PDA

View Full Version : xerox 820 II screen error?



harry
July 10th, 2010, 02:09 PM
I try and regularly boot up most of my machines, and due to finding some more 8" media supposedly for a 820 xerox, I decided to set up my 820 II and take a look at these recent disks. the xerox has been fine, and this time booted fine,with its cpm system disk in the A drive. and I had been looking forward to see what files were on these latest disks. Using the B drive, several disks were unreadable, one of these when inserted in drive B made its usual read noise but nothing from the dir prompt appeared on screen,
what then happened was the rows filled up with the blocks of T and M characters, slowly coming across the screen from right to left? I now have a display,permanent as shown in the attachment, before I start checking things out for component failure, has anyone had this or similar happen, or can anyone please assist me in getting the display back to normal? maybe it was just coincident this has happened while reading the last disk?
.............Harry

saundby
July 11th, 2010, 01:36 AM
That one's new to me. I'd check that the character ROM is being addressed properly to start with.

Chuck(G)
July 11th, 2010, 07:20 AM
Well, the TM symbol is character value 7FH (127) and the reverse attribute is the high order bit, so it looks as if your display is full of hex FF characters--all bits on. I'd start by checking out display memory, since the failure appeared to be slowly progressive. It could also be a failure of the CPU.

When you reset the system (after powering up), is the display full of "TM" or does the display gradually fill up with them?

harry
July 11th, 2010, 10:41 AM
Well, the TM symbol is character value 7FH (127) and the reverse attribute is the high order bit, so it looks as if your display is full of hex FF characters--all bits on. I'd start by checking out display memory, since the failure appeared to be slowly progressive. It could also be a failure of the CPU.

When you reset the system (after powering up), is the display full of "TM" or does the display gradually fill up with them?

Chuck, after powering on, the display is completely full as per the screen picture, , when the reset is pressed, the screen goes completely white for a second,, and then instantly shows the full screen of TM characters. The cpu is still running , and I can type L to load the system, the A drive will operate normally, and when loaded, It will respond to a DIR command ,at least the drive makes the normal responce , it also responds to cntl +c to make it reboot no problem. The monitor program in rom, should be displayed first at boot, and I assume this is happening, although not shown on screen. can the low ram be stuck at this address? if it was the character rom, would it fill the whole screen? I have a spare cpu, there is not much else socketed on the board, most all soldered in. can I force the cpu to increment to another address, or bypass the first low ram? just to see if the screen changes..........Harry

Chuck(G)
July 11th, 2010, 11:53 AM
If the system loads, it's not the CPU. Let me stare at the schematic a bit and see if I can come up with some likely suspects. :)

Do you own any test equipment at all? (e.g. logic probe, oscilloscope, DMM)

harry
July 11th, 2010, 01:01 PM
If the system loads, it's not the CPU. Let me stare at the schematic a bit and see if I can come up with some likely suspects. :)

Do you own any test equipment at all? (e.g. logic probe, oscilloscope, DMM)

Hi Chuck, good of you to look into this for me, I have access to all the items you listed, but I am limited to my own technical ablility,in using them , I can certainly check out anything you suggest , and hopefully narrow things down. the awkward part is getting to the components, as the crt covers most of the front of the m/b, might be able to move the crt up, out of the way while still being connected.............. Harry

Chuck(G)
July 11th, 2010, 02:02 PM
I'd probably start by looking at U54, the large 24-pin DIP that constitutes the 2K screen buffer. A scope would be the
ideal tool for this--just set it so it displays 0-5V so you can see it, select an auto sweep rate of about 10KHz or slower.

Take a look at pin 20, it should be low most of the time. Run a little program that writes to the CRT continuously. You should see pin 20 bouncing and pin 21 bouncing in the opposite direction.

Look at pins 1-8; these are the data output lines. They should be each be generating a fair amount of "hash". Try grounding (momentarily) pin 1. Does the character displayed on the screen change?

harry
July 11th, 2010, 02:39 PM
I'd probably start by looking at U54, the large 24-pin DIP that constitutes the 2K screen buffer. A scope would be the
ideal tool for this--just set it so it displays 0-5V so you can see it, select an auto sweep rate of about 10KHz or slower.

Take a look at pin 20, it should be low most of the time. Run a little program that writes to the CRT continuously. You should see pin 20 bouncing and pin 21 bouncing in the opposite direction.

Look at pins 1-8; these are the data output lines. They should be each be generating a fair amount of "hash". Try grounding (momentarily) pin 1. Does the character displayed on the screen change?

