• Please review our updated Terms and Rules here

New, to me DEC PDP8E

I got a chance to get back to my friends machine shop, where my PDP8E came from. The Kearney and Trecker machine tool was a horizontal mill, model MM100. Apparently an early version of it. He had a manual for it and it had a date of 1978 on it. The tool itself is rather neat. On top there is a rotating tool holder where an arm can grab one of maybe 15 tools and place that tool on the power head. The table, where the material is secured will raise and met the spinning machine tool. The computer then controls both the table and power head movements. They still have a bunch of old paper (actually some kind of Mylar) tapes with zillions of part programs on them. I found a couple of old pictures.

151203_0002.jpg

151203_0009.jpg

Thanks Mike
 
Here are the pictures of the Timing generator signals TP 1, 2, 3 & 4. My schematic of the timing board is rather poor. Most of the Chip ID numbers are to blurry for me to read, so I toke some time to ID each chip and update my print. The TP's wave forms are also rather squiggly, like the TS wave forms were. But it looks like the time periods and relative levels are OK.

This is TS1 and TP1.

TS1_TP1.jpg

This is TS1 and TP2.

TS1_TP2.jpg

This is TS1 and TP3.

TS1_TP3.jpg

This is TS1 and TP4.

TS1_TP4.jpg

Mike
 
Mike_Z said:
I got a chance to get back to my friends machine shop, where my PDP8E came from. The Kearney and Trecker machine tool was a horizontal mill, model MM100. Apparently an early version of it. He had a manual for it and it had a date of 1978 on it. The tool itself is rather neat. On top there is a rotating tool holder where an arm can grab one of maybe 15 tools and place that tool on the power head. The table, where the material is secured will raise and met the spinning machine tool. The computer then controls both the table and power head movements. They still have a bunch of old paper (actually some kind of Mylar) tapes with zillions of part programs on them. I found a couple of old pictures.

View attachment 28136

View attachment 28137

Thanks Mike
Click to expand...

A beauty!

Someday, someone will figure out that these are as historically noteworthy as the computers themselves. Unfortunately, most people won't be hoarding them in collections at home.

(The part where the tools reside is called a carousel, and where they go to do the machining is called a spindle).
 
Mike_Z said:
Here are the pictures of the Timing generator signals TP 1, 2, 3 & 4. My schematic of the timing board is rather poor. Most of the Chip ID numbers are to blurry for me to read, so I toke some time to ID each chip and update my print. The TP's wave forms are also rather squiggly, like the TS wave forms were. But it looks like the time periods and relative levels are OK.

snip horrific photos

Mike
Click to expand...

Gawd, those look awful!
It's been quite a few years since I had to troubleshoot my 8m to get it running, but I don't remember all of that ringing/overshoot on the TS/TPs.
Do you have the termination board installed in the last slot? Is the scope properly compensated? (I don't think I've ever seen an uncompensated probe look that bad, however)
 
KC9UDX said:
Yikes! You're not kidding! It doesn't look like a probe problem. I've seen bad scopes do that, but I doubt that's a bad scope!
Click to expand...

Those traces look exactly like what you might see using a standard passive probe with a 4-6" flying ground lead. The extra inductive ground loop causes that ringing display (which is not really there).

probe.jpg

Where/how you connect the ground lead is also a consideration; it should be ground very close to where you are sampling the signal.

095611_162517.jpg

A probe with a ground sensor lead right at the probe tip would be better (Tek probes came with such an accessory) but I suspect that add-on is long gone.

If your scope has a 50ohm input impedance option, making a high performance passive probe with coax and a 950ohm (for 10:1 voltage divider) resistor at the tip is easy and cheap.
Cut the resistor leads as short as possible, solder it to the center conductor. Tack solder the resistor to the probe point. Tack solder the coax shield to a very close adjacent ground.
You will see much better signal fidelity on the scope compared to the standard probe and the flying ground lead (likely with an alligator clip).

81_6.jpg
 
Last edited:
AK6DN said:
Those traces look exactly like what you might see using a standard passive probe with a 4-6" flying ground lead.
Click to expand...

OK, thank you!

I have not encountered that much ringing that I can recall, but perhaps I haven't had just the right circumstances, or I just don't recall.
 
Since the signals are relatively slow on the 8 I usually just ground the 'scope to the 8's chassis and take the ground leads off the probes - saves the risk of accidentally shorting something whilst poking around. Unless you're really concerned about timing integrity a bit of ringing doesn't get in the way of troubleshooting.

