My ECU and PID

Hi,

I have an ESL ecu, I am having a hard time understanding the turbo dynamics or more so the actual boost control system of this ecu.

here is my original post on scoobynet:

https://www.scoobynet.com/engine-management-and-ecu-remapping-453/1054927-turbo-dynamics-esl.html

Can someone help me map these 3 or 4 options to the PID philosophy.

Any help would be much appreciated, thanks!

What is the engine, what is the vehicle, what problems are you having or what is your goal? I think the "burst" is proportional gain and the continuous is the Integral.

Ej20, subaru v4 uk2000, 4.11 fd sti drivetrain, exedy pink box clutch, exedy lightweight flywheel, td05-16g straight through 3" exhaust with bellmouth decat, 440cc injectors, Chinese brand FMIC, DV delete, stock sti headers and decat upipe heat wrapped with chinas finest lavawrap, orange MAF (good for 350bhp) connected to a samco silicone turbo inlet.

Im having issues with spool and high wg duty, in 3rd gear i should be hitting full boost by 3.3k-3.5k with my mods however im seeing (below target) 1.4 ish bar boost (i want to hit 1.6) at 3.900rpms. Im quite confident there is no preturbo leaks/boost leaks.

My aem xseries uego has ate its 2nd sensor so i still have to order a new one, but since ive had these mods on the car for so long i know im hitting 1:1 targets on my WOT pulls (maf-map blending set to 100% map on WOT)

I know im running rich lowdown and need to rescale my MAF for the KN cone filter

Stock wastegate can hold 1.5 ish bar in the midrange.

My car idles really smooth however if you look at the rev counter it shakes by +/-75 rpm but it fluctuates, holds, then fluctuates for longer than holds for longer.

The only thing i can hear is valve train noise at low load low revs, between 900-1100 and a slight misfire which can be cured by dropping the timing.

Only thing i can think of is mapping, getting a stronger wastegate (potentially masking the problem).

Should be noted i have 2-4mm of preload on the wg.

i'm running really aggressive duties:

AT WOT

RPM, MIN WG, MAX WG

3700, 75% 75%

All my mapping is done in 3rd and 4th.

My Boost Targets

0 - 4300 = 1.5 bar

4300 - 5000 = 1.4

5500 - 6000 = 1.3

6000 - 7000 = 1.1

my (realistic) boost targets per gear are

gear 1: 0.8 bar at 4.5k, 1 bar at redline 7k

gear 2 : 1.2 bar

gear 3 : 1.5 bar in the lower midrange, holding out 1.3 bar till the upper midrange (6.5k), then 1.1-1.2 bar after 6.5k

gear 4 : same as above with a little overshoot

gear 5 same as above with more overshoot.

My aim is:

To bring on boost atleast 3.5k in 3rd, with 1 bar coming in around 3200rpms.

People have been getting 1bar at 2900-3000rpm and full boost at 3300rpms.

Thats with ported headers and upipe, so im going for realistic target of 1.0 bar at 3300 rpms and full boost 3500rpm ish.

I believe the key to improving my map further is to gain a full understanding of PID in relation to my ECU parameters.

should be noted that my ecu can also be run in speed density, which i dont particulary mind, but since its my daily driver, i want that fuel economy when im cruising however i would even sacrifice that bit of economy to do a mafless map, i will practice with a speed density map once my MAF is working correctly!

I've been looking at this video to gain a better understanding:

https://www.youtube.com/watch?annotation_id=annotation_891845&feature=iv&src_vid=UR0hOmjaHp0&v=XfAt6hNV8XM

therefore, based on this video correlating with andres explanation in PID control. I agree with you Richmond.

go WOT, a change of huge proportion whilst cruising = proportional gain

which then brings about the next question

You are cruising, you wanna go WOT, the closed loop system seems the massive proportional change so the ECU applies a function to the value.

Whilst throttle is 100%

IF current duty is insufficient to make target boost then apply {BURST GAIN}

SET BURST UNDERBOOST = (V^o + V^1 + v^2 + v^3)

is SUM OF (BURST + current WG duty) > (max WG duty)

if yes then remove v^o from BURST (giving burst the total compounded value of (V^1 + V^2 + V^3)

If no then add (BURST + CURRENT DUTY) and RUN this value through the {BURST GAIN} function and produce a new target.

IF current duty produces more than target boost reduce the WG duty using the BURST OVERBOOST table and lower untill current wg duty = minimum wg duty .

The piece of 'logic' is how my head thinks it works

The v^0, v^1, v^2, V^3 values represent the values in the burst overboost tables so 11.8, 3.1. 2.0, 0.8

If the system can't mechanically achieve it, PID tuning means nothing. Turn all the gains to 0 and run boost open loop. Set your duty cycle map to 100 percent up to 4300rpm, floor it in 5th gear at 2000-2500rpm to load the engine down and see if it gets there. Make sure overboost fuel cut is on and set to maybe 1.6 bar. Watch knock, be careful of leaning out, make sure its not heatsoaked. Don't be stupid about this.

