klodkrawler05
klodkrawler05 Reader
5/23/18 11:03 a.m.

I've been mulling this over for a while now and have after some time stumbled upon a potential solution? Full disclosure, I am not very electrical techy at all so I may be heading entirely down the wrong path.

One of the ideas for checking the effectiveness of track aero was to drive on a flat surface at a common speed with and without aero, if you monitor ride height changes you could calculate backwards to get the amount of downforce/lift you are making.

The easy although not GRM friendly option is to buy a fancy data logger ($$$) and 4 shock travel sensors ($$x4) These items are ruggedized for motorsports and can be run on track to tell you a whole host of other things about your setup also. 

We're looking for something no less accurate but much less robust simply for gaining knowledge in a test scenario at a much lower price point.

Most every car with HID's have at least a front/rear ride height sensor that spits out 0 to 5v, google shows some candidates that might be easily adapted to any car like this one from range rover:

 

searching all sorts of various key words ultimately lead me to this little $60 box, youtube videos make it appear fairly simple to operate the included logging software.  https://www.dataq.com/products/di-1100/

 

so now I have some sensors, in theory, I will shortly have something that can log 0-5v. next up will be seeing how much the suspension compresses with say, 100 then 200lbs hanging on the splitter and then the wing, that should then give me a reference of how much compression I get and I may be ready to do some logging?

Any glaringly obvious reasons this won't allow me to get some repeatable data? any one else tried this before and succeeded/failed? I found some various FSAE threads discussing this but they were mostly tight lipped about results.

 

 

 

 

adam525i
adam525i Reader
5/23/18 11:17 a.m.

Does your M3 already have similar sensors for the DSC system? You may be able to tap into their signal rather than adding extras. My concern would be whether an extra 200lbs of weight at the front splitter makes a difference in the ride height that can be discerned from all the noise of suspension movement just driving down a normal road with the high spring rates of a track car.

But it can't hurt to try and see what the data looks like, that looks like a cool little logger.

Adam

GTXVette
GTXVette SuperDork
5/23/18 11:56 a.m.

LO-Tech.     Put a Tie wrap around the shock rod(Stem?) on each shock slide it to touch the shock body then load up what you want for weight and see how far up the rod it moved when you take the weight off.  50cent

stafford1500
stafford1500 HalfDork
5/23/18 12:02 p.m.

Those ride height sensors are nothing but rotary potentiometers with arms fitted to convert the motion of the suspension to the rotary the sensor needs. Rotary sensors are very inexpensive and you could make arms that are the right size/shape for your application. You could also use a sensor called a vernier sensor that needs nothing more than a wiper to generate the same sort of suspension displacement output.

I would use measurements at all four corners so that you can determine any roll imparted on the system. Logging speed may be the limiting factor, you will need about 50 samples per second for smooth roads and more than 100 samples per second for rough roads and cornering scenarios. You will have to come up with a way to smooth the raw data and add some sort of signal to identify when you want to start analyzing data (a switch will work just fine). If you log suspension anywhere but the lower ball joint, you are going to have to account for motion ratio differences in both the spring forces and wheel travels.

I have also seen some fairly inexpensive non-contact sensors (proximity sensors) to track suspension motion, but the limited range means a lack of accuracy since they have to be mounted close to suspension pivots and then calibrated to wheel travel. Using the methods laid out will also give you a way to measure how much sway bar twist you generate in corners, which may help dial in the suspension setup.

On the effectiveness of a 200 pound load and it measurement, if the car moves more than 1/2" or so with you standing on the front/back then you should be able to measure the results.

kb58
kb58 SuperDork
5/23/18 12:08 p.m.
adam525i said:

Does your M3 already have similar sensors for the DSC system? You may be able to tap into their signal rather than adding extras. My concern would be whether an extra 200lbs of weight at the front splitter makes a difference in the ride height that can be discerned from all the noise of suspension movement just driving down a normal road with the high spring rates of a track car.

But it can't hurt to try and see what the data looks like, that looks like a cool little logger.

Adam

^This. Roads, and many tracks, are far from smooth and the signal can be hard to see when the noise is as much or more that what you're looking for. Best approach is some heavy filtering. Also, those particular sensors have an "elbow" in them that's not instrumented. You're going to have to be careful how they're mounted else some - or even all - of the desired motion won't be logged.

GameboyRMH
GameboyRMH MegaDork
5/23/18 12:32 p.m.

I think stafford1500 covered it, only things I'd add are that those ride height sensors used for headlight levelling are cheap from a junkyard, and I'd recommend a Race Capture Pro box for data logging on a budget, they give great capability for the money, especially the older models which you might be able to find used.

JBasham
JBasham HalfDork
5/23/18 1:48 p.m.

Slick logger package.  Thanks for sharing.

sleepyhead
sleepyhead Dork
5/23/18 2:03 p.m.

it seems like you're going to be getting really noisy data from position... and that that data is an order removed from getting what you really want: "how much down force am I creating?"

is there a reason to not recommend slipping some threaded "button load cells" between the splitter and its mounting point?  other than the fact that you'd probably need two of those DAQ's and the headache of trying to sync the data between them...

you know what, never mind cheeky

¯\_(ツ)_/¯
¯\_(ツ)_/¯ UltraDork
5/23/18 2:11 p.m.

