Still faulty 80K

[gregctaylor]

After much work and testing, it is not main RAM, I have swapped for new VRAM, swapped M-ROM and CG-ROM. Same error all the time. Have had a good look at the screen and the output is:

code 10100011b (I think) then a space then either code 1100011b or 11100011b then a space then code 10100011b and so on all across the screen and repeated on every line so the screen looks like vertical columns of half blocks (bottom half lit) then space then bottom row lit then space and so on.

I am really running out of ideas on this one. The screen output is regular and the same every time which rules out loads of options. No chips are running hot or cold.

Quick thanks to hlide for all his advice so far and mz-80a (Ben) for burning me a replacement M-ROM.

Hurry up or I shall be bald from pulling my hair out!!

[hlide]

Looks as if your VRAM contains a repeated sequence of $3A $00 $3C $00 or $3A $00 $3C $FF or $3A $FF $3C $00 or $3A $FF $3C $FF. Those characters are valid in the display table.

Was the burned MROM checked to run on a working MZ-80 K before you got it?

Is there any indication the monitor is working? is keyboard working? type 'SG' in prompt to turn on the feature to hear a short beep-tone when pressing a key. If you can hear that beep after pressing keys, then your monitor is working.

If not, either the new MROM is not burnt/set properly or there is an issue with a chip around it (you'll need schematics and an oscilloscope).

[gregctaylor]

Thanks for the ideas. No M-ROM was not tested but my original shows exactly the same pattern so don't think it is that.

Tried SG but no sound, nor out of my fully working 80A either. Perhaps I am doing it wrong? Is it SG then [Return] or just SG?

I shall try the schematic and also check out the actual M-ROM code. The 80K has a static memory map (unlike the 700) so the ROM is always at 0000h. RST 0 then runs the ROM code and obviously something is working as I get the screen output rather that a crash of some kind. Never mind, only 2K of code!!! It might be time I bought an EPROM progammer and write a revised EPROM with some sort of error trapping to see what is going on.

Greg

[hlide]

Tried SG but no sound, nor out of my fully working 80A either. Perhaps I am doing it wrong? Is it SG then [Return] or just SG?

You must type [S] then [G] then [CR] (what you call 'return') just after powering the MZ-80 K. If you cannot hear any beep after pressing keys, so your monitor is not running. Did you check the Z80 CPU is working?

[hlide]

As for MZ-80 A, the commands are different and probably the same as MZ-700: press [ B ] then [CR] to toggle key beeping.

[gregctaylor]

Good idea to check the Z80. I have spares as well which are for a different project but easily purloined.

[TextAdventureKing]

I've been looking at this thread for a bit. I'm hoping you get it working Greg.

[gregctaylor]

Latest gripping update!! I began by checking pin 6 of the Z80, NOT_CLK, at which the scope should have shown a 2MHz clock. I did get something at around 50Hz (bit of mains hum I think) with very low voltage and a very irregular shape. Pulled the Z80 to see if that was at fault. Same poor clock and, out of interest, a series of thin vertical lines on the screen. Obviously, the Z80 is having some effect.
The route for the 2Mhz clock (if anyone is interested) is from pin 1 to pin 2 then pin 11 to pin 10 then pin 9 to pin out of IC7, a hex NOT gate (7404). It then inputs pin 1 of IC9 and outputs pin 8. This is a 7493 4 bit binary counter used to convert the 8Mhz crystal clock to 5Mhz (5/8 speed) and 2 MHz (1/4 speed) the latter for the Z80. From there it goes to pin 11 of IC35, a 74244 transceiver/line driver leaving on pin 9 for pin 6 of the Z80.
Sadly, all three are directly soldered so needs the board out again. My guess now is the 7404 for two reasons, one is I have replaced the same chip in two other faulty computers recently, two, I think designers used the 7404 as buffers to protect the other ICs. You commonly see a line go into one NOT gate then through a second so no logic change at all.
Will replace the IC with a socket first chance I get and try a new 7404. Has to wait a few days though as the kitchen needs the tiles finished.

Greg

[gregctaylor]

Moving forward slowly. Replacing the 7404 and the tuning capacitor means I now have a good clock on the Z80. Still no change to the display though. I will swap out the Z80 a second time (in case the lack of clock was hiding the fact the Z80 had also failed) but, if not, will then attach the video circuitry in case whatever the Z80 is sending is getting lost and a constant pattern sent.

Bit of a delay after my MZ 700 also smoked its power supply. The same capacitor as the others though only one in the 700, a much simpler board. One interesting point was that power comes in at 240V (UK) smoothed by a 500V capacitor. It then goes through a small transformer down to a lower voltage (looks like 115V) then a second 250V smoothing capacitor which is the one which smoked (rather spectacularly). Anyone in the US know what is inside a US spec 700? I wonder if the power is soldered to the transformer output pair with just the larger of the smoothing capacitors (C102). A very clever idea for its day. The video circuitry is also in its own metal container so that is also easy to modify for the 60Hz NTSC market.

Other good news is that my 80B is now fully tested. It looks like I might now hang on to all of them. The engineering is just so fantastic compared with other computers.

Greg

[gregctaylor]

Another gripping instalment! For those not in the know (me included until a little while ago), most home computers of the time used the system clock for everything, dropping clock frequencies by additional chips or by using the processor itself to time things. The Sharps are quite different. They use 555 timers to generate clocks for the screen and cassette. IC13 (555 timer) controls the monitor output and was generating a horrible waveform. Now swapped and I am getting a good clock into the monitor circuit. The next major component is a D8253V programmable interface timer which generates the various pulses to run the monitor. This is outputting the weirdest signals on its output pin so is the next thing to be consigned to the bin. Sadly, I have an 8255 but no 8253 so back to waiting for parts. Good news, the chip is socketed so an easy change.

All the other waveforms now look good so I have fingers crossed this is the last faulty component. Half the machine will be brand new soon!!

My MZ700 is now on ebay (item 263870463688 if interested).

Greg