Base/typical values for closed loop boost PID

[M1tch]

Hi all,

Currently road tuning my car to build up a good open loop wastegate duty cycle base table, next phase will be to then use this as the base table and drop it into closed loop boost control.

Plan will be to have boost by gear and also lower/reduce boost at lower throttle as well as when the engine is not up to full temperature.

Just looking at some of the base maps that have PID enabled as well as the help file I can see that the integral gain is usually set to 0.3 with the derivative gain set to 3.

Looking at a few of the base maps for other cars like the TT and Celica ST205 however the integral gain is set to 0.08 and the derivative gain is set to 0.5.

My question is, what should I start my closed loop PID to as a starting point? Will mainly be working on getting the open loop table sorted out first but would be great to get some basic settings sorted for closed loop.

Engine is currently fitted with the weakest wastegate spring during some shake down running (around 6-7psi) and planning to target something like 10-12psi as the 'higher' boost setting initially although the gearbox is rated to hold the torque of over 20psi so the current boost levels aren't going to really concern the engine if I 'overshoot' the boost to say 14psi although I would doubt it with such a weak spring.

[Adamw]

There are really no "typical settings", even basic stuff like the size of the hole through the boost source nipple or ignition timing has a big influence on the response of the turbo control system.  There are some examples in the help file which I have attached below but I even see big differences in very similar cars so I wouldnt take it as gospel.

I suggest following the guide in the help file here:  

G4+ ECU Tuning Functions > Boost Control > Closed Loop Boost Control > Closed Loop Boost Setup Guide

 

[M1tch]

Thanks for that Adam, will look to perhaps set the initial settings within those ranges and see what the results are, ideally with a well tweaked open loop base map the closed loop part shouldn't need to do much heavy lifting to get to the boost targets etc.

[M1tch]

I have now roughed out the open loop map and am sorting out the closed loop map - and it works most of the time, I have noticed that with stage 2-3 enabled that sometimes the stage 3 or indeed the 'D' part is sometimes too aggressive with the P and I values getting me there in terms of boost level requested.

I am tempted to just run stage 2 rather than stage 2-3, I am running 2 boost levels based on gear, 180kpa in 1st, 2nd, 5th and 6th with 220kpa target for 3rd and 4th (due to my gearbox setup).

I have noticed that on I might need to bump the base map by around 3% duty cycle on the solenoid to get a perfectly stable 180kpa, once I then change gear sometimes the value of D gets pulled over to the higher boost target and means it undershoots then back to the lower boost the D value adjusts the value away and it doesn't boost as high.

Example below is a lower boost pull, I can see that I need to adjust the duty cycle by around 3% on the base table but thats fine:

On a slightly higher boost level I can see that it boosts ok, minor overshoot but thats fine - think I am near the limit of the wastegate spring but again needs about 3% more duty cycle on the base table.

Further on and boosting again in 3rd to 220kpa the D is pulling out 3% overall (the wrong way), I can see the D value is pulling out 3%, 6%, 9% and even 12% in later pulls even at lower boost and its not even hitting target anymore with an undershoot of 40kpa in some of the higher boost gears.

I can only partly guess that it might be fighting the wastegate spring a bit, although it seems to be ok to boost up to 220kpa and has gone higher than that before - only at around 60% duty cycle at 220kpa.

Wastegate trim increase of 35% in 2nd and 3rd to get the higher boost levels

Is it worth me doing the following:

1. Check the logs and add in 2-3% duty cycle to the base table to mean the closed loop system doesn't need to work almost at all for the target boost levels

2. Change over to stage 2 only rather than stage 2-3

Settings for the closed loop are here:

[M1tch]

I have looked back again at some of the maps vs logging for various pulls, I think the issue is that the initial cells when coming onto boost are slightly too high meaning that the derivative gain is pulling duty out of the map to avoid an overshoot but when it actually gets to the cells where the turbo is spooling up they are basically correct.

I have found a fair few pulls where the derivative gain is at 0 when the turbo is on boost but I can see it pulling duty beforehand - I have adjusted the wastegate duty basemap during the spooling area of the map to match off the actual wastegate duty cycle being commanded by the closed loop system which should hopefully mean that the closed loop system doesn't panic as it sees the turbo build boost quickly.

I am seeing stage 1 spool, stage 2 as well as stage 3 operate back and forth so I think my PID gains are ok, it doesn't overshoot and boost comes on quite strongly but smoothly - I can see that the Integral %DC is only up to around 5% at the max so well within the integral clamp - have dropped the clamp to +/- 15% although it doesn't get near it which is good.

