• Please review our updated Terms and Rules here

PDP-12 #435 at the University of Minnesota Duluth

Our M401 has the etch M401H, but there's a stamp on the front that says L... I know that sometimes the same etch is used for different revisions. Is that what the stamp means?
It's an H etch as you identified. I'm not expecting the schematic revision to be an inspection stamp. Instead, the schematic revision is usually pressed into the plastic of the magenta handle.

You have a number pressed into the handle, but I don't see a letter there. So I suspect you've got a revision H. Unfortunately, I don't have drawings for the H, and even the L has a second 7400. This means that you'll have to hunt for which pins of which parts do what. Do you have a schematic for revision H?
I've attached some pics for reference. One of the pins on the SN7400 looks to me like it may have gotten hot -- but I may just be misinterpreting what I'm seeing.
That doesn't look conclusive to me. The solder may have been reheated to clean up a solder joint.
As an aside, if we do have to replace the SN7400, do you folks recommend adding a socket for future maintenance (if there's space)? It doesn't feel very period, but of course it would make future maintenance easier. Just curious.
I'd say it depends on your soldering skills. It they are sketchy, best thing is to clip out the old and install a socket, all the while minimizing the heating time on the pads (while still getting the solder warm enough to flow well). If the available soldering tools and skills are good, there's less concern about damaging the relic, and more about maintaining the original look :-).
 
It's an H etch as you identified. I'm not expecting the schematic revision to be an inspection stamp. Instead, the schematic revision is usually pressed into the plastic of the magenta handle.

Ok, thanks.

You have a number pressed into the handle, but I don't see a letter there. So I suspect you've got a revision H. Unfortunately, I don't have drawings for the H, and even the L has a second 7400. This means that you'll have to hunt for which pins of which parts do what. Do you have a schematic for revision H?

Huh, interesting. I don't think we have a schematic for revision H. We have an original module book with drawing for J, and a print off bitsavers with a drawing for M. We can try to look around online for a schematic.

Are you interested in making a drawing of the H for your own purposes or posterity? If there's anything we can do to assist you in that, please let me know.

That doesn't look conclusive to me. The solder may have been reheated to clean up a solder joint.

Thanks, @vrs42 and @NF6X -- that's encouraging. I won't assume the SN7400 is dead.

I'd say it depends on your soldering skills. It they are sketchy, best thing is to clip out the old and install a socket, all the while minimizing the heating time on the pads (while still getting the solder warm enough to flow well). If the available soldering tools and skills are good, there's less concern about damaging the relic, and more about maintaining the original look :).

I think my soldering skills aren't amazing, but are good enough that I'm not too worried about damaging the relic. But it would be nice to be able to pop that 7400 out and test it in isolation...
 
There is some likelyhood that the relevant portion of your revision wasn't changed before revision L. If that is so, then 7400 pins 1 and 2 are K and J, respectively, and pin 3 is the enable for the oscillator. Also, prints 4, 5, and 8 would be E, and pin 6 would be D. You can check for continuity (with the circuit unpowered!) with the edge connector pins to see if that is the case.
 
Are you interested in making a drawing of the H for your own purposes or posterity? If there's anything we can do to assist you in that, please let me know.
What I'd need is a good photograph, as "square on" as possible, of the component and solder sides of the board. That revision J schematic would also be great, as it is probably closer to what is on your board. Does it show one 7400, or two?
 
Hi @vrs42 , Here's the schamatic for J. Unfortunately for us, it includes two ICs, a 7400N (E1) and an 74H00N (E2). Ours only has a 7400N.
1727453127279.png
I've also attached 600dpi scans of the flip chip. Please let me know if you need anything else.
 

Attachments

  • M401H-front-crop.jpg
    M401H-front-crop.jpg
    1 MB · Views: 15
  • M401H-back-crop.jpg
    M401H-back-crop.jpg
    874.9 KB · Views: 15
Looking at the schematic and our probe trace, @ZachyCatGames and I note that at least one of the NAND gates is functioning, because we get a high output signal on E2 and an inverted signal on D2. How common is it for an IC to fail partially (e.g., for some but not all of the gates on an IC to fail)?
 
