Atom repair progress

discussion of games, software, hardware & emulators relating to the Acorn Atom and Acorn System machines.
User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Atom repair progress

Post by jms2 » Sat Nov 09, 2013 8:50 pm

I have replaced all the electrolytic caps in my Atom, and both crystals X1 and X2. I then decided to fire it up for the first time and... it kind of works. :D Actually it has the problem that Roland told me about, that of random characters appearing on the screen.

Initially it was apparent that certain (not all) ASCII codes were getting corrupted, which suggests one or more lines of the video ram databus are disconnected. However, this went away after a while and I got the following display:

ATOM

GDOS 1.66

930910

>

I was able to type in and run a simple program, so the core machine is obviously working! :D

Now, I am pretty certain that the corrupted characters issue is something to do with the decidedly ropey video ram card that I have, which is very precariously installed. I'll try waggling it about to see if that fixes things, and then ultimately I plan to replace it with a colour card. However, I do have a couple of questions which I hope someone with Atom knowledge might be able to help with:

1) What is the meaning of the six-digit numerical code quoted above? (I might have misremembered it, but it is definitely something like that).
2) The screen blanks completely every second (roughly), and then reappears. Any ideas why? I'm using the RF modulator so that may be something to do with it.
3) Someone has soldered two wires onto pins 9 & 10 of the video connector in order to connect up a composite video socket. Also R6 (82R) has been replaced by a 100uF capacitor. The video connector is where Phill's colour board plugs in I think, so to what extent do I need to reverse these mods to fit the colour board?

User avatar
aerworuld
Posts: 1717
Joined: Tue Sep 25, 2012 8:40 pm
Location: Basingstoke, Hampshire
Contact:

Re: Atom repair progress

Post by aerworuld » Sat Nov 09, 2013 9:49 pm

Nice work with the repairs sir; i hope they continue in success!

Prime
Posts: 2797
Joined: Sun May 31, 2009 11:52 pm
Contact:

Re: Atom repair progress

Post by Prime » Sat Nov 09, 2013 9:57 pm

jms2 wrote:
1) What is the meaning of the six-digit numerical code quoted above? (I might have misremembered it, but it is definitely something like that).
Could be a date perhaps 1993-09-10 ? 10th September 1993 ?
2) The screen blanks completely every second (roughly), and then reappears. Any ideas why? I'm using the RF modulator so that may be something to do with it.
Could be a sync problem or something like that, I know when I had video problems with a mono Atom, it turned out to be the electrolytics, but you said you replaced them. Try making sure everything is seated correctly.
3) Someone has soldered two wires onto pins 9 & 10 of the video connector in order to connect up a composite video socket. Also R6 (82R) has been replaced by a 100uF capacitor. The video connector is where Phill's colour board plugs in I think, so to what extent do I need to reverse these mods to fit the colour board?
The previous versions of the colour board do plug into PL4, but the 4.x version plugs into the 6847 socket and the socket for IC43 so the mods should not have to be reversed. Though it may help with your problem above.

Cheers.

Phill.

User avatar
oss003
Posts: 3122
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Atom repair progress

Post by oss003 » Sun Nov 10, 2013 12:55 pm

jms2 wrote:1) What is the meaning of the six-digit numerical code quoted above? (I might have misremembered it, but it is definitely something like that).
Yep, probably a date. The GDOS interface is dated 1985 and the GDOS rom was version 1.4. The latest verson of GDOS I can find is v1.5 dated 1986. So someone must have updated it to v1.66 but is never published in Acorn Nieuws.
jms2 wrote:2) The screen blanks completely every second (roughly), and then reappears. Any ideas why? I'm using the RF modulator so that may be something to do with it.
Maybe a 50/60 Hz sync problem with the TV? You could try the composite video out pins 9/10 of PL4 to see if it's a TV problem.
jms2 wrote:3) Someone has soldered two wires onto pins 9 & 10 of the video connector in order to connect up a composite video socket. Also R6 (82R) has been replaced by a 100uF capacitor. The video connector is where Phill's colour board plugs in I think, so to what extent do I need to reverse these mods to fit the colour board?
The replacement of R6 was a video signal enhancement and could interfere with RF output:
video.PNG
Greetings
Kees

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Sun Nov 10, 2013 1:10 pm

Thanks Phill & Kees,

Its good to know that the R6 mod is a recognised one and not just some kind of weird mistake! I'll certainly try the composite video output and see what happens.

