When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
DNM you are a genius. I had pretty much instant success using your method. I had to push some numbers around a little. This is awesome!! Thank you!
I'm not trying to be a dick, just some constructive criticism. I think you could gain some accuracy by tuning the min duty% of your solenoid. I doubt your solenoid operates untill 5% so by doing that adjustment and rescaling your bias table, your solenoid won't be as lazy to operate. Also lower your max duty to avoid overshoot. If your max duty at full boost at 7k is 68, then 68 at low rpm full spool will most likely overshoot your target. So by lowering your max duty to say 70-75 ( compensate for air density) you might overshoot less without loosing spool.
Also if you are never getting a output% from 100-120ish kpa, you should try and rescale your boost control targets to reflect that. I have had a lot of success doing this.
Mudflap: You're Welcome. I try to help when I can.
Just a heads up: you are showing the wrong PID "I" (i.e. not for EBC). I suspect your "I" value was around 5, for most of that run. Looks pretty good. Hope you get it all that perfect.
It would be helpful for everyone to state what firmware they are on when posting, because 1.4.0+ are not all the same for boost control. Is anyone successfully using the method DNM explained in 1.5.1?
This is a 4th gear pull because 3rd gear tires spun like they were on bleach.
So I'm super psyched about this one. I'll address the overshoot tomorrow. But the question while we are here is, what is this oscillation I'm seeing? It is not particularly strong, I'm not seeing any bounce in the fuel pressure (which is referenced to the MAP, so I'd think it would bounce along with it).
I wish that this software would let me put markers in spots so you can see what is going on better.
This pull started at 2000rpm and hits peak boost at 4300rpm (28psi). Pardon my predicting (stretching):
measured on a dyno a month or so ago.
PSI, kPA, Torque at 4300, HP at 6000rpm
18 208 320 315
22 235 365 355
so (predicting):
26 280 (400+) (400+)
will fart around to see if I can get that hp program to work.
Keen eyed observers will note my fuel pressure dropping from 75psi to 70psi. This is a worry.
It would be helpful for everyone to state what firmware they are on when posting, because 1.4.0+ are not all the same for boost control. Is anyone successfully using the method DNM explained in 1.5.1?
The only substantive difference between my values in 1.4.0, and 1.5.X were:
1.4.X: I around 60 and Delta around 200 (non - existent)
1.5.X: I around 7 and Delta around 55 to limit the I wind-up by postponing the starting of the PID control.
All demonstrations that include plots or P, I, D as separate traces are on 1.5.X, as that is when that feature was available. So Mudflaps tune, and all of my later posts are 1.5.X. Mine happen to be 1.5.2 Beta 4.
I'm about to tune the EBC on the MS3 Basic per the info in this thread. DNM, you have been great in this thread. Thank you.
And Chiburban's video was really helpful.
I will start by dropping your settings into my 1.4.x setup (edit: going to install the 1.5.x) and will try to do the setup mode for the bias table. Hopefully I don't break anything.
Last edited by sixshooter; 11-29-2020 at 02:36 PM.
Flashed 1.5.1 firmware and warnings popped up. Tacho is pegged at max in TS window and O2 is pegged at minimum value though my gauge still shows 20.whatever. It says there's a conflict for boost controller 2 and CEL output also. It also doesn't recognize when I have turned off the key. The MS is on without the keyswitch now... Da fuq did I do?
Lol, no offense taken. I have been tuning a 1.4 car here for a few weeks, and definitely in the 1.5 camp! The 1.5 time based AE is better, I am very fond of the engine control function, SD logging, and have officially decided I like engine states better. It is still a personal preference thing though.
That aside,
Go through the list of HW conflicts and address each one in-turn. I suspect you have some conflict someplace that is back driving the R/W or keeping the main relay on. There are a few published necessary things that have to be done in the configuration when making this change. I think they are in the 1.5 readme.
I've just read this thread in its entirety and after 213 posts over 3 years, we have a lot of great information, but still haven't quite compiled a definitive guide on setting this up successfully.
I recently had my car tuned on a dyno @ 15 degrees C (~60f) and was achieving 20.5psi of boost. Outside in the cold weather (2c / 35f) I'm making more like 24-25psi using open-loop EBC. My goal is to achieve the same targetted boost level no matter what the weather, and lay the foundations for me to tune boost by gear.
It would be really great to complete a n00b's guide to closed-loop EBC, and it seems to me that we are quite close.
