ESL - Turbo Dynamics - interesting find.
 

OK, so the text below is something I posted a couple of weeks ago on the "ESL Mapping Graduates" forum on Facebook (for those that have attended the course run by Andy and Spencer), but I've had no response on there yet, so I thought I'd see what you guys here thought - as much as anything, because it might be of relevance to you if you're trying to tune this area. Just to note, that since finding this out, I've managed to get my boost map tuned very nicely now, so that when I get overboost, it's very quickly pulled back down to where it should be. I had played with this area of the map configuration in the past, without any luck, and now I understand why.

----------------------------------------------------------------

I've been having a play around the the Turbo Dynamics for a bit now, and contrary to what I though about the terms "over boost" and "under boost" seeming to be self explanatory, I think that either I'm misunderstanding the terms, or the labels are the wrong way around in the ESL software!

In the examples below, I'm trying to tune my 5th gear, low RPM boost using the turbo dynamics to bring the overboost back down. I found that when overboosting, the adjustments that I'm expecting to see being be taken from the 'overboost burst/continuous' tables to apply as a correction factor to the wastegate duty, actually appears to be being taken from 'underboost burst/continuous' tables.

It took me a while to figure this out, as I was making corrections initially to the 'overboost' tables, but nothing seemed to be happening (1st screenshot). I then noticed that the values which were actually being applied matched closely to the 'underboost' tables, so tried swapping them and then started seeing the behaviour I was expecting being applied when overboosting (2nd screenshot).

Is it a case that I'm misunderstanding the interpretations of when overboost vs underboost tables are used (quite probably!) or is it that the labels are the wrong way around in the ESL software?

Screenshot 1:

https://cimg2.ibsrv.net/gimg/www.sco...aaffd76800.png

Screenshot 2:

https://cimg3.ibsrv.net/gimg/www.sco...7a3e9a4794.png

Hi, im just a beginner and haven't even got my ESL yet but looking at the above logic you would appear to be correct, it has adjusted when overboost was detected. I'm looking forward to getting my board and playing I just don't know were I can do all this testing. do you expect to excel? also is the log viewer not he right part of the ESL software.

Are you altering the duty cycle as well or just the turbo dynamics cells ?

Ive thought about altering the turbo dynamics but managed to get the RPM and duty cycle change bang on so i didnt get an overshoot, took some doing tho !

im impressed you are achieving 1.2 bar at 2600 rpm tho , i can only achieve 1 bar at 3000 rpm.

Yes. I've been working on my 95 RON map (aiming for around 1.1 bar boost) and I've got my duty cycles to a point where I'm pretty happy with them in both 3rd and 4th now, I was finding that I could get the TD04 to build too much boost in 5th early on, and if I reduced the duty it was then having a negative impact on 3rd and 4th too (as is to be expected). I could see the turbo dynamics working, but not quickly enough to pull it back down before hitting boost cut, so decided,that was the next area to try tweaking.

Ah a TD04, hence the quicker spool

I wa shaving the exact same issues a while back, i could hit 1.4 at 3200 rpm but it would overshoot and not react quick enough to the duty cycle change for the next rpm range so i , like you , found by alterinbg the suty cycle it then had a negative affect elsewhere

i accepted defeat and just lowered duty cycle so i hit a more stable 1.2 to 1.3 then increased to 1.4 but was annoyed at that .

here is a screen shot of my turbo dynamics, a bit different to yours ( im running on a v1 wrx )

https://cimg2.ibsrv.net/gimg/www.sco...71aa65ff83.jpg



i see you swapped the set of values over, but how did you come to the other set of values in the other 2 tables as they dont match what you originally had ?


Also, the difference column in your excel data log, i dont have that when i do a data log, is it something you have added in yourself ?

The original, 8S base-map settings for the Turbo Dynamics for the TD04 are all exactly the same for each one of the map segments. There's no difference between the Continuous Under Boost and Continuous Over Boost, or the Burst Under Boost and Burst Over Boost. They're all set to exactly the same set of values.

Ignore the fact that the values that I'd swapped aren't exactly the same between the two screenshots. I'd been playing around with different values. It's more the fact that if you examine the logs, you can see that it's the 'under-boost' values which are being applied to the wastegate duty when it's over-boosting, not the over 'over-boost' values.

From reading how the Turbo Dynamics works (detailed description for 16-bit ECU's on RomRaider), the burst tables only come into play momentarily at the switching point from negative to positive error (and vice-versa). For simplicity, I've kept my burst and continuous tables the same, so I'm not trying to work out exactly which one was applied. Plus, that's how it was on the 8S base-map anyway.

I've attached screenshots of the original 8S Turbo Dynamics, and what I've ended up with. What I've tried to achieve is to keep the minor boost error so the adjustment is gentle (which is what's used for everyday, gentle driving) and prevents boost oscillations, but to make it much more aggressive in the amount it can pull on large overboost. I've also tweaked the thresholds for the switching points too. I'm pretty happy with this setup on my car now.

Oh, and that extra column in Excel is one I added to help me see what was going on - it's simply a calculation of the difference between the cell and the one above it.

