search

jotego / jtcores

FPGA cores compatible with multiple arcade game machines and KiCAD schematics of arcade games. Working on MiSTer FPGA/Analogue Pocket
https://patreon.com/jotego
GNU General Public License v3.0
202 stars 37 forks source link

s18: try different length for HS/VS signals #692

Closed jotego closed 5 days ago

jotego commented 3 weeks ago

Some monitors are troubled with JTS18 analogue video output. Try this:

rogersouza-work commented 2 weeks ago

JT, does not work for me either. BTW, would you recommend someone in Spain (I live in Valencia) for recap service? Cheers

PapiFed commented 2 weeks ago

Let's try a shorter HS than the original. This compilation uses 3us for HS width, which is a more typical value. Let me know how it goes.

jts18_6ae69c20.rbf.zip

No luck on the PVM-2950.

jotego commented 2 weeks ago

JT, does not work for me either. BTW, would you recommend someone in Spain (I live in Valencia) for recap service? Cheers

Ask here. Another option is RetroArcadia (Tel 615237811).

rogersouza-work commented 2 weeks ago

JT, RetroArcadia con Ruben. Genial! Thanks a lot, and sorry for my misunderstood about recapping and everything.

erkkav commented 6 days ago

I also have Sega System 18 core sync problems with my Sega Aero City and it's Nanao MS8 -monitor. Every other Mister core works flawlessly on it. If that's valuable info I have Sega System 18 Multi (from Arcade Projects forums) too. It has Clutch Hitter as donor board. Original Clutch Hitter board and Multi afterwards has worked with the same monitor setup without problems.

Fpg-Lee commented 6 days ago

I also have Sega System 18 core sync problems with my Sega Aero City and it's Nanao MS8 -monitor. Every other Mister core works flawlessly on it. If that's valuable info I have Sega System 18 Multi (from Arcade Projects forums) too. It has Clutch Hitter as donor board. Original Clutch Hitter board and Multi afterwards has worked with the same monitor setup without problems.

I'm seeing quite a few owners of original S18 PCB's say they have no problem with S18 sync. . What could be causing such incompatibility from the fpga core, that is not present with the original PCB? Could the problem be coming from the Mister analogue board?

jotego commented 5 days ago

As explained above, the voltage levels are different. That is not set by the core but by the external DAC.

I do not think there is nothing to do here from the core side. I am closing this issue.

Fpg-Lee commented 5 days ago

As explained above, the voltage levels are different. That is not set by the core but by the external DAC.

I do not think there is nothing to do here from the core side. I am closing this issue.

Hey Jose, I scrolled above, but did not see an explanation about DAC voltage levels being different from the original PCB, apologize if I missed/misread anything. I saw your replies regarding you suggesting people get their CRT serviced, and you saying you may add further H+V sync settings to the core, what is the reality moving forward for the Mister DAC issue on S18?

Gracias.

jotego commented 5 days ago

Here it is: https://github.com/jotego/jtcores/issues/692#issuecomment-2167927724

As I said, core-wise I cannot do anything effective. I could add more options but I think it is not going to help. I can modify the core speed, and that may lock the signal better in those screens with problems but at the cost of altering the gameplay.

MiSTer hardware vendors could add a voltage adjustment to the DAC output for this. That is rather common in supergun devices to connect arcade PCBs. This would require a new IO board or an adapter. There are probably things one can do to the signals between the MiSTer and the TV set to fix this. Most notably, changing the voltage levels, changing the line impedance and limiting the bandwidth.

We may try to limit the bandwidth at the core output a bit, as part of the work we are doing on a different topic. That may help but I cannot know for sure.

If I had one of these screens, I would service it and replace all capacitors.

By the way, I can see a similar effect on my CRT with a completely different core. Ghosts'n Goblins video output does not get synchronized until the game gets to the screen full of capital H characters. After that, the TV is able to lock it in. So even the video content can have an effect on the PLL locking the synchronization signal. That clearly points to noise coupling inside the TV set. My point is that video locking is a sensitive analog signal processing and tiny things can disturb it.

jotego commented 5 days ago

Another solution for troubled users is to use MiSTer's composite output instead of component. There is a chance that may work as the signal goes through a lot of processing that will alter the conditions for the PLL synchronization circuit.

Fpg-Lee commented 5 days ago

Here it is: #692 (comment)

As I said, core-wise I cannot do anything effective. I could add more options but I think it is not going to help. I can modify the core speed, and that may lock the signal better in those screens with problems but at the cost of altering the gameplay.

MiSTer hardware vendors could add a voltage adjustment to the DAC output for this. That is rather common in supergun devices to connect arcade PCBs. This would require a new IO board or an adapter. There are probably things one can do to the signals between the MiSTer and the TV set to fix this. Most notably, changing the voltage levels, changing the line impedance and limiting the bandwidth.

We may try to limit the bandwidth at the core output a bit, as part of the work we are doing on a different topic. That may help but I cannot know for sure.

If I had one of these screens, I would service it and replace all capacitors.

By the way, I can see a similar effect on my CRT with a completely different core. Ghosts'n Goblins video output does not get synchronized until the game gets to the screen full of capital H characters. After that, the TV is able to lock it in. So even the video content can have an effect on the PLL locking the synchronization signal. That clearly points to noise coupling inside the TV set. My point is that video locking is a sensitive analog signal processing and tiny things can disturb it.

Sorry I did see that post before, but did not make the connection as to the difference being with the DAC.

Is there a chance that if Directvideo mode is used via HDMI to VGA that output may play nicer with CRT's?

If not then outside of using Composite, which isn't really an option for me, then it is an unfortunate reality.

It's a shame there isn't a solution whereby the Core could increase the DAC so incrementally that the game would only be sped up by say 1HZ or so till the Screen sync's.. I'm guessing this is either not possible? or require a lot of work?

Thanks again.

jotego commented 4 days ago

The core just does not have control over the amplitude of the HS/VS signals on the IO board.

Fpg-Lee commented 4 days ago

The core just do not have control over the amplitude of the HS/VS signals on the IO board.

Gotcha, now I have learned the limitation of the Core-DAC relationship regarding sync issues such as this, this has been valuable for me to learn and hopefully others.