EAE M8340E+M8341D and M8340D+M8341B

[gnupublic]

in some threads here I read about different layouts of J and F connectors of EAE cards. Now I have a pair of EAE cards in my lab/8 that does not seem to fit together. Meanwhile I the got an other pair, that also does not fit, but now I can combine two pairs with equal headconnector layout:




For M8340E + M8341D somebody told me that he uses this combination. But what is about M8430D+M8341B? And what version of other CPU cards are needed?
I only found the engineering drawings of M8340F + M8341E on the web. There you find the remark M8330 REV. B or later.
A very helpful person send me the scan of the engineering https://retrobyte.de/download/pdp8_eae_schema_191010.pdf drawings for versions M8340C + M8341B. There you find near the different layout of the J CONN Pins: " to corresponding Pins on CONN. J of M833 REV. D or later"
(yes M833!)

Are there any further helpful informations?

I really would like to repair these cards and use them.
Have fun
Volker

 

[vrs42]

For M8340E + M8341D somebody told me that he uses this combination.

As do I.

But what is about M8430D+M8341B? And what version of other CPU cards are needed?
I only found the engineering drawings of M8340F + M8341E on the web. There you find the remark M8330 REV. B or later.
A very helpful person send me the scan of the engineering https://retrobyte.de/download/pdp8_eae_schema_191010.pdf drawings for versions M8340C + M8341B. There you find near the different layout of the J CONN Pins: " to corresponding Pins on CONN. J of M833 REV. D or later"
(yes M833!)

One thing to note is that the etch level is different than the "revision" indicated on the schematics. The schematics you point to describe an M8340D with revision C schematic, and an M8341B with revision B schematic. The M8340D card you pictured is schematic revision D, and the M8341B is schematic revision B (as stamped into the handles). So, your M8340D is a little newer than the one you have a schematic for.

The main difference I am aware of is the treatment of "RESTART", which is computed on the M8341, and carried from HH2 to JV2 (and thus to the M8330) on the M8340. These pins are grounded in the earlier cards. Presumably if your cards generate this, you'll also need a CPU revision late enough to receive this signal and act on it. I haven't made a study of which CPU revisions that would be, but the 8/E ECO document:
http://bitsavers.trailing-edge.com/pdf/dec/pdp8/pdp8e/PDP-8E_ECO_LOG_Dec74.pdf
seems to indicate this was added to the M8330 at etch revision C, with etch B requiring an ECO to add the feature.

Hope that helps!

Vince

 

[vrs42]

The main difference I am aware of is the treatment of "RESTART", which is computed on the M8341, and carried from HH2 to JV2 (and thus to the M8330) on the M8340.

RESTART is generated in the middle of the first sheet of the M8341 schematic, by the N8881 at E36. That suggests to me that it is actually active low, even though it isn't called "RESTART L". Since the older EAE cards have the relevant pins at the top connectors grounded, the result I would expect is that if you put an old EAE on a new CPU, the CPU will believe RESTART is constantly being asserted. Since the 8881 is open collector (with a pull-up on the CPU?), I don't *think* you would actually damage things by mix-and-match.

Hey Michael! Does this match your experience?

Vince

 

[gnupublic]

Today I managed to repair the first pair of EAE cards. Now my Mandelbrot program is nearly double the speed then without EAE. That's great!
Is there any other reason to open a bottle of champain today?

Best wishes for the New Year

Volker

 

[gnupublic]

Today I finished repairing these cards and now I have three working pairs of EAE options.

After some strange findings while repairing one M8341, I found that this EAE (with the newer F and J connector layout) came with a CPU with old connector layout (M8330). I didn't expect that and had no look at this before. The M8341 was already repaired after changing two DEC8881 (as I found out later) but if I execute a simple NMI (7411) the complete machine hang. I had to switch off/on. But this was exact the configuration I got this pdp8/e. I then moved the EAE to my lab8/e with M8330 with newer layout, and the EAE worked spot on. EAE Inst. 1+ 2 are OK.
So I plugged the EAE with old connector layout into the CPU with old layout M8330 and this also works! The oldest cards without any error. Yeah!

So this is the final combination of the three working CPU/EAE configurations finally:

card cardletter/handle ..

M8330	B/209A	-/750P	C/353D
M8340	D/142D	E/341F	E/427F	EAE
M8341	B/140B	D/E	D/422E	EAE
M8310	F/337F	B/148C	F/350F
M8300	B/244C	B/23B	B/351D
M837	B	D/320D	D/348E

Greetings,
Volker

 

[m_thompson]

My 8/e has the newer CPU boards, and I have an older set of EAE boards. I plan to disconnect the edge contact from Ground and connect it to RESTART so I can add EAE to the 8/e.

