Atmos/Microdisk - ghost in the machine!?!?!?

If you want to ask questions about how the machine works, peculiar details, the differences between models, here it is !
How to program the oric hardware (VIA, FDC, ...) is also welcome.
Post Reply
omelette
Officer Cadet
Posts: 44
Joined: Wed Mar 10, 2010 9:54 am

Atmos/Microdisk - ghost in the machine!?!?!?

Post by omelette » Fri Nov 29, 2019 8:01 pm

Attached below is a little video of a problem that has been plaguing my Oric/Microdisk setup, like forever. And it's getting worse! Usually I can get 1-2 hours work done on my Atmos, then what you will see in the video starts happening, which if left running, gets worse & worse with more random letters/display attributes being printed, faster & faster. It will eventually lock up completely, and not respond to any more hard-resets.

This happens with an Oric-1 attached as well. Not tied to the Microdisk, the computers will run indefinitely without issues.

First thought is that it's temperature related. Thing is, with the Microdisk's top removed, nothing on the main PCB gets even remotely warm. But, switch off the system for an hour and it can be used again, albeit only for another hour or so.

I thought I'd hold off on contacting an exorcist until I'd run it by the hardware experts here. Has anyone come across this problem in the past?


User avatar
Dbug
Site Admin
Posts: 3049
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Dbug » Fri Nov 29, 2019 9:03 pm

No idea really!

User avatar
ibisum
Wing Commander
Posts: 1131
Joined: Fri Apr 03, 2009 8:56 am

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by ibisum » Sat Nov 30, 2019 11:47 am

>temperature related

Bad capacitors, or bad transistors somewhere? Time to check for slightly-puffy caps somewhere .. maybe they haven't exploded yet ..

User avatar
iss
Squad Leader
Posts: 915
Joined: Sat Apr 03, 2010 5:43 pm
Location: Bulgaria
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by iss » Sat Nov 30, 2019 1:59 pm

Try this fix:
Keyboard-fix.png
Keyboard-fix.png (704.9 KiB) Viewed 166 times
Resistor's value is not critical - from 1k up to 10k should work, experiment with different values in the range.
Maybe changing TR2 with better NPN transistor can help too.

User avatar
mikeb
Flying Officer
Posts: 135
Joined: Wed Sep 05, 2018 8:03 pm
Location: West Midlands, UK
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by mikeb » Sat Nov 30, 2019 7:08 pm

Are the letters being "typed" (with a keyclick, and you can delete them?) or are they just appearing on the screen (not really there, memory corruption, silent arrival) ?

Only if it's memory corruption, rather than phantom typing, does this apply :-

On the Cumana interface board, there is a preset trimmer on there which is used to adjust a critical timing delay (related to the 1MHz Phi-2 clock). It can drift slightly with temperature. It causes random characters/display interruptions as the disc interface starts to tread on the ULA's use of the bus.

Does the Microdisk hardware use this same feature? If so, it might have drifted with age/damp/vibration and be on the edge of working :)

omelette
Officer Cadet
Posts: 44
Joined: Wed Mar 10, 2010 9:54 am

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by omelette » Mon Dec 02, 2019 12:53 am

Thanks for the responses guys.

There have been developments, and not for the good. While checking to see if it could be a drive-related problem, I swapped out the Oric's floppy and installed a 3" Teac drive I salvaged from a Tatung Einstein I once possessed. Long story short, while doing this, the MIcrodisk stopped working, permanently. Nothing shorted etc, it was working with the new drive, then it wasn't.

I did some voltage checking and was shocked to see that the Microdisk's 12v regulator was supplying 16.5v rather than 12v - faulty regulator! After the "Aha" moment that this was probably what was causing all of the problems, I swapped it out with a new one - 12v delivered as expected! The Microdisk was still dead, however I was now more worried that the over-voltage could have also killed BOTH of my 3" drives - the FDC chips are available from China a dime a dozen, whereas the drives themselves are irreplaceable.

Out comes my old Spectrum/Disciple disk-drive setup, and I was quickly able to establish that both 3" drives were still working! So, only the Microdisk is dead.

A quick look at the Microdisk's schematic reveals that the 12v line seems to go only to pin 40 of the FDC chip, which incidentally is a MB8877A (see attached pdf). What is confusing me now is that according to the manufacturers datasheet, pin 40 is no-connection (NC). If this is so, then it can't have been damaged by over-voltage. More importantly, what's the 12v purpose on the main board to begin with? Apart from a small 10uF 'reservoir' capacitor between 12v & Gnd, I can't see any other connection to it, other than to pin 40 of the FDC.
>temperature related

Bad capacitors, or bad transistors somewhere? Time to check for slightly-puffy caps somewhere .. maybe they haven't exploded yet ..
There is only 3 tiny electrolytic caps on the board, besides the ceramic decoupling caps. The electrolytics are fine.
Are the letters being "typed" (with a keyclick, and you can delete them?) or are they just appearing on the screen (not really there, memory corruption, silent arrival) ?

