A tale of three Tubes

Post date: 26-Jun-2010 21:20:34

We had a grand day out at the VCF at Bletchley Park, and discovered that our Tube ULA replacement design didn't work: it didn't work in both a 6502 and a Z80 external second processor, in both cases connected to a Master.

So in a sense we were 2-1 down: our design works in the Master internal coprocessor.

On the plus side, we met some nice helpful people and were able to discuss some possible reasons.

It turns out that there are (at least) three Tube interfaces - all of which will of course work with the original Acorn Tube ULA because they were designed to do that. Specifically, the databus is in the case of the BBC Model B a direct connection to the heavily loaded NMOS CPU's databus, and is helpfully furnished with pulldown resistors. The internal Tube of the Master is again a direction connection, but to a CMOS databus, with rather less loading and without resistors. Finally, the external Tube of the Master is connected to the PBC, a custom chip which bridges the three databusses on that machine, with both pullup and pulldown resistors for termination.

All three busses have variable loading depending on the configuration of the machine: number of ROMs, 1MHz expansion, disk controller and so on. There are quite a few cases to consider. That's why Ed got his ancient scope out of the garage in the hope that study will be rewarded.

Meantime Rich has been taking a tally of various combinations - including a Model B with one of our 816 cards in it - and we have seen some life out of the Z80 second processor, kindly donated to us by Tom at the VCF.

Bletchley Park is an excellent museum by the way.