Shock Sensors

[racinray]

I have been using shock sensors with my FF and I have found them to be very useful. Afterall, the wheels and tires are moving up and down all the time and the data from shock sensors can tell you such things as the amount of roll in the car, whether or not a tire is off the track, how bumps are affecting the handling, the percent of time each shock spends in bump and rebound, the ride height of the car, etc. They will even help you determine if there is any "stiction" in your suspension.

I have an AIM EV04 data system and I expanded it so that I could use shock sensors. Most of the shock sensors that you buy are $250 each so we are talking $1k for just the sensors. That was quite a bit more than I wanted to spend and I figured that others are in the same boat and I thought I would share how I did it. I'm using rotating pots that cost $30 each and radio control hardware that you can buy at a hobby shop. These are the pots that I'm using: http://www.digikey.com/product-searc...-J15L&x=13&y=9 but there are plenty of other ones that can be used. You do want to get ones that have a low tolerance and these are + or - 5%. I had some old AIM cables laying around so I cut them and soldered the wires to the pot and then covered them with epoxy. If you don't have the patch cables then you can purchase just the cable ends (the binders). I made a bracket for the pot and mounted them to the rocker arm. I made another bracket to mount on the other end of the shock and a servo rod and hardware connects the two. The pic shows the idea but I have since gone to a metal arm. So I figure I have $50 in each pot and they seem to be working great!

But there is a tricky part. A true linear sensor is a lot easier to calibrate than this setup. Since this pot is rotating, the amount of movement of the shock does not correlate to the same amount of movement of the shock pot. In order to get around this I created a Custom Sensor for each shock in the AIM software. If you change the sensor type in the software to 0-5v, you can then go to "Online" mode and see the millivolts that the system is reading. Once you take the springs off and disconnect the sway bar, you can then move the shock from full extension to full compression in steps (each step for me was 1/8") and record the shock length and the millivolts at each step. You then enter this info into the Custom Sensor table and save it. Once that is saved you can then change the sensor type to the custom sensor that you just created. As long as you don't disconnect the pots then you shouldn't have to mess with them again. I can change my springs without disconnecting the rod/arm.

I hope someone finds some value in this and gives it a try!

Ray

[Tiago Santos]

Do you find you get enough resolution from the pot with the limited amount of travel? I thought about using a pair of gears (tiny one on the pot, bigger on the shock/upright) to increase the resolution. Meaning, for a smaller movement in the shock, the pot would rotate more.

[racinray]

Do you find you get enough resolution from the pot with the limited amount of travel? I thought about using a pair of gears (tiny one on the pot, bigger on the shock/upright) to increase the resolution. Meaning, for a smaller movement in the shock, the pot would rotate more.

Hi Tiago,

I think you pretty much take care of the resolution with the custom sensor table. If I remember correctly, I got a 7 to 9 millivolt difference per 1/8 inch of shock movement with these pots. Gears are a good idea too. I have seen that setup on other cars.

Ray

[BLS]

The potentiometer specs are for a ten turn pot. Use a single turn and your resolution will jump from 7-9mv to 70-90mv, a much better range. Or, better yet with a limited range of motion, there are 180 degree (or close) pots available. I used to use a throttle position pot on industrial machines for this reason.

There are other ways to improve the resolution. Gears are one way but suffer from losses due to the backlash. Another way when you have a 0-5vdc input and a small range of motion (degree's of rotation) is to use a higher voltage supply. Set the pot so that the resistance range of your movement limits the output to 5vdc. For example. if the actual rotation is 36 degree's (10% of a single turn pot with actual 360 degree movement. most are a little less), you could use a 10 volt power supply and get double the resolution, or 20vdc and get four times the resolution and still not exceed the 5vdc input. You can also add a resistor in series with the pot power supply to change the ohm range, a 5k pot with a 5k resistor becomes a 10K pot with 5k resolution starting at 5k for example.

Pay attention to the potentiometer watt rating. It must be capable of handling the larger power supply. Mechanical resolution of a wire wound pot will be effected by the ohm rating. The current flow through the pot will be effected by the ohm rating. A plastic pot will have a much higher resolution but you will only be able to measure the output when powered up. Trying to measure the plastic pot with an ohm meter only is usually a problem.

[racinray]

The potentiometer specs are for a ten turn pot. Use a single turn and your resolution will jump from 7-9mv to 70-90mv, a much better range. Or, better yet with a limited range of motion, there are 180 degree (or close) pots available. I used to use a throttle position pot on industrial machines for this reason.

