vinayp

Assuming you were able to read and write to fuel and ignition timing maps of whatever EMS system you were using, how would you go about implementing an auto-tuning system?

Modifying the fuel map seems like it would be easier than the ignition timing. If you were running rich or lean in a certain cell, you could just lower or raise the fuel in that cell by a small percentage, correct?

How would you go about doing a similar modification for the ignition map?

Cosmo Donk

Autronic has some kind of auto-tune but I am not sure what parameters it uses. I would imagine you could pull the EGT's and have it auto-tune off of them but you would first need to tune it yourself to give it a baseline of EGT's to look at. Once you gave it a baseline for a specific engine w/ specific mods you could probably have it auto-tune if you wrote the software.

DCrosby

AEM auto tunes, the trick with ignition is that there's no way to tell the ecu, well you lit the fuel on fire 2 degrees too soon... that's subjective to you, when you want to trigger the ignition pulse, and what you can get away with before you start adversely affecting performance, by either igniting too soon, or too late for it to burn properly...

Fuel is something much more measureable, since it has two "Legs" to stand on... 1 Timing is fixed, and 2 You have a reference device... (wideband O2 or Egt) to take readings on what the combustion remainder looks like once ignited... Therefore you also have a common logic, that if there's too much Oxygen, add Fuel, if there's too much Fuel, remove some... It's very clear on what needs to happen. With ignition, fuel and air ratios don't change a whole lot, but when the maximum power is achived does change...

So if you had a dyno input to ignition I suppose you'd be able to let it try something and see if it positively or negatively affected power, but it would take the ecu far more runs than a human. Since he knows what a baseline should be and expects certain results based on experience....
Which a computer won't be able to reproduce... especially mass produced ECU's like Power FC, AEM, Haltech etc... since this "Knowledge" is car specific...

vinayp

Thanks for your responses.

DCrosby, so if the O2 level in the exhaust isn't affected much by small changes to the ignition maps, what do people use as feedback when they street tune? Their butt dyno?

Hyperite

Yes, and that gets dangerous (let me tell YOU!), I'm using the AEM Autotuning and it is VERY good. If it changes a fuel cell at a particular moment and sees you going for the next one, well, it'll change that too (if it hasn't already been changed). VERY good.

vinayp

Hyperite, thanks for the feedback regarding the ignition timing.

I guess I'll just have to stick to the fuel maps.

Or get my butt wired up somehow to the EMS...

vinayp

I've been doing some thinking and maybe we shouldn't give up on auto-tuning of the ignition timing table just yet. We should be able to compute HP and torque from the engine RPM alone.

We can accurately calculate vehicle speed from the engine RPM, the gear and final drive ratios, and the wheel radius. Once we finish a run, we should be able to calculate acceleration as well by taking the derivative of the speed (with some filtering). We can then compute the raw HP by multiplying the weight of the car by acceleration and speed. I have to look up the formulas for this, but I think we can then compute the HP loss from drag using known value of coeff of drag for the FD. Also, we can compute HP losses from tire friction and drivetrain (these are known formulas). Subtracting all these losses from the raw HP should give us a reasonably accurate estimate of HP. Torque is easily derived from this. It may not be completely accurate but it will be CONSISTENT. That is, we can tell if a change raised or lowered the HP. For us, consistency is way more important than accuracy. As long as we are driving on a long, flat road and staying in the same gear, the HP values should be consistent across runs.

Having the ability to measure HP immediately after a run should allow us to automatically tweak ignition timing values between runs (although, this will probably require a LOT of runs to work well).

Does this make any sense or am I completely dreaming?

vinayp

Actually, it's TOTALLY unnecessary to compute the full HP value. At the end of the day, all of the HP calculations come down to one one thing: engine RPM and engine acceleration at that RPM (the speed at which the RPM is changing). If we tweak ignition timing at a given RPM and that results in higher engine acceleration at that RPM, then we are creating more HP (and torque). We don't have to compute what that HP is, but we do know that we are creating more HP. Once again, consistency of conditions between runs is of utmost importance here (same gear, same road, same temperatures, etc...)

So if this works, we can auto-tune the fuel map using A/F ratio, and we can auto-tune the ignition map using engine acceleration.

This all seems too good to be true, so maybe someone can point out where I'm going wrong here...

RETed

EGT and experience


-Ted