How common is it for an IC to fail partially (e.g., for some but not all of the gates on an IC to fail)?
.. it depends.. on how the circuit is designed. ex: If one of 4 gates connects to a bus or external wiring it'd likely see a higher odds of failure compared to a gate that has only connects to other internal logic IC's on the same module. It's also possible that some IC's don't have all gates used in a particular circuit. DEC schematics usually show unused gates on IC's so in a desperate case you /could/ cut the IC pins of the failed gate and rewire in a equivalent spare gate from another IC.
 
There is some likelyhood that the relevant portion of your revision wasn't changed before revision L. If that is so, then 7400 pins 1 and 2 are K and J, respectively, and pin 3 is the enable for the oscillator. Also, prints 4, 5, and 8 would be E, and pin 6 would be D. You can check for continuity (with the circuit unpowered!) with the edge connector pins to see if that is the case.
This does not seem to be the case.

AFAICT, you've got E on pins 1, 2, and 8, D on pin 3, K and J on 4 and 5 respectively. Pins 9 and 10 are the unbuffered oscillator, so if that is oscillating but D and E aren't, then the 7400 is likely at fault. (I think pin 12 should be ramping up and down.)

Also, check that the .01uF cap on S isn't shorted. I think it is the smaller of the two large white ones.
 
I've got a running start on reverse engineering the M401H. To dot all the "i"s I have a couple of questions:

When viewed from the component side with the edge connector on the right...

From the photos, I can't see how the base of the transistor just left of E1 pin 1 is connected to anything. Can you look at the module and tell me what you see?

I see a via hiding under the right corners of the .01uF and 0.1uF caps. It seems to connect to connector pin M on the solder side, but what does it connect to on the component side? I expected to see an RC network there, but I can't see any more components or what connects to M.

Thanks,

Vince
 
Never mind, I think I've managed to sort things out, thanks to some shadow lines visible through the board.

I also found M401C and M401L modules (while looking for an H). I scanned them and put pictures in the repository.
 
This does not seem to be the case.

AFAICT, you've got E on pins 1, 2, and 8, D on pin 3, K and J on 4 and 5 respectively. Pins 9 and 10 are the unbuffered oscillator, so if that is oscillating but D and E aren't, then the 7400 is likely at fault. (I think pin 12 should be ramping up and down.)

Also, check that the .01uF cap on S isn't shorted. I think it is the smaller of the two large white ones.
Hi @vrs42 -- we put the M401 into a loose FC slot and are wiring it up for testing. I hope that we can do some testing tomorrow.

We have never probed a flip chip like this before. I think I understand the main ideas, but:

Assuming we wire up power and ground as described, and the N to S loop for the oscillator selection, we should be able to probe the pins as usual, right? (That is, as if it were simply in the machine?) I guess what I'm asking is: are there any special considerations for testing a chip in isolation like this? Do we need to worry about having any external loads or resistors or anything like that?

If we want/need to provide an input to the M401, can we simply touch the appropriate pin to +5V to simulate a logic high input?

You mentioned trying to prove the unbuffered oscillator pins on the 7400... this makes sense to me as a way to see if the oscillator is working even if the signal isn't making it through the 7400. To do that, can we just probe the IC pins directly (carefully) using the scope or logic probe (assuming we can get decent purchase somewhere)? The Saleae has some dainty probes that I think might work well for this purpose if so. Again, as long as I have the probe or scope grounded, that should be OK, right?

Testing whether the capacitor is short makes sense, and we will do that with our multimeter.

I'll post some images of our rig tomorrow before plugging anything in.

Thanks!
 
I found three M401L spare. If you need one I can send it to you.
Hi @gnupublic -- wow -- thanks! I would not say that we need a spare M401, but it could definitely be very helpful to have a spare M401, since we only have one. We have a small budget for purchases, I would be very happy to discuss purchasing one from you at a fair price if that's something you'd be open to. We can DM about it if you'd like. And thanks again, either way!
 
Do you have a single-wide extender card?

No reason to wire up an otherwise empty backplane slot to test this. You should be able to use the existing slot with an extender to probe the various points carefully with an oscilloscope.
 
This would also be a good opportunity to learn how to write a new test for the Flip-Chip tester and learn how to test cards with the oscilloscope. You will need to pull some existing jumpers and hook up some additional jumpers to configure the frequency (i.e., N to S), but this is easy.

