• Please review our updated Terms and Rules here

Acorn atom

Hmm,

Might be worth looking at the rise time, fall time and duty cycle of the 6847 clock as the beast warms up. There appears to be some MAX constraints on those.

Dave
 
I had decided though to disconnect the video out by removing C4 and FS seems to be stable even after 10 minutes

1731248859745.png

Now, when carrying this out, I had the monitor disconnected, but when I plugged it in, FS shrank, then jumped back when unplugged. The monitor is a standard 75ohm 1V video monitor which should be compatible with the Atom.

The video connector is soldered in just as badly as the rest, so will remake it just in case it has some low resistance paths as I note horizontal sync is available on the adjacent pin 8.
 
Video socket removed and all tracks given a really good clean up. The solder is so hard and has virtually no wetting, the flux is almost solid and stinks like 50's electronics !

Reconnected things up and I now have a screen that lasts for more than 15 minutes but its a bit noisy.

Moving forward.
 
Machine is now stable which is good. Not sure what the issue was, but it seems to have been a temperature related low resistance at the video connector. The display is stable if a little noisy and has run for a couple of hours now without FS glitching. I do note there is a lot of high frequency noise on the +5V lines so I think I need to replace the input filter capacitor (dropping one across it gets rid of the noise so its obviously goosed).

So onto putting the keyboard back together. I have one snapped stem to glue back together and the original N key problem. This one is more difficult. I now have a key but the stem is snapped and its also the wrong one. Someone has obviously had a go at it previously and the sabrecoils are also damaged. While I can get them into a position where they will work, combined with the wrong stem which easily pops out (and then entails a right nightmare to get the coil to sit back) I think I am going to need a more radical solution.

Plan is to take the coils and stem from the space bar and fit them into the N key position (which is a fiddly nightmare), then cut away the spacebar part of the Sabercoil mounting (As it sticks out and is the simplest bit to remove) and replace it with a BBC type 2 angled keyswitch mounted on a suitable adapter. It might not quite have the same travel but that would I think, be less noticeable on the spacebar.

Not something I want to do, but needs must.
 
Now thats annoying

Machine is much more stable but ...

So I thought it had multiple overlaying faults. First was one pin of the VDG address bus was missing, socket fault, second, Field sync kept shrinking, might have been a problem on the video socket. and hasn't come back since the socket was removed and the board cleaned in that area.

Now when its left on for even longer, I seem to be getting breakthrough of the CPU's data bus onto the video data bus. Removed the buffer IC but no change, so removed the socket, cleaned up the board and it 'seems' to have gone. Really this PCB is a mess. It would be a real pain but it might be moving towards removing every component and giving it a real good clean (though that would really mean removing the keyboard and that isn't happening... ever)
 
New socket fitted and not only am I not getting the break through of the CPU data bus onto the VDG data bus, I am actually getting a couple of signals through that I didn't know I was missing :)

Sooooooo. keyboard next. Got the BBC type 2 angled key switch through the post today and its perfect. Not only does it fit to the Atom's keycaps, its travel is spot on and its angle and height seems about right to fit without too much in the way of adaptation.

This next step will need pictures !


I do wonder which socket I need to replace next, so bought a job lot of various sizes from RS.
 
First job is to cut away the space bar switch
20241123_160051.jpg20241123_163611.jpg
Then make up an adaptor for the BBC type 2 switch

20241124_095540.jpg
Then that fits under the keyswitch with the wires running down the channels. The mount is then glued to the keyswitch and the lugs press into the atom pcb and the wires poke through.
20241124_133206.jpg

So a working space bar (which it did before) and now I have some spare sabrecoils

20241123_174615.jpg
Next job is to fit it into the N socket in place of the damaged one.
17324556836982019110392939637934.jpg
 
Sabrecoil installed and N key now working. The keyboard is a bit bouncy and no debounce circuit evident. Read somewhere that a later ROM did software debounce but something to think about later because . . .

After working for a week and even being left running for 6 hours with no problems, the machine suddenly started showing display problems (again, again). Upper and lower blocks would flash with solid lines. Pin 6 of the 2114 RAM shows a minimum of about 1V when cold, rising to about 2V when warm. With every IC on that line removed, Pin 6 shows 5V so there is a low resistance path to 5V somewhere. All VDU RAM sockets replaced and now pin 6 reads 0V.

Ok, so I think this is it, all ICs put back in and it works, for about 10 minutes and now the display corrupts again in yet another new and exciting way. This time in a vertical column of ? then a grey block then a white column, then 7 ?'s then another grey block.


Arrrrrrgh !

Only two sockets left to replace in the video path, so they are coming out next. Good job I bought enough in the last batch.
 
Every socket on the VDG side of the bus now replaced and at last its working (for now) and its back in its case.

20241130_180345.jpg
Time to leave this for a while. It needs a solution for the video port. Might knock up a daughterboard to mount a rca socket but for now I hear the north star calling.

Watch out for a new and exciting thread !
 
You must log in or register to reply here.
Back
Top