ZXmore or the ZX80CORE follow-up

[dessony]

I would like to wait until the device is CP/M implemented. Thanks again.

Bill

[martinjharvey]

On mine, I have noticed that the oscillator based on X1 (the ceramic resonator coloured blue just to the right of the Altera EPM3064 chip) no longer starts at the correct frequency, at least, not as far as my Sharp Aquos TV is concerned

Did you manage to find a fix for this Mark?

I have a similar issue although I've no idea if its to do with the oscillator frequency or not

My ZXmore works great (Thanks Karl) and I've not done a bad job of putting it all together (blows own trumpet) but I have intermittent issues with the display output and TV signal.

At the menu for the Control Instance (instance 0), I get a flickering and loss of signal every now and again which seems like my Samsung TV cant lock on to it. This is also sometimes accompanied by a rainbow type blurring or yellow ghosting effect along the right side of any graphics or on-screen characters.

When I enter a ZX81 or ZX80 instance the same effects or issues remain until I use the DS-C hotkey to deactivate the control instance in the background, and run the ZXmore in compatibility mode, and then the TV/Signal behaves itself. Thinking about it, my ZX81 fitted with one of Karls ZX8-CCB does behave in a similar manner but that can be tweaked and fine tuned with the on-board pixel and sync trimmers to eliminate any graphical or display problems.

Other than that, everything else works great and I'm really pleased with the ZXmore. Looking forward to the updates and the proposed Speccy add-on, but does anyone have an idea or solution to what might be the cause of this problem?

I'll try and upload some photos of what Im talking about if that helps, but it happens quite fast so might be difficult for me to catch on camera

Mart

[PokeMon]

Hi Mart,

there is a small difference between the plain video signal from ZX81 rom (pressed DS-C or using the ROM switch) and the video driver used in rev. 0.9 of the control software. This may behave better with the next revision but not 100% sure as there is a change in video drivers. I am currently working on the update (after an absence of 2 month due to much work in my company) and pretty sure to publish it before xmas. Now working full time (private) on it again.

The main reason for these problems is the very loose video format used from the 80ies computers which are sometimes handled more strictly regarding to the specifications. Old TV's don't care much about timing horizontal/vertical and frame coding (odd/even). It may be, that the Samsung TV has a chipset inside which requires a harder timing than usual, this is depending on the manufacturer and model. So there isn't a rule other than following:

older TVs accept these old video signals better than newer
more expencive TVs accept these old video signals better than the cheaper ones

I have XOMAX small LCD which sometimes doesn't accept any signal from ZX81s (depending on crystals age) while my 10 years old SONY plasma HD Ready TV doesn't have any problem with looser timings at the video input.

Karl

[martinjharvey]

Thanks for the reply Karl

Its no big deal turning the control instance off, as I say that does stabilise the screen on the Samsung TV. I did try the ZXmore on a cheaper 13" Bush LCD late last night and a Commodore 1084s CRT monitor and the picture on them is fine, so as you say its the expensive Samsung TV that's struggling with the signal. The same happens with my ZX81, but I can manage to just about tweak it out of the ZX81 using the trimmers on the ZX8-CCB.

I think I read somewhere in the ZXmore construction guide or user manual that the ZX8-CCB is built into the ZXmore. Is it possible to add in or retrofit the pixel and sync trimmers somewhere to fine tune the TV signal in cases like this?

Looking forward to the update, hopefully that will make the Samsung TV and ZXmore play nice together

Cheers

Mart

[PokeMon]

Announcement - there is a required hardware change for the next release of ZXmaster (control software for ZXmore).
A small 47pF capacitor is needed between pin 27 (/M1) and pin 29 (GND) for timing purposes.
See the picture below.

M1.jpg (191.95 KiB) Viewed 3169 times

The reason for this is, that the NMI routine is moved from ROM to RAM which has a few implications. Moving the NMI routine to RAM was necessary because a switch from many small NMIs to one long NMI (followed by a short sync NMI) and different NMI routines have to be used depending on the software using (ZX80, ZX81, ZX Spectrum, CP/M). This wasn't an easy goal and took some days and nights to get around with all requirements regarding logic and timing. But finally I found a solution.

The NMI detection circuit is quite complex due to internal architecture and missing a special signal like the INT ACK for maskable interrupts and realized in the CPLD. It uses only two signals (/M1 and CLK) instead of a 16 bit address comparator. To avoid additional time for switch back to the rom a special mechanism was used to detect the RET instruction (with /RAMCS and /RD). So this way it would be possible to execute instructions in RAM during the NMI routine but only those instructions with no additional read cycle. So an instruction like LD A,n could not be used because reading n would end the NMI routine and switch back to last instance and probably crash. Even LD A,(nnnn) wouldn't be an option if nnnn is an address in RAM because this address is read first from RAM. But finally I found a way to assure correct timing and a switchable NMI routine in memory.

Unfortunately in practice the /M1 signal is released prior to /RD which would be interpreted accidently as a RAM read and end the NMI routine too fast. When checking the timings /RD should be released first and /M1 after (10ns later) but it is vice versa. So the M1 signal is delayed with this 47pF capacitor to adjust the timing. This could be done with a changed CPLD as well but this smal capacitor does the job very good as well.

Next release can be expected till end of next week (whatever is included, actual revison is 1.5 - the next release to publish would be 1.7).

Whoever need this capacitor may drop me a private message and his address. Maybe you find it in your personal spare parts box as well.

[gemby]

I have issue on first power up my zxmore :-/
Das not boot. No composite video signal. Nothing, checked with few monitors, and double-checked on original zx81 on them, they are all ok.