I did originally have some of the pre spool duty cycle cells higher than they are currently which might have caused this issue over the past week or so - my thought was to get the turbo to spool up quicker but I know stage 1 goes to 90% duty cycle to get the spool up anyway.

[Adamw]

Stage 2 uses P & D only, this stage is to control the approach to target.  Think of "D" as kind of like putting on the brakes as you approach target to prevent it overshooting. The faster you are approaching the target the more DC D will pull out.   Stage 3 disables D and uses P & I only, this stage is to remove any remaining error off target.    

So the main problem appears to be sometimes it is not reaching the stage 3 threshold.  That could be too much D or base DC too low.

The Base DC mode only changes what is used from the base DC table in stage 3.  I would generally leave in in stage2-3.

[M1tch]

Will keep it as stage 2-3, D might be too high although it is preventing an overshoot ok, have pulled back the DC on the spool part of the map and will see if that helps - sometimes its fine though and the DC for the part of the map on boost seems to be about right.

Should I increase the stage 3 on KPa value? I can see all 3 stages being used in the logs so everything is triggering.

[Adamw]

1 hour ago, M1tch said:

Should I increase the stage 3 on KPa value?

You may have to, but 15-20kpa is around where I have usually ended up.  

 

1 hour ago, M1tch said:

I can see all 3 stages being used in the logs so everything is triggering.

In that last log pic above it is stuck in stage 2, so it would never try to correct that last bit of error.  Not sure why it hasnt reached the threshold though.  You have to study logs to find what causes it to not reach the threshold sometimes - it may be gear, air temp, a certain cell in the DC table, a driver related influence such as how quickly the throttle is moved etc...  

[M1tch]

17 minutes ago, Adamw said:

You may have to, but 15-20kpa is around where I have usually ended up.  

 

In that last log pic above it is stuck in stage 2, so it would never try to correct that last bit of error.  Not sure why it hasnt reached the threshold though.  You have to study logs to find what causes it to not reach the threshold sometimes - it may be gear, air temp, a certain cell in the DC table, a driver related influence such as how quickly the throttle is moved etc...  

In that last log where its undershot the target is at 220kpa, kpa at that point is 198 so 2kpa under where stage 3 will trigger (20kpa under the target) - will perhaps bump this to 25kpa for stage 3.

Have adjusted some of the DC during spool up to avoid the D value needing to pull DC to avoid the overshoot - some of these 3rd gear pulls are from a lower rpm rather than shifting at higher RPMs so might be another issue - unable to really rev out much more in 3rd due to road logging and speed limits.

[M1tch]

I have increased the stage 3 to 30kpa to see if it can 'reset' itself a bit, still seem to have the issue that the D value is pulling out too much duty making the turbo quite lazy and missing its boost target and which is cyclic as it undershoots and then never turns off as its undershot.

I have 1/4 mile drag track time booked in for this weekend so I can hopefully get to the bottom of this - car sometimes runs at the boost level I have set but not all of the time, kinda want to only have the 'D' duty cycle below a certain point for the initial spool until it hits its boost target.

[M1tch]

Still battling with this, might revert back to the original settings - boost controller thought it would be funny to just randomly chuck in another 7.5% duty cycle to spike it into boost cut after undershooting again at the RPM where its at maximum torque whilst pulling out duty cycle before hand to make it not actually hit target.

[M1tch]

Dropped the duty cycle back slightly inline with how the system was commanding the turbo, its now just completely ignoring the target - its 40kpa off and not correcting anything?

[Adamw]

Really need the log and tune.  

[M1tch]

I have attached a log file from the drag strip this morning which will give you 2 sets of some full throttle pulls in both normal boost at 180kpa and higher boost 220kpa, I pulled out boost control in 1st near the end of the runs but the duty cycle is the same I believe.

You can see from the start of the week screenshot that at 58% duty cycle it was sitting at 180kpa with boost being controlled at 220kpa at 65% duty cycle.

 

[MagicMike]

Is it still generally accepted that CLBoost is not great and the 'passive open loop' is better?

[Adamw]

The wastegate is tapped out at the max 90% and the boost hasnt reached target.  You need to make some mechanical or plumbing adjustments so that your target boost is achievable with less than about 75% DC.  The Mac valves are generally fully open at about 75-80%, so giving it more DC after that makes no difference.  I would also generally have no zeros in my boost target table.  If 180Kpa is the lowest boost you are going to run then 180kpa is the smallest number I would have anywhere in the target table.  Example table from my car below.  

 

[M1tch]

I will check over all of the plumbing and check the preload on the wastegate as well to see whats happening - it was previously holding 220kpa at 65% duty cycle which is why I was wondering what had happened when it was maxxing out at less than that.

