Mystery Object - ancient electronics?

I would post pictures (especially the insides if you can) to the newsgroup alt.folklore.computers

Those folks are more familiar with 'big iron' and, while they might mostly be from the US and not familiar with particular UK models, might have more experience to draw upon.

Theo

Elecro-reflexology :)

My thoughts were along those lines. The first computer room I worked in had ancillary equipment for dealing with punched cards and there was a device called an IBM Collator. It had two feeds for cards, and 5 output areas of which the centre one could take cards fed from either or both hoppers. The whole thing was controlled by patch boards. By connecting one hole to another by a jump lead and using a collections of such connections it was possible to programme the machine to subdivide and/or merge depending on the outcome to be achieved. The key similarity was that the control boards had a handle like the one in the picture and that allowed the control board to be lowered into a slot and then a lever moved the slot and board into contact with fixed connectors in the machine.

The set-up I worked with had a small library of boards for commonly used jobs, plus a small stock of blank boards and a collection of jump leads of varying lengths so that bespoke jobs could be created and broken down to its components afterwards. My job at the time included making up any bespoke boards needed.

The unit pictured is a lot smaller than the ones I wired up, but the similarity of function with connectors wired one to another suggest a similar function to a standard board for a repeating function. It is not big enough to be for use with punched cards, but it could perhaps provide standard analysis functions for paper tape?

Alternatively, one of the other ancillary equipments was called an interpreter. That similarly had control by boards with jumper connections though I never dealt with that one's boards. Effectively there were 80 columns in a card and 60 print positions and the idea was that the board chose which columns on the card were connected to which print positions, and by choosing the right board for the appropriate card function it was possible to print the values of selected fields into human readable characters printed on the top of the card. This allowed humans with no knowledge of the character coding of the holes to select from a tray of interpreted cards a card with a chosen content.

Having started on the idea of punched cards being the possible target of the pictured board, I started looking at the early uses of punched cards before they became standardised. I found a bespoke type of cards used for analysing US census data:

Ouija board?

Tim

Possibly a program board for a punched card reproducer.

My old Data General Nova 1200 bootstrap loader was 13 words or 26 bytes long. Didn't take too long to toggle in.

Our Elliott mainframe only needed four 24 bit words.

And 18 bytes on a PDP-11 with an RK05 disk.

MANY thanks to you Indy Jess John (and to Geo): this looks very promising. I have. couple more photos, but can't upload them at the moment: I will be back here when I've done so!

AJ

OK I've now managed to upoald the two new photos:

All of this seems to match Indy Jess John's description above. Now I just need to isolate the the machine for which his device was used. As the organisation (the Computing Laboratory at Newcaste University) was founded in 1957, they used paper tape as the I/O medium (cards only came in in 1968 with the IBM S/360-67) - this also fits IJJ's description above. The Lab first had a Ferranti Pegasus, and then an English Electric KDF9. (But this item might not be related to the Lab's computers -- it may simply have been collected by our founder from elsewhere. It is, after all, rather attractive, in a steam-punk kind of way.) Thanks again to everybody for putting us on the trail. I will probably also post this story, as suggested. to alt.folklore.computers to see if anyone can further identify it.

Best wishes all, Another John

The pins appear to be hollow at the bottom to allow links to be inserted. I thinks the links are of a square "U" shape made from stiff wire, which is just pushed in and soldered in place.

One or two pins have solder on the ends but no links, suggesting that the links had been removed.

If you're talking about 80-column punched cards, I was using those on an IBM 1401 in the early 1960's - they pre-dated the S/360.

I think he meant that cards came in with the IBM S/360-67) at the Computing Laboratory at Newcaste University

All of IBM's unit record machines[*] (many with Queen Anne legs) were designed around the 80-column punched card, which dates back to the 1920s.

OK so it looks like it was not meant to be changed often, more like a boot ROM than a patch card for card equipment or an analogue computer. It also looks older than the IBM 360/67 which is when cards first appeared at Newcastle....

.. so the first computer was the Ferranti Pegasus. Whilst I never operated a Pegasus, I did demonstrate the Hartree Differential Analyser at MOSI in Manchester which was next to the Pegasus. I did have a peek inside the Pegasus and don't remember any thing like that.

.. so that leaves the KDF9 which came between the Pegasus and the

.. if you are still reading this, as others have said one of the computing groups, or perhaps contacting Mr Dik Leatherdale at the UK Computer Conservation Society. He could perhaps put a note in their magazine...

Dave Wade

How about EDSAC?

This is in the Newcastle University Museum. EDSAC was Cambridge, Manchester had the Baby/SSEM which was upgraded to the Manchester Mk1 and later the Ferranti MK1.

Dave

yes

Yes, the 029 card punch has a programming drum that you put a punched card on

I dont have any experience on how those were setup

Yep

Someone must know how it did stuff like that

But it looks very digital to me, rather than analog

I know. As an undergraduate, I had to write a program for it as one of our 'Practicals'.

Completely wrong. I also programmed an IBM1401 and all my programs were put on 80-column punched cards which were fed in as the pack on the front of the data cards that the programme would deal with. That was a long time before the IBM360 became available.

The IBM1401 must have had a reasonable user base because the IBM360 came with an additional software option of emulating any existing 1401 programmes at machine code level so that users could migrate to the 360 and continue running their old system on it giving plenty of time to recode it to run as native 360 applications.

This must have been a tricky emulator to write, because the 360 had a fixed word length whereas every 1401 word was of a variable length of characters either defined by the assembler suite to give instructions of