RomRaider

Sorry, was just about to reply that I missed the part about "even if extensive changes have been made. "

I have a 1mm pill I want to install on my stage2 05 Legacy GT and I have a few tuning noob questions.

WGDC
How do you determine the difference or range from WGDC initial and Max? I read that it is typically ~8%, but all of the maps that I’ve seen are way more than that especially in the WOT column.

TD
I read in COBB’s Subie boost control doc that it’s a good idea to have TD slightly high (100 or less for DBW). Looking at the logs posted here topic3870-150.html it appears that 15 was the target value. Am I looking at the correct data or does COBB use a different range? Reason I ask is because 15-100 seems like a large difference.

How do you know if you need to adjust TD Proportional and or integral and how do you determine if your values are correct?

I understand under what conditions TD proportional and integral are active and why they are applied. I also understand that the corrections applied are absolute, but have a cap. What I don’t fully understand how the values are applied. I couple things I could use some clarification on is what is what exactly is meant by accumulate in this statement "The TD Integral correction is different in that it can accumulate over a certain range." and how does TD integral work to correct boost errors over time? If it’s not too much to ask could someone possibly go through a brief scenario using the following tables? Yes I am a very visual learner. A flow chart would be awesome, but I know thats asking a bit much.

Thanks

I'm doing this off the top of my head, so it could use a sanity check... But here's PI feedback in a nutshell:

a) BoostError = DesiredBoost - ActualBoost
b) AccumulatedBoostError = AccumulatedBoostError + BoostError
c) ProportionalFeedback = TDProportionalTable (BoostError)
d) IntegralFeedback = TDIntegralTable (AccumulatedBoostError)
e) Wgdc = InitialWgdc + ProportionalFeedback + IntegralFeedback
f) go back to (a)

Assume for the sake of discussion that the car is in a steady state making 10 pounds of boost with 12 psi target boost, using just the initial WGDC table. Now we add feedback to the picture... Proportional feedback, by itself, will bring you close to target, but not quite to it. The closer you get to the target, the lower the proportional feedback gets.

Notice that in step B we add the current error to the previous value of the AccumulatedBoostError. So, if you have a small error for a long time, the AccumulatedBoostError value grows large, and this pulls the WGDC up, which pulls the actual boost up to the target.

If the proportional or integral tables are too large, you get oscillations because the feedback values will cause boost to overshoot the target.

PI feedback is used for lots of stuff... Cruise control probably works the same way. AF Correction definitely works the same way - the "AF Learning" values are just the AccumulatedAfrError value from the fueling control loop. There might be a PI loop controlling throttle plate angle in DBW cars, too.

EDIT: Fixed the calculation of Boost Error as per Merchgod's reply.

NSFW -> boost error = target boost - actual boost

You cannot always compare Cobb's table values to RomRaiders. It seems Cobb has a habit of using unusual or inconsistent conversions (or raw values) that are not going to be the same as RomRaider in all cases.

What is 'actual boost' ? is this like MRP Direct?

Im trying to use the factory boost control routine on an 02 and targeting 24 psi (on the target boost table). I realize the factory sensor is capping at 22 when logging MRP Corrected, so Im wondering if I swap the map sensor to a 3 bar if this will work out. Is there anyway to target higher then ~25 psi on the target boost table?

Thanks!

Actual boost would be MAP direct, since the ECU deals in absolute pressures as it relates to boost control (your absolute target boost would be table value + 14.7 psi). You cannot target higher than 25 psi with the 16-bit ECU, however, after you rescale for your aftermarket MAP sensor, you can reduce the MAP offset and gain headroom that way (for example, reducing it by 3 psi, would mean logged 25 psig would be the equivalent of 28 psig (same with your target boost), so you can tune with that in mind.

Ok that makes sense. If the target boost table was in absolute pressure, that would be more obvious.

Thanks for the response!

If my max wgdc on the table is 64.1 (330 vs 3200rpm) and I logged primary wgdc at 64.3 with td proportional at 10%, does this mean only .2% was added by td proportional to wgdc? Td proportional +3 error and above is set at 10.

So td prop logged shows maximum correction but what is applied is reflected on logged primary wgdc?

If one is still using the stock turbo, regardless of mods, is it really necessary to make changes to the TD tables? Would it be better just to play with target boost and WGDC tables only?

I'm a little bit interested in the above question as I'm trying to adjust my Cobb Stage1 tune for my GR and I notice they don't adjust the stock tables much, but the IAT WGDC compensation is really flattened out compared to the stock map. I'm thinking this, as well as an unachievable boost target at low rpm, is causing my TD integral windup during spool and overshooting of boost targets.

Since boost comes on slower in the lower gears, how do most people tune integral/proportional so that integral doesn't windup and boost overshoot in the lower gears but still have good boost response in the higher gears?

A few techniques I've played around with:

Set target boost that is actually attainable in low rpm, while hopefully not sacrificing what you can achieve in lower gears (like 3rd on a 5 speed).
This minimizes boost error and hence integral wind-up.

Set your TD proportional positive at max boost error to a large value. This will help prevent the integral from winding up. Set it up so that it it takes up all or most of the gap between WGDC and Max WGDC for large boost error when you stomp on it. If there is no more WGDC to dole out because the proportional is taking it all up to the maximum, the integral won't increase until the proportional starts to reduce itself as you get closer to target boost. The trick is tuning the proportional not to drop out so quick or you start to get boost fluctuations.
I've only done this because I'm using a 3-port and w/o any restrictors. Might not be necessary on a stock BCS setup.

how can I log turbo dynamics, they are not defined.

You need to change your user level in romraider, its probably still set to beginner...

I am using definitions file v308 and have romraider settings set to Highest and don't see the turbo proportional or integral to measure. The only wastegate variable I see is Primary Wastegate Duty Cycle. How do I record the PI terms to know total wastegate value?

EDIT - Worked once I connected to the car. The parameters showed up.