RML380Z

 

1a50742eaaa943789c01aa17ec71b930

Like a lot of people who were at school during the 1970/80 period, the first microcomputer I used was an RML380Z. It wasn’t the first computer I ever used, that was a DEC System 20 (IIRC), but this shiny black box was the first machine that I could sit at and get its full attention.

 

I have been wanting to get a 380Z for a number of years but the silly prices on ebay have made that impossible until now. A chance meeting with a fellow computer collector led to me getting the example you see in the photo.

As you can see, it’s not in the prime of life, it’s covered in stickers and there is no keyboard. When it’s powered on I get a screen full of rubbish but it’s mine and I’m delighted.

53e525645f5aa2c829d66d1d878e69f6

 

Exterior

The case is a 19” type made from pressed steel and aluminium. It has a handle on each side. The front is dominated by the two full height floppy disk drives (8” drives were available IIRC), the Logo and the power switch/reset button combination.

It was possible to lock the machine off, on or locked. Off and on are self-explanitory and “Lock” disabled the reset button. Perhaps thought useful for public installations.

At the back there is a host of connections.

Keyboard. This uses a 15 way D-type socket. The keyboard uses a parallel interface down which the ASCII encoded characters are sent. There is a micro-controller on the keyboard PCB that looks after scanning and translating the key-presses.

Cassette. This uses the DIN socket typical of the time.

TV. The usual TV antenna socket. My unit doesn’t have a modulator and so this doesn’t work.

Monitor. A composite video output. Rather unusually the 380Z uses a SO259 socket much used on amateur radio equipment. This is a high quality connector and is still an expensive item. I don’t know of any other manufacturer of micros using this.

Serial port. A standard issue RS-232 port.

Another D25. I’m not sure about this one. When time permits I’ll go in and have another look.

 

There are plenty of other cut-outs covered by blanking plates so expansion was not a problem.

905120ac67ee5af602f0007b746eb790

Inside the case.

 

The photo above shows my machine with the lid off. The back of the machine is at the bottom of the photo, the front at the top.

As you can see, there is plenty going on.

Power Supply.

To the left you can see the power supply. Unlike later machines, this is not a self contained, switch-mode unit. Instead it is a linear supply made from a collection of components fixed to the nearest piece of case or folded aluminium. Working on it is quite tricky as a fair amount of the machine needs to be taken apart to get the the components. It’s not something that you can take out and hold in your hand.

Card cage.

Like the PSU this term is a bit of a misnomer as the cards stand between pairs of guide rails individually screwed to the bottom of the case. There is plenty of room here. As you can see, I have five cards and there is room for four more.

Floppy disk drives.

On the far right of the case is a pair of 5.25” floppy disk drives. I havent taken these out to check the model numbers but early machines had DSSD drives giving about 84kb per side. Unusually, each side was addressed independently and so A: and B: were different sides of the same drive.

The cards.

I you can see, I have five cards in my machine.

VDU80.

6529f6d02e0556c5885a5f2c392016f7Next to the power supply is the video card. This supports 80 columns and gives composite video out. I noted the high quality connector on the back of the case. There is just a pair of twisted wires running from the connector on the card to the SO-259 on the case. Not the highest quality solution.

CPU.

2813ff65f9d8f5ffb48290dbf2e581e4The CPU card contains the Z80 processor so typical of business machines of the age. It has 32kb RAM in 4116 DRAM chips (the Z80 looks after refresh) and three TMS-2716 2kb EPROMs. It’s unfortunate that they used the TMS 2716 EPROM as these are very hard to get hold of now and tricky to read as they require several power supplies. The “ordinary” 2716s only need 5v and so became far more popular.

32kb RAM card.

 

1c04cc10496f433a822af21e5ad8d5d5Another collection of 4116s gives another 32kb RAM, maxing out the Z80’s address space.

Floppy disk controller.

391d2f4079ff9d7e668fadedabdbd119This really needs a better photo 🙂

Bus terminator card.

fd828ca323c3dfd912897b754c38e4c4Just a collection of resistors and test points really.

380Z – Hmmm

The 380Z seems to be my system of the month and is causing me head scratching. It’s becoming one of those puzzles where you have to slide pieces around but you can’t because something is always blocking them.

 

Memory

I have tried moving the DRAMs on the CPU card around by one socket. I’m not sure that this made any tangiable difference. The screen rubbish doesn’t have as many ‘C’s and ‘0’s in it but left long enough it changes anyway. Did it allways do that? Not sure.

I would really like to read the EPROMS but they’re TMS2716s and they are different to the run of the mill 2716s and so the EPROM programmer I was kindly lent by a friend won’t ready them without an adapter, which we don’t have.

 

Defying logic.

I have borrowed a logic analyser in an attempt to see what’s going on on the bus. As I mentioned before, in a previous post, the bus terminator card has a row of holes and each signal is brought out to one of them. I have soldered some header pins so I can pop the wires of the logic probe on to key sognals and see what’s going on.

The short answer is nothing. nada, nilque. The long answer is that the logic analyser isn’t playing ball. If I use the ‘scope I can see a clock, memory requests, M1 bipping up and down, activity on the data bus, _RD_ is low and the first eight or so lines on the address bus are wibbling as expected. It’s frustrating the analyser won’t work as that would be a great help.

 

More 380Z work.

The 380Z has got as far as displaying rubbish on the screen but no further. It’s quite a predicatable kind of rubbish from a sub-set of characters. Curious.

I have tried running with just the processor, VDU and the passive card that I think is for bus termination. The symptoms are the same. I started removing the chips on the CPU board one by one, using IPA to clean the pins and putting them back. After I had done about a third I just went for lifting the chips, squirting and reseating them.

I did get some different results but I had inadvertently missed a pin on one of the EPROMS. When I put it back where it belonged the machine went back to it’s old behaviour.

It was suggested to me that the EPROMS could have lost their memories by now as their expected life was around ten years so I will try and read them and verift them against some ROM images I have found through the 380Z group on Yahoo.

 

The bus terminator card has a row of holes for (I think) test pins, on every signal. I’m going to solder right angled header pins into these holes to get easy access to them and start probing about.

I also plan to shift the DRAM chips around by 1 socket and see if the fault changes.

Work on the 380Z

ba50d28dd19e2bd2f29366da7ed36617

I’ve been having a look at the RML308Z recently. One of the capacitors in the power supply went pop when I first powered it on. This is not unusual for machines of this age and it’s a good idea to inspect the caps before powering on a new machine. I decided to chance it and POP.

It turns out to be a tantalum on the 12v and a visual inspection wouldn’t have shown anything.

It’s very easy to open the case and look at the boards, even to take them out but to get to the PSU is a right pain. There isn’t a PSU in the usual sense. There is a portion of the machine that takes mains electricity and passes it through a collection of seperate bits, all bolted to different parts of the case.

Anything beyond a board swap is a bit tricky.

 

Anyway. I haven’t got a suitable replacement cap so I snipped of the remains of the old one and re-assembled. A look at the circuit diagram show this is for taking out the ripple and so isn’t vital – I don’t think.

I’ve now powered it on again and it’s nearly working…

53e525645f5aa2c829d66d1d878e69f6