Only if it's memory corruption, rather than phantom typing, does this apply :-

On the Cumana interface board, there is a preset trimmer on there which is used to adjust a critical timing delay (related to the 1MHz Phi-2 clock). It can drift slightly with temperature. It causes random characters/display interruptions as the disc interface starts to tread on the ULA's use of the bus.

Does the Microdisk hardware use this same feature? If so, it might have drifted with age/damp/vibration and be on the edge of working
They are 'really there', and can be deleted from the keyboard.

There is a single adjustable pot.on the Microdisk, RV1. Its exact function I'm not sure of, but definitely not the issue here. Adjusting it with a working Microdisk will just cause the FDC to lose track of data on the disk, adjust it back and everything is fine.

Next plan is to order the FDC chip from China..Does anyone have any information on the 'Test' pin of this chip? The datasheet just lists it, nothing else. Anyway, it could be some time before the Microdisk's up & running again, unless I discover something in the interim. Just as well I have Oricutron to fall back on. :)

Edit:
Just to add, I scoped the clock output - a respectable looking 8Mhz waveform of about 3.2-3.3V.
Attachments
MB8877A datasheet.pdf
(532.63 KiB) Downloaded 4 times

christian
Pilot Officer
Posts: 78
Joined: Sun Nov 24, 2013 9:58 pm

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by christian » Mon Dec 02, 2019 11:47 am

I think the FDC should be a WD1793, not a MB8877.
They are "firmware compatible" but not "pin compatible"

Here is the datasheet: http://bitsavers.org/components/western ... ay1980.pdf

User avatar
Chema
Game master
Posts: 2509
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Chema » Tue Dec 03, 2019 12:16 am

Wait for someone with more knowledge than me, but I think the 12v line is just for the FD motor, not the board itself. Why it is connected to the controller chip, I don't know.

About the RV1 variable resistor... you have to turn it very very slightly and reboot to check if it helps. It controls the deialy of the MAP signal or somehting along that line, so fine-adjusting it may help. Otherwise you may be experiencing the usual problems related to the bus signals. Did this Microdisc ever worked ok with this Oric? If not, maybe you need something such as an Amplibus (a small card to filter signals). I need one to make my disk drive work. Without it it hangs most of the times with garbage on the screen.

omelette
Officer Cadet
Posts: 44
Joined: Wed Mar 10, 2010 9:54 am

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by omelette » Tue Dec 03, 2019 2:11 am

I think the FDC should be a WD1793, not a MB8877.
They are "firmware compatible" but not "pin compatible"

Here is the datasheet: http://bitsavers.org/components/western ... ay1980.pdf
Now that's very interesting, I never thought it could be a pin-compatible replacement. That also means that the original was a WD1793, which would explain why there's 12v going to pin 40, if not what its actual purpose is. Apart from a few US Sellers offering used ones on eBay, (along with the usual ludicrous shipping charges) the WD1793 seems to be a genuinely obsolete part, shame, I'd have got one of those as well when ordering stuff. Guess I'm stuck with the MB8877 - most of which seem to be Chinese clones of the original Japanese part as well, judging by some Aliexpress Buyer feedback comments I've read. Thanks for the pdf link, it contains much more info than the one I linked to.
Wait for someone with more knowledge than me, but I think the 12v line is just for the FD motor, not the board itself. Why it is connected to the controller chip, I don't know.

About the RV1 variable resistor... you have to turn it very very slightly and reboot to check if it helps. It controls the deialy of the MAP signal or somehting along that line, so fine-adjusting it may help. Otherwise you may be experiencing the usual problems related to the bus signals. Did this Microdisc ever worked ok with this Oric? If not, maybe you need something such as an Amplibus (a small card to filter signals). I need one to make my disk drive work. Without it it hangs most of the times with garbage on the screen.
Yeah, that's what I thought as well, though that pdf-link to the WD1793 would seem to suggest otherwise. It can't have been super-critical though if the MB8877 chose to leave it out entirely. It also strongly suggests that my FDC chip is still working.

Yes, the Microdisk has worked perfectly with the Atmos for nearly 30 years! For the most part anyway, like I mentioned above, the random characters appearing on the screen with the Microdisk plugged in has plagued the setup for a long time, but it was more of an annoyance than a show-stopper.

Myself and RV1 have a long history. I may not know what it does, but I'm very familiar with what its effects are. With a working Microdisk, adjusting RV1 either completely clockwise or anti-clockwise will merely cause the FDC to no longer be able to locate data on the spinning disk. Adjust it back to roughly where it was, around half-way on mine, though there's a huge margin for error, and it's back reading & writing disks again.

