Measuring front motorcycle suspension position

[Resilient]

I am wondering if anyone has any good ideas as an alternative to a linear potentiometer to measure the position of the front suspension on a motorcycle.

A linear pot would be ideal, and as I understand it, is what they use for telemetry on factory team race bikes. But I was looking at prices and they are in the several hundred dollar range. I am trying to figure out something that could do the same thing for less that I could use on my bike and maybe package and sell to other amateur motorcyclists who would like to start messing around with telemetry for less of an initial investment.

The travel on the front suspension is 125mm. I don't know how quickly it can compress/rebound, but I have found one example of a linear pot for racing that had a max velocity of 1m/s.

Any thoughts?

[Soeren]

Hi,

A linear pot would be ideal, and as I understand it, is what they use for telemetry on factory team race bikes. But I was looking at prices and they are in the several hundred dollar range.

I assume you are thinking of "slide potentiometers" (when talking potentiometers, the term "linear" refers to a linearly proportional change in resistance for a given displacement - vs. eg. logarithmic or exponential potentiometers).

You can buy very cheap slide potentiometers in "every" electronics store, but the common variety have a slide range of 60mm to 80mm and are not build for anything seeing moisture and dirt, so wouldn't live long in your app - probably only a few kilometers.

You can get any kind of potentiometer, even custom build to your specs, if you're willing to bleed gold for the cause, of course, just look at this page - while I don't know their prices, I am sure they're not cheap and getting anything sturdy enough to survive such harsh environment will be expensive.

I am trying to figure out something that could do the same thing for less that I could use on my bike and maybe package and sell to other amateur motorcyclists who would like to start messing around with telemetry for less of an initial investment.

There are several ways to measure linear displacement. LVDT's are the industry choice, as they are very resilient, but they're very expensive (unless you build them yourself).

The main parameters that sets the price is; precision, resolution, repeatability and MTBF/MTTF.
DIY will be cheap for most types of displacement sensor and you can use a broad range of methods besides resisance measurement. Inductance, capacity, light, magnetism etc.
If you add in a microcontroller, the sensor doesn't even have to be linear, as the equivalent linear value can be calculated or found in a look-up table to get a linear output from a non-linear input from the sensor.

But first, you need to find out which precision etc. you need and being realistic here might save a lot of trouble - if you don't need 0.1mm resolution, it will be money/work thrown out of the window to get it.


If the resolution (approx. 1mm) will do and you can live with the relatively low speed (around 45ms/reading worst case and 38ms typical [i.e. 22..26 readings/s]), you might wanna consider using the Sharp infrared PSD GP2Y3A001K0F as a rugged and fairly cheap solution. Its range is 40mm to 300mm (as is) and I'd assume that it could be fitted on the front stem/fork, measuring the distance to the fender.

Yet another infrared application would be to make a "ruler" with a Grey coded pattern. This can be read extremely fast and a pattern drawn in a CAD program  can be photographically transferred to film with extreme precision (at a repro house or similar - I'd guess around $10 to $20 for a sheet of A4 size) or photo etched into stainless steel foil.
A strip of film mounted on a piece of polycarbonate going through a reader (a screen for the reader could be made on the same film) can be made into an almost friction-less distance reader. If it has to be small to accommodate a high resolution, a CMOS line sensor/photo array chip can be used, while a lower resolution build could be done with conventional IR LED's and photo transistors.

[Resilient]

Your definition of linear potentiometer is probably right but that is the language used by motorsports types to describe these:

http://www.celesco.com/linearpot/

For my purpose +/- 1mm would likely be fine.

A sample rate of 45ms is slower than I would like. Some quick napkin calculations say that the wheel is making one revolution every 20ms if you are going 200mph which is the upper limit of what I would be interested in. I would ideally like 4 samples per revolution at a minimum, so managing to sample at 5ms would be ideal.

I am trying to find prices on CMOS line sensors but am having a hard time. I have found several product descriptions but cant seem to find any retailers for them. That seems like it would potentially be a good solution.

[Admin]

200mph on a motorcycle?!?!

Anyway, although you might calculate 20ms per rotation of the wheel, that's not necessarily the response rate of your shocks. If you have access to data you can probably find an acceptable frequency.

You could also perhaps use two Sharp IRs, and power the second one up 45/2 ms after the first to double their sample rate.

Just a thought . . .

[Resilient]

Talking with some people who do tuning there are two things one wants to look at. One being the behavior of the frame on the suspension. For that a lower rate is fine.

The other is looking at turn in and braking response, so what actually happens as the rider presses on the handlebar to initiate a turn or gets hard on the brakes. In these instances the tire deforms and can cause some unwanted vibration that is due to the rotation of the wheel and sampling multiple times per rotation is really the only way to see this.

