Modifying ABS signals for more slip (Rcomps)?
 

I found nothing (useful) so here goes.

I remember reading about someone local that had some success modifying the ABS sensor rings to enable the ABS calibration to be better suited to track use and Rcomps. I never figured out if he was removing teeth (to fool the ABS to think that the speed was lower) or if he added teeth (to make the speed higher).
Regardless this could just have been an effect on the ABS systems he was playing with, does anyone know if the Miata (the one I have is from 99) would respond well to such mods, and in that case should the ABS speed be lower or higher to allow more slip between the wheels (which is what I want)?

Changing the number of teeth on the trigger wheels seems a bit too mechanical to me and not continuously adjustable. So I was thinking about adjusting the frequency of the signals from the sensors up and down (lets say +-20%). That way you could adjust as you want, and maybe have different adjustments F/R, if that would trigger something desirable from the ABS computer. What circuit etc to use is unknown, but it sounds like it should be doable.

Messing with the signals from the sensors feels more comfortable than hacking the ABS computer itself. I mean, loosing a sensor doesn't kill your brakes completely.

put them on spinners so the ABS thinks the car is still moving!!!!!

Interesting.

I think you can stop faster without ABS if you can threshold brake properly. I guess if you spend a lot of time tracking in the rain it might be worth having. Does it rain a lot in Sweden? I've never been there. It rains a lot in Belgium, been there. But, I'm a noob so I could be wrong on all accounts. Except that I've been to Belgium and it rained a lot. I'm pretty sure I got that one.

Belgium, Sweden, UK, NL, basically we all have rain only. Except on the days that I choose to track my car. Still like the option of ABS (in the rain).

In the racing series who have allowed ABS noone have opted to run without it.
A properly adjusted ABS for the correct conditions (e.g. dry track and R-comps or slicks) will always outperform the best brake specialist.
But then we are focusing on systems like the Bosch Motorsport ABS etc (cost some $$$).

During the last five years of racing I've used my ABS quite a lot, and have no problem accepting the 5% extra weight penalty (in our regulations).
And I can't count the number of times it has reduced the number of flatspots and kept me on track (just a poor corner entry instead of spending the rest of the race in the gravel).

The question is if there are some sneaky way to improve the ABS functionality in the dry (on a wet track it needs no adjustment, you outbrake anyone, especially Caterhams... :)).
The difference between street tires and Rcomps is that maximum grip is achieved at a larger slip with Rcomps, so there is some grip left to be used on the most slipping wheel when the ABS decide to loosen the pressure.
Since it works nicely in the wet there might quite a lot left on the table...

But how to fool the system is probably specific to the implementation, does anyone now if the Miata ABS allows different percentage slip at 30mph than at 80mph? If so it could be interesting to try to fool the system.

Lol at spooky, ya permanent rain over here ;)

Funny that we have code orange and red in most areas right now due to virtually no rain for many weeks.

Without going to a fully programmable system like the Bosch Motorsports package ($10k to get in the door) your best bet would be to modify the valving of the solenoids. Ideally you would try to run them at higher cycles with a shorter stroke, but I don't think that's possible. You might try to reduce the seal they have as an attempt to lessen the edge of their effectiveness, but it would be a trial and error operation and I don't see a way to make it adjustable.

True, true. When I went from my STI with ABS to my Miata without I flatspotted several tires until I learned how to brake with this car. Still not the best brakerer but getting a lot more betterer.

Sorry, can't help you on the technical side so I'll stand down now.

I was thinking this over just the other day.

I am by no means an ABS expert, and I was operating on the idea that the sensor reads every tooth, and as the teeth mover slower and slower, with the brakes applied, or if one stops moving while the others are still in motion the system decides to kick in.

Given that, and that you want more slip before the ABS activates it seems that you'd want to lower how often it read, more or less forcing the computer to react less. So removing some teeth, or having new gears made, that have anywhere from one less to half as many teeth, could be used to fine tune a system. The only downside I can think of is the possibility that the system would ease off the brakes until it reads teeth passing the sensor again, so it knows the wheel has unlocked. In which case having fewer teeth would mean your wheel would have to move farther, and carry you farther down the road, before the system returned to normal.

Either way, an interesting topic, and hopefully some one who knows something instead of assuming can shed some light, I just wanted to throw in some food for thought until then.

It sounds like a micro controller in piggy back may be a good option.

Feed the micro the wheel speed sensors, and have it modify the signals before feeding them to the ABS computer. Much less invasive than hacking the ABS ECU directly, and you could probably set up a pot to control the signal externally.

An external pot could be used to tell the micro to add or subtract signals based on conditions.

--Alex

Id be worried that if you shaved off a tooth the computer would report a faulty gear and not work...


I wonder if you can put a piece of tape over one tooth and try to block the signal and see if that works.

I don't think you'd be able to block the sensor with tape. The magnetic sensor would read through it.

Not sure a micro would work easily either. If the sensor is VR based, you'd need some crazy read circuits and it would be difficult to reproduce to trick the ECU. Would need to scope it to see if its doable.

--Alex

Without reading the signal with a scope, here's an assumption that I have very litte clue how to do:
If the signal is fast enough to be considered a wave with a frequency (not sure how many teeth you need to pass with a similar speed for that to be true), it could be "just" a question about monitoring this frequency and then generate a new wave with a slightly altered frequency (+-20%).
This then need to be done on four channels, fast enough so you don't delay the ABS response (setting the frequency to be unaltered should not change the ABS responsiveness at all). Maybe four teeth is needed, but that is not a big chunk of a revolution.
Could this be done with simple wave with freq->voltage->new voltage->wave with new freq, or does it need to be a Microcontroller involved?
If it's easy and cheap to test, it might get done (preferably by someone that get inspired, sine I seldom finalize my ideas/plans).

All of this assuming that the change in absolute speed makes a difference.
Does anyone know what happens with the ABS brakes if you put crazy different tires on the axles? This might help the analysis. Maybe you can shift the ABS balance by increasing the speed on one axle, at least alter the behavior (probably to the worse, but testing is knowing).