G27/G29 Hall Sensor Mod

[Tjure]

it seems like the side exit is easier for me to install.. but when i see the file located here. ( https://github.com/GeekyDeaks/g29-hall-sensor-mod/tree/master/mod/stl ), i notice the side-exit-mag.stl & side-exit-sensor.stl


I usted side_
You can use this one (https://github.com/GeekyDeaks/g29-hall-sensor-mod/blob/master/mod/side-exit-6mm.scad). You can edit some parametres to adapt the design to yours necesities.

 

[Ozanelm]

Hi! Yes, you are correct, I did say 3.3v, but this is because I am using the G29 which supplies 3.3v. I believe all the previous models are 5v.


It might be better to measure the voltage coming out of the A1324 first to double check things are working ok. The distance and strength of the magnets are quite critical to the design. Can you post a photo of your setup here?

well, when i check voltage on hall sensor pins, it reads exactly as it reads on the arduino side(around 5.03, 5.04), i'm using 6x3 neodum magnets and i have around 100 magnets, tried with bunch of them.

when i tried sensor without housing with bigger magnet there was no problem with readings. Can 3d printed pla part block sensor reading? i printed mine with 100 infill since there was no signicficant time diffrence between 20%.

i didnot use breadboard to test it, i wired it up directly to pedal pins with correct terminals and so, and placed my hall sensor in 3d printed part(it was really tight fit).

Is there anyway i can check sensor readings besides windows calibration tool?

 

[NobyI]

Hopefully not reviving to quiet of a thread, but thanks for all the work y'all put in, and curious any thoughts of this. Mocked it up, it seems like it should be getting around 6-8 steps per mm of travel on the spring perch? If anything seems terribly off please me know! Still working on printing some final mounting pieces before getting some racing in on it. I also joined the RD discord if anyone is there.

 

[blekenbleu]

curious any thoughts of this

That looks usefully simpler than coaxially twisting magnet.

getting around 6-8 steps per mm of travel

.. so graphed Output is digital count (16-bit ADC?) with Distance in meters?

What Hall sensor chip is used? My eyes cannot make out digits from image,
and I do not recall a sensor chip with a dimple.

 

[NobyI]

That looks usefully simpler than coaxially twisting magnet.

.. so graphed Output is digital count (16-bit ADC?) with Distance in meters?

What Hall sensor chip is used? My eyes cannot make out digits from image,
and I do not recall a sensor chip with a dimple.

A1324LUA-T IC? Sorry, I definitely could have provided more useful info haha. X-axis is in inches, measuring the distance between my spring perches. I have around 1-1.5" of travel. The output was just from DIView, the actual output is 10bit, giving around 250-300 steps in the range of travel, or ~6 steps per mm I think.

 

[blekenbleu]

Max (DIView) output reads around 25,000, which would be 100 for each of 250 steps,
but minimum graphed is around 11,000, so 14,000 DIView excursion.
Mapping that to 140 steps, then sensitivity is about 7 steps per mm for 19mm travel.
140 nearly linear steps should suffice for throttle travel and be overkill for clutch.
Given a simple layout avoiding wire flex and anchoring mini magnets to pedal arms
well out of harm's way, this seems a winner.

An experiment for increasing sensitivity could place the Hall sensor on iron instead of polymer.

 

[NobyI]

Max (DIView) output reads around 25,000, which would be 100 for each of 250 steps,
but minimum graphed is around 11,000, so 14,000 DIView excursion.
Mapping that to 140 steps, then sensitivity is about 7 steps per mm for 19mm travel.
140 nearly linear steps should suffice for throttle travel and be overkill for clutch.
Given a simple layout avoiding wire flex and anchoring mini magnets to pedal arms
well out of harm's way, this seems a winner.

An experiment for increasing sensitivity could place the Hall sensor on iron instead of polymer.

Some more data to maybe clarify some and add some more input. This is really using the earlier design that is linear within a small, close range to the sensor. I feel the logitech pedals have pretty little travel though, so maybe it's sufficient. Varied some different combinations of magnets, having to change location based on intensity as well which likely compiles to worse linearity.

Currently printing a new mount for the sensor that should bring it closer to the magnet widening the linear range a bit more.

 

[blekenbleu]

Since real throttle responses are rarely linear, and clutch linearity is a don't care,
I would choose range over linearity

 

[Neilski]

Some more data to maybe clarify some and add some more input.

That's looking far more linear than I would have expected an arrangement like that to deliver - nice!

For me though, the approach of just hijacking the Logitech pot to use it as a shaft for the magnet-holder is so simple that it kinda wins as a retrofit. (Near-perfect linearity is basically for free, though admittedly not vital).

On a custom pedal setup without the Logitech gearing+pot, I'd be keen to try and use something like you've gone with if I could achieve adequate linearity (which looks like a big "yes", given your results).

 

[NobyI]

That's looking far more linear than I would have expected an arrangement like that to deliver - nice!

For me though, the approach of just hijacking the Logitech pot to use it as a shaft for the magnet-holder is so simple that it kinda wins as a retrofit. (Near-perfect linearity is basically for free, though admittedly not vital).

On a custom pedal setup without the Logitech gearing+pot, I'd be keen to try and use something like you've gone with if I could achieve adequate linearity (which looks like a big "yes", given your results).

Yeah, that seems like a pretty fair conclusion. My main reason for getting rid of the pot was just for appearance, although I suppose I could have kept the gears and ran a shaft on the inside.

 

[Neilski]

Yeah, that seems like a pretty fair conclusion. My main reason for getting rid of the pot was just for appearance, although I suppose I could have kept the gears and ran a shaft on the inside.

Ah, yeah good reason then. Mine still live in the original housing with 5 million screws keeping it together.

 

[GeekyDeaks]

Since real throttle responses are rarely linear, and clutch linearity is a don't care,
I would choose range over linearity

You left it a bit late to offer up this bit of wisdom!

 

[blekenbleu]

a bit late

Old too soon, smart too late; had not seen data showing less than half ADC range useful,
but still fairly decent linearity in the worst case. Lacking easy access to a 3D printer,
Hall sensor upgrades to G29 clutch and brake were being put off until pots began glitching...