• Please review our updated Terms and Rules here

DEC design questions

tech58761

Experienced Member
Joined
Nov 10, 2023
Messages
344
Location
Northern Plains, USA
After getting my hands on those Universal Instruments boards (and having moved most along to new homes by now) I got to thinking about two in particular.

I've decided to see about laying out replacement boards and sharing the Gerbers for those who may want to build one for their PDP system.

Looking at the schematics, I see three ICs were used for interfacing the board to the backplane.

Input-only lines are fed through a National DS8837 IC with its control lines wired 'always on'.

Bidirectional lines (the system address / data bus) are fed through a quartet of National DS8838 ICs.

Output-only lines are fed through a 7438 (which is surprisingly still available from TI - is that the last straight TTL part still in regular production??)


First question - are there modern equivalents of the DS8837 / DS8838? I tried looking at the publicly available docs for the UniBone / QBone, but the copies I found seemed to be partial schematics, so am not clear.


Now, one of the two boards I was interested in laying out a replacement is the bootstrap board (A/N 27514) - I could see this being handy to save the time and effort in keying in the bootloader / minimizing wear and tear on punched tape readers.

As is, the board uses a pair of TMS2516 (2716), and it seems that the 2764 / 2864 would be more available these days (in fact, I may even have a tube of unused / lightly used 2764s kicking around).

The design currently uses address lines A0-A9 (A10 tied to ground), and stepping up to the 2764 or 2864 would add lines A11 and A12.


Second question - it would be just a matter of adding a 3-position DIP switch and pull-up resistors to provide for up to 8 sets of bootstrap images, correct? The system looks at only A0-A9, so it shouldn't matter address-wise, right?


The other board I was looking at is the power sequencer board (A/N 4127304), which has been pointed out would allow the use of standard Meanwell supplies instead of the DEC power supply (like the H786 in the 11/03 chassis).

The changes I was looking at included adding the grant continuity traces and relocating the AC power (25VAC center-tap) to a Phoenix connector on the front edge of the board.

The only issue I see is the filter caps for the on-board 7912 and 7805 regulators.


Third question - can someone direct me to the relevant spec for height limits on the Qbus boards? I obviously would need to select appropriately sized electrolytic caps if possible.


Lastly, I was looking at the machine interface board (A/N 18013), I see something that is a little confusing - this is not a board I'd be looking to reproduce, rather, what degree of caution is necessary in using this board.

The issue I see is that one ground trace is run to pad BS1 instead of BT1 as would be expected.


Any thoughts appreciated.
 
First question - are there modern equivalents of the DS8837 / DS8838? I tried looking at the publicly available docs for the UniBone / QBone, but the copies I found seemed to be partial schematics, so am not clear.
Don't think so. Peter Schranz (@cbscpe) uses 74HC4049 operated at 3.3V for receivers & 74(S,F)38 as transmitters.

Now, one of the two boards I was interested in laying out a replacement is the bootstrap board (A/N 27514) - I could see this being handy to save the time and effort in keying in the bootloader / minimizing wear and tear on punched tape readers.

As is, the board uses a pair of TMS2516 (2716), and it seems that the 2764 / 2864 would be more available these days (in fact, I may even have a tube of unused / lightly used 2764s kicking around).
M8048 boards, which are cheap and plentiful, can be hacked to operate in a 22-bit system, so I'm not sure of the need for this one...
 
M8048 boards, which are cheap and plentiful, can be hacked to operate in a 22-bit system, so I'm not sure of the need for this one...

I wasn't aware of this board (my interest had only been on rescuing some PDP systems that were set to be discarded). Thank you for pointing it out.

Edit: Attached a copy of the bootstrap board schematic
 

Attachments

Last edited:
I think they were asking about the maximum height of components on the board. See these attachments for drawings which include component height limits.
 

Attachments

yes, the maximum height was what I was after, as the sequencer card mentioned has a pair of electrolytics that are HUGE in relation to the allowable space.
 

Attachments

  • card.jpg
    card.jpg
    163 KB · Views: 10
The second attachment I uploaded shows a limit for double-width cards. I wonder whether that sequencer card fits within the double-width spec, or would occupy more than two card slots?
 
In the original application, it was set apart from the rest of the slots, because of the caps and the presence of 28VAC on the board.

Looks like a pair of miniature SMPS modules would be a better bet than staying with the 7805 / 7912 (and relocating the AC input away from the backplane).
 
But if the goal is to run MeanWell power supplies, then the regulation happens off the board, doesn't it? I'm not clear on what the purpose of the on-board regulators is, as they can only supply 1 amp - not enough to power the whole processor.

There are lots of existing solutions for generating BPOK/BDCOK/LTC signals with non-DEC power supplies, most of which do not require a backplane slot. I think I've done it 3 different ways myself. But that doesn't mean that another solution would not be welcome.
 
Seems the idea was to power the logic on the board externally without relying on the backplane supplies.

But, if you say there are already existing replacement power sequencer solutions, then I guess I don't need to reinvent the wheel again.
 
Back
Top