There are other ways to improve the resolution. Gears are one way but suffer from losses due to the backlash. Another way when you have a 0-5vdc input and a small range of motion (degree's of rotation) is to use a higher voltage supply. Set the pot so that the resistance range of your movement limits the output to 5vdc. For example. if the actual rotation is 36 degree's (10% of a single turn pot with actual 360 degree movement. most are a little less), you could use a 10 volt power supply and get double the resolution, or 20vdc and get four times the resolution and still not exceed the 5vdc input. You can also add a resistor in series with the pot power supply to change the ohm range, a 5k pot with a 5k resistor becomes a 10K pot with 5k resolution starting at 5k for example.

Pay attention to the potentiometer watt rating. It must be capable of handling the larger power supply. Mechanical resolution of a wire wound pot will be effected by the ohm rating. The current flow through the pot will be effected by the ohm rating. A plastic pot will have a much higher resolution but you will only be able to measure the output when powered up. Trying to measure the plastic pot with an ohm meter only is usually a problem.

You obviously know a lot more about the use of rotational pots than I do! I'm curious about your second paragraph. If the data system is providing 5 volts to the sensor, how would I be able to increase the voltage to 10?

Ray

[BLS]

Ray,

You probably cannot unless the DA system allows for an external PS to be used. If it must supply (possible) the voltage you're stuck with that level. Some inputs only need to have the input voltage and negative supply to make the measurement allowing for an external supply. Some inputs require the use of the internal supply (internally referenced).

In that case, the pot selection is your only improvement avenue.

Regards,
Barry

[racinray]

Ray,

You probably cannot unless the DA system allows for an external PS to be used. If it must supply (possible) the voltage you're stuck with that level. Some inputs only need to have the input voltage and negative supply to make the measurement allowing for an external supply. Some inputs require the use of the internal supply (internally referenced).

In that case, the pot selection is your only improvement avenue.

Regards,
Barry

Ok, thanks Barry. I don't think you can use an external power supply with the AIM EV04 system. But mine seem to be working so I'm not going to change them right away although I now understand why a pot with more limited movement would have been a better choice.

Thanks,

Ray

[CDS]

Our Magnetic suspension sensors get around all of the limitations of pots AND have no moving parts and nothing to wear out. Here is the link to our web page on them:

http://www.competitiondata.com/sensors/magsensor.htm

and here is the data sheet:

http://www.competitiondata.com/downl...splacement.pdf

We are currently running a special on them for ApexSpeed members - set of 4 for $765 which includes your choice of CDS, Aim, or Stack connectors (or pigtails).

[924RACR]

Along the lines of cheap sensors... We've implemented TPS sensors, already robust rotary pots, on our bellcranks, directly on the pivot bolts for the bellcranks themselves (by machining the ends into tabs to engage the TPS sensors, like the end of a throttle shaft).

Inexpensive, but does require some flexibility in the design setup.

I still have to work in sensors (these or others) for my front suspension, which is a rocker setup vs. pullrod. Not impossible, just have to decide the best packaging option for me...

[BLS]

Here is a good choice, available with reduced mechanical travel in a rotary configuration, with spring return built in (CW or CCW) which helps minimize any lost motion. I see them from around $30 to $50 each. Very similar to what I have used in the past on machines running 24/7.

http://ecatalog.beisensors.com/viewi...to-9860-series?

They are IP66 rated for harsh wet environment.

[jcolbert666]

Ray-

Pretty nice little low-buck solution you came up with. I thought I might offer a little suggestion or two of my own.

1) Have you been having any problems with vibration affecting your measurements? While your linkage is clearly about as light as it can be, sometimes that isn't ideal. If your sensor and linkage mounts are not as stiff as they can be it is sort of like trying to measure jello.

2) When you change shocks do you have to remove the rod mount on the bellhousing? If it comes off like it looks like, then you should recalibrate your voltage curves each time.

[racinray]

Ray-

Pretty nice little low-buck solution you came up with. I thought I might offer a little suggestion or two of my own.

1) Have you been having any problems with vibration affecting your measurements? While your linkage is clearly about as light as it can be, sometimes that isn't ideal. If your sensor and linkage mounts are not as stiff as they can be it is sort of like trying to measure jello.

2) When you change shocks do you have to remove the rod mount on the bellhousing? If it comes off like it looks like, then you should recalibrate your voltage curves each time.

Hi John,

There is probably a little bit more "noise" then what I would have with linear or higher quality sensors but the data seems to be good enough for us to make decisions based on what we are seeing and what I'm feeling. I have attached some screen shots of shock traces with shock velocities as well as shock histograms.