The pid tuning is used for duty cycle calculation , it doesn't make the engine spool faster if duty cycle isn't the problem.

I get that raymond, but for example, with a stronger wastegate, i can achieve 1.5 bar in gear 2, however i overboost on the higher gears.

So, to compromise, i have to set my duty lower to compensate for that overboosting. Which im sure (to a certain extent) can be mapped out.

For example everything turned off, min and max wg set to the same.

I can spike to 1.55 bar in 4th at around 3800rpms. However once i apply the turbo dyanamics table and the min wg duties, i will no longer spike to 1.55 but more so to 1.35-1.4

I now have a stronger wastegate which u have fine tuned in closed loop to spike

1bar in 1st

1.25-3 in 2nd

1.35-4 in 3rd

1.50 in 4th

Havent checked 5th but i dropped my min wg duty by 5% to compensate for that

Ive also dropped some timing at the full boost rpm of 3.8-4k to 16 degrees

Can you concisely state your question?

I would like to fully understand the meanings of the 4 values placed under each turbo dynamic table:

boost error

burst underboost gain,

burst overboost gain,

continuous overboost gain,

continuous underboost gain.

there are 4 values in each of the above tables.

id like to know:

Are these values applied at a certain speed? certain gear?. are all 4 values applied in succession?

Are these values most likely a wastegate percentage value or are they for example, a PSI value?

What do each of the individual tables actually do; how does each table correlate to the PID control system(I want to gain understanding)?

Boost error = target boost - actual boost; so boost error can be positive or negative? so do you increase this value even if you undershoot your target?

Cobb and RomRaider basically explain how the boost control system works. Your reflashing vendor uses slightly different names for the tables. It also depends on the specifics of what model year ECU you have.

http://www.romraider.com/forum/viewtopic.php?f=40&t=1814&sid=f9869797c40983d8b40c15fd08e4d142

So in a closed loop system using PID control ,or in this case P + I control, you have to tune the feedback gain. In a very simple system you have one gain value for the entire controller (one P gain, one I gain to cover all situations). In Subaru's case they allow you to adjust the gain based on the error. This means that, if you are slightly underboosting you can have a different gain from when you have big underboost. The same applies to overboosting. It's a way to allow more stable control.

Thanks for that, so just to confirm I have understood it right, whenever the criteria for the activation of burst/continuous boost is met, the whole table is applied in an attempt to smooth out the boost correct?

So on a gear 3 or 4 pull from 2k rpm, I would be using the continuous tables correct?

as the ecu is targeting 1.5 bar at 2k (impossible) would the continuous under boost table be applied?

as I reach, for example, 4.5k and my target boost is 1.5 bar, however in 4th I overboost to 1.6 bar. would the values from the continuous overboost table be applied to bring down the duty to try and control the boost?

assuming I've understood the above correctly, how would I utilize the boost error table? because in 1-3 I have no overboost and in 4th I get overboost to 1.55+ and in 5th, I will hit boost cut with a spike of 1.7bar ?

assuming my current boost error table:

BOOST ERROR = 0.20

BOOST ERROR = 0.12

BOOST ERROR = 0.05

BOOST ERROR = 0.02

would I then increase the boost error table to 1.7-1.5 = 0.2

old error + calculated new error = new error

0.20 + 0.20 = 0.40

so then, would I make my table

0.40

0.20

0.10

0.05

So to summaries the question, what should be the best way to control the boost:

tune the actual Turbo Dynamics tables?

update the boost error table ?

lower min wastegate duty ?

lower max wastegate duty ?

The first thing you need to do is fix your target table. Don't request a boost that is impossible to achieve. That's closed loop tuning 101.

how can I do that with a boost table without any gear or speed compensation?

I can achieve 1.7 bar+ in 4th at 3.5kish

I can achieve 1.5bar in 3rd at 3.9k ish

I can achieve 1.35bar in 2nd at 4.2kish

yes, I can pull my targets and my duties down to gain more control in higher gears, but that also kills my response in the lower gears.

So advise me, what shall I do, how can I gain the best response, which values should I tweak after setting the target boost levels to your recommendations.

You can't get what you want without a trade off because you have the old logic for 16 bit ECUs.

You can try the hybrid mbc / EBC method.

just an update.

Ive managed to get

1.1bar in 1st

1.4bar in 2nd

1.5bar in 3rd

1.6bar in 4th

i dont bother with 5th.

i currently have a 1.5bar mamba wastegate, i have some new fitting coming for it so should be able to take 1mm of preload off it, if not ill throw on the 1.1 bar spring.

MAX Figures

100% wg duty to 3.4k

80% duty to 3.7k

60% duty till redline

MIN Figures

100% to 3.4k

60% to 3.7k

flat 30% to Redline.

in 3rd and higher i taper out to 1.25ish bar at redline which is fine by me.

Its a shame i cant really understand the turbo dynamics table for the sake of knowledge and fine tuning the perfect control.

I will prob put the restrictor pill that i was told to remove back in the 3 port and get my boost more consistent and saferish for the upcoming cold weather