Closed Loop boost control not working. Help please

[Nige]

I have successfully run open loop boost control for several years. I have Map1 and Map2 with different Boost By Gear targets.

I am trying to get MAP2 to be Closed Loop instead.

To aid testing, I`ve set the target to be 1.05Bar everywhere (205kpa on the log)


Driving on Map1, the Turbo Solenoid matches the MAP1 base duty table, here is MAP2 base duty table


Full throttle, duty rises to 100% initially to aid spool then starts dropping once boost is made, 0.8Bar at 3,500rpm.

On the lower trace, MAP2Switch is Yellow, Turbo Duty Cyan.


As SOON as I switch to Map2, the duty starts dropping to Zero which obviously causes the boost to fall. As soon as I switch back to MAP1, the duty rises again to match the Map1 settings and the boost rises.

This is repeatable everytime I switch to MAP2. Obviously, I have something wrong in the PID settings, but I have no idea what.

Does anyone have working Closed Loop Boost control and you could share your Turbo Parameters please ?

[David Ferguson]

Since positive PID parameters seems to go the wrong way, have you tried using negative factors (ie. PID Prop Factor = -50, PID Int Factor = -10, PID Diff Factor = -10). You will want to plot MAP and PID Target Pressure on the same graph to see how the control is doing, relative to the target.

[Nige]

I haven't tried negative setting. I didn't know that was a thing. Is this from experience or just a suggestion for something to test ?

I'll give it a try, thank you

I can`t see where you can log target pressure.

[Nige]

Since positive PID parameters seems to go the wrong way, have you tried using negative factors (ie. PID Prop Factor = -50, PID Int Factor = -10, PID Diff Factor = -10). You will want to plot MAP and PID Target Pressure on the same graph to see how the control is doing, relative to the target.

Have you actually tried negative pid on a working open loop setup ?

It just caused it to maintain turbo solenoid duty and was producing 2.1bar by 4,000rpm so I abandoned that

[stevieturbo]

Does the PID setup use a base lookup table ?

Or is it only used when you select the base lookup table ?

[Nige]

Yes it uses the base table as a reference starting point and then the pid loop from there

[David Ferguson]

I have not tuned a DTA boost control, but I've done plenty of PID tuning on other platforms.

Based on your results, I would cut the proportional in half (so -25), and continue to do so until it begins to control the boost.

To tune a PID control loop, start with the Integral and Derivative to 0, and work with just the Proportional term. Get it to where it responds but doesn't overshoot or continue to oscillate. Then increase the Derivative term to improve the response and prevent the overshoot.
Finally, determine the Integral term that will help reduce any small remaining error.

[Nige]

It's not fine tuning at this point, but the fact it drops to 0% and stays there as soon as you switch to closed loop.

[David Ferguson]

But it's only doing that because the parameters for the PID are incorrect for your setup. I'm sure it can be fixed, by tuning using the procedure I describe above.