Have you thought about the risk of getting an unstable, oscillating correction mechanism if you make things remarkably faster? There is always a delay between signal and answer in this case, so the system might be quite prone to such a behaviour.

Hubert

This could be more important for ecus running in automobile environments. Differently to cars a motorcycle engine is running very quickly through the map. Quick accelerations, throttle cut, wide open again, shift up, shift down, that's in no way comparable to what you do with a car. That's one of the reasons that most if not all M/C ecus are based on this simple alpha/n model. No MAP sensors and other gimmicks. Just quick and dirty guessing.

That's what I already wrote in another thread: for me it's most important to have a quick or even real-time working ecu. Looking into too many maps and doing too many calculations would be counterproductive.

The art of ecu programming then is defined as to know what "Too many" means. If you don't know a good answer for that questions you better leave out the one or other feature

Hubert

You're right, this speaks against closed loop operation. And in fact I feel the ecu behaves crispier when run in open, in standard mode. Of course you need a more or less usable map for this. The map Cliff supports with every My15M is good enough for that.
Talking about the time delay: let the bike idle, whack the throttle, then wait for the O2 value to change and come back. It takes a lot of time!

Now don't come over and say that's a problem having to do with my equipement. The point is that an ecu must work with everybody's stuff and, most difficult , with everybody's preferencies.

Hubert

Luhbo wrote on 08/08/09 at 02:44:37:

I noticed that too. I had a very good closed loop map, then when my sensor started to act like a woman I went open loop - and the bike felt better than *ever*. This is one of the reasons it would be nice to be able to switch between open and closed loop using the Optimiser. My goal is to have a perfect map for my bike before my sensor dies of age - and from that point I will run open loop.

_Cliff_ wrote on 08/08/09 at 10:44:58:

I'm not sure I understand that fully. How does it know if it should adjust by 8 or 1 or something else? Does it 'learn' how much a deviation from the target should be treated?

_Cliff_ wrote on 08/18/09 at 11:49:43:

So it makes assumptions on what the o2 value means. That means if I knew what that assumption is, and set my LC-1 output level as close as possible to that assumption, I should get quicker closed loop reaction. Or if we look at it another way, if I could tell the ECU what the sensor level means, that should also yield quicker homing.