[Beta] - BackForceFeeder : feeder vJoy FFB pour racecabs

[Tathan]

Oops, je te fais ça. Pardon.

En bidouillant un peu, je constate que le RAW Input 16 me donne toujours un message d'erreur, il fonctionne, mais je ne peux pas le "nommer".

Si on sélectionne la Raw Input 16 je reçois toujours le message d'erreur "Error : Please select a raw input first".

Peut-être lié au fait que ça compte de 0 à 15 et pas de 1 à 16?

Je sais pas si ça peut avoir un lien avec l'erreur, mais je n'avais pas ce souci dans l'ancienne version.

(c'est le monnayeur sur ma borne)

[njz3]

J'ai fait un recovery à partir de ton fichier de backup json, mais je crains qu'il n'ait déjà été "réduit" au plus simple car je n'y vois que 3 control sets dedans.
Du coup ce soir tu m'a motivé à ajouter une fonction de Backup/Restore de tous les control sets sous forme d'archive zip.
Tu peux télécharger la v0.7.0.28 sur github.


Tu as du recevoir par email le fichier d'archive que j'ai reconstruit à partir de ton backup json, et je t'invite à d'abord faire un Backup avec le bouton, puis à essayer la fonction Restore avec l'archive que je t'ai envoyé.

[njz3]

Petites avancées concernant le support natif des servoboards midi et RS485 directement dans le BFF.

En réalité, le protocole de commande avec les retours a déjà été documenté dans  ces lignes de code de Flycast et le code du FFB Plugin de Boomslangnz.
Une trame est sur 4 octets, et se termine par CRC calculé pour chaque trame (voir le code de Flycast pour sur la lecture d'une trame).
Les commandes sont faciles à obtenir en démarrant un jeu naomi comme initial D ou ClubKart sur Flycast et en se connectant au port 8000 qui est la sortie de flycast en mode MAME Network Output (sortie texte).
On récupère ainsi toutes les trames au démarrage qui servent à faire probablement l'initialisation de la servoboard, puis ensuite on a les effets en jeux ou en mode test.
J'ai déjà dumpé quelques jeux et c'est pratiquement toujours la même chose, ce qui fait que je devrais pouvoir l'intégrer assez facilement dans le BFF (j'ai déjà pratiquement tout le code de fait)
Niveau connection entre le PC, il y a plusieurs options. Je suis parti au plus simple pour l'instant avec un convertisseur USB vers midi avec optocoupleurs. A voir sur mon banc si cela fonctionne...
Affaire à suivre !

[BigPanik]

 

[Tibal]

GROSSE avancée ! Bravo !

[frankrizzo2]

I am currently having an issue with my pwm2Happ and Sega ffb board setup. To clarify I am running a standard happ motor from rush 2049 with a 24v meanwell power supply and the current ffb boards from Aganyte. I have successfully calibrated the steering wheel and pedals and am able to play a few teknoparrot racing games with force feedback for a few minutes at a time. The issue that I am having is that midway through a game the force feedback in the wheel will aggressively go to the right and firmly hold that position until I kill the game. If I go back to the Backforce feeder software and go into diagnostic test slider, it forces the wheel to the right no matter which way the slider is positioned. The only way I am able to play another game is to completely shutdown the entire cabinet and reboot. I am then able to use the diagnostic mode correctly in Backforce feeder and the ffb in the games work correctly until there is am aggressive movement in the game causing it to hold the right side position until reboot. I am so close, I just need to figure out this issue and I can move on to the next.

[njz3]

Do you have any message in the BFF log that could to help see if there is a communication issue between the PC and the arduino board ?

[frankrizzo2]

The log doesn't show any errors that I can see. I will upload some pictures and hopefully a video to explain the issue.

[frankrizzo2]

[frankrizzo2]

[njz3]

Strange bug, and I cannot see why it would be related to the BFF software.
Any way, if you have an Arduino IDE with a serial monitor connection to the pwm2happ board (configure it at 500kbps), you might type the following commands that performs low-level pwm output to the pwm2happ board. Each commands is confirmed by a press on "Enter" for newline.

Get version info: V
no pwm (no torque): P800
80% pwm in one direction: P200
80% pwm in the other direction: PC00

And let's see whether you see the same behavior. If that's the case, they the issue is related to the pwm2happ board.

[frankrizzo2]

The P200 side which turns clockwise draws about an amp more than the PC00 direction (counter clockwise)
P200 3 amps
PC00 2 amps

Serial monitor will only communicate with the board at 1000k. At 500k it displays weird characters and won't communicate

[frankrizzo2]

[frankrizzo2]

[njz3]

Sorry, my mistake.
Serial monitor is indeed 1Mbps and it should be PE00 instead of PD00 for 80% pwm in other direction. The command is actually PXXX where XXX is an hexadecimal number between 000 and FFF, with 800 being the center of the pwm value.

After performing the test for a long time, have you seen the same issue happening where the direction does not change when typing P200 then PE00 several times?

[frankrizzo2]

So today I did a lot of testing with different settings and scenarios. First thing I did was remove the LED strip and Audio amplifier from the shared power supply to rule out interference from those. It seemed like it was the cause of things as I was able to play for a bit longer than normal. I thought I was onto something but it eventually went into fault and pulled the wheel to the right once again. I did a full system reboot and opened the serial monitor and set 100% ffb to clockwise and set the amps to 1.5amps draw and was able to play for an hr without issue but with weak effects. I have now turned it up to 2 amps and it seems to work okay and without any issues. The issue is that most games still feel fairly weak on the force feedback so I still feel like I need to be in that 3 amp area for this motor to feel correct and arcade accurate. I still don't trust it without more testing. I think it is a thermal issue as most games are fine but the Batman game is very ffb intense and is a constant force with the machine gun feedback and the crash effects. I think I either need better cooking or a more robust h bridge chip.