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Sun Nov 10, 2013 8:23 pm

I have experimented a bit and had some success.

Waggling the video card about a bit has made the display reasonably stable and free of corruption. Also, I switched to composite video and whilst the display isn't brilliant, it is a bit better than before.

I definitely have GDOS 1.66, and yes the code is a date - 30th Sept 1991.

Here's an odd thing though - the screen does still glitch occasionally, but when it does it glitches in colour, green and magenta. Does this suggest that the homemade board I have in the machine is actually a colour board?? :?

There are more oddities as well. I tried out some of the other screen modes, eg CLEAR 4. In this mode I get a solid white border around the screen, but the actual text display is compressed down to just one line, as if the characters are still being entered into screen memory as single bytes. Maybe this is normal and these are just graphics-only modes? If so I've failed to spot it in ATAP!

Also I tried to find out the equivalent of PAGE by typing PRINT &?#18, expecting to get something like #2800 as a response. Instead I got #BF00, which was a bit of a surprise! Surely that's too close to rom to be comfortable?

There are two switches on the Combikaart. One is a read/write switch but the other one should switch in some alternative roms as I understand it. However neither seem to do anything.

User avatar
roland
Posts: 3480
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Atom repair progress

Post by roland » Sun Nov 10, 2013 10:31 pm

The 6847 cannot write characters in graphics mode. So if you type or print something in any graphics mode > 0 then you see only a few pixels changing. This is by design :D to use graphics and text simultaneously you need additional software which draws the characters like grmod, vdu2440 or screenrom.

The equivalent of page is P.&?18 (without the #) and it should respond with 29.

I will search for your atom's pictures (if any posted) for more info about the other questions.

Greetings,
Roland
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Sun Nov 10, 2013 10:44 pm

Aha - that's solved two of my questions, thanks Roland. I didn't consider for a moment that memory location 18 would be decimal. #-o That's my BBC-centric thinking! I never did understand why other machines (eg the C64), used decimal for memory addresses. In the Atom's case, its probably just to save a byte when programming.

I need to take some better pictures of the video card. But now I understand what is going on with the graphics modes, I can write a program to see if colour works.

One other question - how can you find out what roms are fitted without taking out the rom physically and reading it with a EPROM programmer? The Atom doesn't seem to have a *HELP command. I'm thinking like a BBC user again! :lol:

User avatar
roland
Posts: 3480
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Atom repair progress

Post by roland » Sun Nov 10, 2013 10:54 pm

There's no easy way to find out which roms nare in the eprom. The only way is to read them in some way, but of course you don't need an eprom programme for that.

Many roms have their statementtable at #A032. So if you do a P.$#A032 you often get a list of statements an real experts can tell you what rom it is :) p-charm has its statement table at #A300. You can also write a small hexdump program in basic to reveal the roms secrets.

You can switch between roms by writing a number to ?#BFFF.
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

User avatar
richardtoohey
Posts: 3720
Joined: Thu Dec 29, 2011 5:13 am
Location: Tauranga, New Zealand
Contact:

Re: Atom repair progress

Post by richardtoohey » Sun Nov 10, 2013 11:06 pm

jms2 wrote:I didn't consider for a moment that memory location 18 would be decimal. #-o That's my BBC-centric thinking! I never did understand why other machines (eg the C64), used decimal for memory addresses. In the Atom's case, its probably just to save a byte when programming
I'm a bit confused by what you mean here (also learning about Atoms, and coming from a BBC background.)

Atom #18 means BBC &18; but the location you wanted is 18 decimal.

On ATAP page 194, "25.4 Block Zero RAM Locations" it has "#12 18 Text-space pointer".

In a couple of places in ATAP (e.g. page 168) it talks about peeking/poking (decimal) location 18.

Ah, I think I understand now! You meant that if you were writing some code, putting ?18=X saves one byte compared to ?#12=X

And whenever you look at Spectrum/C64 stuff they seem to POKE decimal,X whereas most BBC stuff would be ?&hex=0

Not trying to be a smart-arse [-X , I wanted to check my own understanding of things Atomic and understand your point! Which I think I do.