I'd agree with the earlier post though that even with my setup you'd not usually see so much ringing just due to an ungrounded scope probe and I'd suspect that the omnibus isn't terminated which is why there's such a regular lowish-frequency ring. I'd suspect that the termination card isn't in place or maybe the inter-omnibus connectors or some bad connections in there.

Still, the signals look correct which is the main thing just now!
 
I also thought that the wave forms were rather poor. So I looked at a youtube '#68: Oscilloscope Probe Ground lead length affects on signal quality'. This guy shows what happens with longer ground probe leads with high frequencies. My ground lead is at least 6" long. Maybe I was a little sloppy. Later, I'll try some of this guys ideas and re post the results. This morning I have to help my son install a new refrigerator, requires some carpenter work to get it in under some lower cabinets.

Thanks Mike
 
The pdp8 doesn't have any high frequencies comparatively speaking - only around 10MHz maximum and generally much less, but it's fast edges rather than frequency which gives rise to ringing. In fact though the pdp8 uses standard TTL which has (relatively) slow edges. I habitually only connect the 'scope and '8 grounds together with a couple of feet of jumper and don't use a ground on my 'scope leads at all. As mentioned earlier I'd check the omnibus terminations (this is the bus loads card referenced above, M832/M8320, which should be in the last slot of the omnibus).
 
I would check all the diodes on the bus load board, a lot of the omnibus signals are pulled up by 12 volts then clamped to about 5v, this gives a much faster rise time than just pulling up to 5, I once found some lines on the bus topping out at 10V !! due to diode failure , got away with it surprisingly.
Dave H
 
Well.... I tried using a shorter ground lead on my oscilloscope and it does make a difference. I shorted it to about 3 inches and this is the TS1 waveform. Much better.

TS1_ShortLeads.jpg

Here is the TP4 with the 3" ground lead.

TP4_ShortLeads1.jpg

Then just for fun (didn't turn out to be fun, because it was a pain to get this connected), I tried leads that were about 3/4" long.

TP4_Ultra.jpg

An improvement, but probably not worth the effort of tying in these very short leads each time I want to look at a wave form. By the time I got it connected, I'd forget what I was looking at and why.

I have the bus load card in the last slot, just behind the memory cards. I'll pull it out and check all the diodes. There could be a problem here also. Who knows until you look.

Thanks Mike
 
I checked all the diodes on the bus load card. I found about 20 that had about 137 ohms across them in both directions. These diodes are paralleled by many smaller resistors. So, I'm not sure if they are OK or not. I'd have to pull a lead and probably would ruin it in the process. If I did find a bad diode, what kind would I use to replace it? I'd like to have a few replacements, prior to pulling anything. Thanks Mike
 
Jack, I looked at this website and there is a lot of interesting items. I believe I also saw the 32k memory on some other website. I could not find any place where you could purchase a kit. Maybe that has expired, seems that most of the dates on the site are old. Suppose that I could put one together on my own, there are prints.

As far as the bus loads go, how could I test it to see if it is actually working? Could I remove all the boards, except the front panel and the bus load card, then look for 3 3/4 volts on each bus line? M Thompson mentioned rise accelerators on the bus load board. It that what the small transformers are? I was wondering what their purpose was, how to they work?

Thanks Mike
 
When testing diodes in circuit I use the diode test function on my multimeter as it uses a voltage high enough to forward bias a silicon diode , a resistance test may not work as the voltage is often not high enough, I just look for a difference in the reading between testing the diode in both directions, if there is its probably ok, if not its suspect. For replacement of a small signal diode I normality use a plain old 1N4148, iv even used thees to replace a failed steering diode on a core plane, these are in a dip package but if they fail open as has happened to me, you can simply solder a replacement diode to the relevant pins, a bit messy but it works.
BTW my "general purpose " small sig npn replacement is the 2N2222, again Iv used these to replace failed core driver pulse drives on G227's and G233's
they are in a DIP, 4 to a pack, does not get left "naked" as on the pic but gets a bit of heat shrink as a protection against shorts.
Dave H

2N2222 on G233.jpg
 
I pulled all the boards except the front panel and the bus load board. The voltage on the Omnibus was anywhere between 3.58 and 3.80 volts on all points of the bus load board, except, power ok (4.19v), Initialize (-0.55v), load addr (-0.55v). So I think the for the most part the bis load board is OK.

Mike
 
You must log in or register to reply here.
Back
Top