Closed loop boost control - or lack thereof

[Thom]

Hi,

I have read the Help files upside down and I still cannot figure out how to fill the WG DC%1 Table.

As mentioned in the set up guide, I have set the Boost Target 1 Table as a 2D table with RPM on the X-Axis

I have a 170 kPa spring in my wastegate, so I filled the Boost Target 1 Table with values of 180 kPA.

As mentioned in the set up guide, values for PID are all zero'd.

By then I can pin point the required DC % in the WG DC%1 Table to reach the 180 kPa target. Let's say this required DC% is 20%.

Now this is where the problem begins. When I aim at reaching a peak boost of 200 kPa, even if I put values of 30% in the 200 kPa row, the tracer will stay on the 180 kPa row all through the rev range at WOT.

The tracer seems unable to "jump" from the 180 kPa row to the 200 kPa row in the WG DC%1 Table, absolutely regardless of the value I put in the Boost Target 2D table, be it 180 kPa or 200 kPa or whatever, and I will still reach a peak boost of 180 kPa. I tried more resolution in the WG DC%1 Table (adding a 190 kPA row), makes no difference.

-> What am I supposed to do?

Side question : why is it necessary to put values in the Boost Target Table when PID settings are zero'd since te Boost Target table is mentioned to be disabled when PID settings are zero'd? This set up is described as "Advanced Open Loop boost control" in the Help files.

I'd be feel grateful for help because I have wasted hours trying to figure out how this works, as what's mentioned in the Help files does not work in my case.

Thanks.

[Dave Kriedeman]

Firstly,

Does your boost control work if using OPEN LOOP Control.

Forget about CLOSED LOOP for a minute.

You need to establish that your OPEN LOOP works before stepping into Closed Loop.

Also, you mention you are using RPM on your X axis, what do you have on your Y axis.

[Thom]

1) Yes, Open Loop boost control works, if I put the exact same value everywhere in the WG DC%1 Table.

I have not tried DC% values varying from one row to the the next one though, but I suspect I would find myself in the same situation described above, i.e. the tracer won't be able to "jump" from one row to the next one.

Actually, if I also proceed by filling the whole WG DC%1 Table with the same value when running Closed Loop, I will reach whatever peak boost I reached when running Open Loop with the same WG DC%1 Table.

2) On the Boost Target 1 Table, I have no dimension on the Y-axis, as mentioned in the Help files. Just RPM on the X-Axis.

[Jason]

try following this old guide from Ray: http://users.tpg.com.au/adsl0uxv/Closed ... ontrol.pdf

[Thom]

Thanks Jason, it helped a bit.

Since then, I tried the 2 following methods for setting boost :

Method 1

-> Follow precisely the directions in the Help File for setting up closed loop boost control.

To identify the DC% values required to reach boost targets from 170 kPa to 220 kPa, I proceeded with full load accelerations in 3rd gear.

The DC% values I found for each boost target are :

180 kPa : 15%

190 kPa : 20%

200 kPa : 25%

210 kPa : 30%

220 kPa : 31%

Pros :

- filling precisely each target boost with its corresponding DC% makes TPS-based boost control usable.

Cons :

- if I choose a target boost of 220 kPa regardless of TPS, in 3rd gear the boost won't get past 190 kPa unless running in 4th and 5th gear. Since I identified the DC% values while accelerating in 3rd gear, I do not understand this contradiction.

- when boost gets higher than 190 kPa, which would therefore be only in 4th or 5th gear, spool under full load remains "slow" : I'm not seeing 200 kPa before 3800 rpm.

- the boost overshoots when engaging next gear at full load, regardless of PID settings.

Method 2

-> Fill all of the WG DC% Table with the value (31%) needed to reach a target boost of 220 kPa.

Pros :

- better spool (seeing 200 kPa at 3400 rpm), most certainly explained by the lower boost target lines being filled with higher DC% values than they "should" be.

Cons :

- cannot use TPS-based boost control

- the boost overshoots when engaging next gear at full load, regardless of PID settings.

Now, as a side note :

The WG is normally fully closed if the WG DC% is set at 100%.

So, in order to get good spool at lower boost levels (180, 190 kPa, etc) before reaching the target boost of 220 kPa, I should have a "resulting" WG DC% of 100% when using PID settings.

The help file says "With a base duty cycle of 50%, a proportional gain of 5.0 and a current error of 10kPa, the base duty will be trimmed to 55% = (50 + (5.0 * -10 * 0.1%)".

Firstly, this basic equation is wrong, as (50 + (5.0 * -10 * 0.1%)" makes 45%, not 55%.

Secondly, if I try to apply this equation to my case, without taking account of the "-" sign :

if during spool I'm at 180 kPa, DC% is 15%.

The error with the target of 220 kPa is therefore 40 kPa.

With a P of 10, the Maths say : (15 + (10 * 40 * 0.1%) = 55% <- That is a pretty long way from 100%, isn't it?

Given that it is not possible to use values higher than 10 for P (and D), how to keep the wastegate fully closed if it's not possible to reach a calculated WG DC% of 100% during spool up?

What am I getting wrong?

