Extended and modified Atom (was "Faulty Atom")

discussion of games, software, hardware & emulators relating to the Acorn Atom and Acorn System machines.
Post Reply
User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Extended and modified Atom (was "Faulty Atom")

Post by DutchAcorn » Sun Sep 08, 2019 3:19 pm

I have been working on the Atom that I got yesterday (viewtopic.php?f=41&t=5594&start=5370#p247342).


Regrettably it does not work yet. I have replaced the socket for IC31 and also replaced the IC for a known good one but the symptoms have not changed. I also replaced C4 and I’m using a known good PSU. See here for what I get:
https://youtu.be/QWLWe_g-dXk

If I touch the top part of IC31 with my finger the video behaves erratically (as shown in the youtube video).

I checked the clock signal, which is fine.

It seems something is floating. Any ideas on what to try / check?
Last edited by DutchAcorn on Sun Sep 15, 2019 9:49 am, edited 1 time in total.
Paul

Image

User avatar
1024MAK
Posts: 9400
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Faulty Atom

Post by 1024MAK » Sun Sep 08, 2019 5:05 pm

So if the displayed screen data is staying the same (in between the erratic behaviour when you touch IC31), then the 6847 is able to read the same data from the video RAM. That does not mean that the CPU is doing anything, or doing anything useful. It also could be that the CPU can’t access the video RAM.

Do you have a logic probe, logic analyser or oscilloscope? It may be worthwhile testing the various control signals on the CPU to see what, if anything, it is doing.

The erratic behaviour when you touch IC31, looks like either a dry solder joint, or a bad socket contact, or possibly a cracked PCB track. Does it only occur when you press this chip, or does it occur if you press other areas of the board, or flex or tap other areas of the board?

The other possibility is that the circuit is picking up induced mains voltage from you when you touch IC31 via inputs that are floating (and which should be driven) due to an open circuit fault somewhere else on the board.

Mark

User avatar
oss003
Posts: 3131
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Faulty Atom

Post by oss003 » Sun Sep 08, 2019 6:04 pm

There are no 2114's in IC42/43 and you need that at #8000, the text screen.

Greetings
Kees

User avatar
roland
Posts: 3505
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Faulty Atom

Post by roland » Sun Sep 08, 2019 6:18 pm

If you need some extra 2114's just give me a sign
Or even better: install Hoglet's noise killer board.

For now you can borrow some 2114's from the program space, just not IC51/52 (IIRC) because those are the zero page.

Does it beep when you press CTRL+G ?
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Sun Sep 08, 2019 7:49 pm

roland wrote:
Sun Sep 08, 2019 6:18 pm
If you need some extra 2114's just give me a sign
Or even better: install Hoglet's noise killer board.

For now you can borrow some 2114's from the program space, just not IC51/52 (IIRC) because those are the zero page.

Does it beep when you press CTRL+G ?
Thanks for the offer, Roland! Actually it came with a fully loaded video memory, I just removed some when I took the picture. It behaves the same with memory installed in IC42 and 43.
1024MAK wrote:
Sun Sep 08, 2019 5:05 pm
So if the displayed screen data is staying the same (in between the erratic behaviour when you touch IC31), then the 6847 is able to read the same data from the video RAM. That does not mean that the CPU is doing anything, or doing anything useful. It also could be that the CPU can’t access the video RAM.

Do you have a logic probe, logic analyser or oscilloscope? It may be worthwhile testing the various control signals on the CPU to see what, if anything, it is doing.

The erratic behaviour when you touch IC31, looks like either a dry solder joint, or a bad socket contact, or possibly a cracked PCB track. Does it only occur when you press this chip, or does it occur if you press other areas of the board, or flex or tap other areas of the board?

The other possibility is that the circuit is picking up induced mains voltage from you when you touch IC31 via inputs that are floating (and which should be driven) due to an open circuit fault somewhere else on the board.

Mark
Thanks for leading me onto the CPU, Mark! Using another Atom I found that I can replicate the symptoms by removing the CPU.

The CPU does not seem to be very active. The reset on pin 40 goes low when I press break and stays high. But so do all the data and address lines. Clock signals are ok.
Paul

Image

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Sun Sep 08, 2019 7:51 pm

Paul,

Can you post a decent resolution photo of the current configuration you are testing.