With regards to my boost targets, I am limiting boost levels by gear due to having slightly stronger 3rd and 4th gear, I have worked out that I think the max torque the other gears is at around 180kpa - this is why the base table is at 180kpa with a boost by gear modifier for 3rd and 4th. I also left some cells blank below say 50% throttle as I don't really need the turbo to spool up when on cruise or around town etc, have also seen that this is also the case so that it doesn't try and spool up mid corner on lighter throttle.

My plan for boost by gear is:

1st - wastegate spring pressure to allow for easier launches

2nd - 180kpa

3rd - 220kpa+

4th - 220kpa+

5th - 180kpa

6th - 180kpa

To your point though, I might look to set the whole base table to 180kpa, unsure what spring I currently have in the wastegate at the moment but might see if I can run something like a 9psi spring to then boost up to 12psi easier (but not overshoot which could damage the gearbox) but make it easier to get to higher boost levels where I can.

I think I have worked out that I can run around 220-240kpa maybe to 260kpa at higher RPMs in 3rd and 4th in terms of torque rating but currently I am just after 180kpa base bumping up to 220kpa for 3rd and 4th.

[M1tch]

Will check the preload once the engine cools down a bit, have changed over the BOV, checked all leaks for both exhaust and boost hoses, not allowed to attach the log file as its too large it seems but its still pulling duty even when its still far under target and not really increasing that fast.

[M1tch]

Slight update on this topic, have adjusted the preload on the turbo now, base pressure is set at around 9.5psi, its also hitting the lower boost target of 180kpa at around 25-30% duty cycle, which is great as its not needing to do a lot to get up to the 180kpa lower boost levels.

Issue comes with the slightly higher boost level when aiming for 220kpa, I have set the wastegate trim to increase by 30% and again at 50% and its not really getting there - might have to bump it much higher. I guess my issue is that because the preload is now at a higher level I need less wastegate duty to get to the lower target figure. The wastegate trim is based on percentage of the base table so an example:

Old table needing 40% to get to 180kpa, 30% trim increase would give me 52% 'base' duty

New table needing just 25% to get to 180kpa, 30% trim increase would give me 32.5% 'base' duty or 37.5% 'base' duty at 50% trim increase.

I guess I would need 100% trim (think that's the max) to get just 50% 'base' duty which might not be enough to run the boost to 220kpa - worth noting that the current base table is now dialled in to hit the 180kpa spot on so don't really want to increase the duty cycle on the base table anymore.

Is there a way I can run 2 boost duty tables and run it dependant on gear? Eg a 3d table or perhaps just have gear vs RPM with duty cycle?

Wonder if this would work:

• I have put the gear status vs RPM in the active boost wastegate table
• Zeroed out 1st gear so it runs spring pressure
• Gears 2, 4 and 5 have the duty cycles that work for 180kpa
• Bumped gears 3 and 4 by 50% (although might need to go higher but can now be adjusted independently)
• Base boost table at 180kpa
• Wastegate duty trims are zeroed out (as the table does the bump in duty cycle)
• Boost offset in 3rd and 4th are there to the desired amount above the base boost target table
• Boost control will activate at 80kpa (might drop this slightly) but will mean that nothing will activate during 'normal' driving around town.

[Adamw]

Yes, just put gear on 1 axis of your base DC table.  Or if you want more control variables you can switch between 3 different base tables, say have 1 for 1st & 2nd, change to table 2 for 3rd & 4th and table 3 for top gear.

[M1tch]

12 hours ago, Adamw said:

Yes, just put gear on 1 axis of your base DC table.  Or if you want more control variables you can switch between 3 different base tables, say have 1 for 1st & 2nd, change to table 2 for 3rd & 4th and table 3 for top gear.

Thanks Adam, have just run this table and it seems much better in terms of control - don't think I need any further granularity for these, mainly just to split out the different duty cycle curves needed for each gear. Stage 1 runs at 90% until stage 2 starts 50kpa below target so the turbo spools ok at the moment but might need some fettling. Basically the 'low boost' 180kpa setting it working well so far, gradually creeping up on the 220kpa mid boost levels - only bumped to around 200kpa on about 40-50% duty cycle so hopefully can make it to 220kpa with a bit more.

Needs around 23-32% to hold 180kpa, its needing double that to get to around 200kpa, spring pressure to 180kpa is around 3psi, 180kpa to 200kpa is another 3psi so guessing bumping up to 220kpa is another 3psi so needing an extra 20-30% on top maybe (thinking about 20% every 3 psi increase). As a guess I would need something in the region of 60-80% duty cycle to get to 220kpa so just about within the limits of the MAC valve.