Potentiometer Selection? - Accelerator and Clutch axis... Which to use?

[Phill Routledge]

Hi guys,

Hello fellow racers,

I am about to start my build on a custom set of load cell based pedals using this pedal base from TAS Autosports / Revolution Race Gear (images below) >

I already have a 60kg load cell as used on the CST F1 pedals, and I have a Leo Bodnar BU0836-LC and or a DSD 12bit controller w/- AMP at my disposal for use on the brake pedal axis. (So I have the brakes completely sorted).

What I really need help with is pot selection for the clutch and accelerator axis.

A quick look at Leo Bodnars website reveals very little as his descriptions of each pot is limited to poor and his providing of part #'s and make / model information somewhat withheld / protected for his business gain.

Now guys, don't get me wrong, I am in Australia and I would like to try purchase these pots locally to save a few quid... I simply can't justify the the hefty freight cost around the globe for something so small, that surely I can buy locally. That and the spend locally will help keep another Aussie in a job.

If I had the full make and model number info on the items I could get pricing locally from http://australia.rs-online.com/web/.

Can anyone give me some advice and the complete part #'s + brand make/model information on which pot is the the best option to use for this application?

 

[RichardHessels]

Just use standard 10kOhm linear potentiometer and you will be fine.
Should be available at any electronics (online)shop. They last a millions or so revolutions..

 

[Phill Routledge]

Just use standard 10K linear potentiometer and you will be fine.

Easy enough thx Richard,

How many turn should it be? It won't see a full rotation so I guess a single turn one would be fine? However it looks like the one Leo mentions is a 10 turn so I guess it doesn't really matter?

Should I use a sealed one (for keeping out the dust and the crap)?

 

[Kevin Cordice]

I to am about to build a custom set of pedals but i'm going to use 100ohm units

I think the seal option is the best way to go

 

[Grzegorz Koguciuk]

Easy enough thx Richard,

How many turn should it be? It won't see a full rotation so I guess a single turn one would be fine? However it looks like the one Leo mentions is a 10 turn so I guess it doesn't really matter?

Should I use a sealed one (for keeping out the dust and the crap)?

You should use as long working distance as you can. P260 from Leo's website has 295* of mechanical rotation.
You would like to use sealed one but that's not necessary.

Even better than potentiometer is hall sensor, it does't wear out, doesn't care too much about dust etc.

 

[RichardHessels]

You should use as long working distance as you can. P260 from Leo's website has 295* of mechanical rotation.
You would like to use sealed one but that's not necessary.

Even better than potentiometer is hall sensor, it does't wear out, doesn't care too much about dust etc.

A potmeter with 295 degrees of rotation is not really a good option for pedals.
It means the whole range of the potmeter is divided by the amount of steps available at the controller, mostly between 256 to 4096 steps. So the amount of rotation you don't use is lost in resolution, with most pedals you use about 120 degrees. So you would loose about 2/3 of the available resolution.
With a linear potmeter you loose much less of resolution.

A hall sensor is also a very good option...

 

[Grzegorz Koguciuk]

That's why I said you should use as much range as possible. Standard potentiometers have range between 180-300 deg. The main reason why you would like to use standard pots is their price and availability(Try to find potentiometers used in Logitech G25/27 wheels - it's impossible(Pots from thesimshop are not working correctly)).

Resolution is not only affected by potentiometers, but also by controller that "read" them.
Leo Bodnar's says his board can work with 1kOhm to 100kOhm pots but for example Derek Speare is suggesting to use 100kohm pots.
Another thing is real use of this resolution and I don't think it's so important to have 4096 steps in clutch and throttle because our feets are not so precise. Such high resolution is not needed even if you want to smooth out the signal.
There is article on RD about that: http://www.racedepartment.com/forum/threads/why-you-dont-need-many-pedal-calibration-steps.55558/

 

[Phill Routledge]

Ok so if a Hall Sensor is a good option, which one is compatible with Dereks board and which one do you recommend?

Keep it simple guys...

 

[posimosh]

Use the frex HAL pot (look it up) for at least the accelerator. Better relyability, and more sturdy than any pot I have used. Furthermore, no need to worry about cat/dog/human hair, dust, grease, or anything else for that matter causing inconsistency or spiking signals to Leo's rather sweet Joystick controller board.
As for having increased resolution, it never hurts, and you are going to have the 4k steps pretty much any way you choose (if you are going to seek out and buy a purpose built joystick controller PCB). Although, as you say, our feet don't have the dexterity of our fingers, it is sometimes nice to be able to roll on the throttle extra slow and smooth coming out of an increasing radius turn after a braking zone. This has more to do with pedal travel though, and as your article (and others) have pointed out, 8 bits is sufficient resolution for just about any sim racing implementation.
With that said, sims are able to incorporate an increasing number of variables, at smaller increments as time and computing hardware progresses. Also, every sim deals with filling in the blanks between the steps of resolution that controller modules output differently (ok so some may be more alike than others, and some likely use the same logarithm to predict proper readings between steps... but you get the point... I hope). Increased resolution just fills in those blanks more often than lower resolution does.
There are hardware and software solutions that assist in this approximation and prediction process, but there is no substitute for more precise control provided by increased resolution. Plus, the increased expandability and quality, means, with a better joystick controller, you won't have to replace hardware as often when you go crazy with button boxes, dashboards, additional toggles, and dials, and, they tend to be made with smaller manufacturing tolerences which means it will stay functional and accurate longer than the cheaper stuff. Oh, and the cheaper stuff isn't really that much cheaper anyway.

 

[Tankslapper]

Hall Effect sensors are the way to go Honewell 495A or Alegro 1302 both are around 180deg turn, this makes it easier to utilise the full resolution.
Tho extra resolution doesn't help much in clutch or accelerator it doesn't hurt much either.
You could probably reduce the sampling time and increase sample accuracy at the expense of resolution.
I'm currently building my own Arduino Leonardo driven custom pedal set, Im playing with oversampling to acheive higher resolution on brake pedal which is hydraulic and uses pressure sensor and millivolt amplifier.