I have kicked around the idea of trying to do a cheap datalogging setup, with wheel position, steering angle, throttle and brake positions as the main targets- then I start going off into load cells/strain gauges on all sorts of components, tons of engine data, damper temperatures, tire temperatures... and I realize why these things are expensive.  Can I get something like that dataq module with just a E36 M3 ton of inputs?

klodkrawler05
klodkrawler05 Reader
5/23/18 2:28 p.m.
sleepyhead said:

it seems like you're going to be getting really noisy data from position... and that that data is an order removed from getting what you really want: "how much down force am I creating?"

is there a reason to not recommend slipping some threaded "button load cells" between the splitter and its mounting point?  other than the fact that you'd probably need two of those DAQ's and the headache of trying to sync the data between them...

you know what, never mind cheeky

Any reason not to put strain gauges on the splitter/wing mounts and measure tension/compression then at that point? 

Also they make these little boxes with up to 32 inputs so it could be done with one sync'd unit.

edit: after a quick google, button load cells are pricey, at that level of cost I'd be going for the motorsports shock travel jobbies

klodkrawler05
klodkrawler05 Reader
5/23/18 2:30 p.m.

In reply to ¯\_(ツ)_/¯ :

they have some that do more channels yea, they get order of magnitudes more expensive once you get into 32+ channels though

https://www.dataq.com/data-acquisition/voltage/

sleepyhead
sleepyhead Dork
5/23/18 2:54 p.m.
klodkrawler05 said:
sleepyhead said:

it seems like you're going to be getting really noisy data from position... and that that data is an order removed from getting what you really want: "how much down force am I creating?"

is there a reason to not recommend slipping some threaded "button load cells" between the splitter and its mounting point?  other than the fact that you'd probably need two of those DAQ's and the headache of trying to sync the data between them...

you know what, never mind cheeky

Any reason not to put strain gauges on the splitter/wing mounts and measure tension/compression then at that point? 

Also they make these little boxes with up to 32 inputs so it could be done with one sync'd unit.

edit: after a quick google, button load cells are pricey, at that level of cost I'd be going for the motorsports shock travel jobbies

yeah... leftover thoughts from my old Simulator CFM DAQ days... little different when a company is purchasing things

alfadriver
alfadriver MegaDork
5/23/18 2:56 p.m.
¯\_(ツ)_/¯ said:

I have kicked around the idea of trying to do a cheap datalogging setup, with wheel position, steering angle, throttle and brake positions as the main targets- then I start going off into load cells/strain gauges on all sorts of components, tons of engine data, damper temperatures, tire temperatures... and I realize why these things are expensive.  Can I get something like that dataq module with just a E36 M3 ton of inputs?

How many do you need?  Other than the commercially available ones, in theory, you can chain together many arduinos plus a raspberry pi.  All talking to the R-Pi and recording happens there.

¯\_(ツ)_/¯
¯\_(ツ)_/¯ UltraDork
5/23/18 4:11 p.m.

In reply to alfadriver :

Probably thirty something?  Throttle, brake, steering, X, Y, Z, travel on each corner of suspension, load on two or three attachment points for each corner, probably temperatures for a bunch of stuff... it's hard to pin down where to stop, I would mainly want a better understanding of what forces are really going where and under what circumstances.

alfadriver
alfadriver MegaDork
5/23/18 4:45 p.m.

In reply to ¯\_(ツ)_/¯ :

So an arduino mega has 16 analog inputs.  And I've seen a project that had 3 arduinos on a R-Pi- so you only need 2.  And you could likely digitally input GPS data.

clshore
clshore New Reader
5/23/18 6:10 p.m.

I'm a longtime user of the DataQ products.

You can homebrew a Rasberry solution, but then you would miss the spectacularly useful interface and software that is already available for the DataQ.

Works like a multichannel recording oscilloscope, each channel has it's own trace.

You can record as much data as you have file storage, set triggers, setpoints, do searches, forward, reverse, slow, stop, fast, zom in, zoom out, whatever you need.

The interface is open, and so custom hooks to spreadsheets, etc are pretty easy to do.

As for position sensors, the common throttle position sensor is cheap and available, rugged, reliable, and adaptable.

Linear slide potentiometers can also be easily adapted, surplus and eBay are your friend.

alfadriver
alfadriver MegaDork
5/23/18 7:45 p.m.

In reply to clshore :

The software you look at is important.  Very important.  I wish what we used at work was easily available to all- it's the easiest plotting program ever.  So I asked them if it was....

clshore
clshore Reader
1/23/21 6:58 p.m.

In reply to sleepyhead :

The Dataq are 4 channel units, and mated with a laptop or handheld can capture as much data as you have storage for.

The interface to review/observe the logged data is very intuitive, like a an Oscilloscope and a VCR combined.

Our Preferred Partners
9QJ4sDJQfrc4LPzqhNM1P2SxrK9cPrVok7oCZ1mnoB3mecNHErqFfh99UEoZ6oNM