how to measure the speed of wheels? Which algorithm?

[mr_g]

dnb, I hear you... ;)

[mr_g]

Here is the TCM hardware... :D





I have made the algoritam. Very simple without timers, without dividers. Just a simple counter... And it works perfect...

The only issue is false traction slip detection in hard short turns...

4x LM1815, ATmega128, 4x20 LCD, ULN2003D, A/D inputs, SPI, UART, and 2 expansion slots.. :D

[gunni]

very interesting,

Explain the operation and adjusting if you could.

[dnb]

Well done - told you it didn't need a divider :)

False slip readings could be compensated by a steering wheel position sensor or 3 axis accelerometer - with that you could quite easily calculate what the cross axle speed should be for a given turn.  There may be a way to calculate it without this.

After my autotest today, I think I need traction control on the TVR...  Are you making boards you're willing to sell?  I could provide a reasonable input to algorithms if you want too.

[mr_g]

I think I could detect turning by counting impulses from left and right rear wheel and detecting difference in their speed...

[dnb]

How will you tell that from one of the rear wheels starting to slip?  (Not that you can do a lot about that without something to control the brakes)

[mr_g]

My rear wheel wouldn't slip... :lol: It is FWD.. :D

[dnb]

lol.  These people with wrong wheel drive... ;)

[mr_g]

As soon we have an stable and tested operating version, I will sell the board to anybody who needs it... ;)

Now I need some interface  to the VEMS to tell it to lover the torque for a moment of wheel loosing traction.

[multiplex]

very interested in your developments here  ;D

subscribed

[mr_g]

For those who have joined the discussion later, I'm working on TCM (traction control module) which will act as a g-tech (measuring acceleration in meters and km/h) and the OBC (on board computer) that will measure trip times, fuel consumption average, current, etc. ..

It has the 4x20 text display, where you will get all the information, and the acceleration measurement:
*automatically (as soon as you start from standing)
*manual (if you don't want to brake to zero, but you want to measure acc for example from 100km/h to 200km/h)

And later I will develop an application in VB, for those who have carPC, that will show it graphically on the screen with graphs, history, best times, etc. ..

And of course the primary thing is to lower the torque at the start when you got too heavy foot.. :)

It meassures:
N*100m acc,
0-60km/h,
0-100km/h,
0-160km/h,
0-200km/h,
60-100km/h,
100-160km/h,
160-200km/h and
100-200km/h...

[gunni]

As soon we have an stable and tested operating version, I will sell the board to anybody who needs it... ;)

Now I need some interface  to the VEMS to tell it to lover the torque for a moment of wheel loosing traction.

This would be perfect with drive by wire. Which VEMS will have to upgrade to eventually.

Otherwise cylinder fuel cut would probably be best for NA vehicles and boost lowering along with cylinder fuel cut if boost has been lowered as much as possible.

This sounds very promising. Can you use VR sensors or are Hall a must

[mr_g]

VR sensors are standard ABS sensors.... So it uses VR.. ;)

[peter_jensen]

This is a great project.

I am joust wondering if it could communicate with Vems by canbus.
This will give us the possibility to make a lot of things like fly by wire and so.
Also maybe you could use some of the data there are already in Vems.

Looks good 

[dnb]

In the short term there is the nitrous retard feature.  This could be used to pull out a few degrees of timing when a trigger is applied.  You could also construct a simple PWM interface connected to the VEMS PS2 port to simulate a wheel speed sensor - you can then use the VEMS launch control settings to cause a limit in torque for a given input.

I know neither is ideal, but it gives you a quick way of testing your control algorithms now.