how-to: ECU-based direct boost control
#152
Are those changes for 1.30+ or 1.31+? The reason I ask is I've been using EcuFlash 1.30.1095 with the old (endian="little") scalings for some time. I may get around to upgrading EcuFlash but since it ain't broke I've never bothered.
#153
boost control
Are those changes for 1.30+ or 1.31+? The reason I ask is I've been using EcuFlash 1.30.1095 with the old (endian="little") scalings for some time. I may get around to upgrading EcuFlash but since it ain't broke I've never bothered.
#155
No problem. I should be thanking you though, you're making this way too easy. As long as I'm correcting you I'll add that where you mention adding to the evo9base.xml for the 94170008 ROM when I'm pretty sure you mean the evo7base.xml file.
Last edited by juyanith; Dec 19, 2007 at 11:22 AM.
#157
That's exactly my setup: Direct boost mod, stock solenoid with a single #65 pill near the turbo (no other pills). I'm still doing some tweaking since I'm adjusting my tune for the colder weather but the boost control works very well.
#159
Ok, maybe this has been asked before, but I didn't see it... The MWGDC tables in my stock ROM (88590015) are at 100% in all fields. Is that normal? I'm pulling them down to about 45% to start tuning with direct boost control, which I think will be safe.
Another question... does the boost limit table still work? My guess is no, but I didn't see it specifically mentioned.
Figured I'd upload a log while I'm on here. Stock boost control is UGLY. I've been playing with MIVEC, but other than that, a JDM MAP sensor, 2-byte load mod, and a drop-in filter, the car is bone stock.
The log is in excel. Rename to .xls to see it.
Another question... does the boost limit table still work? My guess is no, but I didn't see it specifically mentioned.
Figured I'd upload a log while I'm on here. Stock boost control is UGLY. I've been playing with MIVEC, but other than that, a JDM MAP sensor, 2-byte load mod, and a drop-in filter, the car is bone stock.
The log is in excel. Rename to .xls to see it.
Last edited by GTLocke13; Dec 27, 2007 at 08:07 PM.
#160
Ok, maybe this has been asked before, but I didn't see it... The MWGDC tables in my stock ROM (88590015) are at 100% in all fields. Is that normal? I'm pulling them down to about 45% to start tuning with direct boost control, which I think will be safe.
Another question... does the boost limit table still work? My guess is no, but I didn't see it specifically mentioned.
Another question... does the boost limit table still work? My guess is no, but I didn't see it specifically mentioned.
Do you have an upgraded pill or different solenoid? With the 100% stock system you won't be doing much going to ecu-based direct boost control.
#161
I want to get a feel for it before I install the GM 3-port that's sitting on my shelf. That and get rid of the boost wiggles you can see in that log. I'm a controls engineer. Oscillating systems annoy me. They usually mean that the people who designed them *cough* *cough* didn't do their homework.
Last edited by GTLocke13; Dec 27, 2007 at 08:53 PM.
#162
your attached file gives me an error.
Log WGDC and you will probably see the stock system pulling WGDC and running less than 100%. The stock system basically uses the BDEL and BCLO to regulate boost since its running at 100%.
If you raise the BCLO a little and message the BDEL a bit it will get rid of the up and down of the stock boost system since it is most likely hitting error correction.
Since I can't see your log, if you are talking about the tight zig zaggy look of the logged boost curve with the JDM sensor, that is normal as there is no smoothing involved.
Does it look like this?
Log WGDC and you will probably see the stock system pulling WGDC and running less than 100%. The stock system basically uses the BDEL and BCLO to regulate boost since its running at 100%.
If you raise the BCLO a little and message the BDEL a bit it will get rid of the up and down of the stock boost system since it is most likely hitting error correction.
Since I can't see your log, if you are talking about the tight zig zaggy look of the logged boost curve with the JDM sensor, that is normal as there is no smoothing involved.
Does it look like this?
#164
Did you change the filename extension to .xls?
Its actually oscillating, over and above any noise. Its controlling boost by hitting the boost cut. You can see the oscillations much better in 2byteload than in the MAP sensor, but it basically looks like an underdamped system. Lots of overshoot and then it oscillates its way to the target. If you manage to get the log open you can see the WGDC changing once per second and causing all those oscillations.
I'm slowly figuring out that BDEL = Baseline Boost, BCLO = Boost Adder, TBEC = Boost Error Correction, just with different scalings. Other than that it just switched pressure for load in the WGDC output calculation, right? Not to say it was easy... simple concept, complex implementation.
The other thing I like about this mod is that you use absolute pressure as the feedback signal instead of a calculated load that no one can really give a good definition for (that I've come across at least). And I think I discussed this in the 3-port thread, but manifold absolute pressure, not gauge pressure, is what matters to the motor. I know we all think in psig, but psia is really the correct feedback to use for boost control.
One last question. Why with the stock control does my WGDC only get up to 78%? Especially as boost is ramping up, the MWGDC is asking for 100%, the load is well under the BDEL+BCLO and its nowhere near the boost cut.
Its actually oscillating, over and above any noise. Its controlling boost by hitting the boost cut. You can see the oscillations much better in 2byteload than in the MAP sensor, but it basically looks like an underdamped system. Lots of overshoot and then it oscillates its way to the target. If you manage to get the log open you can see the WGDC changing once per second and causing all those oscillations.
I'm slowly figuring out that BDEL = Baseline Boost, BCLO = Boost Adder, TBEC = Boost Error Correction, just with different scalings. Other than that it just switched pressure for load in the WGDC output calculation, right? Not to say it was easy... simple concept, complex implementation.
The other thing I like about this mod is that you use absolute pressure as the feedback signal instead of a calculated load that no one can really give a good definition for (that I've come across at least). And I think I discussed this in the 3-port thread, but manifold absolute pressure, not gauge pressure, is what matters to the motor. I know we all think in psig, but psia is really the correct feedback to use for boost control.
One last question. Why with the stock control does my WGDC only get up to 78%? Especially as boost is ramping up, the MWGDC is asking for 100%, the load is well under the BDEL+BCLO and its nowhere near the boost cut.
Last edited by GTLocke13; Dec 27, 2007 at 10:20 PM.
#165
Its actually oscillating, over and above any noise. Its controlling boost by hitting the boost cut. You can see the oscillations much better in 2byteload than in the MAP sensor, but it basically looks like an underdamped system. Lots of overshoot and then it oscillates its way to the target. If you manage to get the log open you can see the WGDC changing once per second and causing all those oscillations.