Xeno Crisis Mega Drive AES/MVS DC etc...

Neo Geo Latest - Physical

In the last update we spoke of some issues with the PCB we had that were causing the odd rare reset on some AES/MVS machines when it read some P ROM data incorrectly. Unfortunately the changes we made back then didn’t cure those issues totally; it actually became a bit of a needle in a haystack trying to find the issue to be able to resolve it. So we took the decision to contract a professional electrical engineer to take a look at the PCB.


We’re pleased to say we have now found the issue, we’ve been able to reproduce it consistently, we understand why it is happening and we now have a fix. We’re waiting on a sample of our revised PCB design to arrive next week. Assuming all is well and as we expect, we will then be placing the main PCB manufacturing order!


For those of you that would like a little more insight into what we had to track down and how we went about it, here is a brief overview of the main discoveries:


First phase; we had an issue with our Neo Geo Devkit for flashing the PCBs which was not allowing us to fully verify the P ROM when flashing but would for all other ROMs. So we couldn’t be 100% sure that all data had been written correctly, even if the PCB did generally run fine with just the odd reset and Unibios CRC failure.


This proved relatively easy to track down, there was a voltage level issue that was causing the Flash memory not to be read correctly by the devkit, fortunately write commands were being done correctly as we expected. Now that we could properly verify the P ROM is correct we could move on to attempt to find the issue with the PCB.


Second phase; an initial comparison of our signals vs an original PCB showed we needed a small tweak to the FPGA logic for the ROMOE behaviours (there was a slight signal glitch as you can see on D7). This made the PCB behaviour better but it still was reading the P ROM incorrectly at random points.


After weeks of numerous tests prodding and probing all aspects of the PCB, we realised we would have to find a way to reproduce the issue and pin-point the fault. We could prove it was a bad read by running the Unibios CRC check but the values would change and be inconsistent, and it would pass the CRC most times without any errors. So we decided to create a custom “CRC” BIOS for the NeoGeo, so we could try to control reading the P ROM in sections, detect read errors as soon as they occurred and hopefully all in a reproducible and reliable way.


In the end we read the P ROM in 128 byte chunks and compared each chunk against a known CRC state for that chunk and then repeated this through the whole ROM. We also were able to halt the moment a bad chunk was read - we could re-read this chunk to compare and see if we could spot a pattern.


This worked extremely well and we are able to capture bad reads in a reproducible and consistent manner, as you can see above a capture from a bad vs good read we’re getting some sort of interference to the signal at a 16Mhz rate. We also noticed it was always an 0xFFFF read that would randomly fail.


From here we were able to analyse the PCB and find the issue that had been eluding us. Under the conditions of an 0xFFFF read, if the floating levels were in the wrong place the level translator was glitching and jumping between high and low states. After testing some modifications to our previous prototype, we now finally have what we believe are the necessary modifications to fix our PCB design!!!!


Obviously the PCB is just one part of the physical items we need, so we’ve been taking the time whilst it’s being fixed to prepare and ready all of the other physical items. Above are just a few of the various bits we needed that have been steadily making their way to the warehouse ready for assembly.


We’ve also decided to manufacture a US and JP style release for Neo Geo CD, as you can see in the preview above (we will contact you to confirm your preference). Also note that both NGCD versions like original games will come with spine cards. With more space in our new warehouse we have also taken the opportunity to order some more plastic shells and snap cases - we do have some excess stock at present so you can still pre-order your copy now.


Neo Geo Latest - Software

The last few months have seen a lot of refinement and fixes to the actual game code of Xeno Crisis, with some of the most notable points being:


The Z80 audio logic assembly code is now complete - we had to do nearly a full rewrite, but a huge thank you goes out to Blastar (https://blastar.citavia.de/) for all of his kind support.


The MVS, AES and NGCD integrations are also complete - one of the key things for us has been ensuring that our games behave and act just like original games on those platforms did. So we’re pleased to say that we’ve now implemented what we believe to be the correct behaviours when interacting with the bios in the various combinations of machine set-ups, supporting MVS/AES standard and also Unibios/NeoSD modified behaviours. This means that all of the standard features you expect like software dips, memory card support, backup memory, and book-keeping all behave as expected.


The remaining bits of software development are shrinking by the day, most focus at the moment is on Neo Geo specific polish, things like scaling sprites, extra frames, more palettes, better colour depth than the Mega Drive (like in the HQ above), more sprite details. We expect to go gold very soon, and we will let you know when we do!


We will also finally be confirming details with all collectors’ edition pre-orders, as we have to lock down hostage names to embed in the game (and serial numbers for production). Please expect an email in the next couple of weeks with details.

emixam a écrit:Alors Hans il avait pris des chevilles 60x4mm a expansion progressive pour beton cellulaire creux, mais la vis en titane mou de chez bricoflex etait pas compatible du coup, on s est pris l etagere de jeux sur le coin de la gueule....

emixam a écrit:Bon au moins quand ils ont du retard, ils le justifie assez bien
Il y a une vraie envie de transparence, ils ont eu un pb avec la pcb et visiblement ils en ont bien chie pour trouver le pb
J imagine nos 2 baltringues de ngdev qui prennent les photos de leur etagere qui s est pete la gueule....
Alors Hans il avait pris des chevilles 60x4mm a expansion progressive pour beton cellulaire creux, mais la vis en titane mou de chez bricoflex etait pas compatible du coup, on s est pris l etagere de jeux sur le coin de la gueule....
A oui et sinon, on fabrique nous meme les coques AES en faisant fondre le plastique a la bougie, alors c est un peu long...

Pxl a écrit:J'en avais jamais entendu parler avant.

zouzzz a écrit:

Pxl a écrit:J'en avais jamais entendu parler avant.

"Qu'est ce que tu ferai si on t'annonçait qu'il ne te restait que 6 mois à vivre ?"
"J'irai vivre en Haute-Vienne, ils ont 3 ans de retard."

Pxl a écrit:

zouzzz a écrit:

Pxl a écrit:J'en avais jamais entendu parler avant.

"Qu'est ce que tu ferai si on t'annonçait qu'il ne te restait que 6 mois à vivre ?"
"J'irai vivre en Haute-Vienne, ils ont 3 ans de retard."

Attends c'est pire, je viens de débarquer en haute vienne. Avant j'étais 4 ans en Creuse et encore avant 11 ans dans le.cantal lol