Open loop or closed loop tuning?

[Chris Blanchet]

Hi everyone,

My team is in the process of tuning our F4i with a PE3 ecu. I stumbled upon the target lambda tables in PE and did some digging to find that it was based on closed loop fueling. Are there any suggestions regarding using this type of tuning strategy? Would closed loop mapping be sufficient for the overall map or would open loop tuning, though more time consuming, be more effective?

[Jay Lawrence]

My understanding is that closed loop Lambda control is typically employed on road cars during periods of sustained constant throttle (and I would think that this control would apply a trim, rather than adjusting the base fuel map). As FSAE cars are very rarely at the same throttle position for very long, closed loop control is not particularly effective. You would find that there would be Lambda overshoots everywhere and it would not work very well (and could be quite dangerous for your engine).

So I would say go open loop. Just tune at 100% throttle, then come down in increments (5,10,15,25%, whatever suits) on the dyno, and then fine tune on the track (and adjust your trims as required).

I'm not sure how the PE works, but with MoTeC you can use the aim lambda table to tune fuel automatically for a given site. This doesn't then imply ongoing closed loop tuning, however.

[Kirk Feldkamp]

Yup, that's pretty much it. The time delay of the lambda sensor doesn't usually track well with fast transients. Most street cars generally have a open loop map that is then "trimmed" with longer-term, closed-loop adjustments when there are periods of steady throttle position. As Jay says, you likely won't see much, if any, difference on a FSAE-style course.

Also, I'd *highly* recommend that you work up , rather that down, when you build your fuel map from scratch. This will mitigate the risk of building in rich misfire areas, as well as keeps things safer because you won't accidentally jump in to un-tuned (read:lean) high load areas.

Do a search of the forum for a fairly recent thread about starting and tuning strategies. It had some OK rundowns.

-Kirk

[Dash]

Our team uses MegaSquirt and used the closed loop tuning mechanism to get a base tune on the car. The tables were then altered manually to fine tune the car. Basically we jumped in the car with a fuel map that the powertrain guy made, drove the car around for about 20 or 30 minutes while it was doing the closed loop tuning, and the car made amazing leaps in power. The fine tuning was then done by hand on a dyno to speed up the process ( it made really big jumps in fuel based on how far off it was from a target value ).

[Kirk Feldkamp]

I have also been very impressed with the MegaSquirt auto-tuning capabilities in the past. It's definitely a benefit of going that route. It's not a "regular" closed-loop tuning algorithm like you would normally think about though. My understanding was it takes large amounts of logged data, averages and compares it to the target map, and applies widespread modifications to the whole map while the engine is no longer running. It sounds like there's probably a new piece of software that allows that modification to happen on the fly. "Normal" closed-loop fueling that I'm used to is just a realtime trim based on current operating conditions, without any permanent modification to the base map(s).

-Kirk

[Mbirt]

Check out what the latest Vipec ECU's are doing to edit VE map cells with WBO2 feedback. I think this is pretty cool: http://www.youtube.com/watch?v=jel2P1slIag

[Dash]

Originally posted by Kirk Feldkamp:
I have also been very impressed with the MegaSquirt auto-tuning capabilities in the past. It's definitely a benefit of going that route. It's not a "regular" closed-loop tuning algorithm like you would normally think about though. My understanding was it takes large amounts of logged data, averages and compares it to the target map, and applies widespread modifications to the whole map while the engine is no longer running. It sounds like there's probably a new piece of software that allows that modification to happen on the fly. "Normal" closed-loop fueling that I'm used to is just a realtime trim based on current operating conditions, without any permanent modification to the base map(s).

-Kirk


The version we used allows the updates on the fly. It would re-write the data every so often but it was quick enough the motor wouldn't die during the writing process.

[Warpspeed]

Originally posted by Chris Blanchet:
Hi everyone, Would closed loop mapping be sufficient for the overall map or would open loop tuning, though more time consuming, be more effective?
They are virtually the same thing.

Manual tuning on a dyno just means you get to edit the injector duration numbers yourself in each tuning cell, via the keyboard, to reach your target air fuel ratio with the engine held at a constant steady speed and load on the dyno.

Closed loop tuning does the same thing, except you first enter the target air fuel ratio into each tuning cell, then run the engine on a dyno, hold the speed and load constant, and the software makes the air fuel ratio converge with the required air fuel setting you entered.

This is only used in "closed loop tuning mode" during steady state dyno operation.
It makes dyno tuning very fast and easy, and it is just as good if not better than you could do it yourself manually.

During normal engine operation the system is run open loop in the normal way, for fastest possible response.

[Chris Blanchet]

Thanks everyone for the responses. This clears up quite a bit.