12 out of 12 members liked this post.
ATR Boost Control
Ok I won’t mention reps… oops.
Shout out: Most of the this info I’ve posted before and in the beginning had the help of Carl Morris, Dfv2, Mangler, and some others knocking around ideas… and I think we helped some "certain people" in their understanding of reading logs and understanding 1 table in particular… they will never admit it though. Guess I’m fishing for negs with this comment, but it’s true.
Here we go…
We tune by setting a load, which is basically MAF at a certain engine speed. This is determined by MAP and temperature. Boost per load decreases as rpms increase… actually boost per load has a certain correlation in the bottom end, then increases significantly around 3500rpms(IIRC), and then drops rapidly soon after and finally will decrease more steadily with rpm… not very fun if you are a precision kinda person. Anyway, we don’t have direct control over boost… we control req load and some of the WGDC parameters used in hitting these targets. When I say “boost to load” I’m talking req boost to meet req load… but this is NOT stable either as actual load will stray farther from req load “when meeting boost targets”. Likely this is due to Cobb’s interface with the DME not being complete and/or various hidden limits.
How to read a log: WG base DC is before PID. Potentially WGDC (temp) is included into this value also. Sorry I reviewed this a long time ago and don’t remember my conclusion… but it doesn’t really have any effect in tuning. Mainly make sure WGDC after PID is adding DC and within its limits… too much, or maxed PID could indicate some issue with the hardware. WGDC Bank X is the final output to the solenoids and should be used if comparing to other people, tunes, potential boost issues, run to run, etc.
Base: a direct correlation to WGDC output in %. But we only have a narrow range to work with as many have experienced. Req MAF is the calculated desired flow through the engine… simply, dependent on rpm/PR. Factor is req MAP / baro. You can log both of these channels for reference and to point you in the right area of the base map.
Pre-control: basically off boost WGDC… potentially useful in reducing rattle. More closed, faster response. If part throttle is too responsive you could also try reducing it. Zeroing out the entire table will NOT effect DC during spool since this is only at lower loads, but it can effect response. The more closed increases psi available pre-throttle. MAP log channel is pre-throttle pressure. Boost log channel switches between pre/post throttle MAPs.
Spool: Tough to tell when the DME is using spool tables. You can see spool in the logs. Do a run starting at 3k rpms and take note of the load/MAP channels around 4-4.5k… then do a run starting at 4k and observe the same rpm range. You’ll notice boost to load values being different… IIRC req MAP will be lower at the same req load… it’s pretty easy to see, I just don’t remember the specifics. I assume this is spool to the DME. Doesn’t really effect our tuning at this time with the current ATR knowhow/access.
Adder (Airflow): adds global WGDC. Reduce for tight, increase for loose WGs.
Ceiling (Adder): this is the max WGDC that can be added by PID to the base. If maxed you need to increase it. Be careful though, depending on your load targets you could get a substantial increase in boost.
D-factor Multiplier: in my opinion its easier to adjust the multiplier table then the factor table. D is the anticipated future error… lower will be smoother, but will stray more from the setpoint. Higher will try and keep closer to setpoint.
NOTE: I’m NOT a PID expert and this info is based on testing one car (mine) and some reading… some info could be incorrect.
Dynamic: no idea, but a global adjustment of some sort.
I-factor: factor added in calculating MV based on historical error. Basically reduce value to reduce oscillations and increase value to keep closer to setpoint.
P-factor: largest contributor to MV… response to immediate error. Same effects as I-factor to oscillations but to a greater degree.
NOTE: since the DME can over react with overshoots, it could be beneficial to have a higher PID factor with positive boost errors.
**The last 2 tables I’m not sure… I would be more or less guessing. I asked Cobb in an earlier thread.
To control the boost based on IATs refer to limit table “load target limit factor (CAT)”. This a multiplier to req load based on IAT. Keep in mind that timing will increase when load is reduced so pay attention to your CAT compensation factor and multipliers.
A table set I’m very curious about is the “cylinder temp comp”. I believe these have to do with WGDC based on temperature. This could be a very easy way to reduce the overall WGDC if you can’t easily get there by the base table. But I’m only really guessing as I have not tested.
Now after reading all of the above you could be overwhelmed and much of it maybe due to my poor organization/writing… but I only have so much time to put this together. SO very simply do this:
Loose WGs: you don’t really have to do much as it should work fairly well with OTS mapping. To improve results increase the “WGDC Ceiling (Adder)” by an additional value of 5 to 10 in the last 2 or 3 cells. As long as you don’t max after PID channel you are basically OK. To further improve you can rescale your WGDC base column to max of around 370 to 400gs and maybe globally increase the base by 20% in the final 2 to 4 columns and rows.
Tight WGs: much tougher, but first action would be to reduce p-factor and i-factor… maybe start with global multiplier of 0.5 to 0.7. You can also reduce base by 20% at first in the cells mentioned above and then reevaluate.
NOTE: if you really get into part throttle control then you will have to adjust a larger qty of cells in the base table. Not really necessary with loose WGs.
Couple of points. I prefer to tune at a lower load range, to ensure that I can hit my targets consistently and smoothly then once I’m successful here I can “turn it up”. The reasoning is that with higher targets you can’t observe the DME’s full reaction, cause it will never hit the targets. Once you can hit targets as desired the more aggressive tuning will perform better… especially at part throttle. The ATR interface has us restricted below a certain max curve in load/boost and its not necessarily just a maximum; it’s scaled based on your setpoints. For example if I target 16psi in top I may hit 15, but if I target 17psi I may hit 15.5. Actually very irritating.
A comment on boost overshoots. This will happen. An example is once you have built some boost, or holding boost during a shift. A good test to see how severe is stepping the throttle… like 40, 60, 100%... it will really show what areas you need to fine tune. Likely you have to be ok with some overshoot at times. If Cobb can work out the ignition retard torque limits then the overshoots will be less troublesome. Throttle closure is definitely preferred in my opinion over reducing ignition.
By no means do I know this stuff 100%... I mostly use a stacked combination of Procede/ATR, so I may not have near the tuning/seat time with ATR as others do. So please explain where I could be misleading or incorrect. And definitely add anything I left out, expand on the info, question it, etc.
phew, that was long.