Microbee restoration

[aladinsane]

Hi,

New to the forum and new to Microbees - so in general just a newbee (sorry, couldn't resist...)

I recently got hold of two Microbees without any accessories besides a big stack of games. The two machines was in a dirty state and after having found out PSU needs and the wiring of the DIN-connector, they also it turned out quite dead  

The machines are almost identical. Both have a revision H mainboard. One as a 32k core board and the other a 16k core board. An one has a RF-modulator and some alteration to the cassette-circuit.

Luckily the batteries were intact and hadn't leaked any acid. On both machines one of the two LM7805 had caused a short, so I replaced them with a pin compatible 5v-dc-dc module. I gave both machines a good scrub (IPA, vinegar and soap) and after that I have +5v on the 5v-rail, +10.5v om the unregulated 10v-rail and 11.5v on the 12v-rail. 

Z80 cpu is confirmed working (I tested it in a Sinclair ZX81) and I have checked for life on RESET and CLK but IORQ (pin 20) is dead. CRCT (6545) gets 12Mhz clock signal but does not provide and SYNC signals on pin 39 & 40. So there is life, but no screen output

I have also dumped the EPROMS and they seem to be valid. 

Are there any common errors with aging Microbees? Or does anyone have any ideas as to what I should try/look for? I suspect the 6116 RAM chips, but it requires desoldering and I'm a bit afraid of the risk of damaging the board by removing so many 24p chips.

   

   

[MbeeTech]

Welcome.

Quick question, just to make sure, when you were checking signals (RESET, CLK, IORQ etc.) did you have both boards (coreboard & baseboard) plugged together?
The Coreboard has the system reset circuit on it. The RESET Key is connected to pin 25 (front most pin) of the right hand 25 way socket strip which then goes to the reset circuit on the coreboard. This generates a Keyboard reset (hold the RESET key down for 1 second to activate) and a 'Power-on Clock '. These get ANDéd together (active low signals) to provide a CPU reset signal.
The left hand side 25 way socket strip (on the coreboard / baseboard) have the Z80 Address Buss, ROM chip select and Z80 data buss on them.
The right hand side 25 way has the Z80 control signals IORQ, MREQ, RD, WR, Clock, JumpLatch + more.
At power on, the CPU should cycle MREQ RD cycles from address 0 to Address 0x7FFF and execute NOP instructions.
When it reaches 0x8000 (first address of the BASIC roms) the ROMcs going active will set JLATCH active, and open up the Z80 data buss to the ROM data and the Z80 should start executing the code in the BASIC rom.
IF you don't see the Z80 doing these cycles at power up then there is something wrong with
* the CPU not getting reset
* BUSREQ active where it shouldn't be
* NMI active where it shouldn't be
* WAIT active where it shouldn't be.
If you do see MREQ RD cycles, but no IORQ cycles, it means that it's not accessing code from the ROM most likely due to
* incorrect addresses (an address line or 2 not pulling all the way up or down due to a faulty output driver or shorted line)
* data buss transceiver (74ls245) not being enabled by the JLATCH signal / ROMcs detection.
* a fault on the data buss, with one or more bits incorrect.

I notice the boards are from Swedish configured Microbees.

The dot clock for the 6545 / screen section should be 13.5 Mhz, not 12Mhz.

Do you have the schematics?

P.S. the first thing I do when looking at these old machines is to get a fine file and ' clean ' the inside and outside edges of the coreboard's 25 way pin strips.
Make sure they are nice and bright, removing any tarnished surface. This seems a little brutal, but the newly cleaned surfaces of the pins on the pinstrip are then
a bit like mini files that after a few insertions and removals into the baseboard sockets, end up cleaning the socket contacts usually enough to get the bee going.

[aladinsane]

Thank you for your reply and pointers.

I just want to add that I have measured the signal on the solderside with the coreboard attached. Also I was mistaken with the crystal, it's a 13.5Mhz and the readings matches 13.5Mhz (it's a Danish configuration with Danish keyboard and character rom).

I have a schematic, but only for the original kit, so it's not accurate for the main board and for the core board it's very different. Unfortunately I haven't been able to find one for Rev H or similar.

I did clean all the socket but I did not think to clean the coreboard pin strips. I've done that now and measured all >0.6ohm on all pin connections between coreboard and mainboard now.

Status now is that I do see MREQ RD cycles (on both CPU and 25 pin connector) but no IORQ activity. JLATCH (pin 6) is high and there's activity on the 74ls245 pins. But as I'm self taught in these matters, I'm not quite sure what to look for next. Would it make sense to replace IC19 (74LS245) and IC33 (74C74)

[ChickenMan]

We would be very interested in the all the ROM dumps being a Danish model. We have the Character ROM dump but nothing else.

