to load/plug in?
Nah, you're wasting your time there. Especially because it'd all be forgotten again with no CMOS RAM (the unplug settings are stored on the CMOS RAM...). Plus the configuration settings would all be askew. It's really, honestly a waste of time to try and reset a CMOS RAM chip without a delete-power-on.

User avatar
SarahWalker
Posts: 1303
Joined: Fri Jan 14, 2005 3:56 pm
Contact:

Re: A5000 repairs

Post by SarahWalker » Wed Sep 16, 2020 6:04 pm

Should emphasise that delete-power-on will not work if the CMOS RAM chip is faulty, as is quite likely.

User avatar
myelin
Posts: 951
Joined: Tue Apr 26, 2016 10:17 pm
Location: Mountain View, CA, USA
Contact:

Re: A5000 repairs

Post by myelin » Wed Sep 16, 2020 6:56 pm

IanJeffray -- are you absolutely sure that you've been able to get to the desktop without a CMOS chip installed at all? That's not my experience (or danielj's or SarahWalker).

When I was repairing my A5000, I found that it would boot to Supervisor without the CMOS chip installed (as in, literally cut off the motherboard, not just failing to work), but suffered from the same issue in this thread: most modules unplugged. Here are my notes from the time -- I thought the failure to get to the desktop was because of the MEMC test failing, but it was actually (see the following post) because of the missing CMOS chip.

My understanding of what happens during RISC OS boot is that the CMOS contents are reset, then read back and used. So if there's no CMOS chip, holding down DEL has no effect, and you won't get a desktop boot unless you *insert everything by hand.

Here's the CMOS reset code: https://gitlab.riscosopen.org/RiscOS/So ... Reset#L849

Then the CMOS contents are re-read here: https://gitlab.riscosopen.org/RiscOS/So ... eset#L1836
Last edited by myelin on Wed Sep 16, 2020 7:07 pm, edited 2 times in total.
SW/EE from New Zealand, now in Mountain View, CA, making BBC/Electron hardware projects for fun.
Most interesting: Arcflash, FX2+PiTubeDirect Tube/Cartridge adapter, USB cart interface.

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 7:07 pm

myelin wrote:
Wed Sep 16, 2020 6:56 pm
IanJeffray -- are you absolutely sure that you've been able to get to the desktop without a CMOS chip installed at all?
Yes. Absolutely certain. Just recently I had an A4 where the chip had died. I was surprised. Removing it, and booting with DEL (which was a rotten PITA on the A4 as you have to reconnect the internal keyboard - can't use the external PS2) showed that everything with the machnie was all working fine and it was the CMOS RAM chip that was faulty.

User avatar
myelin
Posts: 951
Joined: Tue Apr 26, 2016 10:17 pm
Location: Mountain View, CA, USA
Contact:

Re: A5000 repairs

Post by myelin » Wed Sep 16, 2020 7:08 pm

I wonder how that possibly worked! Was this with RISC OS 3.11?

I just tried disabling CMOS access in Arculator (adding a "return;" at the start of i2c_change in cmos.c, which simulates getting no response at all from the chip), and doing a boot with DEL held down. This gave me a red screen to indicate the failed CMOS chip, then a black screen with a red border to show that it's attempting to reset the CMOS contents, then it continued to the Supervisor prompt.
SW/EE from New Zealand, now in Mountain View, CA, making BBC/Electron hardware projects for fun.
Most interesting: Arcflash, FX2+PiTubeDirect Tube/Cartridge adapter, USB cart interface.

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 7:14 pm

myelin wrote:
Wed Sep 16, 2020 7:08 pm
I wonder how that possibly worked! Was this with RISC OS 3.11?
RISC OS 3.10. I ran it like this for a week or so, while waiting for the new CMOS chip to arrive and patching-up other issues. How it worked, that's a different matter. YMMV etc etc and feel free to ignore any advice you find on the Internet, etc :lol:

User avatar
myelin
Posts: 951
Joined: Tue Apr 26, 2016 10:17 pm
Location: Mountain View, CA, USA
Contact:

Re: A5000 repairs

Post by myelin » Wed Sep 16, 2020 7:24 pm

I wonder if there's some special magic on the A4 (something else on the I2C bus maybe?) that makes this work, or if you were just very lucky :) If anyone else out there has an A4 they wouldn't mind experimenting on, or perhaps one that probably needs its CMOS chip replacing anyway, it would be very interesting to attempt a reproduction of this!
SW/EE from New Zealand, now in Mountain View, CA, making BBC/Electron hardware projects for fun.
Most interesting: Arcflash, FX2+PiTubeDirect Tube/Cartridge adapter, USB cart interface.