Chuck, thanks for this, its bed time for me , tomorrow I'll check things out , and let you know what I find......great stuff ..... Harry

harry
July 12th, 2010, 01:12 PM
I'd probably start by looking at U54, the large 24-pin DIP that constitutes the 2K screen buffer. A scope would be the
ideal tool for this--just set it so it displays 0-5V so you can see it, select an auto sweep rate of about 10KHz or slower.

Take a look at pin 20, it should be low most of the time. Run a little program that writes to the CRT continuously. You should see pin 20 bouncing and pin 21 bouncing in the opposite direction.

Look at pins 1-8; these are the data output lines. They should be each be generating a fair amount of "hash". Try grounding (momentarily) pin 1. Does the character displayed on the screen change?

Chuck, I had to remove the crt and leave it in position just out front , pin 20 on u54 is constanly low all the time, no change, pin 21 is high (5v) constant
no change on these pins during reboot, reset,or loading sys , the same goes for pins 1-8 no change to the outputs regardless of actions from diskloading or reset.
the signals on pin1,2and 3are looking not clean as the rest, infact it appears as a double trace with no verticals on these pins, the others show0v to 5v cycles which increase in frequency up to pin 8. nothing is moving or changing regardless of input, at least not on these pins I have checked for you.
..............Harry ( oh nearly forgot..... no change when I ground out pin 1)

Chuck(G)
July 12th, 2010, 06:35 PM
Harry, let me stare at the schematic a little more and see if I can come up with some likely candidates.

harry
July 15th, 2010, 04:29 AM
Harry, let me stare at the schematic a little more and see if I can come up with some likely candidates.

Chuck, I have found some schematics on line, but the writing on them is mostly unreadable, am I right in assuming u57
is the graphics rom, and u58 is the character rom, if so I am going to swap these around and see if the screen changes,
if not then I am following your instinct and replace u54, the only video ram chip. I take it u55 and u56 are just the flip flops to read in and out the two character roms, as required, It would be good to try something to make the screen change, or possibly indicate what is stuck at the present location, what are your thoughts?? many thanks ...........Harry

nige the hippy
July 15th, 2010, 05:50 AM
Hello there,
O/T
Glad to know the boot disks were fine - quite jealous of the ICL branded disks though!

On Topic - you won't do any harm by swapping the two roms round, I'd go with the 2016 though as a first.
I'll have to revert back to the original post, but didn't you say the characters were appearing one after the other? it's not something daft like something o/c in the keyboard making characters appear? as it's a parallel keyboard j2-9 could be being strobed by mains hum??

Chuck(G)
July 15th, 2010, 09:34 AM
Harry, readable 820-II schematics here (http://bitsavers.informatik.uni-stuttgart.de/pdf/xerox/820-II/Xerox820-II_Schematics.pdf).

What bothers me is that (Sheet 4, near U54) DATAWR/ isn't changing state when the system is supposed to be writing to the 2016P SRAM. That says to me that (a) your Z80 isn't writing to the display at all or (b) something in the video RAM access timing is very wrong (see the bottom half of sheet 5). You may want to trace things back from there.

Also, you may want to check on pin 11 of U56, the LS374 octal flip-flop and the logic in that area. If that's not toggling, or all of its outputs are stuck high, that would also cause the screen to be filled with "TM" characters.

Just some starting points for your probing...

harry
July 15th, 2010, 03:08 PM
Hello there,
O/T
Glad to know the boot disks were fine - quite jealous of the ICL branded disks though!

On Topic - you won't do any harm by swapping the two roms round, I'd go with the 2016 though as a first.
I'll have to revert back to the original post, but didn't you say the characters were appearing one after the other? it's not something daft like something o/c in the keyboard making characters appear? as it's a parallel keyboard j2-9 could be being strobed by mains hum??

Hello Nige, thanks for interest, do you also have a xerox ? I did disconnect the keyboard to see if anything changes, when this first happened, but the screen remained as per the photo , and will boot to this screen without k/b connected,or with. .....Harry

harry
July 15th, 2010, 03:29 PM
Harry, readable 820-II schematics here (http://bitsavers.informatik.uni-stuttgart.de/pdf/xerox/820-II/Xerox820-II_Schematics.pdf).

What bothers me is that (Sheet 4, near U54) DATAWR/ isn't changing state when the system is supposed to be writing to the 2016P SRAM. That says to me that (a) your Z80 isn't writing to the display at all or (b) something in the video RAM access timing is very wrong (see the bottom half of sheet 5). You may want to trace things back from there.