User avatar
CMcDougall
Posts: 6858
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Atom repair progress

Post by CMcDougall » Sun Nov 10, 2013 11:21 pm

John wrote:That's my BBC-centric thinking!
this trips me up as well! 8)
jms2 wrote:I need to take some better pictures of the video card
yeah, the 'inside growing tent' ones where a bit :shock: LOL
http://www.stardot.org.uk/forums/viewto ... 167#p72081

regarding eproms, the proggy in my thread (Dave further up) helps lots :
http://www.stardot.org.uk/forums/viewto ... 198#p72885
ImageImageImage

Prime
Posts: 2797
Joined: Sun May 31, 2009 11:52 pm
Contact:

Re: Atom repair progress

Post by Prime » Mon Nov 11, 2013 12:05 am

jms2 wrote: And whenever you look at Spectrum/C64 stuff they seem to POKE decimal,X whereas most BBC stuff would be ?&hex=0

Not trying to be a smart-arse [-X , I wanted to check my own understanding of things Atomic and understand your point! Which I think I do.
Slightly OT:

And on various Microsoft basics (Dragon, GW etc) &HFF would be bbc &FF or Atom #FF, and then use the HEX$() function to go the other way.

Cheers.

Phill.

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Mon Nov 11, 2013 7:16 am

richardtoohey wrote: Ah, I think I understand now! You meant that if you were writing some code, putting ?18=X saves one byte compared to ?#12=X.
Yes, that's what I meant. :D

I found the reference in ATAP to the fact that the Atom can't directly display characters in graphics modes. It is so brief and matter of fact that I had missed it completely. It's fascinating how this kind of feature was normal at the time, and yet became completely obsolete within just a few years. As a BBC person I'm familiar with the concept of a text only mode (7) but not a graphics only mode!

User avatar
richardtoohey
Posts: 3720
Joined: Thu Dec 29, 2011 5:13 am
Location: Tauranga, New Zealand
Contact:

Re: Atom repair progress

Post by richardtoohey » Mon Nov 11, 2013 7:34 am

I really like the ATAP - it's so hands-on and right into the nitty-gritty and dirty details! =D>

But I guess the Atom was also available in kit form, so the target audience would have been up for it.

The BBC manual did have lots of low-level detail, but packed away nearer the back.

And these days we don't even get a manual - it's find the PDF on the CD if you are lucky and when you do find it they'll be obsessed with hiding their IP and making sure you don't do anything that means you need to make a support call.

Arrrgh, I'm digressing again, sorry. #-o [-X :oops: :-# Good luck with the Atom!

User avatar
oss003
Posts: 3122
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Atom repair progress

Post by oss003 » Mon Nov 11, 2013 8:02 am

jms2 wrote:I need to take some better pictures of the video card. But now I understand what is going on with the graphics modes, I can write a program to see if colour works.
If this is your video card: http://www.stardot.org.uk/forums/viewto ... then it's no colour board. Only the video buffers and video RAM are on the board, no colour hardware.

A program to test colours is:

Code: Select all

10 CLEAR4
20 F.I=0 TO 3
30  COLOUR I
40  F.X=I*32 TO (X+1)*32
50   MOVE X,0;DRAW X,191
60  N.
70 N.
80 END
To use GDOS, you need a special disc controller board with a WD1793 or WD1793 controller chip. This board was designed by G.Hillebrand and I'm not sure if a decent PCB was designed. The advantage of GDOS is that you'll be able to read/write 40 and 80 tracks drives with single or double density.

http://www.acornatom.nl/atom_nieuws/198 ... 862043.htm
http://www.acornatom.nl/atom_nieuws/198 ... 862044.htm

A simple dump program to investigate the first 256 byte of a ROM is this one:

Code: Select all

10 F.I=0 TO 7
20  ?#BFFF=I
30  P.$12"ROM NR."I'
40  F.A=#A000 TO #A0FF
50   IF ?A>32 AND ?A<>127;P.$?A
60  N.
70  P.'"PRESS A KEY TO CONTINUE";LI.#FFE3
80 N.
90 END
Greetings
Kees

User avatar
roland
Posts: 3480
Joined: Thu Aug 29, 2013 8:29 pm
Location: Born (NL)
Contact:

Re: Atom repair progress

Post by roland » Tue Nov 12, 2013 9:21 am

Hi Kees,

I think I have a small indication there was a decent pcb for the gdos controller:
image.jpg
image.jpg
@JMS: one of the switches on your combicard is for write protect Axxx and 1xxx. The other one is for switching between two versions of CDEF-rom if you have a 27256 installed. That switch is useless if there is a 27128 on board.