Dave
Last edited by hoglet on Sun Sep 08, 2019 7:56 pm, edited 1 time in total.

User avatar
oss003
Posts: 3131
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Faulty Atom

Post by oss003 » Sun Sep 08, 2019 7:53 pm


User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Sun Sep 08, 2019 7:58 pm

hoglet wrote:
Sun Sep 08, 2019 7:51 pm
Paul,

Can you post a decent resolution photo of the current configuration you are testing.

Dave
Attachments
89A85AA5-C272-48C4-A5C9-483F018D75CB.jpeg
0037510E-BE7E-4F0E-8C3D-26F8D85E7F2F.jpeg
Paul

Image

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Sun Sep 08, 2019 8:00 pm

oss003 wrote:
Sun Sep 08, 2019 7:53 pm
Maybe this helps: viewtopic.php?f=44&t=6318&p=60505&hilit ... ion#p60505

Greetings
Kees
Thanks Kees, I'll try those.
Paul

Image

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Sun Sep 08, 2019 8:06 pm

IC23, which does the ROM select decoding, looks like it's been modified significantly.

I'd try to understand what's that's trying to achieve.

Any chance of a close up photo of that area?

Dave

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Sun Sep 08, 2019 8:19 pm

hoglet wrote:
Sun Sep 08, 2019 8:06 pm
IC23, which does the ROM select decoding, looks like it's been modified significantly.

I'd try to understand what's that's trying to achieve.

Any chance of a close up photo of that area?

Dave
Attachments
D61319C5-3A4D-48DF-B2BE-E6051148DC3C.jpeg
45E65BC1-9D59-4B79-8A9E-5B43C9662AA6.jpeg
043CEEE4-6D9A-4B62-9369-666476BF2AD1.jpeg
9F9E0300-CFCD-4AA8-9BCD-BB247145DCA9.jpeg
Paul

Image

User avatar
oss003
Posts: 3131
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Faulty Atom

Post by oss003 » Sun Sep 08, 2019 8:21 pm

Maybe it's better to swap IC23 with a standard LS138 for testing.

Greetings
Kees

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Sun Sep 08, 2019 8:31 pm

I think at the very least you should try to trace/document the obvious modifications to the address decoding around !C23.

It's not a modificiation I've ever seen before.

Dave

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Mon Sep 09, 2019 4:14 am

Tonight I’ve been connecting some dots. Rom selection logic + CPU not processing = wrong OS selected?

So I thought it might be a variation of the bootstrap logic that I have encountered earlier:
viewtopic.php?f=3&t=10789&p=133077#p132767

Trying shift-break gets me this:
F4533838-4E5A-47C3-A1C5-48515ED4896B.jpeg
ACORN ATOM is in lower case but I’m not complaining :D
Paul

Image

User avatar
roland
Posts: 3505
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Faulty Atom

Post by roland » Mon Sep 09, 2019 9:23 am

DutchAcorn wrote:
Mon Sep 09, 2019 4:14 am
ACORN ATOM is in lower case but I’m not complaining :D
Some people did this just for fun. There are also Atoms that only display the word ATOM (inverted) because that freed up some bytes in the reset routine. Also the memory check for RAM at #2900 was often removed (together with the characters "ACORN " - note the space) to make some room for some additional stuff in the reset routine like automatically enable the DOSrom and enable some bank switching software.

However, this is a remarkable situation because you have an original ROM so that is not modified. My best guess is that the extras on the 74LS138 disable that ROM and enable an eprom on one of the memory boards.
Last edited by roland on Mon Sep 09, 2019 9:29 am, edited 1 time in total.
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Mon Sep 09, 2019 9:43 am

roland wrote:
Mon Sep 09, 2019 9:23 am
However, this is a remarkable situation because you have an original ROM so that is not modified. My best guess is that the extras on the 74LS138 disable that ROM and enable an eprom on one of the memory boards.
If you look carefully at IC24 you'll see it has a "F" on it, which is a bug clue.

It also looks like it's chip select pin is benr out, and has been wired to pin 7 of IC23, which is the 74LS138 address decoder output for #Fxxx

So I think the #Fxxx region of the original ROM is disabled, and is overlaid with by IC23.

I would be very interesting to somehow get a dump of the #F000-#FFFF. Or even just #FF00-#FFFF.