From what I have read, these are the steps towards successful closed-loop control on MS3:
Step 1. Figure out the minimum and maximum duty of your solenoid valve. The initial post suggested listening for action and setting the duty at points where you could hear the valve operating. Subsequent posts suggest a narrower operating range to help ensure the software can control the valve in a more efficient manner. It would seem like a sensible idea to use the closed-loop 'setup mode' to raise duty cycle to a point where you start to increase boost significantly and set that duty cycle value as the minimum Duty % in the boost control settings. How do we suggest finding out a sensible maximum DC without blowing stuff up? One idea I had was to flip the output polarity of the valve to 'inverted' and repeating the same test, but with a bias table starting at 100 and working downwards until we see boost start to rise. *** I have not tested this, so if you're going to try this, BE SURE OVERBOOST PROTECTION IS SET UP ***. Does that sound like a sensible way of working out the valve operating window?
Step 2. Set the tuning mode to 'Setup Mode' and set the X&Y axis of the Bias Duty table to sensible numbers, to give more resolution in useful areas. For example, re-binning low RPM columns where you're not making any meaningful boost. Populate the Bias duty table to achieve your desired boost level. You can do this by working out what Duty cycle is required to achieve the target boost throughout the cells. There was a good video of this
The advice seems to be to reduce the numbers in the boost control bias duty table across the board, so that it's underachieving the boost target. You want to do this to avoid boost overshooting your target level in cold weather conditions. PID can help pull the boost up, but in wintertime, will struggle to correct boost overshoot quickly enough.
Step 3. Once you have a populated bias duty table, you'll want to fill in the Boost Target table- simple enough.
Step 4. Set your 'Boost Control Lower Limit Delta' to a sensible number. Around 40 seems to be a reasonable number, but with an under-achieving bias table, maybe this can be lowered to increase turbo response without the high risk of boost overshoot?
Step 5. Now this is where things get a little foggy to me and I could do with some clarity. There are a few key areas up for discussion... The first is: should we be choosing Basic Mode or skipping straight to Advanced? There is a 'closed-loop sensitivity' slider. This seems to have an effect on both Basic & Advanced. Is that correct? From what I can tell, most people seem to nearly max this slider to have any meaningful effect.
What PID starting values should we be trying, and what is the best method for testing the system? Should we be stabbing the throttle in a boost 'zone' (for example 4500rpm) or should we be rolling into boost from 2000rpm? Should we be logging both? I think this is where a tried and tested, sensible dummies guide to dialing this in is missing. I'm being a bit of a keyboard tuner right now, because the weather outside right now is hardly conducive to 4th gear WOT runs (GTX2860 @ over 20psi).
It would also be a good idea to have some common issues and suggested fixes. For example:
Issue: when stabbing the throttle at high rpm, the boost overshoots initially.
Possible solution: lower values in the bias table in the problematic areas, or increase the Boost control Lower limit delta to a higher value
Issue: when rolling into boost, the boost target is achieved, but tapers off
Possible solution: increase value XYZ etc
I think we are very nearly there with a complete beginners guide and would be invaluable to newcomers like me to MS closed loop boost control.
Further to the last post, I had some time this evening to go and test this out. My first time playing with closed-loop and I've got to say - I'm really happy with where I've got today in BASIC mode.
I started by doing a pull at wastegate pressure (around 12.5psi).
Then I configured the boost control method to SETUP mode and populated the bias table a little at a time until I could hold around 210kpa (~16psi with the aim of using the boost target table to reach 225kpa (~18.2psi).
I scaled this table and filled it in, and using setup mode I was achieving around 210kpa across the rev range in a 4th gear pull. You can see from the logs I was battling a little wheelspin (spikes in VSS and RPM).
After I was achieving 210kpa, I set my target table to 225kpa and entered basic mode. I set the lower limit delta to 30kpa and moved the slider quite a lot to the right after seeing other people's experience with that.
This is the first session after entering basic mode. As you can see, there is a 10kpa overshoot initially to 235kpa, which quickly settles down and then the system seems to be doing its thing reasonably well, keeping MAP within 3kpa of the target (lowest was 221.8 at 6800rpm), highest after the initial spike was 226kpa.
I'll have a play about with the slider sensitivity and boost control lower limit delta, but from what I can tell so far, it won't be necessary to enter advanced mode at all.
Once the weather gets better I'll continue to experiment with the ultimate goal of around 22.5 psi (255kpa).
06-07-2019, 08:53 PM
0
