Remapping the HP150 keyboard

For Rik and anyone else who's still folling the saga of the HP120 keyboard, I've now hacked an HP150 keyboard to make it HP120-compatible. Here are some details... As we all know by now, the HP150 keyboard has the right interface to plug into an HP120, it's just that the keys are sent in the wrong order. The scanning is done by a 4024 7 bit counter inside the keyboard, my idea was to replace with with a circuit that was essentially a counter with a non-standard count sequence to select the keys in the order needed by the HP120. And the obvious way of doing that is simply a 7 bit counter followed by a ROM acting as a lookup table to do the code conversion. Of course it's not quite as simple as that (things never are!). The main problem is that HP120 keyboard runs at 12V internally (it's all 4000-series CMOS logic chips). And most common EPROMs run at 5V. I ended up with a board containing 5 chips and a 7805 voltage regulator. This links by a short length of ribbon cable and a 14 pin DIL header to a socket which I fitted in place of the scan counter chip (a 4024 at location M2) on the HP150 keyboard PCB. All the new chips run off the 5V output of that 7805. Here's a brief circuit description : 2 sections of a 4050 hex buffer takes the clock and reset signals from the HP150 keyboard PCB (these are the signals that would have controlled that 4024). It doesn't seem to be generally known (but it's in all the data sheets I've looked at) that the 4050 (and its inverting brother the 4049) can take inputs up to 15V even when run off a lower supply voltage. They are intended for level-shifting from CMOS voltages to TTL. And unlike the older 4009 and 4010 (which they replace) they only need the 5V power line, not a second power line corresponding to the input voltage. The outputs of that buffer go to a 4024 7 bit counter in the obvious way. The outputs of that go to the bottom 7 address lines (A0...A6) of a 27C64 EPROM. A7 goes to a switch and pull-up resistor so I can select between 2 key mappings (HP120 and HP150 mode). The remaining address lines, CE/ and OE/ are all grounded. Why use such a large EPROM? It was the first CMOS one I found in the junk box :-). The outputs of the EPROM are, of course at TTL levels, and need to be level-shifted back to 12V to drive the CMOS logic in the keyboard. I used 7 sections of 74LS07 (so 2 packages total) with 3k3 pull-up resistors to +12V The EPROM cotnains 2 128-byte tables. The first one (locations 00-7f) gives the remapping for the HP120 mode. The second one (locations 80-ff) cotnains the values 00-7f in order, thus keeping the keys in the original sequence for the HP150 mode. Anyway, I soldered it up. I worked out the correct data for the first half of the EPROM, I programmed the EPROM. And I tried it out. Most of the keys worked correctly, but some, including f8 did nothing. While fiddling around, I found f8 _sometimes_ worked. Oh great, I've got an intermittant fault. So I grabbed the LogicDart. While looking at the scan drive signals (the outputs of the 74ls07s), I noticed some glitches on some of them, doubtless due to the fact that the 4024 is a fairly slow ripple counter. Thinking that was the problem, I decied that maybe a 74HCT4024 would help. Alas my junk box doesn't contain much in the way of HCT chips, but I did find a 74HCT4040 which is a 12 bit ripple counter of a totally different pinout. One header plug, One DIL socket, and lots of wire later and I'd kludged up an adapter. Plugged it in. No change. Glitches still there, f8 works intermittantly. I did discover (by comparison with the HP262x keybaord I'd already got working) that the effects of some keys are not noticeable in 'Local' mode unless you turn 'Display Functions' on (this displays control codes on the screen rather than executing them). I did this and found that essentially only 2 keys on the HP150 keybaord were dodgy. Hmmm Finally I put the LogicDart on the signal I should have looked at in the first place -- namely the data line back to the HP120. And I found that when the key didn't work, there was nothing on that line. No glitches, just nothing. The output of the scan decoder (4028) in the HP150 keyboard was hehaving correctly, as far as I could tell, all the time. The input to the row multiplexer (4051) wasn't. Between them, of coruse, are the key switches and their diodes.] And yes, you guessed it. The dodgy keys had dodgy keyswitches. All my hackery was working find. Dismantling the keyboard assembly (lots of little screws), a squirt of contact cleaner, and it seems OK now. All I have to do is box it up. The new board will have to go in a plastic box screwed to the top of the keyboard. But I want to get things just right before I start drilling holes. -tony