Stock Boost Control - Closed loop PWM or just Boost On / Off - Page 2 - 3000GT/Stealth International Message Center

SHiVeR-91VR4 · 07-09-2004, 12:17 PM

There are some MBC's out there that use two regulators. One to open up after a certain set psi so the boost creep problem does not happen, and then another one to set the maximum boost. There is also a check valve installed in the setup I saw to allow the pressure to escape out of the wastegate when boost is cut.

EBC's should be able to beat the piss out of MBC's, but my gut tells me that the mfgr's of the EBC's don't really put all this control in there. On my sbc-id they put in a boost limiter so if you spike it automatically pulls boost for a second to a decent level. There is also a way to set the boost by psi and the "fuzzy logic" will do all the limiting etc you need. But most people say the fuzzy logic doesn't work to well, and that the manual Gain/Ratio settings work the best. There has also been a secrect setting that people have found on the boost controller to set how fast boost builds.

The sbc-id is an incredible controller, it's just that most people (including me) have no clue on how to best set it. I'm sure no one pulls 20 dyno runs on varying Gain/Ratio settings to determine the best curves either. Most people set it so it hits and holds a certain boost. But this may cause overboosting etc.

The other problem is the $500 price tag for the dual solenoid design (solenoids are big $$).

cbatters · 07-09-2004, 01:27 PM

I do not (yet) understand the engineering benefit of dual solenoids unless they are put in series for redundancy in a fail safe configuration. Only three variables I can see for the solenoid. Open time, close time and flow. (And a good control system would compensate for actuator response time by using an adaptive PWM.)

In terms of product reviews, I have yet to see any comparisons with hard data from an engine control/boost simulator or MAP logs from a test car. All of the reviews like the one below are very subjective.

http://www.sportcompactcarweb.com/te..._boostcontrol/

Excerpt:

For weeks we've had the EVC IV running on our externally wastegated test car with no problems. Installing it on the Eclipse was straightforward; it went in just like the EVC EZ. We could get through the learning procedure just fine. Once a boost level was dialed in however, the EVC just seemed to sit back and watch the show. Boost would build quickly, and just keep on going right past our set point. Peak/Hold recorded only a fraction of the pressure run, which was probably the reason the alarm didn't activate. Luckily, we had good fuel in the car, and it was able to survive the 25 psi of boost. We tried reprogramming, swapping out the harness, even using the actuator from the EVC EZ. All the settings were checked by a number of people. Replacing it on the other car proved the unit works fine. It just didn't like our Eclipse.

Biggest surprise was how poor most of the auto-tuning worked. (Or maybe the manufacturers know that users want to twiddle with the controls to "get the maximum performance" out of their cars. Everyone knows that a human can tune things better than some stupid computer chip!")

Clint

cbatters · 07-09-2004, 02:08 PM

EBC only needs setpoint plus one of three PID control settings:

1. Over damped (never exceeds boost setting)

2. Critically damped (ideal control - fast response and rarely exceeds boost setting)

3. Under damped (where you are willing to tolerate overboost in exchange for faster response time. Underdamped control is typically expressed as factors of the critical damping control.)

NON LINEAR CONTROL
There are also non-linear control methods where fast control response is learned over time but any changes can result in considerable control overshoot.

HI boost alarm could quickly close the boost solenoid but there is a finite time for control pressure to build and for the wastegates to respond resulting in addtioonal boost overshoot. HI-HI boost threshold alarm could be used to initiate an immediate fuel / spark cutoff to protect the engine. (For those that have not felt fuel cutoff before - sufffice it to say that it is an unpleasant experience. Feels like you hit something with your car as you are thrown forward when the engine shifts from full power to full braking force in < 100 mS. Not fun but it may save an engine.)

Still looking for hard data where someone has characterized the boost control loop requiremetns for a particular turbo/wastegate combination. (Wastegate actuator could easily be the limiting factor in boost control but I have not seen much written about it.)

Clint

SHiVeR-91VR4 · 07-09-2004, 06:31 PM

Yes, but there are also ways of having self-learning filters (or transfer functions) created to follow a trend you would like. These are probably the best way to do it, and would take time to learn the setup of the car. But once learned it would be able to spool the fastest with the least amount of overboost.

Also.....cbatters. I have been trying to find info on flashing our ECU's. I saw a page where you can re-flash complete fuel and timing maps on the subaru ECU's. I asked a dealership a while ago and they didn't think you could, but the 98+ cars have a flash connection that we could possibly use to do the ecu shit.