My recommendation is to look at existing tests, and perhaps Vince or Doug will chime in with more info as well.
 
Assuming we wire up power and ground as described, and the N to S loop for the oscillator selection, we should be able to probe the pins as usual, right? (That is, as if it were simply in the machine?) I guess what I'm asking is: are there any special considerations for testing a chip in isolation like this? Do we need to worry about having any external loads or resistors or anything like that?
The output signals should not require any special loading, unless you want to stress test them.

If we want/need to provide an input to the M401, can we simply touch the appropriate pin to +5V to simulate a logic high input?
It's probably safer to use the 3V output on pin V. That has resistors to prevent shorting the supply rails, which can happen if you try to drive something high with +5V, but it's already being held low somewhere.

You mentioned trying to prove the unbuffered oscillator pins on the 7400... this makes sense to me as a way to see if the oscillator is working even if the signal isn't making it through the 7400. To do that, can we just probe the IC pins directly (carefully) using the scope or logic probe (assuming we can get decent purchase somewhere)? The Saleae has some dainty probes that I think might work well for this purpose if so. Again, as long as I have the probe or scope grounded, that should be OK, right?
Yes. In my experience the thing to be careful with is not to short adjacent pins.
 
Hi @gnupublic -- wow -- thanks! I would not say that we need a spare M401, but it could definitely be very helpful to have a spare M401, since we only have one. We have a small budget for purchases, I would be very happy to discuss purchasing one from you at a fair price if that's something you'd be open to. We can DM about it if you'd like. And thanks again, either way!
I could probably part with one also, and domestic shipping is bound to be cheaper.

Assuming my spares work :-).
 
Do you have a single-wide extender card?

No reason to wire up an otherwise empty backplane slot to test this. You should be able to use the existing slot with an extender to probe the various points carefully with an oscilloscope.
We actually do not have an extender... that's been something on our list for a long time, but haven't yet needed one.
This would also be a good opportunity to learn how to write a new test for the Flip-Chip tester and learn how to test cards with the oscilloscope. You will need to pull some existing jumpers and hook up some additional jumpers to configure the frequency (i.e., N to S), but this is easy.

My recommendation is to look at existing tests, and perhaps Vince or Doug will chime in with more info as well.
This is a great idea, Kyle, thank you! We will do this, regardless of how we actually end up debugging our current problem. We'd love to contribute to the set of tests.
 
The output signals should not require any special loading, unless you want to stress test them.


It's probably safer to use the 3V output on pin V. That has resistors to prevent shorting the supply rails, which can happen if you try to drive something high with +5V, but it's already being held low somewhere.
Thanks. Yes, see, this is exactly the kind of newbie mistake I would make. This makes sense, thank you.

We have a power supply with (I think) two channels... if a flip chip did not have a predictable 3V output, would there be anything wrong with having one channel of our power supply set to +5V and one to +3V, using the +5V channel for power and ground, and using the 3V channel for inputs in a situation like this?

Yes. In my experience the thing to be careful with is not to short adjacent pins.
Yes. Thanks -- that's exactly what I would be worried about.
I could probably part with one [M401] also, and domestic shipping is bound to be cheaper.

Assuming my spares work :).
That is also a very kind offer, thank you! More generally, I'm interested in building up a set of spares for the FCs in our machine, as we have basically none. We are also happy to repair FCs, which would be good practice for us.

I'm also hoping to test the M401 later with Zach, so hopefully we will have more info... maybe we will end up being able to diagnose and fix our current M401 quickly.

Thanks again to all of you for your patient and generous help...
 
We have a power supply with (I think) two channels... if a flip chip did not have a predictable 3V output, would there be anything wrong with having one channel of our power supply set to +5V and one to +3V, using the +5V channel for power and ground, and using the 3V channel for inputs in a situation like this?
That could work, but you'd want to set the current limit on the 3V channel for 16mA or less.

The thing you are trying to avoid is to provide large currents in the event that an actual fight occurs, which can happen if you get the wrong pin, or even if things are just not as you expected. I suspect many of us have melted a trace and/or fried a component this way, and you definitely don't want that in your vintage gear.
 
Back
Top