MoTeC Global Forum

this system is on a motorcycle with front and rear wheel speed sensors. and using those to calculate % of slip.



EDIT: I read a pdf file on tuning PID's that someone posted here, and i feel i have an understanding of what the parameters do.

Although, no my question is what would be a typical starting value for the P,I,D? would 1 be effective? would .5 be effective or do i need to be starting with bigger numbers like 5 or 10?

this is a parameter that I am hesitant to experiment with much without someone's input as the results of incorrect settings could be, well, a little dangerous.

I have some data logs with wheel slip but im not 100% sure im affecting what's happening and need to know what values would be an acceptable range..

PID depends on engine type, number of cylinders and torque output and also what cut type you are using to a certain degree.

Have a look here....

viewtopic.php?f=39&t=1240&start=0&st=0&sk=t&sd=a

FWIW in the main car I work on we are using a P of 4, with no I or D.

IDP: understood, that all motors require differing levels.... but for reference i'd like to know what a motor uses.. any motor really, so i can start to narrow down where i should be. At least in the ball park. or even the the stadium.

Sean: thanks very much, that is helpful to know. and very helpful to actually know a value that's being used in the P even if it's for a car it's at least a reference point.

currently i too am only using a value in p with the rest zero'd, and i think I am experiencing some oscillation right now and think that the P may be too high and i need to start adding in some D.

it appears if i get hits of TC and it returns the tire to FULL traction. and then the process starts over again with slip then a hit of TC and then full traction.

do you think that by adding some d that I could maintain a certain slip instead of bringing the tire back to full traction. because once it hooks up it unsettles the chassis a bit. The point I'd like to acheive is a consistent wheel slip and slowly bring it back to full traction.

I seem to recall that when we started with the TC we had a P of 10, with I and D at zero, then had to start to add I and D to maintain some control. We then gave up on that and went to the P of 4 , and I and D zero and had no problems since. I don't say it's ideal, but it works and there is always something more importanat to test

You looked at the sample file in the other thread, should give you some ideas I think? What is your slip filter set to, we like 4 at present? You are logging in the ecu I guess, you are logging ign cut total and fuel cut total at reasonable log rates?

yes, that file they were using 2 in the p and the rest of the settings were the same as what i have. I am unsure of what bike they were using it on, it was an inline 4 i can only guess it's a 1000cc... i didn't look too far at the file but am going to look again later to see if there's anything that might tell what size the motor it was.

yes 4 for filtering.

what do you reccomend logging the cuts at? 100hz enough?

btw, what ecu are you using and on what type of car?

I generally get the customer to install a multi position switch for a gain compensation. You can then start with a P (no I or D) of 1. You gain table will be 0 - 200% stepping up with each position of the switch (we use 9 position switches).

You should start with low values of Slip Aim so that the system with activate with very little provocation. Actual slip values should be logged first with no TC enabled. Log Slip and Longitudinal G force as a guide but remember having the front wheel off the ground will affect the G force reading.

The initial tests are more to make sure the system is working, with a low value of P you should be able to nail the bike in a straight line and hear/feel the engine cutting. Obviously you need to be careful at this stage. If nothing much is happening you can wind up the multi position switch....saves coming into the pits all the time to connect the laptop.

I would also set up some Power Reduction Delay (maybe 150-200msecs) which will give a filtering effect to the TC cut calculation. Also some Power Reduction Decay (maybe 50%/sec) which will slowly bring the power back in.

Get a feel for what the calculation is doing first with your logging. A softly-softly approach would be advisable on a bike. Don't try to set up P, I and D all at once, stick with the P only until you are more familiar with the system.

MarkMc wrote:I generally get the customer to install a multi position switch for a gain compensation. You can then start with a P (no I or D) of 1. You gain table will be 0 - 200% stepping up with each position of the switch (we use 9 position switches).

You should start with low values of Slip Aim so that the system with activate with very little provocation. Actual slip values should be logged first with no TC enabled. Log Slip and Longitudinal G force as a guide but remember having the front wheel off the ground will affect the G force reading.

The initial tests are more to make sure the system is working, with a low value of P you should be able to nail the bike in a straight line and hear/feel the engine cutting. Obviously you need to be careful at this stage. If nothing much is happening you can wind up the multi position switch....saves coming into the pits all the time to connect the laptop.

I would also set up some Power Reduction Delay (maybe 150-200msecs) which will give a filtering effect to the TC cut calculation. Also some Power Reduction Decay (maybe 50%/sec) which will slowly bring the power back in.

Get a feel for what the calculation is doing first with your logging. A softly-softly approach would be advisable on a bike. Don't try to set up P, I and D all at once, stick with the P only until you are more familiar with the system.

mark, thanks for the input, this helps alot. I am using a a multi position switch.

I sent you a pm.

Hi Ben

The main thing which affects the setup of the TC is the available level of grip. This is usually defined by the tyres (make, slicks or road legal etc) and the bike type (Superbike or Supersport).

Attached is a basic example file which contains a setup for a Supersport bike running with Pirelli tyres.

I would anticipate that the P term is pretty close at 2.0, even if you have a Superbike with different tyres, but the aim slip may need to go up for a Superbike.

The rider trim is on an aim slip comp table rather than a gain table but this can be swapped depending on what you want to achieve.

The other aim slip comp table has throttle position on its axis (you can also add gear). The purpose of this table is to give a rough compensation for lean angle (there is a correlation between lean and throttle position) in the absence of a lean angle input (swap throttle position for lean if you have it available). This will generally have to be tweaked depending on the rider and the make of tyre (aspect ratio differences).

There is also a gain table with fork position, this gives some control over the severity of the inherent anti-wheelie power reduction.

You will also notice that the maximum power reduction table is at 0% at WOT, you may want to change this for a Superbike to allow some full throttle power reduction.

This is a pretty basic setup but should get you going.

FoxPerformance wrote:what do you reccomend logging the cuts at? 100hz enough?

btw, what ecu are you using and on what type of car?

Ben, I'm no expert, 100hz sounds fine.

M800 in a 800hp RWD Gt car.