Voltages are ok, i triple checked all components, i quadriplechecked all soldering points for bed joints and shorts. It looks like it is oscillating ( i hear on radio ) i have no osciloscope to check ( but i will borrow one ). ROM selector is behaving weird, on 0 position it lits white, and after 10 or so power on / power off's lits blue, and become black in minute or so. Is there any schematics, or troubleshooting tutorial?

PS: found schematics, but some troubleshooting guide is welcome.

[PokeMon]

Hi gemby,

sorry to hear issues with your ZXmore.

First:
The latest schematic is available in this post:
http://sinclairzxworld.com/viewtopic.ph ... tic#p17474

Most known issues have been wrong or forgotten setting jumpers, be sure to use the BaseAddr set to A7 and be sure that you not put a jumper on the reset pin header (one guy really did as it was spare ). Check also the capacitors soldered at the back side and the wanted short between two pins. You may also remove IC22 and IC24 to avoid wrong latch settings. It should start up but LED should not be set anymore to white (led off). When removing the latches (IC22/24) it is not possible to switch to other instances by software (but with the switch).

You may also startup in switch position 1 or 2 or 3 to see if the ZX81 instances are working. This is a single mode and the corresponding color will be set through the switch instead of the latches. You should avoid switching around during power on. I think the flash rom is more protected against accidently beeing overwritten than the EEPROM but you never know. So switch during power off only to be more safe.

On default position 1,2,3,4,5 show ZX81 image and 6,7 show ZX80 image while 0 shows the control software.
You should also check that the wire on the background is still in position, connecting pin 1 of IC15 with pin 11 of the CPLD - see picture below showing pin 11 position on a 44 pin CPLD.
http://circuits.datasheetdir.com/389/L6515-pinout.jpg

These are the first steps to check. Unfortunately I am in a short holiday from tomorrow but are other people here to help as well I think. Points to check could be clock at pin 6 of CPU, NMI generator (NMI pulses on pin 17 CPU). Removing RAM and ROM and IC5, the CPU executes NOPs - so this way you can measure around many signals. The address bus counts upwards, giving half frequency from every address line from A0 to A15 and the control signals can be measured as well to check any possible shorts or mysterious signals.

And check if all pins are correct in the sockets and no pin accidently bended during inserting - just to be sure.
If you never find out the problem I offer you to ship the board to me and let me find out (next year).
So you can be sure it will run finally - so or so. Money never lost. Just have to pay the postage both way in that case. Up to you.

Good luck,

Karl

[gemby]

Hi gemby,

sorry to hear issues with your ZXmore.

First:
The latest schematic is available in this post:
http://sinclairzxworld.com/viewtopic.ph ... tic#p17474

Most known issues have been wrong or forgotten setting jumpers, be sure to use the BaseAddr set to A7 and be sure that you not put a jumper on the reset pin header (one guy really did as it was spare ). Check also the capacitors soldered at the back side and the wanted short between two pins. You may also remove IC22 and IC24 to avoid wrong latch settings. It should start up but LED should not be set anymore to white (led off). When removing the latches (IC22/24) it is not possible to switch to other instances by software (but with the switch).

You may also startup in switch position 1 or 2 or 3 to see if the ZX81 instances are working. This is a single mode and the corresponding color will be set through the switch instead of the latches. You should avoid switching around during power on. I think the flash rom is more protected against accidently beeing overwritten than the EEPROM but you never know. So switch during power off only to be more safe.

On default position 1,2,3,4,5 show ZX81 image and 6,7 show ZX80 image while 0 shows the control software.
You should also check that the wire on the background is still in position, connecting pin 1 of IC15 with pin 11 of the CPLD - see picture below showing pin 11 position on a 44 pin CPLD.
http://circuits.datasheetdir.com/389/L6515-pinout.jpg

These are the first steps to check. Unfortunately I am in a short holiday from tomorrow but are other people here to help as well I think. Points to check could be clock at pin 6 of CPU, NMI generator (NMI pulses on pin 17 CPU). Removing RAM and ROM and IC5, the CPU executes NOPs - so this way you can measure around many signals. The address bus counts upwards, giving half frequency from every address line from A0 to A15 and the control signals can be measured as well to check any possible shorts or mysterious signals.

And check if all pins are correct in the sockets and no pin accidently bended during inserting - just to be sure.
If you never find out the problem I offer you to ship the board to me and let me find out (next year).
So you can be sure it will run finally - so or so. Money never lost. Just have to pay the postage both way in that case. Up to you.

Good luck,

Karl

Thank you Karl, yes, this is good starting point, will try to fix it my self first and let you know how it went.

[1024MAK]

Hello gemby, welcome to the forum

Do ask for help here if you still have problems.

If you don't get access to a 'scope, you may want to get a logic probe. A logic probe is much simpler (and so costs a lot less) but can tell you about static and slow speed digital signals, and can give you a "feel" for some "medium" speed signals (watch the colour/intensity of the LEDs). It is much better than a normal multimeter (which are far too slow).

Mark

[gemby]

Hello gemby, welcome to the forum

Do ask for help here if you still have problems.

If you don't get access to a 'scope, you may want to get a logic probe. A logic probe is much simpler (and so costs a lot less) but can tell you about static and slow speed digital signals, and can give you a "feel" for some "medium" speed signals (watch the colour/intensity of the LEDs). It is much better than a normal multimeter (which are far too slow).

Mark

Yep, thank you, if i got stuck, will try to make one. I am almost certain that is something so obvious, but i must find it first