Set up is GTX3076R 0.82, Tial 38mm wastegate with 170 kPa spring (yellow) in dual port with an AEM solenoid.

Thanks.

[lekonna]

The reason why you are seeing more boost on higher gears is due to the fact that the engine is under more load with higher gears at the same rpm. you can compensate by making a gear compensation table for the waste-gate.

though i never use open loop for final tune, i just find the approximate dcs for different boosts with open loop and then switch to closed loop for final tune, on closed loop the

pre-spool stage keeps the waste-gate DC at max dc to help spool faster.

[Thom]

Thanks for your quick reply.

If I reduce "Stage 2 ON" from 50 to 10 kPa the "Stage 1 Duty" will be whatever I have set it at but it will result in not being able to use the DC% values needed for TPS-based boost control.

I am now considering using an additional WG DC% Table triggered by TPS, as the problem, the way I see it, is that the "rate of change" of TPS is not taken account by the closed loop algorithm. Let's say I put 100% values in this additional table and taper the high side more or less progressively down from 100% to 31%, I should get the best spool possible when I floor the throttle, while retaining the other WG DC% Table when I want TPS-based boost control...

[lekonna]

Thanks for your quick reply.

If I reduce "Stage 2 ON" from 50 to 10 kPa the "Stage 1 Duty" will be whatever I have set it at but it will result in not being able to use the DC% values needed for TPS-based boost control.

I am now considering using an additional WG DC% Table triggered by TPS, as the problem, the way I see it, is that the "rate of change" of TPS is not taken account by the closed loop algorithm. Let's say I put 100% values in this additional table and taper the high side more or less progressively down from 100% to 31%, I should get the best spool possible when I floor the throttle, while retaining the other WG DC% Table when I want TPS-based boost control...

I don't get it, how does the TPS based boost control not work for you ?

I would do it the following way:

Boost target: table with X axis TPS, y axis Gear and value is the boost target. (to get some traction control for lower gears)

wastegate DC: x axis, RPM, y axis Boost target, value WG DC%

Boost gear wastegate trim table: trim for wg values for each gear.

Then just use closed loop for the adjustment, the boost target would come from the TPS table and the boost PID setup along with the DC table should take care of handling the boost.

I can try this out later this week. So far i haven't bothered with the TPS based boost limits since the throttle angle limits the amount of air the engine gets anyways, thus limiting the power output (i might not be able to get the boost i'm requesting with a given TP angle).

[Thom]

Well, this is already how I'm using TPS-based boost control.

The problem, as I described earlier, is that PID settings do not help at all in my case. P in particular does nothing but kill spool compared when running a WG DC% value of 100% in the lower boost target lines.

The idea is therefore to run "Advance Open Loop" (PID zero'd) with multiple boost tables.

Maybe the equation in the help file is incorrect, but it seems impossible to get a resulting WG DC% value of 100% when using PID settings. I would appreciate some input from the staff on what the actual mathematical formula is, as the one mentioned in the help file does not make sense to me.

[Thom]

So I did some more maths and this is what I found.

For any given value of P below or equal to 5 I will never get the boost to climb through the different boost lines because the resulting WG DC% increases with the Target instead of decreasing. This should explain why I saw P "killing the spool", at least for values <= 5.

With P set at 10, as I said 3 posts above, I still can't get a resulting WG DC% higher than 55%

-> if I use a much softer wastegate spring, the base WG DC% values will be higher than now, so the P trim should be higher thus allow for resulting WG DC% values higher than 55%. Does that make sense?

[lekonna]

something seems to be way off on the basic setup here, if you want send the map to [email protected] and i'll take a look.

for the closed loop control, the PID control is not supposed to spool your turbo, its not activated in stage-1, it should be only used when you are close enough of the target to actually take it into control with the PID.

so in the stage-2 and stage-3. also as it is an integral loop, it means that it is not supposed to hit the correct target WG% with the first iteration, lets say that you have the setup that is in the first of the tables

with P=1 and target = 220, and actual 180, so it will make the wg DC 19% for the first iteration, then for the next round it will notice that its only getting 190kPa and will change the DC accordingly to ~23% and so forth

"For any given value of P below or equal to 5 I will never get the boost to climb through the different boost lines because the resulting WG DC% increases with the Target instead of decreasing." this line is something that i don't understand at all, if the WG DC increases, so should the _actual_ boost, _not_ the target boost! your TPS should set the target, not the WG DC!

[Thom]

I am not confusing actual boost with target boost.

The algorithm doesn't seem to care where I have the target boost set at (100% TPS = 220 kPa here) because the boost won't reach 220 kPa in lower gears when P is different from zero and when I have each boost line filled up with its corresponding DC% value.

[lekonna]

I am not confusing actual boost with target boost.

The algorithm doesn't seem to care where I have the target boost set at (100% TPS = 220 kPa here) because the boost won't reach 220 kPa in lower gears when P is different from zero and when I have each boost line filled up with its corresponding DC% value.

so after the pre-spool stage the PID does not cause the boost to creep up at all?

can you take a screencapture with a log viewer that shows the following variables: Boost status, WGate DC%, MAP, Boost Target, TP (MAIN)%, Engine Speed ?

it would be more easy to understand than the textual description

[Thom]

so after the pre-spool stage the PID does not cause the boost to creep up at all?