Right now, the Microdisk's appearing dead - push the reset button on the Microdisk and there's little or no response on the screen. I've also seen this condition before, normally like I mentioned in my opening post, when the Atmos/Microdisk was left running for a few hours, but it would always return to normal operation when left switched off for an hour or so.

So what I'm wondering now why, when I press the Microdisk's reset button, i do not instantly see that "insert system disk" message being displayed. This code is being read from the Microdisks eprom, and executed by the 6502, not the FDC. In fact, I would have thought that even with the FDC chip removed entirely, that "insert system disk" message should still be displayed when I press the reset button, presuming of course that all of the rest of the circuitry was working. Though it could be anything, I think I'll start by looking at those 74LS244/245 drivers for the address & data lines. All the chips except for the FDC and eprom are soldered in, which sucks, but I should be able to check if the address line chips are ok, before having to resort to desoldering stuff.

omelette
Officer Cadet
Posts: 44
Joined: Wed Mar 10, 2010 9:54 am

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by omelette » Thu Dec 05, 2019 1:31 am

Well, that proved to be far easier than it had any right to be. Turns out the second address line of IC2 (74ls244) that I checked was faulty, holding the poor Old Oric's A1 line low. Replacing it proved a real pain. I ended up having to cut the faulty chip's pins with a dremel and take them out individually. I left it on soak-test for 24hrs and no more ghosts - weee!

You've got to wonder though why the guys that designed the Microdisk didn't go mad and just buffer all of the address lines. Leaving nearly half hanging in the breeze seems dumb. I might try knocking up a pcb from that Microdisk schematic sometime, for kicks, and to give that other 3" drive a purpose.

User avatar
Chema
Game master
Posts: 2509
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Chema » Thu Dec 05, 2019 9:02 am

Glad you fixed that Microdisc!

By the way, do you have any means of writing 3" disks from disk images (dsk)? I know there is a way of connecting a 3" unit to a PC, but it seems nobody has such a setup nowadays.

User avatar
Dbug
Site Admin
Posts: 3049
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Dbug » Thu Dec 05, 2019 9:23 am

Chema wrote:
Thu Dec 05, 2019 9:02 am
By the way, do you have any means of writing 3" disks from disk images (dsk)? I know there is a way of connecting a 3" unit to a PC, but it seems nobody has such a setup nowadays.
I've personally gave up with that, but the combo of HxC (or Gotek I guess) + Microdisc works well.

You basically convert the DSK to HFE and store them, on the HxC, and you copy them on the Oric from the HxC to the physical floppy.

User avatar
Symoon
Archivist
Posts: 1750
Joined: Sat Jan 14, 2006 12:44 am
Location: Paris, France
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Symoon » Thu Dec 05, 2019 1:56 pm

I still have a PC with 3" disk drive attached but gave up trying to make a 3" disk of FT-DOS for someone a few years ago (after something like 15 tries, drive change, and so on).

I do think it's time we save the remaining 3" disks (especially originals), and give up using 3" disks.
Drives are failing, disks are failing (never been able to load my 3" copy of Athanor despite I know Thierry and Eric checked them), none is produced anymore... I still look at my pile of unreadable original disks with sadness.

User avatar
Chema
Game master
Posts: 2509
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Chema » Thu Dec 05, 2019 2:49 pm

Yeah, this seems to get more and more difficult as time goes by :(
Dbug wrote:
Thu Dec 05, 2019 9:23 am
I've personally gave up with that, but the combo of HxC (or Gotek I guess) + Microdisc works well.

You basically convert the DSK to HFE and store them, on the HxC, and you copy them on the Oric from the HxC to the physical floppy.
Well, no way of making a copy of Blake's 7 or any demo disk created with FloppyBuilder, as they don't have any DOS format. Maybe it is time to start thinking again on a low-level diskcopy tool. I am quite sure we discussed that some time ago.

I really wanted someone to test the game on a double-sided 3" disk drive. Did I ever mentioned that the Jasmin compatible boot loader, which worked perfectly under emulation, completely failed on real hardware? But it took almost a year to find someone with the equipment to test. And there is no way to debug that, unless I steal the unit from the owner :lol:

User avatar
Dbug
Site Admin
Posts: 3049
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Atmos/Microdisk - ghost in the machine!?!?!?

Post by Dbug » Thu Dec 05, 2019 3:16 pm

I'm still going to try :)

If only because a complete Oric system looks nicer that anything recent that has been produced.

Sure, Erebus and Cumulus and practical, more user friendly and reliable than tapes and floppies, but there's something satisfying in just managing to load a game from a physical media made of rust particles glued on top of some oil derived substance moving over a magnetic field reader :)

Post Reply