And your right, if I remove the shocks then I do have to recalibrate but I actually have the numbers for my custom sensor table written down and as long as I set the millivolts to what they are in the table for full droop then I'm right back to where I need to be. I do check the data output while the car is on the setup pad though. I usually run the car for 20 or 30 seconds so that I can then look at the data and make sure that the shock lengths are actually what the data system says they are.

I do like the sensors that BLS mentioned -- wish I had found those before I bought mine!

Ray

[Jean-Sebastien Stoezel]

A true linear sensor is a lot easier to calibrate than this setup. Since this pot is rotating, the amount of movement of the shock does not correlate to the same amount of movement of the shock pot.

What about using a log potentiometer to compensate for the degradation of resolution and linearity when the servo arm moves away from the 90 degree position (decreasing measurement rate)? This might work if the total angle of rotation of the pot is less than 90 degrees.

[racinray]

What about using a log potentiometer to compensate for the degradation of resolution and linearity when the servo arm moves away from the 90 degree position (decreasing measurement rate)? This might work if the total angle of rotation of the pot is less than 90 degrees.

Hi Jean,

I don't know what a log potentiometer is so please educate me.

Just so everyone knows, my background is not in electronics or engineering -- I have a Finance degree and a lot of people have referred to me as a bean counter. So a lot of this stuff is frankly over my head but the more I read about it the more I learn. I'm very analytical and love the data acquisition side of racing. I can't afford all of the trick stuff so sometimes I just figure out a different way to do it, buy what I can afford and what I THINK will work. This is basically what I did with these shock sensors. They probably aren't good enough to measure small increments but the data seems good enough to see bigger shock movements, roll, pitch, histograms, etc. One of the points of this post is to get people to be creative and try different ways of gathering data. It might work or it might not but if you haven't spent a lot of money on it then there isn't a big loss. Maybe you try a cheap method first and use that for a season or so and then upgrade when you see there is more potential value there. I actually really like the CDS magnetic sensors but there are other items on my list that I should spend money on first.

Ray

[Jean-Sebastien Stoezel]

A logarithmic potentiometer has an approximated log transfer function, not a linear one. The rate of change of the resistance of the potentiometer changes in a non linear fashion.
Like you said the mechanical part of the setup you have does not have a linear response, you lose and gain degrees of rotation per travel as the RC arm moves around the 90 degree position with the push rod. I believe the relationship between travel vs rotation is an arctan function.

So using a log would not linearize your output since the mechanical part is modelled as an arctan and the pot is log, but it would somehow compensate for the loss of resolution pot rotation vs travel. The response of the mechanical part fades as the electrical part increases.

Hi Jean,

I don't know what a log potentiometer is so please educate me.

Just so everyone knows, my background is not in electronics or engineering -- I have a Finance degree and a lot of people have referred to me as a bean counter. So a lot of this stuff is frankly over my head but the more I read about it the more I learn. I'm very analytical and love the data acquisition side of racing. I can't afford all of the trick stuff so sometimes I just figure out a different way to do it, buy what I can afford and what I THINK will work. This is basically what I did with these shock sensors. They probably aren't good enough to measure small increments but the data seems good enough to see bigger shock movements, roll, pitch, histograms, etc. One of the points of this post is to get people to be creative and try different ways of gathering data. It might work or it might not but if you haven't spent a lot of money on it then there isn't a big loss. Maybe you try a cheap method first and use that for a season or so and then upgrade when you see there is more potential value there. I actually really like the CDS magnetic sensors but there are other items on my list that I should spend money on first.

Ray

[mblanc]

I have successfully used string pot sensors for suspension Sensors. After being told by several people that's not good enough to work, I improved the design. I added a small pulley to the Bell crank and lengthened the strings, and returned it back to the sensor itself. Essentially doubling the resolution of the sensor. Using high tech fishing line spiderwire that my fishing buddy recommended, there's zero stretch in the string. Doubling that resolution really did not significantly improve my read out or results.

Vibration or noise of the string is a concern but the results seem pretty good. I have also heard that it's not uncommon for the motorcycle crowd to use string pots on suspension, but I have never seen it in person. The string pots have plenty enough string tension to keep up with the deacceleration while returning the string.

I also love the idea of the CDS magnet sensor, simple is good. Wish I had started there. But I still wonder how linear they are as the sensor moves away from the magnet in an arc, would it take a custom table?

[David Ferguson]

I have installed several of the CDS Sen15 -- I love their flexibility, reliability and small size.