But you will need the $10,000 MUT-III to do it, and then hack the code out.

iguana · 07-09-2004, 11:49 PM

i'd be surprised if our geriatric ECU technology can be flashed. It's a 68000 derivative processor and the idea of flash updates was not around in 91. maybe OBDII could but i doubt it. i'd have thought we'd have heard by now if it could be done.
questions

1 Is there a pcmcia(?) card and software for laptops that act like a scope? i.e. fast sampling and voltage protection.
2 has anyone cracked open various EBC's to see if they are really software driven or just feedback amp systems with variable gain adjustment.

cbatters · 07-10-2004, 05:42 AM

Quote:

Originally Posted by iguana

i'd be surprised if our geriatric ECU technology can be flashed. It's a 68000 derivative processor and the idea of flash updates was not around in 91. maybe OBDII could but i doubt it. i'd have thought we'd have heard by now if it could be done.
questions

1 Is there a pcmcia(?) card and software for laptops that act like a scope? i.e. fast sampling and voltage protection.
2 has anyone cracked open various EBC's to see if they are really software driven or just feedback amp systems with variable gain adjustment.

ECU CPU
Agreed. OTP and MASK parts were much more common in the early 90's - you were pushing the envelope back then to get 256 bytes of EEPROM built into a microcontroller. Low cost / small geometry FLASH technology for program storage did not come along until later.

Still willing to bet this is an 8 bit 68HCT micontroller part from Hitachi that licensed 68XX cores from Motorolla and Z80 core from Zilog (for their 64180 series of miocrocontrollers) I can't think of anything the ECU is doing that could not be done by an 8 bit part with ~16K of assembly code.

I'll have to look at the DSM thread again to see why some folks think it was a 68K processor.

EBC
Very sure they are uP based and not discrete analog control. (uP much cheaper / more reliable.)

Clint

Jeff Lucius · 07-10-2004, 06:34 AM

Quote:

Originally Posted by cbatters

All of the reviews like the one below are very subjective.

http://www.sportcompactcarweb.com/te..._boostcontrol/

Well I can't contribute much to this discussion technically. But I can say the Sport Compact Car article mentioned above appeared in the July 2002 issue and it was not subjective. Their online article is incomplete. The original article has lots of data. They used an Edelbrock QwikData Digitizer/Logger to record TPS, boost, and rpm at a 500 Hz sample rate. They tested each BC at 8, 12, and 16 psi settings. A DynoJet was used with the '97 Eclipse GS-T in 3rd gear (Frank Stage 3 turbo with a 3" exhaust). Two tests at the 3 boost levels were done: throttle floored at 2000 rpm and floored at 4500 rpm. I'm not sure why they didn't include the many graphs in the online article (unless I just missed them).

cbatters · 07-10-2004, 07:29 AM

Quote:

Originally Posted by Jeff Lucius

Well I can't contribute much to this discussion technically. But I can say the Sport Compact Car article mentioned above appeared in the July 2002 issue and it was not subjective. Their online article is incomplete. The original article has lots of data. They used an Edelbrock QwikData Digitizer/Logger to record TPS, boost, and rpm at a 500 Hz sample rate. They tested each BC at 8, 12, and 16 psi settings. A DynoJet was used with the '97 Eclipse GS-T in 3rd gear (Frank Stage 3 turbo with a 3" exhaust). Two tests at the 3 boost levels were done: throttle floored at 2000 rpm and floored at 4500 rpm. I'm not sure why they didn't include the many graphs in the online article (unless I just missed them).

No mention in the online article about test methodology nor hard data. Back issues are not available so I will ping the mag to see if they can scare up a link to the data/charts. (If this fails, I will be begging you for some scanner time / PDF's!)

Clint

SHiVeR-91VR4 · 07-10-2004, 12:52 PM

I didn't get a read of that article....what did they say about the comparison between the devices? Which was better?

I'm going to take apart my sbc-id when I switch cars, but I'm 100% sure they use uP inside of them. I'm pretty sure it's going to be a derivative of the 6800 as well since it was so widely used. The sbc-id though has two things that I can take apart. There is a main unit, then a display unit. You need both to function, so the uP could be in either portion of the device.

Jeff Lucius · 07-10-2004, 01:27 PM

I'll scan the article in make it available. It is afterall 2 years old and reprints are unavailable right?