Greetings Roland
FPGAtom: much better than Atom2k15 which was even better than the real thing.
MAN WOMAN :shock:

Prime
Posts: 2797
Joined: Sun May 31, 2009 11:52 pm
Contact:

Re: Atom repair progress

Post by Prime » Tue Nov 12, 2013 9:46 am

Hi all,

Humm going on the schematic for the GDOS controller, re-designing with a WD2793 would allow removal of about half of the LS chips as the WD2793 is compatible with the WD1793, but has the data separator built in. It also has an option to operate at 2MHz with a data rate equivalent to high density....it was designed for 8" drives but HD 3.5" drives use the same data rate :)

Cheers.

Phill.

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Sun Nov 17, 2013 8:19 am

Thanks for explaining a bit about GDOS Kees, I was planning on asking a question about it at some point! my intention is actually to use AtoMMC rather than go down the route of real disk drives, but with the combikaart I can retain both sets of functionality which is great. :D

I found that waggling the power lead into my Atom had the effect of making the screen corruption go away, in fact just tilting the motherboard to the correct angle would do it, so I decided to fit a new socket. The old one looked really dirty and corroded.

Anyway, I have done it and, if anything, the corruption is now worse! :lol: I haven't had time to experiment with it very much but I don't get a readable display at all now. So I must have affected something, but I'm not sure what!

I also tried levering out the 6847 and re-seating it. It fits into a socket made of 2 turned pin SIL strips but it seems a curiously loose fit.

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Tue Nov 26, 2013 10:03 pm

I have levered out and reseated all the ics on the video ram board, and this has basically fixed the display! :D It looks as if adding Prime's new colour board (when it's available) will provide a robust solution.

I have also tried examining some of the roms. One that I recognise is Salfaa, but the others are unfamiliar to me.

I think AtoMMC needs to be my next port of call, and I plan to use the #E000 rom.

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Wed Jan 15, 2014 9:13 pm

Since my last post on this thread I have fitted an Atommc2 board and the E000 firmware. Video problems persist though - and here's the really strange thing: I get much worse screen corruption with Atommc2 connected than without.

Here's what the screen looks like without Atommc2. Sometimes it is a bit better than this (more black). The white fuzzy stuff ripples and moves around on the screen.
P1040809.JPG
If I plug Atommc2 into PL8, I get this :shock: (which also moves):
P1040807.JPG
I also get random characters, but not too many.

Atommc2 works if I type LINK#EFCC, but it doesn't get detected on Break.

Do these symptoms suggest what might be wrong? It looks like quite a lot of things! :lol: My plan has been to get the machine basically working before investing in a colour board, but given that the machine contains a hand-made video ram board (which plugs into numerous ram sockets simultaneously) would I be better going straight for the colour board option and eliminating all the messy connections?

The above screen shots were taken after I fitted new turned-pin sockets for the video ram board to plug into. I also took out the 6847 and cleaned its pins. None of this seemed to help. :(

User avatar
oss003
Posts: 3122
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Atom repair progress

Post by oss003 » Wed Jan 15, 2014 9:44 pm

Hi jms2,

just some questions:

AtoMMC:
- Where is the #Exxx ROM located, on the expansion board?
- Is it in one ROM with #C/D/E/F?
- Did you patch the #Fxxx ROM?

Video:
- Are you using the video output, pin9/10 PL4?
- Can you check the RF output to see if the problems are the same?
- Did you replace C2/C3 and C4?
- What is the voltage on the Atom board?

Greetings
Kees

Code: Select all

#Fxxx Patch for V 2.x AtoMMC2-2.x-E000

Offset in kernel.rom  Original value  New value
    f8ea                  0d              cc
    f8eb                  e0              ef
Changes execution address of DOS command from $E00D to $EFCC

Offset in kernel.rom  Original value  New value
    ff95                  b2              00
    ff96                  c2              e0
Changes JMP at FF94 from JMP $C282 to become JMP $e000

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Wed Jan 15, 2014 9:56 pm

oss003 wrote:Hi jms2,
Hello Kees, thanks for replying so quickly! :D

just some questions:

AtoMMC:
- Where is the #Exxx ROM located, on the expansion board?
- Is it in one ROM with #C/D/E/F?
- Did you patch the #Fxxx ROM?
Yes, the #Exxx rom is on the expansion board, in a single C/D/E/F rom. I didn't patch the #Fxxx rom myself, but I started with Roland's C/D/E/F rom which I understood had been patched. I am pretty certain I checked this actually, looking at your list of changes below - they seem familiar!
Video:
- Are you using the video output, pin9/10 PL4?
- Can you check the RF output to see if the problems are the same?
- Did you replace C2/C3 and C4?
- What is the voltage on the Atom board?
Actually I am using the RF output, but last time I checked the pin9/10 output on PL4 was the same.

