[Diy_efi] Traction Control System (update 2)

Nic van der Walt nic at unibell.co.za
Sat Apr 27 22:19:30 GMT 2002


>Using the LM339, it has 4 input comparators that will transform the ABS
>signal (frequency from each wheel) to a usable square wave. I will then
need
>4 x LM2907 units to convert each channel from a frequency to a usable
>voltage. (the LM2907 has a built in op-amp?) Then I will need an AtoD
to
>produce a DC signal from which I can then send via the RS232 using an
UART?

Once you have a usable frequency converting it to a voltage, and then
digitizing
the voltage again introduces a lot of unnecessary complexity and
unacceptable error.

You can and should read the frequency directly. A PIC can easily measure
the frequencies
you are talking about and send a host the counts via RS232.

Hook the outputs of the LM339 to four input pins on the PIC. Check the
microchip site
for their application note that describes how to do a frequency counter.

Realise that to measure frequency you have to approaches:

1) Count number of pulses in a given time period. Easy, but it means
your minimum update rate is
very dependent on minimum frequency. Assume you want to measure down to
5km/h. Rough estimate puts that
at 75Hz input. For reasonable accuracy you want to count at least 20
teeth ( 5% error ). So you need
to count for about 250mS. That gives you a measurement 4 times a second.
But not being able to measure
below 5km/h doesn't even allow the pretence of launch control or
traction control. All the fun is over
by the time you reach 5km/h.

2) Measure the time between teeth. More difficult, but it gives you an
update every 1/50 revolutions.
Most ABS systems work this way, because it gives you a more stable
control loop. It also makes your
wheel to wheel accuracy very good. For traction control you need this
level of accuracy.

A small micro ( PIC, Atmel AVR, Motorola ) can do either, depending on
you coding skill.

The important thing to note is that the minimum frequency is more of an
problem than the maximum. On the
early bike ABS systems the computers could only measure speed accurately
down to about 5-10km/h. So ABS
would work well till you are almost at a standstill and the wheel would
lock suddenly.

Check the datasheet on the frequency to voltage converter, they are very
inaccurate, and they have a small
range, i.e. it wont do the low frequencies properly in this application,
and that is where you need accuracy.

N.



_______________________________________________
Diy_efi mailing list
Diy_efi at diy-efi.org
http://www.diy-efi.org/mailman/listinfo/diy_efi



More information about the Diy_efi mailing list