Yeah I know technically a copyright infringement .... unless it is used for educational or academic purpose and no fee is charged. .....

We'll see if SCC gets mad at me.

This answer as to which is the best is not entirely clear. You'll have to read the complete article. Some BCs do spike bad and others control spike well.

07-09-2004, 12:17 PM	#11 (permalink)
SHiVeR-91VR4 Forum Member Join Date: Apr 2001 Location: Ohio Drives: 99 VR4 Trader Rating: (88)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off There are some MBC's out there that use two regulators. One to open up after a certain set psi so the boost creep problem does not happen, and then another one to set the maximum boost. There is also a check valve installed in the setup I saw to allow the pressure to escape out of the wastegate when boost is cut. EBC's should be able to beat the piss out of MBC's, but my gut tells me that the mfgr's of the EBC's don't really put all this control in there. On my sbc-id they put in a boost limiter so if you spike it automatically pulls boost for a second to a decent level. There is also a way to set the boost by psi and the "fuzzy logic" will do all the limiting etc you need. But most people say the fuzzy logic doesn't work to well, and that the manual Gain/Ratio settings work the best. There has also been a secrect setting that people have found on the boost controller to set how fast boost builds. The sbc-id is an incredible controller, it's just that most people (including me) have no clue on how to best set it. I'm sure no one pulls 20 dyno runs on varying Gain/Ratio settings to determine the best curves either. Most people set it so it hits and holds a certain boost. But this may cause overboosting etc. The other problem is the $500 price tag for the dual solenoid design (solenoids are big $$). __________________ Need a datalogger? http://automotive.miragecorp.com Few parts left! (stock forged rods & pistons, EGT sensor, Driver motor mount)

07-09-2004, 01:27 PM	#12 (permalink)
cbatters Member: Burned valve club Join Date: Jun 2002 Location: Stoughton, MA Drives: 91 Stealth RT/TT Trader Rating: (12)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off I do not (yet) understand the engineering benefit of dual solenoids unless they are put in series for redundancy in a fail safe configuration. Only three variables I can see for the solenoid. Open time, close time and flow. (And a good control system would compensate for actuator response time by using an adaptive PWM.) In terms of product reviews, I have yet to see any comparisons with hard data from an engine control/boost simulator or MAP logs from a test car. All of the reviews like the one below are very subjective. http://www.sportcompactcarweb.com/te..._boostcontrol/ Excerpt: For weeks we've had the EVC IV running on our externally wastegated test car with no problems. Installing it on the Eclipse was straightforward; it went in just like the EVC EZ. We could get through the learning procedure just fine. Once a boost level was dialed in however, the EVC just seemed to sit back and watch the show. Boost would build quickly, and just keep on going right past our set point. Peak/Hold recorded only a fraction of the pressure run, which was probably the reason the alarm didn't activate. Luckily, we had good fuel in the car, and it was able to survive the 25 psi of boost. We tried reprogramming, swapping out the harness, even using the actuator from the EVC EZ. All the settings were checked by a number of people. Replacing it on the other car proved the unit works fine. It just didn't like our Eclipse. Biggest surprise was how poor most of the auto-tuning worked. (Or maybe the manufacturers know that users want to twiddle with the controls to "get the maximum performance" out of their cars. Everyone knows that a human can tune things better than some stupid computer chip!") Clint __________________ 1991 Stealth RT/TT - Firestorm Red - 193K Miles / Original Owner Maint: Castrol 10W40 / Mitsu Filter / Rislone Replaced: 2X clutch / 3X brakes / 2X front rotors / 2X ABS HU / 1X NGK Plugs / 1X Ball joints / 2X Antenna mast / 1X Active exhaust cable / 1X Windshield / 1X Fuel FIlter / 1X Struts / 1X rear rotors / 1X Front Head / VSS / 1X rear wheel bearings / 1X Transfer case (warranty) / 1x Tie rod end (drivers side) Overhauled: Starter, IAC, BISS, CV Boots Upgrade: SilverStar Headlights / ebay "Xenon" Fog Bulbs (no more yellow!) A/F Gauge / MAP - Electronic Boost Logging / MMCd Data Logger Pending Projects: Steering Rack

