Disabling Closed Loop

[Yellow Rose]

From a popular engine controller's website:

It is best to disable closed loop operation while tuning. Otherwise what commonly occurs is that the ECU will alter the mixture using the long term adjustment while the car is idling between dyno runs, which means that the mixture is not repeatable between dyno runs.

*****************

When tuning boosted NSXs, how is this done? Is there a disable feature within the tuning software or is the factory ECU shunted somewhere?

 

[D'Ecosse]

I saw similar advice Andy, whether it was on the same forum or not (AEM?), I don't recall - I don't think you can do this with the stock ECU however, it is feature of the s/ware I believe.

 

[true]

Sorry Andy, you can't 'disable' closed loop with the factory ECU unless it has been internally modified.

There is an external hack floating around the net that jumps to open loop to try and refrain from getting boost in closed loop ; but the idea is not.....ideal. Someone has a circuit diagram that uses a relay and pressure switch to send 5v (WOT) to the ECU TPS which immedietly makes the ECU run open loop. That's as good as I've heard...and may actually be helpful for you if you are having problems transitioning from closed to open loop with boost...but it's not disabling closed loop, just making it open up sooner. With some turbos we can generate a good amount of boost in closed loop. This design may not help you at all if you don't boost in closed loop. I think some people actually run a constant 5v to the ECU's TPS input for full openloop as well.

[This message has been edited by true (edited 01 January 2003).]

 

[BryanZublin]

This does not make sense. If you can't allow the car to run closed loop during idle when tuning, then what will happen on the road when the car WILL run closed loop and adjust fuel trim values? I think repeatability on the road is more important than repeatability on the dyno.

Your fuel injector hardware modifications + ECU map changes should NOT be altering the response at idle, therefore running in closed loop mode at idle is unchanged.

For example, if you install injectors with 1.5 times the flow rate, then the injector duty cycle should be scaled down by 0.666 at idle so that the flow rate is the same. In general, this static scale factor would be performed over the entire range from full vacuum to 0 psi boost so that the stock characteristics are maintained when not on boost. Then increase the duty cycle when on boost to maintain proper fuel ratio.

The above example applies to a black box that would be inserted inline with the stock ECU and fuel injectors.

 

[true]

Originally posted by BryanZublin:
This does not make sense. If you can't allow the car to run closed loop during idle when tuning, then what will happen on the road when the car WILL run closed loop and adjust fuel trim values?

I dont understand what you don't understand. It makes plenty of sense....simply tune light load situations to 14.7. As you mentioned, when you turn closed loop back on, the ECU fuel trim takes over. As long as you're not +or- 30% off, closed loop operation will be unaffected long term. It's extremely nice to just build a map where you are 100% sure that narrowband o2 sensors aren't dynamically adjusting your fuel injection calculations. Once you get your tune mint, flip closed loop back on and it will take care of all the standard adjustments.

 

[sjs]

The stock MAP sensor provides ~3v at zero vacuum/boost (atmospheric). Anything over about 3.1 will throw an error. But you don't need anything fancy to trick the ECU. Why not just T into the vacuum line to the MAP sensor and vent it open between tuning runs?

As for the rest of the argument, I think both sides have merit. I can see turning closed loop on and off during tuning to see the impact it has under various conditions and to keep it from distorting some readings. But it seems clear that final tuning needs to work with it functioning as designed.

 

[true]

The stock MAP sensor provides ~3v at zero vacuum/boost (atmospheric). Anything over about 3.1 will throw an error. But you don't need anything fancy to trick the ECU. Why not just T into the vacuum line to the MAP sensor and vent it open between tuning runs?

Hi Sjs, I don't understand what you are getting at here. Can you please expand...are you saying run an open vacuum line to disable closed loop??? Hrm. A lot of people run check valves off of their MAP line. This vents boost to the map (no check engine) and keeps from vacuum leaks in atmospheric -situations, but I dont see how this will do anything to change loop style.

But it seems clear that final tuning needs to work with it functioning as designed.

Exactly...that's why I call these ways of tricking the ECU into open loop a hack. It works, but it's far from ideal. A management system that can legitly run open loop with 6% and even 0% TPS is much more ideal then 100% if you plan to tune each cell.


[This message has been edited by true (edited 04 January 2003).]

 

[sjs]

Originally posted by true:
Hi Sjs, I don't understand what you are getting at here. Can you please expand...are you saying run an open vacuum line to disable closed loop??? Hrm. A lot of people run check valves off of their MAP line. This vents boost to the map (no check engine) and keeps from vacuum leaks in atmospheric -situations, but I dont see how this will do anything to change loop style.

Well, it seems to me that zero vacuum is where you get the full 3v from the MAP sensor to the ECU, and zero vacuum is interpreted as high load and/or WOT. So if you bleed off the vacuum on the line to the sensor then it sends the full 3v and I presume (but don’t know) that’s enough for the ECU to stop closed loop activity. Perhaps it must also sense a certain physical throttle position (I believe there are two sensors for that) or other parameters, I’m not sure, but simply inducing a well placed vacuum leak should cover the MAP requirement. Admittedly a small valve that closes off the line ahead of the sensor is more elegant, but you’d need it to also vent the sensor side to keep from trapping vacuum in that part of the line. However, such a tiny leak shouldn’t be a problem since it would only be left open between runs.