• Please review our updated Terms and Rules here

PDP-11/35 - The real fun has begun

Thanks for going to that trouble. I'll spend some more quality time with the available schematics, to see what I can come up with for when you get the extenders.

Any place where we can see current pics?

BTW - Are those cables indeed shielded? They don't look it.
 
Twylo said:
This may also be a source of some confusion...

I have been saying that I have the KY11-D console. This is incorrect, I don't! I have the KY11-K console. My console has the BERG connectors on the right, and the power connector on the left.
Click to expand...

Ah yes, thank you - this changes things considerably. You know, I’ve often wondered how DEC made the profit they did considering all the product variations they supported early on.

Good luck sleuthing!
 
KM11 said:
Ah yes, thank you - this changes things considerably...
Click to expand...

Any idea if the documentation [mostly schematics] of the KY11-D will be close enough, or should I start digging through my stash?

...You know, I’ve often wondered how DEC made the profit they did considering all the product variations they supported early on.
Click to expand...

It was indeed a concern to them throughout the company history. Products like the PDP-10 evolved so much that DEC found itself buried under all the engineering "support weight" of it's success. The same thing was happening in the PDP-11 line. They began to realize it had to be solved when orders for new product typically turned out inoperable "as configured".

Their ultimate solution was to use product evolution to drive sales into "Pre-Configured Systems". The MicroPDP-11 and Decsystem 20 lines of the early 80's were the result. As an additional discouragement, they would not "support" PDP-10, 20 option configurations they didn't "Bless".

Ultimately, DEC had to [forcibly] cease production of the 36 bit lines altogether to make them "go away". There was no technical reason for this, and perhaps not even a marketing one. Many PDP-10 systems were still in use when DEC was later acquired, and higher performance "Clones" and replacement parts were available, most customer engineered out of frustration at the neglect .

I find this very ironic in light of DEC's mid 80's declaration that they "would rather have 2% of IBM's business than all the customers they had" [Olsen]

Just what kind of example did they think it was for a target market so see how they treated their last, loyal, "Large System" customers?

...and the PDP-11's? They sold the line to Mentec to eliminate the engineering "distraction". I would like to see the justification logic for this decision. DEC was deliberately creating a competitor in smaller systems, rather than evolve a line that could easily outstrip the commercial success of the company that would eventually acquire DEC (Compaq).

It is a wonder, indeed.

Sorry, my sour grapes are showing again.
 
Last edited:
Just a very quick note. I got my board extenders yesterday (yay!) and I'm in the midst of writing up a test plan for tracing the signal from the HALT/ENABLE switch. The maintenance manual and engineering drawings are absolutely fantastic, and I've been making good use of them.

RSX11M+ said:
Any idea if the documentation [mostly schematics] of the KY11-D will be close enough, or should I start digging through my stash?
Click to expand...

I think I'm alright. The 7410 logic seems to be identical between my KY11-K and the KY11-D schematics. The only difference seems to be that J1 and J2 are flipped over to the other side of the board, and the shield is AWAY from the board on the KY11-D, but TOWARD the board on the KY11-K. I guess the KY11-K was always paired with the 11/35 in the 10.5" chassis.
 
Well well well well well. I think I have discovered a very big part of the problem.

Back when I originally fixed my power supply, I adjusted all of the voltages with no load. I _thought_ this was the correct procedure based on my scan of the H750 manual. But tonight when I finally measured Vcc directly on one of the ICs, I discovered to my embarrassment that it was reading around 4.4V. I know TTL should be able to run as low as 4.7V, but 4.4V seems like it would be well into the "flakey" territory.

Unfortunately, I didn't get much of a chance to test this theory. As I was adjusting the 5V level on the supply, the regulator failed :(

I've become very close friends with the H750 power supply, so I was quickly able to isolate the problem: The 15A pico fuse (F1) and the bridge rectifier on the 5409728 regulator had both failed. That pico fuse on the +5V/+15V regulation circuit is blown by the crowbar circuit if the +5V circuit is adjusted over 105% rating. It's quite possible that I accidentally did this, and I guess the pico fuse must not have blown fast enough to protect the rectifier.

So, alas, I need a new rectifier. I have ordered a couple of 200V 35A diode rectifiers in GBPC-W packages. One of these should be more than enough. I hope nothing else on the board was damaged. I'd also better check that crowbar circuit to make sure I didn't damage it before the rectifier blew. There's no visual indication that anything got too hot, and certainly no magic smoke got out, so I hope that will solve it.

Once that's sorted out (and I hope it will be) I can get back to trying to fire up the logic at something better than 4.4V!

-Twylo, always causing trouble
 
The plot thickens.

I was wrong. The rectifier is NOT shorted. I pulled it out and tested it outside the circuit. It seems totally fine.

So, SOMETHING is shorted on the board, specifically something in the +5V rectifier circuit. What it is, I have no idea. I am going to have to track it down. A real pity, finding shorts in-circuit is such a pain :(

-Twylo
 
OK, my last note until I get some spare parts in. It looks like D11 on the 5409728 regulator board was shorted. Not only that, but it had CLEARLY been replaced before, and whoever did it did a really lousy job - the traces are lifting where it was badly re-soldered. I'll have to tread carefully when I put in a replacement.