07-09-2004, 02:08 PM	#13 (permalink)
cbatters Member: Burned valve club Join Date: Jun 2002 Location: Stoughton, MA Drives: 91 Stealth RT/TT Trader Rating: (12)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off EBC only needs setpoint plus one of three PID control settings: 1. Over damped (never exceeds boost setting) 2. Critically damped (ideal control - fast response and rarely exceeds boost setting) 3. Under damped (where you are willing to tolerate overboost in exchange for faster response time. Underdamped control is typically expressed as factors of the critical damping control.) NON LINEAR CONTROL There are also non-linear control methods where fast control response is learned over time but any changes can result in considerable control overshoot. HI boost alarm could quickly close the boost solenoid but there is a finite time for control pressure to build and for the wastegates to respond resulting in addtioonal boost overshoot. HI-HI boost threshold alarm could be used to initiate an immediate fuel / spark cutoff to protect the engine. (For those that have not felt fuel cutoff before - sufffice it to say that it is an unpleasant experience. Feels like you hit something with your car as you are thrown forward when the engine shifts from full power to full braking force in < 100 mS. Not fun but it may save an engine.) Still looking for hard data where someone has characterized the boost control loop requiremetns for a particular turbo/wastegate combination. (Wastegate actuator could easily be the limiting factor in boost control but I have not seen much written about it.) Clint

07-09-2004, 06:31 PM	#14 (permalink)
SHiVeR-91VR4 Forum Member Join Date: Apr 2001 Location: Ohio Drives: 99 VR4 Trader Rating: (88)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off Yes, but there are also ways of having self-learning filters (or transfer functions) created to follow a trend you would like. These are probably the best way to do it, and would take time to learn the setup of the car. But once learned it would be able to spool the fastest with the least amount of overboost. Also.....cbatters. I have been trying to find info on flashing our ECU's. I saw a page where you can re-flash complete fuel and timing maps on the subaru ECU's. I asked a dealership a while ago and they didn't think you could, but the 98+ cars have a flash connection that we could possibly use to do the ecu shit. But you will need the $10,000 MUT-III to do it, and then hack the code out.

07-09-2004, 11:49 PM	#15 (permalink)
iguana little Boss man Join Date: Jan 2000 Location: Longwood FL,Usa Drives: 93 VR4 Trader Rating: (8)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off i'd be surprised if our geriatric ECU technology can be flashed. It's a 68000 derivative processor and the idea of flash updates was not around in 91. maybe OBDII could but i doubt it. i'd have thought we'd have heard by now if it could be done. questions 1 Is there a pcmcia(?) card and software for laptops that act like a scope? i.e. fast sampling and voltage protection. 2 has anyone cracked open various EBC's to see if they are really software driven or just feedback amp systems with variable gain adjustment. __________________ Longwood Florida DR500 turbos, ProfecB Spec II, No cats, Modified Stillen DP, 3" pipe with dual N1 cans, Walbro FP, TurboXS BOV, Stillen pads, Rotors, DXD clutch, Fidanza flywheel. MAFT, bigass KNN filter, 550cc injectors. 3SX lower mounts. Tien S springs. Jays93VR4 rear upper arms. Aluminum intake pipes, 3SX light pulley. Still to install, Pre cat eliminators.

Sponsored Links
3000GT/Stealth International Message Center Advertisement

07-10-2004, 12:52 PM	#19 (permalink)
SHiVeR-91VR4 Forum Member Join Date: Apr 2001 Location: Ohio Drives: 99 VR4 Trader Rating: (88)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off I didn't get a read of that article....what did they say about the comparison between the devices? Which was better? I'm going to take apart my sbc-id when I switch cars, but I'm 100% sure they use uP inside of them. I'm pretty sure it's going to be a derivative of the 6800 as well since it was so widely used. The sbc-id though has two things that I can take apart. There is a main unit, then a display unit. You need both to function, so the uP could be in either portion of the device.

07-10-2004, 01:27 PM	#20 (permalink)
Jeff Lucius Senior Fanatic Join Date: Jan 2000 Location: Colorado, USA Drives: 1992 Stealth R/T TT Trader Rating: (0)	Re: Stock Boost Control - Closed loop PWM or just Boost On / Off I'll scan the article in make it available. It is afterall 2 years old and reprints are unavailable right? Yeah I know technically a copyright infringement .... unless it is used for educational or academic purpose and no fee is charged. ..... We'll see if SCC gets mad at me. This answer as to which is the best is not entirely clear. You'll have to read the complete article. Some BCs do spike bad and others control spike well.