Page 1 of 5

ZX81 Clone Tynemouth Minstrel 3

Posted: Sun Dec 22, 2019 6:29 pm
by Mustermann
Back in 1982 I bought my first computer: An issue 1 ZX81.

A month ago I noticed Tynemouth Minstrel 3 ZX81 clone.

Details at:
http://blog.tynemouthsoftware.co.uk/201 ... r-kit.html
https://www.tindie.com/products/tynemou ... puter-kit/

It seems to be the perfect basis for tinkering around on this retro style device as all of the parts can be replaced easily if broken and it would fit into the original ZX81 case.

I ordered one of this batch of PCBs and brought it to life yesterday.

Great work! Thanks to Tynemouth,

What I changed from the build instruction:

Power supply:
I am using an external 12V switching power supply.
Also using a switching regulator to generate 5V (It will never save the amount of energy to be economic but will save some CO2)
As I noticed that there are external power supplies with different pin assignments I added an diode to protect the circuit.

Rom:
I did some rewiring at the socket to connect the original ROM out of my 1982 device as I botched up the old board.
So I own a clone with a licensed operating system.

Ram:
I replaced the 32k by a 128k chip using a 32 pin socket but wired A15 and A16 to VCC for now. (Planning to enable 56k or 64k in future)

Type of 74 chips:
To keep compatibility with LS and reduce power consumption I choose HCT chips for all of the 74xx

Other:
Based on my current stock I choose BAT48 instead of the 1N4148.
As I didn't found a vendor for 6.5MHz crystal I took a 6.55MHz one.

And then TATA!! alive on first try :D . BUT the nose has moved to the right ear. :o
Device is up and running but all of the characters are shifted to the right for some pixels.

Surprised Max

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Sun Dec 22, 2019 6:59 pm
by 1024MAK
Hello Mustermann and welcome :D

We do have a welcome area if you want to introduce yourself ;-)

The character bit data being offset indicates a timing error with the shift register that sends the pixel data to the TV / video display. It looks like a signal is being delayed too much.

Mark

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Sun Dec 22, 2019 9:28 pm
by Mustermann
Hi Mark,

thank you for the welcome and the first hints to get minstrel's nose on the right place.
First I was thinking about run-time issues within the digital circuits.

Then I noticed that there are at least 3 RC elements existing in the minstrel (beneath audio part).

One of them is located at clock generator (so may affect , the other ones are within video generation.
Tynemouth told that minstrel will work with LS and HC chips.
I decided to use HCT.

May that be the root cause?

What is the difference between LS, HCT and HC?
At least input and output levels.

HC and HCT are going up to VCC for logical H at output.
LS is typical going up to 3.4V for logical H at output.

LS and HCT are switching from low to high at 1.4V input voltage.
HC does change from low to high at 2.5V input voltage if I am right.

So 5V/2.5V and 3.4V/1.4V is similar (HC and LS)
5V/1,4V is not (HCT)

Ok that is only an rough estimation and I need to do some integrations to verify that, but this may be ab indication that an RC element that work with LS AND HC may show a different behavior using HCT.

So it seems, changing the RC parameters or the type of logical gates may be a way out.

Will keep on trying......

Max

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Sun Dec 29, 2019 3:26 pm
by Mustermann
After some experiments I am able to tell, that there are at least three ways to fix the pixel shift I noticed with 74HCT chips:

- Replacing IC9 with 74HC08
See page 3 and 4 within
https://d3s5r33r268y59.cloudfront.net/d ... v3.5.2.pdf


- Changing R8 from 330 to 196 Ohm and R9 from 470 to 274 Ohm to adapt the RC circuit to lower switching voltage of HCT
See page 4 and 8 within
https://d3s5r33r268y59.cloudfront.net/d ... v3.5.2.pdf

- Using 74HC chips only

All of them had been tested and worked well.

At the end I choose "Using 74HC chips only" for myself to stay compatible with other minstrel 3 boards.


