[adam-]

So these are controlled by a PID inside the ECU. Basically it has a target and it's programmed to know what duty it'll need to hit that target. They're almost always designed to overshoot - it makes the car appear faster when you get a big surge of torque, but the boost should then follow target. Sometimes it undershoots, which is what yours is doing. You're getting basically no deviation once it settles into steady state (note that req follows act).

Your RPM slams off when duty does - so you've lifted off the throttle. Remember that VCDS is aids for logging so you're seeing about 5 samples a second, useless for fine tuning.

Depending on how many samples you have between flooring it and actual meeting req, I'd be inclined to say you've got a leak - you're not hitting req and DC is low, so the vanes are asking for all she can give. Depends on sample count though, it might just be inaccurate.

Looks fine IMO - could tweak the vanes in an attempt to smooth out the undershoot.

 

[adam-]

It's very easy Gerry.

 

[adam-]

You don't have a leak at high revs, or the turbo is able to mask it with more duty, it's fine. It's that initial punch where act doesn't follow req even though DC is high, but there isn't enough data points for me to be able to say for definite.

I'd be trying to smooth out the undershoot so I'd tweak one way and see if it improves, if the deviation gets worse, go the other way. PIDs will NEVER be perfect, there's too much variation between motors and they don't calibrate them to be that accurate, so always expect some over/undershoot.

 

[adam-]

It's worse than the first time. PID can't regulate well and you have a huge hole at the start of the run. Request is following actual well in the first run.

Not sure what you're expecting between 3500 and 4000...

 

[adam-]

ECU hard limit is 95% on 1.8ts - I'd have thought the diesels would have had a similar limit. It's to stop driving the solenoid at 100% duty which is hard on them. The 5% off time helps keep the coil cooler.

Anyway, 95% duty means the ECU is trying to open the vanes to reduce boost. Assuming the limp code is for overboost?

Adjust rod to bring it back inline. There's no real reason to deviate from stock.

 

[adam-]

You're aiming for the ECU to control it. Trying to fixate yourself on a target duty number is not how the ECU works.

Put the rod to the factory setting and log it. All you want is that the requested boost matches the actual. The first over/undershoot doesn't matter unless it's massive.

Sample rates for VCDS are shit so they can easily make the graph look sloppy when it's not that bad.

 

[adam-]

Worth noting that an actuator is just a diaphram with a spring. Spring tension ultimately dictates actuator response which has to be correct for the ECU. If you put non-OEM actuators on a turbo that have a different spring rate, the ECU will need a recalibration on the dyno which takes hours/days.

Best checking actuator is OEM then compare against a built turbo and starting there. From memory your first graph was fine.