User avatar
IanJeffray
Posts: 181
Joined: Sat Jun 06, 2020 3:50 pm
Contact:

Re: A5000 repairs

Post by IanJeffray » Wed Sep 16, 2020 7:32 pm

Nothing special on the A4. You've got me doubting myself now. But it was such a surprise that the machine even cleanly passed POST without a CMOS chip, that it was memorable. I'll try and give it a whirl on my hack-a-A4000 board.

User avatar
8271
Posts: 88
Joined: Sun May 24, 2020 1:20 pm
Contact:

Re: A5000 repairs

Post by 8271 » Wed Sep 16, 2020 10:13 pm

I have exactly the same fault with my A5000 right now.

Boots to supervisor after reading out the faulty CMOS floppy drive code. Del start-up doesn't clear it.
All the modules are unplugged apart from a few low level modules.

*status shows huge memory allocations for things such as spritearea, screen memory, filer etc. If I rmreinit desktop,filecore, adfs i get an error box but at least it tries to draw the screen (although in mode 0).

I have bought the cmos podule chip from CJE and will cut out the motherboard one. My thoughts are the cmos is completely stuffed, desktop cannot start as it has memory configurations from the cmos that it cannot allocate.
Serial computer hoarder & econetophile
2 x BBC-B, 4 x Master128, Master Viglen, Master Domesday, A310, A4, A4000, RiscPC700, RiscOS_Pi, Econet & Ethernet
A7000 with Econet & Ethernet as AUN Fileserver

philpem
Posts: 524
Joined: Fri Apr 04, 2014 6:42 pm
Contact:

Re: A5000 repairs

Post by philpem » Wed Sep 16, 2020 11:22 pm

myelin wrote:
Wed Sep 16, 2020 7:08 pm
I wonder how that possibly worked! Was this with RISC OS 3.11?

I just tried disabling CMOS access in Arculator (adding a "return;" at the start of i2c_change in cmos.c, which simulates getting no response at all from the chip), and doing a boot with DEL held down. This gave me a red screen to indicate the failed CMOS chip, then a black screen with a red border to show that it's attempting to reset the CMOS contents, then it continued to the Supervisor prompt.
I suspect the failure mode on the CMOS chips isn't that they stop responding (unless the SDA/SCL lines are open). I'd expect some kind of functional failure -- a dead bit or byte in the RAM array, perhaps. As long as that didn't hit something important, the machine would reset and run fine.

The other possibility is that maybe a "no CMOS detected" situation prevents the CMOS RAM clear (delete-reboot) from happening?

EDIT: Just seen Myelin's post. Writing to the CMOS (to reset it) then reading straight back will work - provided the chip can remember what it's been told for a couple of seconds on boot. If you have no CMOS chip then the relevant OSByte will time out without reading anything -- meaning the memory buffer will contain undefined data.
Maybe whatever Arculator initialises the RAM to (all zeroes?) differs from what happens on POR with DRAM -- and that's changing the behaviour. It'd also account for the differences in success rates on physical machines!

nikoskon74
Posts: 15
Joined: Tue Sep 15, 2020 9:12 am
Contact:

Re: A5000 repairs

Post by nikoskon74 » Thu Sep 17, 2020 12:00 am

Thanks for all the posts everyone, this is turning out to be a really good learning experience for me. A lot of great knowledge out there!:)

I tried what @8271 tried, RMReinit desktop, filecore and adfs, but still the same error.

In regards to the CMOS chip, as it currently is, pins 5 thru 8 are not actually connected to the board at all, due to corrosion, pads or totally gone, so the SCL, SDA, /INT & VCC lines are not connected presently.

What i might do, while waiting for parts to come in (still need to place an order), is wire those 4 pins up with wire, as see what happens when it is connected. Not sure if it will work, depending on how bad he CMOS chip is, but it can't hurt to try i suppose. I have a copy of the schematics, so shouldn't be hard to trace the connection points.

But wont be able to do this until i get back home on saturday, but i'll give it a go, and let you know the outcome.

User avatar
DutchAcorn
Posts: 2281
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: A5000 repairs

Post by DutchAcorn » Thu Sep 17, 2020 12:25 pm

You need 5,6 and 8 connected to get the chip to work. Only some of the battery battered A5000s I have seen needed a CMOS chip replacement. Mostly fixing the tracks does it.
Paul

Image

Post Reply

Return to “32-bit acorn hardware”