Also, you may want to check on pin 11 of U56, the LS374 octal flip-flop and the logic in that area. If that's not toggling, or all of its outputs are stuck high, that would also cause the screen to be filled with "TM" characters.

Just some starting points for your probing...

Chuck, Thanks for the link to the much better schematics, at least I can make out some of the pin no's, I have spent some time checking around, and you are right about the U56 outputs being all high? pin 11 has its clk pulsing away, all the inputs on u56 are stuck at 1.6v. The corresponding I/o lines on u54 (pins 9-16) match the same at 1.6v, the address lines 1-8 on U54 when checked with the logic probe all pulse, any Ideas why these I/O line are floating at 1.6v?
thanks for your help so far, much needed and apreciated.............Harry

Chuck(G)
July 15th, 2010, 04:28 PM
With OE/ grounded, the output of U56 should not float under any circumstance. The question in my mind is "are these truly floating?". Try grounding pin 12 of U56 briefly. Does the display change?

If not, try pulling first one then both character generator EPROMs, U57 and U58 and then look at the output of U56 and see if it looks any cleaner. The only thing that U56 is driving are the address inputs of the EPROMs. If they're stuck at 1.6V, we have two possibilities, it seems to me. Either one of the EPROMs has gone bad and is loading U56's output lines or U56 itself is bad.

You should also check the input lines to U56 to make certain they look normal. They should.

harry
July 16th, 2010, 02:45 PM
With OE/ grounded, the output of U56 should not float under any circumstance. The question in my mind is "are these truly floating?". Try grounding pin 12 of U56 briefly. Does the display change?

If not, try pulling first one then both character generator EPROMs, U57 and U58 and then look at the output of U56 and see if it looks any cleaner. The only thing that U56 is driving are the address inputs of the EPROMs. If they're stuck at 1.6V, we have two possibilities, it seems to me. Either one of the EPROMs has gone bad and is loading U56's output lines or U56 itself is bad.

You should also check the input lines to U56 to make certain they look normal. They should.


Chuck, I tried grounding OE on the 2016 but this has no effect on the inputs or outputs of u56, it is the lines going to the address inputs u57 and u58 that are stuck high, the lines (bw) coming from the 2016p to u55,u56 ,& u66,are all stuck at 1.6v. But grounding pin 12 of U56 changes the screen, to a full screen of
tilde characters. I have removed the roms, u57 does not effect the screen, when U58 is removed the screen is completely blank, . I can refit U57 on its own, put cs to ground and the screen fills with graphic characters of some sort.
with both roms removed, the i/o lines from the 2016 are still fixed at 1.6v, they are also tied to U66 so it seems this voltage has to come from either U66 or u54, , I
am going to try and source a new 2016p, so far they seem difficult to find?
any further thoughts a great help!.........many thanks .....Harry

Chuck(G)
July 16th, 2010, 07:55 PM
The 2016 is/was made by various manufacturers (TI TMM2016, Dallas DS2016...etc.), but the good news is that the same pinout and functionality is shared by the 6116 SRAM. Those should be plentiful as cockroaches (I've got a couple of drawers full of them in my parts bin). Futurlec has them, but you can probably find a bunch closer to home.

harry
July 17th, 2010, 12:25 PM
The 2016 is/was made by various manufacturers (TI TMM2016, Dallas DS2016...etc.), but the good news is that the same pinout and functionality is shared by the 6116 SRAM. Those should be plentiful as cockroaches (I've got a couple of drawers full of them in my parts bin). Futurlec has them, but you can probably find a bunch closer to home.

Chuck, thanks for info, I'll be keen to get some parts and see what progress I can make, many thanks for your help and time, ............Harry

nige the hippy
July 17th, 2010, 04:13 PM
I'm afraid I didn't read your later posts, so no, - I was barking up the wrong tree. Chuck seems to be pointing you in the right direction so, I won't chip in with any more confusing suggestions.
I have an 820 II, minus keyboard. when I got it it had a dram chip down. I only got the disk drives recently, and the main machine was packed away. It now would probably work if I could find another keyboard. I may have something which would work, but I might put a shout on here...

I think I may have a couple of 6116s around, probably still on old boards, I'll have a look tomorrow.

nige the hippy
July 19th, 2010, 02:14 PM
I do in fact have just a couple, so I'm prepared to let one go. (at least for a while)
although there seem to be a few on epay at the moment.