KME-8 (Memory Extension M837) - A little HELP please!

[AteEm]

While my 8m passes all the MAINDEC processor diagnostics, it appears I've got problems with the Extended Memory control board in that while Field 0 is good to go (passes the checkerboard testing), I can't access Field 1 (the second 4K). I have the 3 board all-in-one 8K core memory, not two 4K sets.
While I can load both the IF (instruction field) and DF (data field) registers from the front panel using EXT ADDR LOAD, and confirm the proper EMA bit is set, as soon as I try to DEPosit, the EMA bit clears. According to the KME-8 Acceptance Test Procedure this is NOT suppose to happen when using DEPosit; the EMA bits are suppose to remain to what was set with EXT ADDR LOAD. Also, loading the IF after the deposit and attempting to EXAM Field 1 also results in the EMA bit being cleared.

Any ideas?


Thanks,

Mike

 

[DaveH]

Hi, no thats not right, EMA should stay set, have you run the M837 Maindec thats 8E-D1FA ?

do both IF and DF clear or is it just one ? also what do you get if you set them both to 7 then do a deposit, do you get the same thing with a examine ?

Dave

 

[AteEm]

Thank you! I was not aware of the 8E-D1FA; I will track that down.

Both IF and DF clear. I have used all eight combinations, including both to 7, and they set fine with EXT ADDR LOAD, but all bits always clear when either DEP or EXAM is pressed.

Hi, no thats not right, EMA should stay set, have you run the M837 Maindec thats 8E-D1FA ?

do both IF and DF clear or is it just one ? also what do you get if you set them both to 7 then do a deposit, do you get the same thing with a examine ?

Dave

 

[DaveH]

Hi, OK , the IF and DF can be changed by there respective IOT commands but are also both cleared both at power up and at the start of an interrupt service, I would start by looking at the logic responsible for this clear signal, its a few gates on the M837,
the signel is "IB,IF,DF CLEAR L", IC E16 seems to do some of this . Now this may NOT be where the problem is, its just where I would start looking and the maindec may give us more information.
How are you for documentation, extender board and test gear, as I can see a session involving schematic, scope and coffee approaching.
Dave

 

[AteEm]

LOL! Yeah, coffee, for sure
I'm kinda set up old school: All I've got is an old Tek 7000 series analog storage 'scope. But, ten years back I managed to troubleshoot several bad ICs on various processor boards to resurrect the machine, so it should be adequate!
All my 'documentation' are downloads over the years; the schematics all seem to be poor scans of photcopies of photocopies. I have to blow up the image on the screen until I can decipher (barely) signal names and reference designators
Thanks for the pointer to signals to check. That's how I like to approach this: understand the circuit to a point where there's some logic to the signals probed that will hopefully give a clue to what to check next. Exactly the kind of help I was looking for! Thank you.
When I get home tonight I'll try to find the MAINDEC diagnostic, download, and post my results.

Hi, OK , the IF and DF can be changed by there respective IOT commands but are also both cleared both at power up and at the start of an interrupt service, I would start by looking at the logic responsible for this clear signal, its a few gates on the M837,
the signel is "IB,IF,DF CLEAR L", IC E16 seems to do some of this . Now this may NOT be where the problem is, its just where I would start looking and the maindec may give us more information.
How are you for documentation, extender board and test gear, as I can see a session involving schematic, scope and coffee approaching.
Dave

 

[Al Kossow]

the schematics all seem to be poor scans of photcopies of photocopies.

They are Xeroxes from DEC archival aperture cards of blueprints. That's the way they came from DEC.
Over the years, they got worse since they appear to have not gone back to the aperture cards.

PDP-8e drawings, in particular the CPU drawings, are ALL bad. The scans on bitsavers have been picked from
the best of the surviving drawings.

 

[DaveH]

They are Xeroxes from DEC archival aperture cards of blueprints. That's the way they came from DEC.
Over the years, they got worse since they appear to have not gone back to the aperture cards.

PDP-8e drawings, in particular the CPU drawings, are ALL bad. The scans on bitsavers have been picked from
the best of the surviving drawings.

Al, most of my 8e drawings are origonal DEC, could this help,? I can scan a few as a trial if you like
DaveH

 

[AteEm]

Back to an 8K machine!

I didn't even have to drag out the 'scope. I pulled the M837 board and took a detailed look at it under an illuminated magnifier...the board has been ECN'ed with a half-dozen wire wrap wire 'corrections' and I noticed that one which went to an edge pin looked like a poor solder job. When I tugged on the wire it pulled right out of the solder 'fillet'! The bare end was badly corroded. Cleaned it up, tinned it, and resoldered to the board. Problem solved!

Edge pin was B2M, LD ADDR ENABLE, and the ECN wire runs to the clock pins on both the IF and DF registers. So, pretty much makes sense that the M837 wouldn't 'remember' the field settings.

Mike

 

[tradde]

Back to an 8K machine!

I didn't even have to drag out the 'scope. I pulled the M837 board and took a detailed look at it under an illuminated magnifier...the board has been ECN'ed with a half-dozen wire wrap wire 'corrections' and I noticed that one which went to an edge pin looked like a poor solder job. When I tugged on the wire it pulled right out of the solder 'fillet'! The bare end was badly corroded. Cleaned it up, tinned it, and resoldered to the board. Problem solved!

Edge pin was B2M, LD ADDR ENABLE, and the ECN wire runs to the clock pins on both the IF and DF registers. So, pretty much makes sense that the M837 wouldn't 'remember' the field settings.

Mike

Those are the kind of fixes I like. Easy.

 

[Al Kossow]

Al, most of my 8e drawings are origonal DEC, could this help,? I can scan a few as a trial if you like
DaveH

Maybe. I have gone through at least a dozen Major Registers drawings and they were all poor copies. If you have the time, compare
the 8e drawings you have with the ones on highgate or bitsavers and see if you have any cleaner copies.

One of the big problems is the original 8e drawings were VERY large, and they didn't reduce well when they were archived.

A clean copy of the TD8E would be good to find also.

 

[DaveH]

Back to an 8K machine!

I didn't even have to drag out the 'scope. I pulled the M837 board and took a detailed look at it under an illuminated magnifier...the board has been ECN'ed with a half-dozen wire wrap wire 'corrections' and I noticed that one which went to an edge pin looked like a poor solder job. When I tugged on the wire it pulled right out of the solder 'fillet'! The bare end was badly corroded. Cleaned it up, tinned it, and resoldered to the board. Problem solved!

Edge pin was B2M, LD ADDR ENABLE, and the ECN wire runs to the clock pins on both the IF and DF registers. So, pretty much makes sense that the M837 wouldn't 'remember' the field settings.

Mike

Well done, its nice to get an easy one now and again. My trusty m went down a week or so ago with a popped breaker, that bit was easy, blown bridge rectifier, them when doing load/ voltage check before re installing I found the 15+ all over the place, after much prodding and removing, checking transitors ect I found it was "howling " at unpteen kHz, ended up with my own ECO, an extra 0.1uF slugging the feedback loop, that stopped it!!
Tonight start on my RX02, wish my luck
Dave

 

[DaveH]

Maybe. I have gone through at least a dozen Major Registers drawings and they were all poor copies. If you have the time, compare
the 8e drawings you have with the ones on highgate or bitsavers and see if you have any cleaner copies.

One of the big problems is the original 8e drawings were VERY large, and they didn't reduce well when they were archived.

A clean copy of the TD8E would be good to find also.

Ar, I see , I will have a look anyway, we may get lucky. Hope to have the 8a maindec listing scans done soon ( Lou,s 8a thread ).
DaveH