Jump to content

New Pin2DMD Issue


Recommended Posts

  • Content Provider

Hi guys,

I've been having an issue with a few games lately with the Pin2Dmd behaving strangely. I first started to notice it on DE Last Action Hero when testing out Dante's colour files. Sometimes frames would lag, fixed issues would come back or represent differently and some scenes that definitely have had keyframes were no longer triggering. I thought maybe it was just the single Pin2DMD hardware, the SD card or the firmware. It wasn't until tonight that I was getting the same issues on a different machine that I decided to do some tests.

I have tested multiple Pin2DMD 128x32 units (Evo 1.3) with firmware ranging from 3.17 all the way up to 4.0. The issue presents exactly the same in all firmware versions. Tonight I was using a 128x16 Evo 1.0 and the problems started to happen on this too.
I initially thought it was the SD card so I swapped that to a different brand and different size. I have tried Kingston 2GB, 4GB, 8GB as well as Sandisk 8GB and 16GB all genuine cards verified. No difference. I then updated the firmware incrementally from 3.17 to 4.0 and the problem did not go away.  I also know there was an SD card delay fix in the firmware but that doesn't seem to make any difference.

The biggest sign is when it is power cycled. You can tell after a power cycle almost straight away if it is having an issue because for LAH the "Last Action Hero" Attract animation lags, parts in game lag and some sequences don't trigger. This is a big issue because bugs I have been reporting as a colour file problem have actually been I believe, the hardware/firmware. I've been working on testing an x16 game and once in every 3 or 4 power cycles it will show the same issues, I added some key frames previously as some were not triggering and when the unit is not working correctly these will not trigger. Replaced / redrawn animation can clip in and out showing the original art that it replaced behind and other weird things.

The only thing in common that I can tell is both scenarios are using a variant of the Data East Device setting (Data East / Sega for LAH and Data East x16 for the other game). I am not sure this is the common factor because I haven't extensively tested other platforms recently but I did do a very long test session for both FT and IJ and with FT the issues did not present after modifications were made to the project files by the author. IJ is not presenting in the same way as this either and is consistent.

The Pin2DMD hardware in question were from different batches and also different form factors. 

To reproduce this you can Install the Pin2dmd in your machine of a Data East game (LAH verified) and just power off the machine and back on again until you start getting lagged animation and skipped frames etc.  It's almost like the hardware can't keep up?

What I might do is try and downgrade to a lower firmware and see if I can find the exact version it was introduced in but that could take some time, and then it may have always been there.

@lucky1 If you have any suggestions of what I can try to help diagnose this better please let me know.

Thanks

 

Link to comment
Share on other sites

  • Content Provider

I would say that this is neither a hardware issue nor a firmware problem. I don´t expect any changes going from 3.17 to 3.20 or even 4.00 as these versions use exactly the same code for your 16 colorLAH  project. If you want you can also try 2.68 which is also in the stable release list (https://github.com/lucky01/PIN2DMD/releases), but I don´t expect any difference in the behaviour there either.  

Sounds more like a timing issues which come up when dumps generated by a vpin are used in a real machine or vice versa. There have been several tricks posted here by others authors which adress such problems. For e.g replacement scenes you should edit the delay of the last frame that it stays on the display until the next trigger comes to avoid that the original art is shown. For Colormask Scenes Single frame triggering or using LCM Scenes instead could help. Unfortunatelly it is part of the job to adjust some things to make it work fine on both platforms.

Other authors like @slippifishi or @NetzZwerg from which the other projects, which work fine are, have more experience in correcting these glitches in the project.

Maybe they can help.




 

Link to comment
Share on other sites

  • Content Provider

One more thing. ColorMask Sequence (or Replace Sequence) are theoretically ideal as it should keep the sync to the frames in the background but in practice it has shown that single frame ColorMask triggers  work better.

Link to comment
Share on other sites

  • Content Provider

Thanks for the input. I will speak to the authors of the colour files that I am testing and see if they are able to use these methods to adjust the project.

I’m glad it’s not a hardware/firmware issue then.

Thank you

Link to comment
Share on other sites

  • Content Provider

@TerranigmaThe consistance of the triggering depends on the way you trigger the scenes. In general my experience is for real pins, that you will definately need to use single frame triggers or work with the  layered color mask if this makes sense. The real pin is not very consistant in delay times for the single frames, so I would recommend to use a very high delay time for each frame (500 or above). That will keep the mask active until the next frame gets triggered and will prevent you from "fallbacks" to pal0 what will let the frame lose its color. If you try to use the simple color mask with multiple frames, the mask will run out of sync. Replacement scenes with multiple frames will run in a different timing (delay time per frame) and the result is that the animation does not run all the way to the end, because it gets disrupted by a new trigger.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
  • Create New...