Dave
Last edited by hoglet on Mon Sep 09, 2019 9:55 am, edited 2 times in total.

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Mon Sep 09, 2019 9:49 am

Could you run the following program, and take a clear photo?

Code: Select all

10 PRINT $12
20 FOR X=#FF10 TO #FFFF
30 A=?X
40 LINK #F802
50 NEXT
60 END
We can then see whether there are other OS ROM modifications.

Dave

User avatar
Multiwizard
Posts: 1676
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Faulty Atom

Post by Multiwizard » Mon Sep 09, 2019 11:38 am

hoglet wrote:
Sun Sep 08, 2019 8:06 pm
IC23, which does the ROM select decoding, looks like it's been modified significantly.

I'd try to understand what's that's trying to achieve.

Any chance of a close up photo of that area?

Dave
Hi,

these modifications looks almost exactly the same as my modified CP/M IC23... :shock:


Greetings, Wim... :-)
Attachments
P1290579.JPG
P1290582.JPG

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Mon Sep 09, 2019 5:38 pm

hoglet wrote:
Mon Sep 09, 2019 9:49 am
Could you run the following program, and take a clear photo?

Code: Select all

10 PRINT $12
20 FOR X=#FF10 TO #FFFF
30 A=?X
40 LINK #F802
50 NEXT
60 END
We can then see whether there are other OS ROM modifications.

Dave
This is the result:
Attachments
A047BA5D-16E0-4729-846F-F01C140ECBAB.jpeg
Paul

Image

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Mon Sep 09, 2019 6:58 pm

Multiwizard wrote:
Mon Sep 09, 2019 11:38 am
Hi,

these modifications looks almost exactly the same as my modified CP/M IC23... :shock:


Greetings, Wim... :-)
Do you know what the modification in your computer is for?
Paul

Image

User avatar
Multiwizard
Posts: 1676
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Faulty Atom

Post by Multiwizard » Mon Sep 09, 2019 7:16 pm

DutchAcorn wrote:
Mon Sep 09, 2019 6:58 pm
Multiwizard wrote:
Mon Sep 09, 2019 11:38 am
Hi,

these modifications looks almost exactly the same as my modified CP/M IC23... :shock:


Greetings, Wim... :-)
Do you know what the modification in your computer is for?
This modification had something to do with the switch board...

And if the switchboard is working it needs a different FP rom too,,, FP 1000 or something.


Greetings, Wim... :-)

Edit, I found a pic (from 2016) too:
Attachments
12322405_10205495857772093_3450046030593682311_o.jpg
Last edited by Multiwizard on Mon Sep 09, 2019 7:57 pm, edited 1 time in total.

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Mon Sep 09, 2019 7:17 pm

Hi Paul,

Thanks for posting that screen shot.

Here's a quick disassembly of the updated reset code, which is quite interesting (and a clever optimization).

Code: Select all

FF3F   AD FF 9F   LDA $9FFF
FF42   29 4F      AND #$4F
FF44   8D FF 9F   STA $9FFF
FF47   8D FF BF   STA $BFFF
FF4A   AD 01 B0   LDA $B001
FF4D   10 03      BPL $FF52
FF4F   6C FC 1F   JMP ($1FFC)
FF52   A2 33      LDX #$33
FF54   A9 FF      LDA #$FF
FF56   9D EB 02   STA $02EB,X
FF59   BD 98 FF   LDA $FF98,X
FF5C   9D 02 B0   STA $B002,X
FF5F   BD 9A FF   LDA $FF9A,X
FF62   9D 04 02   STA $0204,X
FF65   CA         DEX
FF66   10 EC      BPL $FF54
FF68   9A         TXS
FF69   E8         INX
FF6A   86 EA      STX $EA
FF6C   86 E1      STX $E1
FF6E   86 E7      STX $E7
FF70   A9 0A      LDA #$0A
FF72   85 FE      STA $FE
FF74   20 D1 F7   JSR $F7D1
FF77              EQUB $06,$0C,$0F
FF7A              EQUS "acorn atom"
FF84              EQUB $0A,$0A.$0D
FF87   A9 82      LDA #$82
FF89   58         CLI
FF8A   8D 01 29   STA $2901
FF8D   CD 01 29   CMP $2901
FF90   D0 03      BNE $FF95
FF92   4C B2 C2   JMP $C2B2
FF95   4C B4 C2   JMP $C2B4
FF98   07         EQUB $07,$8A
The original reset code can be seen here:
http://www.acornatom.nl/atom_handleidin ... m/f000.txt

The extra code that has been squeezed in is:

Code: Select all

FF3F   AD FF 9F   LDA $9FFF
FF42   29 4F      AND #$4F
FF44   8D FF 9F   STA $9FFF
FF47   8D FF BF   STA $BFFF
FF4A   AD 01 B0   LDA $B001
FF4D   10 03      BPL $FF52
FF4F   6C FC 1F   JMP ($1FFC)
It basically hands control over to a ROM at #1xxx unless the shift key is pressed.