I have a bunch of parts on the way (for the grand total cost of about $19, thanks Mouser!) that should get me through any other scary discoveries. This regulator board has clearly seen better days. I'm certainly not the only one who's repaired it.

Of course, the wiser man might throw in the towel and just get a replacement 5409728 regulator board (not that I know where I would find one even if I wanted one), but I am not a wiser man.
 
I'd say you were the wiser man for making the component level repair. You will learn that power supply well and be able to help the next guy that has a problem with his.

While you are in the H750, of course check all the big electrolytic caps. When I built my 11/04, I had an H745 with a shorted cap.

Lou
 
We're Back!

We're Back!

It took long enough, but I fixed my dang power supply. I'm embarrassed to admit that I made it much harder on myself than it needed to be. The 5V regulator really is a pretty simple circuit with not very many parts. Anyway, it's fixed now and hopefully it'll stay that way assuming no other parts go bad.

I've made some progress on the CPU as well. I won't rehash the entire thing here because details are on my blog, but in summary:

  • Having correct +5V voltage did not magically fix the CPU. It still acts as it did before.
  • I have traced the HALT/ENABLE switch to both the M7235 and M7234 modules, and verified without a doubt that the signal is getting through.
  • I found a bad 7404 in the signal path of the HALT/ENABLE switch (position E75 on the M7234) and replaced it.
  • I watched the power-up waveforms (see page 3-11 in the KD11-A Maintenance Manual) on a logic analyzer and an oscilloscope, and found at least one timing problem.
  • I think I have identified a bad 74123 related to the power-on waveform problem, and will replace it as soon as I get a new one.

For more gory details, my weblog is here: http://www.loomcom.com/blog/

Now, I have two questions for y'all!

  1. I'm using 74LSxx parts to replace original TTL 74xx parts. The 74LSxx parts are obviously going to be faster and use less power, but should have the same load capacity and otherwise be a drop-in replacement. If this is not true, please please correct me!
  2. The KD11-A uses some 74Hxx parts. So far both the failed IC's I've identified are plain old 74xx parts, but I bet I'll find some failed 74Hxx parts along the way. Can I use 74ALSxx and/or 74Fxx parts for these? Again, I think I can, but I'd like to be corrected if that's wrong.

-Twylo
 
Twylo said:
I'm using 74LSxx parts to replace original TTL 74xx parts. The 74LSxx parts are obviously going to be faster and use less power, but should have the same load capacity and otherwise be a drop-in replacement. If this is not true, please please correct me!
Click to expand...

I believe that his is not true in general. Specifically, I believe the IOL of an LS04 is 8mA, where the
7404 is 16mA. About half the load capacity, in this case.

Twylo said:
The KD11-A uses some 74Hxx parts. So far both the failed IC's I've identified are plain old 74xx parts, but I bet I'll find some failed 74Hxx parts along the way. Can I use 74ALSxx and/or 74Fxx parts for these? Again, I think I can, but I'd like to be corrected if that's wrong.
Click to expand...

The ALS04 is still an 8mA part. The F04 is a 20mA part, so that might work. So is the S04. A main difference between an 74xx and a 74Hxx is the speed, though, so you'd want to check the speed of SSI parts for 6ns or better. (The S04 is 5ns, the F04 is 6ns, and the regular 04 is 12ns.)

Hope that helps!

Vince
 
vrs42 said:
I believe that his is not true in general. Specifically, I believe the IOL of an LS04 is 8mA, where the
7404 is 16mA. About half the load capacity, in this case.
Click to expand...

Damn :(

I just found my copy of "The TTL Cookbook", I should have thought to look in there first. It has a nice chart on page 31 of what families can drive how many of what other family's inputs. I'm going to have to be very careful here in deciding what chips I can and cannot replace with LS parts.

Thanks for the warning!

-Twylo
 
Still Going!

Still Going!

I'm still debugging, and for now I'm focusing all my attention on one particular start-up oddity. Namely, the "PWERUP INIT L" line, which is supposed to go low for 20ms, but which almost never does.

I've recounted the problems on my blog: http://www.loomcom.com/blog/2012/08/06/new-theory/

Something is keeping that 74123 from behaving the way it's supposed to -- and not all the time, just most of the time. I have no idea if my wild guess about the spurious 74H04 output is a good guess or not, but I'd rather not to replace any more ICs unless I'm absolutely sure that I'm doing the right thing. I hate to think that the original 74123 was alright, but I replaced it anyway. I don't want to do "shotgun debugging", as they say.

-Twylo
 
vrs42 said:
I believe that his is not true in general. Specifically, I believe the IOL of an LS04 is 8mA, where the
7404 is 16mA. About half the load capacity, in this case.



The ALS04 is still an 8mA part. The F04 is a 20mA part, so that might work. So is the S04. A main difference between an 74xx and a 74Hxx is the speed, though, so you'd want to check the speed of SSI parts for 6ns or better. (The S04 is 5ns, the F04 is 6ns, and the regular 04 is 12ns.)

Hope that helps!

Vince
Click to expand...

http://www.ti.com/lit/an/scla007a/scla007a.pdf

Is a nice summary of the various families
The AS and F were high current, fast (at the time) rise time parts and you could get in trouble with ringing on long unterminated trace runs.
 
You must log in or register to reply here.
Back
Top