 

[gnupublic]

Do you think rerouting of JV2 of M8340 is the only thing to change to make an old EAE working with new CPU (M8330)?

 

[m_thompson]

Do you think rerouting of JV2 of M8340 is the only thing to change to make an old EAE working with new CPU (M8330)?

I think that it is worth a try. The EAE boards are useless to me otherwise.

 

[gnupublic]

In the newer version RESTART comes from M8341 E36 (Pin 1) going to HH2. HH2 is connected to M8340 and there going to JV2.
Unfortnately my shematic for the old version is not complete and is missing one part of M8341.... But on M8340 HH2 is also going to ground like JV2.
I'm not at home at the moment and will have a look at this in the evening.

 

[gnupublic]

In the old M8341 RESTART comes from E36 Pin 1+4 and goes to CL1.
CL1 on new M8330 goes to E17 pin 1 (DEC8881). RESTART is expected from JV2 going to E37 (DEC7417) pin 1. So I combined the old EAE wit new M8330 in my lab8/e. On the M8330 I put some tape over JV2 and connected E37 pin1 with E17 pin 1. I also removed one pull up resistor near E37, because near E17 is also one.
Then I tried loading the stepcounter from memory and back to AC (singlestep):
7403
0000
7641

I saw a 11 instead of an expected 37. Does not work.

Also SHL:
7301
7421
7301
7413
0000
5003

singlestepping this loads 0001 in MQ and AC an does shift left. But that also does not work.

So I think combining old EAE wit newer M8330 needs a bit more modification, or I did something wrong.

 

[gnupublic]

I have to correct myself!
After bringing the lab/8 in the normal configuration, I had a lot of trouble. Finally I found a bad H851 connector. After changing this, the old configuration passed CPU Inst 1+2 , EAE Inst 1+2 tests.

So I asked myself if this connector may have disturbed the test new M8330 with old EAE?

So I again prepaired the cards and configuration. And now stepcounter and shifting is fine with singlestepping. EAE Inst 1 (MAINDEC-8E-D0MB-D-PB.bin) runs fine some minutes.

But EAE Inst 2 (MAINDEC-8E-D0MB-D-PB.bin) crashes the CPU, it's not responding to HALT.

Thats a quite different result, not complete satisfying, but a bit better.
Now it's late. Tomorrow I will change the M8330 and repeat the tests.

Greetings,
Volker

 

[Roland Huisman]

What is wrong with the H851? Contacts or pcb? I had problems with these H851 too

http://www.vcfed.org/forum/showthread.php?54312-Dec-H851-pcb-redesign

Regards, Roland

 

[gnupublic]

Three bridges where open. After soldering the pins it is OK again. I marked the H851 as suspect. Klemens Krause gave me the tip to measure all H851 before usage. I did. But this one became bad after some usage. Very annoying.
Maybe I should have a look at your gerber files.
Greetings,
Volker

 

[gnupublic]

Testing with an other M8330 (old layout) shows same behaviour, so there is somthing wrong with this EAE cards.
I'm working on that.
The EAE Inst 1 test runs since 15 Minutes in A Mode only. In A + B mode it stops after some minutes.
EAE Inst 2 freezes the CPU.
Every time you do MVY or DIV the CPU hangs.

So I come back when the card is debugged.

Greetings,
Volker

 

[gnupublic]

Today I was able to find the last remaining bad IC on the board. I had some luck since the freezing stopped when I put the IC clamp over E05 for doing some measurements with my LA. Removed the clamp, freeses again.
Now the card passes both EAE INST1 and EAE INST2 test several times or 20 minutes each. (In the old layout environment).

To come back to the mixed layout configuration I changed the M8330 against the one in new layout. Put some tape over JV2 and connected E37 pin1 with E17 pin , and removed one pull up resistor. Then I also run the tests. EAE INST2 was OK with now problems. But EAE INST1 was not very stable. Had some problems starting the test, sometimes it stopped with ACS. When it runs several minutes, it stops with SAM error.
So I would say, in general it might work to combine the cards. Maybe the timing of my card is slightly incompatible with this EAE, that may vary from card to card. Maybe the pullup needs to be optimized.

After putting the card back in the new layout environment, it works without any error.

Greetings,
Volker