That's it.

My ECU logs do not show WG DC%, P, I and D. They are mentioned as "not logged" datas. How do I get them logged?

[Thom]

"For any given value of P below or equal to 5 I will never get the boost to climb through the different boost lines because the resulting WG DC% increases with the Target instead of decreasing." this line is something that i don't understand at all, if the WG DC increases, so should the _actual_ boost, _not_ the target boost! your TPS should set the target, not the WG DC!

What I mean is that I see little sense in having a DC% profile that decreases suddenly (as Stage 1 ends) then increases again during Stage 2.

In order to get the cleanest and best spool, the way I see it is that DC% should decrease more or less progessively not only when the PID algorithm kicks in at the beginning of Stage 2, but also all the way during Stage 2 until applying the DC% value required to reach for (peak) Target Boost.

On my set up, I therefore see more sense in using values of P higher or equal to 5, as I'm sure that the DC% decreasing sharply from say 90% (during Stage 1) to increase again during Stage 2 messes things up.

If you look at the first grid, with P = 1, let's say I have Stage 1 at 90 % or therabouts. As Stage 2 begins at 180 kPa, DC% suddenly falls fom 90% to 19% then increases progressively to 23%, 27% (etc) as it closes up to the peak boost - that feels wrong, doesn't it?

RH's Supra in the guide he made for setting up closed loop boost control reflects this :

[lekonna]

That's it.

My ECU logs do not show WG DC%, P, I and D. They are mentioned as "not logged" datas. How do I get them logged?

i log with the PC not with the ECU, you get more data that way.

[Thom]

Pleased to say it finally works, if I use high values for P - anything lower than 6 or 7 won't work.

Thanks for everyone's input.

[lekonna]

thats good to hear, does the boost stay on target, any overshoot or oscillation ?

[Thom]

I ran P/I/D at 10/0.20/3 but had slight oscillations around the target during Stage 3 so I reduced I from 0.20 to 0.10 and it's now a bit more stable.

I am still surprised that the behaviour is so poor with values for P lower than ~6, and as I said earlier I suppose I can put this on the account of the base DC% values being too low (wastegate spring "too stiff"), resulting in a P Gain trimming % not being high enough to make a difference.

Would you mind sharing on your own settings (PID, base DC% Table, wastegate spring stiffness) so that we could see if this theory works with your application?

[lekonna]

Sure thing,

my WG DC% for 1.5bar of boost is from 25% around spool up to 30% closer to redline.

The spring on my WG is around 0.8bar, so not too stiff.

P = 3,

I = 0.1

D = 3

With I = 0.2 had some oscillation in stage 3. With this setup i'm overshooting the target a tad bit on spool-up.

here is a log pull which also shows how the pids change during different stages:

[Thom]

Interesting, thanks for sharing.

Could you add WG DC% along with PID?

Here is a PC log. That was with P/I/D set at 10/0.20/3, before I reduced I from 0.20 to 0.10.

[lekonna]

now they are in the same graph, btw added IAT as well, it seems to me that your IC is not working quite as well maybe as it should, 50C+ IAT starts to be quite high..

[Thom]

Thanks.

Yes I'm not sure of the performance of my IC, but I have the cone filter in the engine bay and this log was done with an ambiant temp of 30°C.

Also, I have the IAT sensor in the intake manifold right after the throttle body.

Where do you have your filter, the IAT sensor, and what was the ambiant during your log?

[lekonna]

Thanks.

Yes I'm not sure of the performance of my IC, but I have the cone filter in the engine bay and this log was done with an ambiant temp of 30°C.

Also, I have the IAT sensor in the intake manifold right after the throttle body.

Where do you have your filter, the IAT sensor, and what was the ambiant during your log?

filter is in the engine bay, IAT is in the pressure pipe just before the throttle body. ambient was around 20C. (previously had IAT in the intake mani, but it heat soaked and didn't show even remotely accurate numbers)

with higher gears i can see some IAT increase, but not up to fourth gear, in fact the IAT goes down a bit around there because of increased airflow.

just a bulk 600x300x75 china IC on the front, though quite well in the airflow and ducted.

[Thom]

so after the pre-spool stage the PID does not cause the boost to creep up at all?

Does anyone from Vi-PEC have an idea about this behaviour? Scott or Simon please?

The only way I understand I have to use such high values for P is because my wastegate spring is too stiff, which causes factor P to NOT add enough % trimming to the actual WG duty cyle, because WG duty cycle values are just too low to begin with. Is this correct?

Isn't 31% a remarkably low percentage of WG duty to reach a peak MAP of 220 kPa?

Another point :

My target boost is 220 kPa, I use a 170 kPa WG spring, and I have "Stage 2 On" set at 50 kPA.

That means that Stage 1 ends when MAP reaches a value of 220-50=170 kPa, which is the rating of the WG spring.

-> When "Boost Target 1" - "Stage 2 On" ≤ WG spring stiffness, doesn't that mean that Stage 1 Duty has zero effect at all in how boost builds up?