PDP8/M M8330 Problem

[MauriceH]

For couple of day's I am struggling with my PDP8/M and the Timing Control Board M8330.
T's definitive this board , because with an other M8330 the machine is working fine.

The problem is, the PDP8 will not read in From Asynchrone M8655 (1200Bd), Terminal VT220.
Did all the toggle in tests 20 till 37. (The known toggle in written test paper).
It reads in the M847 Bootloader card.
It print characters on screen,
But it fails to run prog: 33 the 03/04 adress Echo test and the
test 37 Keyboard read in test and display on AC the typed character.

All other tests like microinstruction test 1&2, operate instruction, ISZ, JMS,JMP,Vissual AC Incr.,Checker Board runs perfect as aspected.

What causes that the M8330 wil not start read in Data from M8655?

Also it runs the maindec-08-dhkaa without errors (The read in I did with the working M8330),shut down machine,
swap boards,start pdp8 and run the program.
At display the code printed "HC" with SW=0003 (fast test), Same as with working M8330.

Any idea's where to look for the failure.

 

[vrs42]

For couple of day's I am struggling with my PDP8/M and the Timing Control Board M8330.
T's definitive this board , because with an other M8330 the machine is working fine.

The problem is, the PDP8 will not read in From Asynchrone M8655 (1200Bd), Terminal VT220.

The M8330 is basically a big ring counter, which shifts a pulse around every 50 ns to a set of gates and flops that set and reset at various times to generate all the timing signals.

Since your machine mostly works, except during I/O, I'd focus on signals like io_pause that are unique to that situation.

Vince

 

[MauriceH]

Vince,
Thanks Your so Right! I/O Pause, followed the signal and stopped at E17 .
A DEC 8881, think a Replacement N7439 , May be a 7401 would work for now, But a bit low on Current output draw.
The Port's 3 & 4 go to outside LINK LOADL , so think these needs to be lot or current and 7401 would not be sufficient.
Have to look for 7439 , and see if M8330 will work after this fix.
So far defect E17 found.

 

[vrs42]

Vince,
Thanks Your so Right! I/O Pause, followed the signal and stopped at E17 .

Interesting. So, you are generating io_pause_l , but not getting bus_strobe_l? What are you getting instead at bus_strobe_l?

It seens wierd to me that both E17 gates connected in parallel are failed. Hopefully there are not problems with R20, R21, or with the buffered not_last_xfer_l from E27.

Is link_load_l during RTF (from the other two gates of E17) also broken? (Then I'd worry about the supply at E17.)

The current you have to sink is coming from the 150 ohm resistors R35 and R20, which I think will dwarf any leakage from the other OC outputs and the couple of receivers listening. The 7401 would be pretty marginal there, but the 7439 should be fine.

Vince

 

[MauriceH]

I ordered some 7439's , will be a coulpel of day's.
Output Bus_Strobe = '1' ,Due to input 6="1"should be the invert of input 5.
Second 3&4 also parallel the Link_LoadL ,output is "0" , Input 9= Puls high, Input 8='0'' So Output constant '1', as aspected.
As output 4 is High, I for now assume Buffer E37 and resistors are OK.

Problem could have been occurred when debugging a defective M8650, that card has got me quite by the fingers, what a struggle.(weeks)
Have a working M8655 so at that point system is working correct.
I am now that far that an echo test (33 toggle) is working and get the right char at the VT220 back.(Now with the 2nd working M8330 that is).
But at print Char test (toggle 32) I got just random characters on screen.
Couple of failures I already digged out.(E03(7474),E12(7400), E43(8251,now 7442 conversion) (Board has never worked as long as I have it, so no Idea what happened previous with it)
Thought would be a sludgy STOP MARK CNT,(8271) But no.changed with a 74179
This last session of problem must be some kind of timing issue.

Thanks for the help.,Maurice

 

[vrs42]

I am now that far that an echo test (33 toggle) is working and get the right char at the VT220 back.(Now with the 2nd working M8330 that is).
But at print Char test (toggle 32) I got just random characters on screen.

If it correctly echos slowly, but doesn't get sequences of characters right, I'd suspect the number of stop bits generated doesn't match the receiver's expectations.

Vince