Cal switch WG duty adder

[TLicense]

Hmmmm, strange thing here....

I've just been taking a look at quantifying my wastegate duty vs MAP setup and have noticed something a bit strange....

I've set the base wastegate duty table to 50% everywhere. I've then turned off the closed loop feedback by setting the wastegate integrator gain to 0% across the range and setting the manifold pressure error demand table to 0% everywhere.

I've then set the wastegate control duty adder as follows:-

Pos 1 = 0%
Pos 2 = +10%
Pos 3 = -10%
Pos 4 = +20%
Pos 5 = -20%


During an outing, I've then done a run at pos 1 at max load, then backed off and switched to pos 2 and gone to max load, then backed off and gone to pos 3 and gone to max load, etc etc. I was expecting to see that as I incremented through each position I would see differing max MAPs. I could then use the information to correctly setup the wastegate duty table.

However, the final duty for all five runs is 50%. It appears that the wastegate control duty adder isn't doing anything.

I've attached a screen grabs of the calibration showing the wastegate control duty adder settings and the logged data. If anyone could explain what's going on here, I'd be really grateful.

Thanks,

Tony

[pavlo]

The answer is in your own screenshot, assuming of course you're not running the old LMP strategy for boost control. Without a cal file to look it it's difficult to see what you're doing.

Really you don't need to alter the WG setting, seeing as you have a duty table that is target specific.

What are you trying to achieve in the end with your boost control? Are you trying to measure boost vs WG duty for incremented duty levels so you can use the into to plug into your final WG base table?

[TLicense]

My final goal is to try to improve the closed loop control. I'm seeing that the control is very slow to home in on the target. It undershoots and then the integrator slowly pulls it up.

What are you trying to achieve in the end with your boost control? Are you trying to measure boost vs WG duty for incremented duty levels so you can use the into to plug into your final WG base table?

Exactly!

I looked at the base wastegate duty table and saw that it was basically set to crack up the duty to 80% (which relates to the max MAP target I'm looking for). As you mention, I then reset the table to hold the solenoid closed until the MAP approached the MAPtarget.
However what I found was that the MAP undershot the target by about 100kPa.
Looking into why, it appears that the wgFinalDuty was held at 0%, so I figured it could be that the polarity of the wastegate is incorrect and 0% is fully open.

I thought an easy test would be to turn off the closed loop control and then add to the duty by +/- 10% then 20%, but then obviously saw that it wasn't adding anything to the duty.

[pavlo]

Well the first thing you need to check is that you phase duty table is how you want it. I would guess that it currently is backwards which is used if you plumb the valve differently.

Your base WG duty table should be higher where the boost is below target and lower where the boost is above target. Although if your WG is pretty good your try of starting it with a slope in the Y axis is a okay start. BUt to get the control better you will want to adjust the shape of the table. The default PD table can produce odd results if you have fluctuations in MAP.

On the S8 only, the PD table (Manifold Pressure error demand) is backwards, I don't know why!

To do what you want, you should put WG values in different boost targets, and use the cal switch to change the target boost. If you use boost targets that match the rows in the WG base duty table you will get where you want to be.

To "correct" the operation of the valve, go to Standard Strategy\Output Linerisations\PWM Output Linearisation and set it to 0=0% and 100=100%

[TLicense]

Ah cheers!

As I suspected the base wastegate duty table was the wrong way around. 0% = Fully open, 100% = Fully closed.

The phase duty table looks right (increasing proportionally with final duty) so I think that's all OK.

I'm not quite seeing what you mean by the PD table being the wrong way around. It looks right to me, the same as the S6...

I've also gotten to the bottom of the reason why the cal switch duty adder appeared not to be working. The wastegate control duty adder only affects the OLD LMP strategy base duty default.

[pavlo]

You don't get me.

The default base WG duty table is not the wrong way around. YOUR table is the wrong way around.

If you're phase duty is set correctly, you should get increasing boost with increasing WG duty. With your current setup the strategy will not properly, you need to double check the operation of the valve, the plumbing of the wastegate OR set the phase duty to be inversely proportional to solenoid duty.

When tuning the boost you will also need to log your base, final, integral and PD duty.

I did mention the OLD LMP strategy in my first post....

[TLicense]

Sorry, I have got it sorted, just not explained myself particularly clearly. I'll try to be a bit clearer...

Yes, the base wastegate duty table is (or rather was) the wrong way around. I've corrected it now so that 0% = wastegate fully open and 100% = wastegate fully closed, so it's as you describe. i.e. I get increasing boost as my WGduty increases.
I've proven this by doing as you suggested with regards to setting different MAP targets and logging the duty required to achieve those. When I'm below the target I've now set the base duty to 100%, and then ramped down to the duty required. When it's over target the duty is set less to help bring it down. I can now tune the integral to be a little bit more aggresive to get it to home in a bit quicker, but obviously the closer I get the base duty table to being accurate the less work the integrator will have to do.

I hadn't fully twigged your comment about the OLD LMP strategy meant it only worked with that. All OK now though.

Many thanks for your help!

[pavlo]

You need to work on the "Manifold Pressure Error Demand" table, not the Integral table. An integral correction IS slow (relatively), because it is time based by its very nature, and in this case it is linked to the Manifold Pressure Error Demand output.

It is this "Manifold Pressure Error Demand" table that is backwards. It consists of boost error and mapDT (rate of change of manifold pressure). A simple way to start tuning this is to add duty when the boost is below target and subtract duty when boost is over target. You can do this by making the changes the same for all values of mapDT, and this will add a simple proportional element to your control.

The same table also allows a derivative element to the correction by looking at what manifold pressure (MAP) is doing at the time. If the MAP is rising fast you may want to reduce the WGduty in order to prevent an overshoot, however if you're a long way off the target you may infact want to add even more WGduty. HOwever once your boost is close to target be aware that minor fluctuations in MAP may result in fluctuations in the PD element of the WG duty, so you will often need to flatten your table a fair bit at low MAP error values.

Having checked, the input variable into the Integral table is boost error percentage.

Really I need to put together some proper learning material, responding to odd posts on the forum is not a very efficient way at giving out detailed information.

Sorry, I have got it sorted, just not explained myself particularly clearly. I'll try to be a bit clearer...

Yes, the base wastegate duty table is (or rather was) the wrong way around. I've corrected it now so that 0% = wastegate fully open and 100% = wastegate fully closed, so it's as you describe. i.e. I get increasing boost as my WGduty increases.
I've proven this by doing as you suggested with regards to setting different MAP targets and logging the duty required to achieve those. When I'm below the target I've now set the base duty to 100%, and then ramped down to the duty required. When it's over target the duty is set less to help bring it down. I can now tune the integral to be a little bit more aggresive to get it to home in a bit quicker, but obviously the closer I get the base duty table to being accurate the less work the integrator will have to do.

I hadn't fully twigged your comment about the OLD LMP strategy meant it only worked with that. All OK now though.

Many thanks for your help!