Irem92 no pixelclk

[Franck 78]

Hi,

I have this game to fix since some time now. I give it a shot when inspired.

There is no display at all. Not helping with messages.
But with some video found here I think I can say that the v33 is booting
and testing because there is activity on various addresses/data LS244 LS245
then it stops. Nothing. As in displaying a message and going to HALT
There is no V33 datasheet so not easy to confirm.

But the main thing is no pixel clock to animate IC9 IC10.
Sadly we don't know which GA is providing it.
GA22 pin155
GA23 PIN 62
IC42 is also clocked with this pixelclk

When I apply 5v I see the line going up so it is not a gate shorted.

what is necessary to get the GA (i think GA22) working ?

 

[ShootTheCore]

When he produced his MiSTer FPGA core, @wickerwaka produced some KiCad schematics for Irem M92 that may be helpful for you:
https://github.com/wickerwaka/Arcade-IremM92_MiSTer/tree/main/docs

Here's a screenshot of the clock generation circuit - it looks like the clock runs through a 74LS04 (IC54) and a 74LS74 (IC53).
The pixel clock signal for pin 155 on GA22 comes from Pin 11 on IC53. The GA22 expects two more clock signals on Pins 31 and 151 as well.

 

[SiQux]

Have you checked the crystal oscillators?

My board behaved similarly. No display, activity on the address/data lines for a few secs before stopping.

In my case, the 26.666Mhz crystal at X2 was dead. Replacing it brought everything back to life.

 

[Franck 78]

Hello,

I do have the 26Mhz and 13Mhz.

The clk_pix is missing.
IC53B produces CLR_CLR not CLK_PIX.

I think strongly the signal comes from GA22. When I lift the pin, the line goes to some floating level.
There is no other 74LSxxx driving the line (IC9,IC10,IC53,IC42(second board), are inputs.
Only GA22 GA23 can provide CLK_PIX

If main board A produces garbage on screen without B, then the clock is from GA22. But this is odd because
HSYNC, closely related to CLk_PIX is generated by GA23

CLK_PIX is a fraction of 13.3MHz. If I'm lucky I can rebuilt it but I need the exact frequency (from the Rigol scope....)

I bet on 6.65MHz pixelclock!

pixlclk = refresh_rate(50 or 60) x Hres x Yres

 

[Franck 78]

I'm a genius

I'm complaining about missing pixel_clock. Well, let's refabricate it, after all. It must not be more complicated than another divide by 2 or by 4 from 13MHz.

And here is the result. I was right about guessing the cpu V33 is testing then halts. First display from this board.

Now I have to figure if the GA22 is completely dead or believe the RAM error. It says RAM NG 02 03 04 05 06 07 08

If some knows the author of KiCad schematics, tell him he can put 'output' on GA22.155

 

[ShootTheCore]

If some knows the author of KiCad schematics, tell him he can put 'output' on GA22.155

I started up KiCad to update the schematics myself and noticed in the Symbol Editor that Pin 155 on GA22 was already labeled as an Output - I just did a bad job with the screen shots in my earlier post.

Great job figuring it out on your own though @Franck 78 - sounds like you're well on your way to fixing that board. Hopefully you can track down another GA22 from a donor.

 

[Franck 78]

There is more to the story. Last time I 'checked' the board was in 2021.
pixelclock was present.
I don't understand why the RAM error. If I remove IC43 or IC44 or both, the board starts and display the demo. Except no sprites.

If I hot-plug in the RAM, then sprites appear, with the fault I remember now.
Also one out of ten times, the board starts as it should (eg displaying ROM OK RAM OK).

I think the GA22 is dying.

 

[Franck 78]

dying (more) or not, it needs to be replaced.
If somebody have it. NANAO GA22 !