[aladinsane]

We would be very interested in the all the ROM dumps being a Danish model.  We have the Character ROM dump but nothing else.

Sure thing Should I just post them here or send them directly to someone?

---

Okay, so there is some progress. Following a recommendation, I tried removing the T2 and WordBee roms as well as the PIO chip. And now one machine boots to a Basic prompt

The machine still boots to Basic with all roms inserted, but not when the PIO chip is in. Furthermore, I have now been able to validate that both core boards works.
However, the other machine still does not boot - regardless of whether the PIO chip or the extra roms are in or not. On the other hand, it seems that the 74C74 (IC33) does not forward the MREQ signal from the CPU, so I suspect that that chip is dead.

I have tried swapping the two PIO chips, but it makes no difference. Of course, I cannot say whether they are both defective or whether something in the circuit around the PIO is faulty.

The keyboard does not work, but I guess it's due to the missing PIO chip.

The image is not completely stable. It looks fine enough but every now and then there are some purple artifacts that fill the entire background. They come in flashes and furthermore the background is not "pure black" but there is a little "noise" in the black. Maybe a faulty clock divider?

[MbeeTech]

The keyboard is scanned by the 6545 CRT controller, so the PIO being out of it's socket will not affect the keyboard.

It's rare, but not impossible, that all the keys are faulty. You can check the 'On' resistance of each key with a multimeter.
For a key to register the ON resistance should be lower than 100ohms. There are 2 types of key switches used in these models - the early
production units used a switch with a black shaft (that the cap sits on) and later a white shaft. Although different inside, they are
both carbonised rubber on metal switch plates designs rather than metal on metal. Sadly, over time, the rubber deforms and becomes
brittle, and the metal plates tarnish.
A simple way to check if the keyboard circuit is working is just to short the contacts of one of the switches to see if the key press appears
on screen.

In terms of display stability, it sounds like you are using a reasonably modern TV with it's AV input at a guess.
Most (post 2000 era) TVs have trouble syncing with the Microbee's mono video composite output and what you are seeing is most likely
just that incompatibility.

[ChickenMan]

We would be very interested in the all the ROM dumps being a Danish model.  We have the Character ROM dump but nothing else.

Sure thing Should I just post them here or send them directly to someone?

---

Thats great, thanks, I have sent you a Private Message.

[CheshireNoir]

It's a rite of passage to have to rebuild your Microbee keyboard at least once. :-D

There's an excellent guide on the basics of popping apart the keys, cleaning the contacts and reassembling them in the Repository.

Cheers!

John

[aladinsane]

I just want to share an update on my restoration project. Thanks to all your great feedback, help and ideas the status now, is that both machines is more or less alive.

The 16K model had a defective Z88PIO which is replaced now. All of the keyboard keys - except "x" - are non-functioning They measure between 50k to 500k ohm when pressed down and 6M Ohm unpressed. But shorting the pins on the solder side result in the corrosponding character being printet. Also the 6545-chip is partly defect - it displays an image but it tends to get distored and has these purple artifacts. However if I replace the chip with a 68B45 I had lying around, the image is stable. But the 6845 apparently does not have the registers needed for the keyboard scans, so it only helps in diagnosing the defective 6545.

The 32K model also have a defective Z80PIO and a fully defective 6545. The MREQ line wasn't present on the core board even though it was present on the CPU. I replaced the IC33 and IC19 just in case. With the 68B45 chip it now turns on but sometimes the basic promt disappears or random characters appears. Could that be due to another bad logic chips in the video circuitry - maybe one of the 74LS157 or the 74LS175? I have checked the videoram.
The keyboard seems to be as dead as on the other machine.

Besides from sourcing 2 new 6545 it would seem I have to start cleaning the keyboard switches. Is it correct that I need to desolder each switch, dismantle it and clean the parts before refitting it? It's the black variant and I tried to remove a random key and clean it. Now it shows ~1k ohm when pressed.

[MbeeTech]

Besides from sourcing 2 new 6545 it would seem I have to start cleaning the keyboard switches. Is it correct that I need to desolder each switch, dismantle it and clean the parts before refitting it? It's the black variant and I tried to remove a random key and clean it. Now it shows ~1k ohm when pressed.

The best thing you can do in terms of having the keyboards fully working and reliable, is to get a keyboard replacement kit.
https://microbeetechnology.com.au/store/...emium.html

While you can sometimes get a good rate of success in repairing the old switches, they don't tend to last. The keyboard replacement kits have a different type of switch that is still compatible with the key caps
but is fully metal on metal switch contacts, and should outlast the machines. It is quite a job doing the replacement, but actually less work than trying to repair the old individual switches.
You can download the instructions on the product page before purchase to see how it all goes together.