This work for me, BUT:

As the high level in HC logic is minimum 3,5V instead of minimum 2V for LS and HCT the level of the tape output need to be VERY high to load programs.
My old tape drive was able to create sufficient level for original ZX81 and minstrel using 74HCT but not for minstrel using 74 HC.

So an amplifier is needed that is able to give at least 8Vpp. An audio amplifier based on TDA2002 did it when VCC is at least 12V.

So at the moment my minstrel 3 is working and is able to load programs from tape and smartphone and run WRX based high-res games (tested with pacman.p and invader.p)


Ready for take off,

Max

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Mon Dec 30, 2019 4:16 am
by 1024MAK
Hi Max

Nice to hear that you have sorted out all the problems. Thanks for posting what you did / found and that you got it working :D

Have fun :mrgreen:

Mark

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Thu Jan 02, 2020 11:20 am
by tynemouthsoftware
Thanks for posting the findings, interesting to see how that worked out.

I had tried to remove as many of the analogue bits of the design like those RC filters, but had to leave those in as I couldn't get the timing right using the synchronous 74166 rather than the giving a short pulse to the non-synchronous '165.

Most of the original design work was done in LS, and I only had to make a couple of tweaks for HC to work as well, and both ways have had lots of testing.

So I stick by the recommendations of HC or LS. The only issue with HC as you say is the higher level for cassette input, and you can get around that by replacing the two 257s near the bottom of the board with LS (or possibly HCT). I did look at adding a transistor buffer or a comparitor and various other things to the tape input, but I still had mixed results from real tapes and just didn't have the space. I was planning a tape amplifier with a notch filter as a separate board, but with Minstrel ZXpand out now, I don't think there will be much need.

Thanks,

Dave

Re: ZX81 Clone Tynemouth Minstrel 3 56k

Posted: Sat Jan 04, 2020 11:16 pm
by Mustermann
The way to 56k:

As already mentioned I replaced the 32k x 8 SRAM(IC2) by an 128k x 8 SRAM.
As this doesn't fit into a 28 pin socket, I took a 32 pin socket and cut of the legs 1,2,31 and 32.
Pin 1 is Not Connected at the 128k chip, pin 2 is A16. As I decided to use up to 64k I connected both of them to VCC(pin 30 of 128k = pin 28 of board).

CE2 is high active and if connected to VCC it would enable the chip if /CE is low, So connecting pin 30 of the 128k to pin 28 of board that is connected to VCC does not cause any trouble.

Temporarily I connected A15(pin 31 of 128k) to VCC using a long wire.
I also connected pin 32 to VCC using a long wire.

For further use i cut of leg 22 of the socket(pin 20 of board) and connected the board as well as the socket to separate wires.
For now these wires were connected to each other some cm away from the socket.
/CS of the socket is connected to /CS of the board, but can be disconnected when needed.
32k.jpg
This was tested, working and provide the Minstrel 3 with 32k of RAM as expected.

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Mon Jan 06, 2020 2:54 pm
by Shaun_B
I treated myself to one of these bad boys for Christmas. Wonderful stuff. Although I'm having some difficulties in loading from cassette tape.

The tape player works fine on my Sinclair ZX Spectrum and has also worked perfectly well on my Minstrel ZX80 clone.

Any clues anyone?

Many thanks,

Shaun.

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Mon Jan 06, 2020 4:58 pm
by 1024MAK
Which 74 series chips are you using? 74HCxxx, 74LSxxx or 74HCTxxx?

With the 74HCxxx CMOS versions, they need a higher voltage on their inputs (compared to 74LSxxx) to register as a logic one / logic high, so either turn up the volume or use an audio amplifier.

Mark

Re: ZX81 Clone Tynemouth Minstrel 3

Posted: Wed Jan 08, 2020 3:00 pm
by Shaun_B
Just a thought, does the Chroma Interface work with the Minstrel 3 clone? If so, is the Chroma still available to purchase?

Thanks,

Shaun.