q) about "accurately road tuning" the base duty cycle table for your boost control

[cbDrift]

goal : to get a boost curve that has no spikes or dips and tracks target boost accurately.
can it be done well, using "road tuning" - or do you need a dyno to "get it right"?

eg:
If you datalog the PID's and the "wgate DC (%)" and all the boost correction tables and all their axis's, so you have data to work with, and then you do some test driving and check your datalogs - to try and "road tune" your wgate dc tables....

Then if your datalog after a drive around says, say, at 4500rpm that your logged "wgate dc (%)" value is 20% to reach your boost target (stably without any fluctuations at that boost)
...and the correction table "boost iat wastegate trim" table is "-6%" due to the current IAT (no other trims applied)

...should the "wastegate dc%" tables be changed to be 20 at that point, which after road testing, results in boost reaching target pressures better than before?

or - should you set it to 20 + 6% to take into account that a correction is being applied to trim off 6% to result in a logged output of the "wgate dc%" of 20% that achieves your boost target?

What if adding the extra 6% to the table value causes boost spikes and fluctuations instead of a stable rise in boost and hold at the targeted boost level?
Could it be the base dc table needs changing?
...or the correction table applying its trim?
...or would you want to look at changing the pid's again to try and control things at the new wgate dc table values?

If I "get it working" ok at the moment with IAT's in the 10c-20c range due to lovely christchurch winters (a bonus for turbo cars 😁) - what will boost do in summer time if I don't touch anything else? Would I need to re-adjust the tables again or will they "just work" still?

Will iterative tuning and tweaking let me get the nice smooth unwobbly boost curve and control that I want - or am I fighting a losing battle and should give it to someone with a dyno to fine tune things?

Does LinkECU have a dyno down here in Chch? 😁

[Adamw]

You would typically first tune the base DC in open loop mode, or in closed loop mode with the PID zeroed out, to get a boost curve somewhere close to the desired target.  To keep it simple, start with a simple single row table with minimal cells - probably just 4 RPM breakpoints at say 2000,4000,6000 & 8000 would be enough for most.  Put say 20% right across, do a pull then look at logs, adjust iteratively from there - bump it up a bit at a time until boost is flat for the whole RPM range.  Keep the test pulls as similar as poss - same gear, WOT from same RPM each time etc.  Once the base DC is calibrated, then you can start to add some closed loop action and tune that under stable conditions.  You then will likely need to add gear and IAT trims later as you start to operate and test over different conditions.  

[cbDrift]

Cheers - makes perfect sense

I'm almost starting to think I should have asked for the dyno guys to setup the boost control perfectly for me instead of saying I'd have a fiddle around with it myself if they do the fuel and ignition for me (almost)
Still - learning is good - and that's an awesomely simple way to get it sorted out - thanks heaps Adam

...  yeah - I know the Link help file says to set it up that way as well - but "it was so close, so just tweaking things should be fiiiiine right?"
(insert its all fine meme picture here)