This is consistent, I think, with what you have observed.

Dave
Last edited by hoglet on Mon Sep 09, 2019 7:17 pm, edited 1 time in total.

User avatar
hoglet
Posts: 8692
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Faulty Atom

Post by hoglet » Mon Sep 09, 2019 9:15 pm

Last edited by hoglet on Tue Sep 10, 2019 6:06 am, edited 1 time in total.

User avatar
Multiwizard
Posts: 1676
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Faulty Atom

Post by Multiwizard » Tue Sep 10, 2019 6:21 am

Thanks Dave =D> ,

I was already looking for these pages (again) but couldn't find them... :(


Greetings, Wim... :-)
Last edited by Multiwizard on Tue Sep 10, 2019 6:22 am, edited 1 time in total.

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Tue Sep 10, 2019 8:17 am

hoglet wrote:
Mon Sep 09, 2019 7:17 pm
...
It basically hands control over to a ROM at #1xxx unless the shift key is pressed.

This is consistent, I think, with what you have observed.

Dave
Thanks Dave! Good to know that it behaves as intended.
Paul

Image

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Faulty Atom

Post by DutchAcorn » Fri Sep 13, 2019 9:06 pm

I had some time to look into stuff. Thanks to the documentation Multiwizard pointed to I found out how to get the RAM/ROM board to work.

Also found a disc (attached) with alternate OS images. Loading one of them (SOSHS3) at #1000 and pressing break (without shift) gets me this:
atom bootstrap.JPG
So that works as Dave explained from the disassembly.
Attachments
Disc 9.zip
(42.78 KiB) Downloaded 5 times
Paul

Image

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Extended and modified Atom (was "Faulty Atom")

Post by DutchAcorn » Sun Sep 15, 2019 9:58 am

The modifications and software are described in this article in Acorn Nieuws (in Dutch). SOS stands for Switch board Operating System and was designed to be able to access all the commands in the support roms that are on the ram/rom board without manually having to select the specific rom (something that the BBC does as standard).

The extra 9 pin socket turns out to be a joystick interface.

Still to find out: the Big Benny RTC; I have not managed to get a time out of that yet...
Paul

Image

User avatar
CMcDougall
Posts: 6865
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Extended and modified Atom (was "Faulty Atom")

Post by CMcDougall » Sun Sep 15, 2019 11:10 am

/\ benny needs 2x1.5v AA batts to keep time, but think you will be using its port for a AtoMMC !
ImageImageImage

User avatar
DutchAcorn
Posts: 2140
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Extended and modified Atom (was "Faulty Atom")

Post by DutchAcorn » Sun Sep 15, 2019 8:17 pm

CMcDougall wrote:
Sun Sep 15, 2019 11:10 am
/\ benny needs 2x1.5v AA batts to keep time, but think you will be using its port for a AtoMMC !
Check, got the batteries installed. But I only read the first page of the article describing the RTC board so I did not have the actual code to set the clock :oops:

After typing in the code on the second page I found the RTC works perfectly and keeps time. :D
Paul

Image

User avatar
roland
Posts: 3505
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Extended and modified Atom (was "Faulty Atom")

Post by roland » Sun Sep 15, 2019 9:19 pm

=D>

Big Benny is a nice extension for the Atom and it works quite well. I even added a PL8 compatible connector on my Atom2k15/2k18 extension boards for Big Benny and I got it running on those models as well. It's a bit complicated to read and write the clock IC but with the software in Atom Nieuws you can get it working.

There is also a program that makes it work with interrupts so the time is always on screen. I wrote that a long long time ago....

If there are other people that want a Big Benny ... I might have a few ones left. Just drop me a PM when interested.
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

Post Reply