And my motorcycle tops out at around 120mph. But people who would be interested in this certianly have motorcycles that can get up to 190 stock and with a little work can get over 200.

[Admin]

I got a "tiny" 150cc for city driving . . . supposedly it can do 55 on level terrain . . . but my fear of mortality kicks in at about 45mph . . . so even 100mph is 'insane' for me lol . . .

If your goal is to measure vibration, you could use an accelerometer in addition to your other shock travel distance sensor. That will give you a very fast sample rate . . . but I'm not sure if that's what you really want.

[Resilient]

I was planning on using one mounted to the frame and filter out the noise... I never considered on mounted to the suspension specifically to measure vibration. I will have to try this and see how it works.

[Soeren]

Hi,

For my purpose +/- 1mm would likely be fine.

Then a 7-bit resolution (0..127) will do fine.

A sample rate of 45ms is slower than I would like. Some quick napkin calculations say that the wheel is making one revolution every 20ms if you are going 200mph which is the upper limit of what I would be interested in. I would ideally like 4 samples per revolution at a minimum, so managing to sample at 5ms would be ideal.

With a linear Gray encoder read by photo transistors, you could sample a hundred times that.
Even better, when using Gray code, only one bit at a time toggles, so if set up with pin change interrupts on a microcontroller for reading it, you can get the exact time it igets to any given number and as long as this doesn't change (stays within the same 1mm), nothing happens. Each time the reading changes, you store the time since last change (and the number) and reset the timer in an interrupt routine. This will give you a much higher precision than with polled samples.

I am trying to find prices on CMOS line sensors but am having a hard time. I have found several product descriptions but cant seem to find any retailers for them. That seems like it would potentially be a good solution.

Line sensors are slow as well!
7 photo transistors, 4..7 IR-LED's and a bit of work will give you high speed at low cost.
Using a photo transistor you can get times as short as 10µs (just don't let them saturate).

What I don't know is, whether you are able to construct something rigid and stable (as it would be bad if parts start flying, when you cruise the local neighborhood at 200mph

If you feel up to the construction, I can post a Gray encoder, that you can print on a LASER printer and glue to a thin piece of clear plastic for initial tests (in a finished version, a lithographic film bonded to a piece of 2mm polycarbonate would be my choice, short of photo etching the pattern onto a sheet of stainless steel foil).

[Resilient]

Luckily for me, I live next door to some mechanical engineering students who are always up for a challenge.

I know I have seen some Gray encoders online in the past, but if you have a template handy that would be great.

[Admin]

I should have asked this in my first reply . . . why do you want to measure the position of the front suspension on a motorcycle?

[Soeren]

Hi,

Luckily for me, I live next door to some mechanical engineering students who are always up for a challenge.

Heh, lucky you

I know I have seen some Gray encoders online in the past, but if you have a template handy that would be great.

Attached.
I made the center distance 6mm wide to go well with 5mm photo transistors, but if you can find some 3mm devices it could be made more narrow. It could be changed to fit eg. 2/10" if you want to use matrix board for mounting the PT's.
If it needs to be as narrow as possible, 1mm plastic fiber optics can be used.
I made a couple of "windows" as well. One at each side will work for masking from false light.

On screen, the thin lines may look disturbingly bold, but when printed, they'll either be very light or missing completely (depending on the printer), as I defined them to be only 10nm wide (so they really shouldn't print). I only put the lines to be able to follow the pattern by eye.


Another thought that just pops up...
Hacking a digital caliper might be another option, although at least some of them won't accept high speed movement, but I have a carbon fiber caliper that takes quite a bit of knocking (yes knocking) back and forth to confuse it and that would give you 150mm of movement with either 0.1mm or 0.01mm resolution.
I have no idea of how fast it can be read out though.

[Resilient]

I should have asked this in my first reply . . . why do you want to measure the position of the front suspension on a motorcycle?

It is very useful is tuning said motorcycle. You want to make sure you are generally using as much of the suspension travel as possible. Then looking at how it extends and compresses under acceleration and breaking can help figure out what kind of compression and rebound damping you need to always have the best traction.

Another thought that just pops up...
Hacking a digital caliper might be another option, although at least some of them won't accept high speed movement, but I have a carbon fiber caliper that takes quite a bit of knocking (yes knocking) back and forth to confuse it and that would give you 150mm of movement with either 0.1mm or 0.01mm resolution.
I have no idea of how fast it can be read out though.

Yeah, I considered that too. I saw them pull one apart in a Sparkfun video. I may end up giving that a shot as well. Or at least trying to find someone who has hacked one to find out how quickly it can be read from.

Thanks for the encoder!