The calibration is very nearly linear if the magnet and the sensor move in an arc that rotates about the centerline of the magnet and the sensor (in other words each points to the center of the arc). Any other option (linear motion where the magnet doesn't turn), and you will need a non-linear calibration.

[CDS]

I have installed several of the CDS Sen15 -- I love their flexibility, reliability and small size.

The calibration is very nearly linear if the magnet and the sensor move in an arc that rotates about the centerline of the magnet and the sensor (in other words each points to the center of the arc). Any other option (linear motion where the magnet doesn't turn), and you will need a non-linear calibration.

David, is non linear calibration still a problem with some systems? I thought that horse died in the 90's

[924RACR]

As a mild sidebar, since we're talking about all these fun logarithmic etc calibrations... I'd like to highlight one pitfall I noticed with the RaceTechnology DL1 custom sensor setup.

If you're entering voltage values to create a calibration curve (in the config software), the software defaults to creating a polynomial curve fit equation.

That can mean some significant error when your sensor is logarithmic, etc!

Note that many temperature sensors are exponential.

So when I set up my oil and water temp sensors, I found that I was able to get much better accuracy in my logged gauge reading by making up a curve fit in an excel spreadsheet, then directly entering in the resultant equation... instead of allowing the software to make up the equation.

[David Ferguson]

David, is non linear calibration still a problem with some systems? I thought that horse died in the 90's

It's not a problem, but just not as straightforward as linear pots mounted in parallel with the shock. As Vaughan mentioned, some systems do generate equations instead of interpolating between a table of data points -- and those are the trickiest to use with the SEN-15.

[racedata]

BLS definitely knows his stuff !

Awesome info in here! I have mostly used shock pots on GT / touring cars so I've used traditional linear pots...but the info in here will definitely be helpful for future reference!

I've previously used TPS sensors off a Ford Ranger (4 cylinder) as shock pots on an open wheel car. They cost around 40$ each and are made to last and survive vibrations, cold, hot temperatures...pretty good deal for what they cost!

[Mike Scanlan]

I switched over to the CDS sen15 during the season and found the data a bit cleaner than the shock pots. They also take up a lot less space

[924RACR]

Additional amusement over this past weekend - got my car back out on track... getting a nice smooth, continuous curve of ride height from the RR shock pot (TPS sensor)... but the LR isn't. It's a bit noisy due to a nearby voltage regulator... and even worse, it seems to work like a step-function; it doesn't rail-out, but will hang-up at different levels in the travel.

Looks to me like the axis of rotation of the sensor doesn't quite line up with the axis of the bellcrank, resulting in a bit of stiction...
Apparently a touch more adjustment needed on that one...

[George Main]

Additional amusement over this past weekend - got my car back out on track... getting a nice smooth, continuous curve of ride height from the RR shock pot (TPS sensor)... but the LR isn't. It's a bit noisy due to a nearby voltage regulator... and even worse, it seems to work like a step-function; it doesn't rail-out, but will hang-up at different levels in the travel.

Looks to me like the axis of rotation of the sensor doesn't quite line up with the axis of the bellcrank, resulting in a bit of stiction...
Apparently a touch more adjustment needed on that one...

If it's a stepped graph, could be pointing towards a bad sensor...stepped graphs aren't usually associated with interference but rather bad spots in the sensors... look at it in live calibration and rotate it slowly to make the numbers climb and descend properly, if they jump around at all..either clean out the pot, (electrical cleaner) or replace it...
Alternately, could be a bad wire not supplying a constant 5 volts, would also effect the sensor in the same fashion...
BTW, your comment on reading suspension "ride height", you aren't... you are however reading "ride change"... useful in pitch, roll, heave, warp reading... the true ride height can only be accurately read with a ride height sensor..Laser for instance. : )

One of the biggest problems using throttle sensors is that dirt and wet stuff get into the sensor, as most are not sealed properly to prevent it..

[924RACR]

I don't mean electronic/RF interference... I mean, literally, physical interference with the sensor and bellcrank shafts not lining up, causing stiction.

Otherwise, yes, certainly. But I don't, at this point, care about what the tires are up to - that's for the driver to manage while behind the wheel. I'm interested in suspension articulation, therefore the bellcrank angle, while in motion, that I may, while parked, pursue necessary fixes to any problems that arise on-track.

Once I get a full data set without reservations I'll be able to see if there are in fact any issues to resolve...

PS - the noise is clearly evident as a separate influence on the signal, obviously much higher frequency.

Good point re: contamination... but I'm surprised that TPS's wouldn't be sealed enough. I guess we shall see how it goes. Worst case I'll migrate to MR suspension pots...