Original 8S Turbo Dynamics:
https://cimg6.ibsrv.net/gimg/www.sco...1758f90fde.png

My current Turbo Dynamics settings:
https://cimg5.ibsrv.net/gimg/www.sco...81b7aa56e8.png

does the documentation explain burst under and over boost?

Ok bud cheers for clearing that up.

As i say i stopped trying to adjust the turbo dynamics as i went the duty cycle route, here in this log you can see i hit target boost and maintain it throughout, it took a while and the rpm in the duty cycle / target tables took some time to get it right like this, some of the next cells are only a 100 rpm different etc but it made hell of a difference in not getting any overboost, if that makes sense.
i couldnt get my head around the turbo dynamics as the manual doesnt really explain anything, you are lucky you get the advice ( or not lol ) from the closed facebook group none of us original buyers of the ESL board get.

This run was in 5th from 1500 rpm WOT , manifold pressure the 6th column,this way i could find any duty cycles that were causing overboost as the engine load generated is very quick ( i do most my tuning in 5th as thats where i mainly do the booting, i dont really floor it in say 3rd or 4th as i broke my gearbox last time ha ha

https://cimg0.ibsrv.net/gimg/www.sco...9ebbdf3ba6.jpg

Interesting read thanks,I've never thought of touching the turbo dynamics,I'm getting overboost but I have a very different set up(twin scroll) I've fitted a 3 port bcs recently but not had much chance to rework the wgdc,I lowered across the board by 20% which stopped the overshoot but spool has suffered,I need more time to work on it but this info is great and something I'll now look at.
Ben are you still running the stock 2 port bcs?

What i forgot to say was, as a result of the not having the reaction time quick enough i only get 1.4 bar if i do a full WOT from low rpm and keep it held all the way, as the duty is too low for say 3rd or 4th gear pulls, its a fine balance of having the boost target achievable for where you most generally use it is the philosphy i went for ha.

i generally get around 1.35 to 1.38 normally (1.4 being target) due to this setup.

I dont know what cell on the turbo dynamics tables do what function, so with my setup below, what cells/figures would i alter to increase the speed of the reaction to overboost ?

https://cimg0.ibsrv.net/gimg/www.sco...8b30d504ac.jpg

OK - my understanding of the way Turbo Dynamics works, is as follows. For ease or understanding, ignore my comments above about the Overboost and Underboost names being the wrong way around. Assume they're the correct way around for this explanation as it'll make it easier to explain. Obviously keep this in mind though if you do make alterations, as you'll want to substitute one for the other, assuming my theory above is indeed correct.

-----------------------------------------------------------------------------------------------------------------------------------------

The Target Boost table is the main point of reference for how much boost the engine should be trying to make at any point in time. However, the actual amount of boost being produced will very rarely be the exact value that Target Boost is aiming for (Target - Actual). This means that there is always a Boost Error value, which will either be positive or negative depending on whether the actual boost being made is above or below the target boost. On the very rare occasion that it's exactly the same, the Boost Error will be 0. The Turbo Dynamics are based around the Boost Error to try and achieve the Target Boost. Unfortunately, ESL doesn't provide a way of logging the Boost Error (that I know of).

The values in the Continuous/Burst Overboost/Underboost tables are correction factors which are applied to the Duty Cycle in order try and achieve the Target Boost. The actual correction factor that is is applied depends on the current Boost Error value, and how it relates to the previous Boost Error value.

If the Boost Error has swapped from a positive value to a negative value (or vice-versa), then it is considered a "Burst" correction. Most of the time though, it'll be a "Continuous" correction that is being used, as there will be many 'cycles' of the ECU with it trying to attain the target boost, whilst the error remains positive or negative (not switching).

The Boost Error Threshold values determine which cell is chosen from the relevant table. The thresholds themselves are the Boost Error values (Target - Actual), therefore a positive Boost Error would represent underboost, and a negative Boost Error would represent overboost. Threshold 1 represents the left-most cell, whereas Threshold 4 represents the right-most cell in the Continuous/Burst Overboost/Underboost tables.

So - based on us knowing what target boost the ECU is trying to make, and how far from it the actual boost is, you can control how much of a correction factor is applied to the wastegate duty cycle in order to try and obtain the Target Boost. Obviously the Min/Max wastegate duty cycles come into play as well, as they are the absolute lower and upper limits that the ECU can permit, but the Target Boost and Turbo Dynamics control the way that the ECU is actually making adjustments to try to obtain those boost values.

-----------------------------------------------------------------------------------------------------------------------------------------

I hope the above makes some kind of sense. It's a bit of a "brain-dump" as I've written it down pretty much, just how it comes to mind!

Ben.

Ok cheers for that

So, we can say Burst corrections are for when you suddenly go from cruising boost level to say suddenly creating loads of engine load and booting it to the max boost level , whereas contnuous is for the "gentle" throttle useage/ gradual boost build up ?

im still unsure of what the values actual mean, percentage increases/decrease of the duty cycle ?

So if i overshoot by 0.20 ( threshold 1 ) it uses the far left cell and reduces the current duty by 11.8 % ?