I have been round the board and replaced all the electrolytic capacitors.

The Atom is 5V only, no regulators. For a power supply I am using a spare 5V supply of unknown origin, but which works fine with a Master Compact that I have.

User avatar
oss003
Posts: 3122
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Atom repair progress

Post by oss003 » Wed Jan 15, 2014 10:05 pm

jms2 wrote:Yes, the #Exxx rom is on the expansion board, in a single C/D/E/F rom. I didn't patch the #Fxxx rom myself, but I started with Roland's C/D/E/F rom which I understood had been patched. I am pretty certain I checked this actually, looking at your list of changes below - they seem familiar!
You can easily check this by typing P.&!#F8EA,&!#FF95
Can you also type P.&!#E000 to see if the AtoMMC ROM is seen?
Actually I am using the RF output, but last time I checked the pin9/10 output on PL4 was the same.

I have been round the board and replaced all the electrolytic capacitors.

The Atom is 5V only, no regulators. For a power supply I am using a spare 5V supply of unknown origin, but which works fine with a Master Compact that I have.
Hmm .... video output on PL4 should definitely be of better quality. Maybe there's some noise on your power supply, can you check another one?

Greetings
Kees

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Wed Jan 15, 2014 11:10 pm

I could try the genuine master compact psu if I borrow a power lead from an arc. It'll have to wait a bit though!

User avatar
CMcDougall
Posts: 6858
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Atom repair progress

Post by CMcDougall » Wed Jan 15, 2014 11:49 pm

^save time & space John, if you have one of these at hand :
http://static.scan.co.uk/images/products/2146744-a.jpg
can be found with a camera/tablet/mobile charger (5v 1 or 2amp) :)
much better than looking like these :
http://www.stardot.org.uk/forums/downlo ... &mode=view
https://fbcdn-sphotos-c-a.akamaihd.net/ ... 9677_n.jpg
Thought my ginger atom was a bad wee beastie (now tamed & Very nice), but yours must be a strawberry blonde beastie! :lol:
ImageImageImage

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Thu Jan 16, 2014 7:32 am

Col, did you find using a different psu helped with your Atom? I do have some of those little usb ones, but the one I am currently using is actually fairly similar.

User avatar
CMcDougall
Posts: 6858
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Atom repair progress

Post by CMcDougall » Thu Jan 16, 2014 11:50 am

It never helped mine at all :lol: But the USB 1Amp charger was best (still use now), as was closer to the 5 volt mark (5.15ish), and the mCompact & PC was 5.36v.
Still to check the three 2Amp ones with the volt meter that appeared at xmas (they are even smaller :o ), but would anticipate it to be around the 5.1volt too.
ImageImageImage

User avatar
oss003
Posts: 3122
Joined: Tue Jul 14, 2009 11:57 am
Location: Netherlands
Contact:

Re: Atom repair progress

Post by oss003 » Thu Jan 16, 2014 11:56 am

jms2 wrote:I have been round the board and replaced all the electrolytic capacitors.
If the capacitors C2/3/4 are replaced, the only thing left are Q1 and Q2 ...... :)

Greetings
Kees

User avatar
jms2
Posts: 2313
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Atom repair progress

Post by jms2 » Thu Jan 16, 2014 2:49 pm

Just photocopied the schematic and I've come to the same conclusion!

User avatar
1024MAK
Posts: 9379
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Atom repair progress

Post by 1024MAK » Sun Jan 19, 2014 6:13 pm

The first photo you posted of the Atom screen reminds me of problems with poor earth/0V/ground connections. So check that all 0V ground connections are okay. The next thing to check is that the video circuits have the correct supply voltage.
The second picture is very strange - the streaks go the wrong way to what I would expect! It again could be a problem be a problem with the voltages in the video section. Almost as if something is charging / discharging (and causing the video level to change). Did you get poor video before you renewed the capacitors?

Mark

Post Reply