search

Yohannfra / Touche

An open source and open hardware wireless fencing system
GNU General Public License v3.0
60 stars 12 forks source link

Collaboration #21

Open wnew opened 4 months ago

wnew commented 4 months ago

Hi @Yohannfra,

I have been looking at this project and love what you have done. I am the author of the https://github.com/wnew/fencing_scoring_box.

I have been upgrading and creating a couple of PCB designs for my design and would like to collaborate. Would you be interested in having a chat sometime?

Regards

Wes

nova0205 commented 3 months ago

this colaboration gonna be awsome

wnew commented 3 months ago

@nova0205 unfortunately it doesnt seem that @Yohannfra is supporting this project anymore.

I have some big updates coming for the project I linked above and have started working on a wireless system as well.

Regards

Yohannfra commented 3 months ago

Hi @wnew,

Thank you for your interest and this collaboration proposition.

You are right, I didn't support this project for years and all I did since the initial release was some small fixes here and there across the project.

The main reason I stopped working on this project is that I was stuck. I couldn't find a solution for the 'common ground' feature. This is the most important feature for such a system but I overlooked it and it doesn't work and won't work with this current implementation with the arduino 'CapacitiveSensor' library, it's too imprecise and not reliable.

I'm mainly a software engineer and don't have enough knowledge in electrical engineering to find how to solve it.

I'd be really interested in a collaboration and willing to get back to this project if we can find how to solve this first.

Last year I also started a V2 based around a NRF52 dongle to improve the fabrication cost and complexity. I now have this V2 that is much much simpler and smaller but I can't release anything if the common ground is not solved first.

V1 client box compared to v2: IMG_6350

Let me know if you want to collaborate so we can discuss how it can be done.

Regards

Yohann

wnew commented 3 months ago

Thank you very much for your reply @Yohannfra

You are right, there are 2 major hurdles to the wireless scoring apparatus. The main one being the lack of a common ground, which could be solved in a couple of ways, but most of these are really complex to implement with the required precision. The second one is the synchronisation of the 2 "client" devices to within a millisecond. This is vital as wireless communications can be flaky and often require a resend of the packet, so we cant rely on time of arrival at the server side.

I see 3 possible solutions to the common ground problem. 1. Capacitive touch sensing, this takes time to detect and also requires calibration. 2. Time domain reflectometry, highly complex and costly to implement at the nano second scale. 3. Some type of RF signalling and detection as used in radio transceivers. (My initial research leads me to believe this would be a good solution)

I have also played with the idea of designing a modified tip, which would allow for the detection of a touch on a conductive surface. This on its own would only be a partial solution.

I am an electrical engineer so am happy to tackle these challenges.

Anyway, I would like to arrange a zoom call at some stage, with a couple of other similar project developers. How could I get hold of you to arrange this? Otherwise, you can find my email address in the readme here: https://github.com/wnew/fencing_scoring_box

Regards

Wes

wnew commented 3 months ago

I love your upgraded solution with the NRF, it is a great little device. The external antenna version is also amazing. I have gotten about a km line-of-sight in ideal conditions.

It might also be worth while looking at using the ESP32.

nova0205 commented 3 months ago

@wnew i just sended a email to you

Yohannfra commented 3 months ago

It might also be worth while looking at using the ESP32.

Well I don't think esp32 are a good idea, the power consumption of these is too high for a battery powered device like this. I did a first version with it and esp-now (you can find it in the Archive/ directory).

I will send you an email so we can arrange this zoom call.

nova0205 commented 3 months ago

I'm interested in being part of that Zoom call if you'll allow me.

METhOphetamine commented 1 month ago

Hi, I would like to collaborate on this as part of a school project. ~I think a possible solution for the common ground would be sending PWM signal and detecting it with a tune decoder as is described here: https://forum.arduino.cc/t/wireless-fencing-scoring-system-detecting-a-pwm-signal-without-common-ground/686963~

edit: That solution is not a solution. Any signal above 1kHz will be everywhere with basically zero attenuation.

~The OP of the thread got stuck because the PWM was also coupling to the inner epee wires. Because of that his solution canceled hits where the blade touched the opponent's guard, not just hits directly on the guard. I would like to try to find a frequency range where the coupling is not that bad. The LM567 tone decoder has voltage values which will be detected and values which will not. So I want to try to find out at which frequency is the coupling attenuation good enough for the LM567's spec.~

@Yohannfra, it looks like in the v2 version you are using the Nordic NRF52 dongle. Are you still using platformio then? I know it's not officially supported but there are some workarounds on the internet. Would you be willing to share the unfinished V2?

Could you also please explain the capacitive sensor to me? I know the original library is charging and discharging a capacitor and measuring the time it takes to charge it to find out whether an object is connected to it. But how does that work for sensing hits on the guard? I'm guessing the metal guard takes different time to charge than the fencing jacket?