[BLS]

The TPS's I have used on machinery were IP66/67 rated. They are used in a washdown area without issue.

[dsmithwc04]

You obviously know a lot more about the use of rotational pots than I do! I'm curious about your second paragraph. If the data system is providing 5 volts to the sensor, how would I be able to increase the voltage to 10?

Ray

I know this is an old thread, but some really good ideas are here. I just wanted to say that with the AiM system, I believe the analog channels use a 10-30v signal on pin #3 (on the 4 pin connector) I know that the brake pressure sensors for sure use a 10-30V signal on #3 so surely all analog connections must be the same. This might be able to help get cheap pot sensors to be more accurate?

[romoman]

I know this is an old thread, but some really good ideas are here. I just wanted to say that with the AiM system, I believe the analog channels use a 10-30v signal on pin #3 (on the 4 pin connector) I know that the brake pressure sensors for sure use a 10-30V signal on #3 so surely all analog connections must be the same. This might be able to help get cheap pot sensors to be more accurate?

The AiM outputs are 5v and 12v. The pressure sensors take 12v.

[dsmithwc04]

The AiM outputs are 5v and 12v. The pressure sensors take 12v.

Thanks for the confirmation. I knew it was somewhere between 10v-30v from looking at the brake pressure sensor pinout page but never actually measured it to see what it was. 12V makes sense as that is also what CDS uses.

[TimW]

The AiM outputs are 5v and 12v. The pressure sensors take 12v.

And more importantly, the 5V is regulated and the 12v is essentially Vbat. You want a regulated power supply for exciting a potentiometer for accuracy. a potentiometer is essentially voltage divider. If you don't now how much you're dividing, you the signal will not be very useful.

[dsmithwc04]

And more importantly, the 5V is regulated and the 12v is essentially Vbat. You want a regulated power supply for exciting a potentiometer for accuracy. a potentiometer is essentially voltage divider. If you don't now how much you're dividing, you the signal will not be very useful.

I think the 12V is regulated internally in an AiM system. Atleast 8n the Evo4 logger.

[TimW]

I think the 12V is regulated internally in an AiM system. Atleast 8n the Evo4 logger.

Documentation is unclear at best, frankly, the VB in the pinout implies its essentially the stabilized voltage of the battery. the Vref is the regulated 5V.

For measurement, you want a very stable voltage level, and the +12V stabilization is likely less noisy than the cars +12V, but likely not suitable for precise measurement.

Power is provided by connecting the appropriate cable to the battery of your vehicle. It is not necessary to use special attention, because there is a stabilizer that regulates the internal power supply used by the system. So you can connect EVO4 without any problem to a NOT stabilized power source from 8 to 24 Volts.

I think this statement makes it a bit unclear, as I think they're overusing the 'regulates' word here. One way to tell is to set a fixed voltage divider to drop the 12V to 4.5V and then measure while the car (and alternator) are running and see if it drifts and/or fluctuates. I wouldn't rely on the +12V being a reference quality power supply until actually testing for that.

http://aim-sportline.com/download/do...O4_101_eng.pdf
http://aim-sportline.com/download/do...O4_115_eng.pdf

[romoman]

The 12v is for power supply, not to use as a reference.

[Marshall Mauney]

Sensors that are powered with 12V generally use it solely as a power source, not as a stable reference. The 1-5V output pressure sensors that we use for brake pressure, for example, require a power supply in the range of 8-30VDC; this is internally regulated to a fixed 5VDC that the sensing element uses.

As a result, in most places where you see +12VDC from a data acquisition system, it will be filtered and/or isolated, but probably not regulated. This is the case in all current CDS systems; 12VDC is filtered and isolated, but not regulated; 5VDC is always regulated.

This is a throwback to the days when the only power widely available in a car was a poorly regulated 12VDC. It made more sense to generate the reference Voltage inside the sensor than to hope that the main input Voltage was stable.

[dsmithwc04]

When wiring to a potentiometer, does it matter which leg the Ground and Vref connect to? Am I correct in assuming the center pin (pin 2) is the analog signal output but pins 3 and 1 should connect to ground and Vref correct in any fashion since it's only detecting the relativity of the two voltages? For reference, I'm using a Bourns 6639 10k pot.

[TimW]

When wiring to a potentiometer, does it matter which leg the Ground and Vref connect to? Am I correct in assuming the center pin (pin 2) is the analog signal output but pins 3 and 1 should connect to ground and Vref correct in any fashion since it's only detecting the relativity of the two voltages? For reference, I'm using a Bourns 6639 10k pot.

Yes it just changes the direction of change of the voltage.