|
MazdaSpeed 3/6 - ECU Computer Tuning Have an ECU tuning issue? Do you want to discuss Mazdaspeed 3/6 tuning solutions? Up to date info on the latest! Piasini Reflash, Cobb Access Port, Xede, Procede, Standback, Superchips & others supported here! |
Welcome to Mazdaspeed ForumsMazda Forums / MazdaSpeed Forums - Mazdaspeed Forum with High Performance Discussions, Car Gallery, Mods Information, Street Racing Stats,Street Racing Videos & more. The Archived MazdaspeedForums! Mazdaspeed Mazda MSF MazdaSpeedForums.org. You are currently viewing our boards in read mode only. We are here to preserve the data and info that is useful to Mazda gear heads. If you'd like to donate to keep this info alive, your donation will go to hosting, SSL, and domain fees. | |
Search Forum |
| LinkBack | Thread Tools | Search this Thread | Display Modes |
|
10-31-2008, 04:30 PM | #1 |
Join Date: Mar 2008 Location: Earth
Posts: 1,590
(View Stats)
iTrader: (1)
Rep Power: 0 Thanks: 6
Thanked 97 Times in 70 Posts
Groans: 0
Groaned at 0 Times in 0 Posts
(Thread Starter) | Not Ranked : 0 score Here are 3 runs/log I did while driving home from work... Let the brain storming begin. Run #1 was done in 5th gear... I don't know why I did it in 5th, but I was curious to see what it would look like. Notes: Max Knock Retard: 1.8 (I don't know why DashHawk does not display the 5th Y-axis column; anyone know how to display it?) Direct Link: Run 1 5th Gear Run #2 done in 4th gear... Notes: Max Knock Retard: 1.1 Direct Link: Run 2 4th Gear Run #3 done in 4th gear... Notes: Max Knock Retard: 1.8 Direct Link: Run 3 4th Gear One observation I have done is that it makes my AFRs too long to reach 12s... So, what AFRs should we shoot for? Safe tune? Extreme tune? What about timing? How do I know where to set it?
__________________ 2008 Stage 3 Roush Mustang 302E 5.0L Stroker 647rwhp-620 rwtq Last edited by AutoXRacer; 11-04-2008 at 09:10 AM. |
11-03-2008, 06:35 PM | #2 |
Approved MSF Vendor Join Date: May 2008 Location: Beltsville, Maryland
Posts: 1,728
(View Stats)
iTrader: (0)
Rep Power: 2499 Thanks: 425
Thanked 4,675 Times in 997 Posts
Groans: 10
Groaned at 115 Times in 51 Posts
| Not Ranked : 0 score
Here's some information on PID boost control tuning. I wrote this a while ago and never got around to formatting it and putting it on the website, so I figure that I might as well post PID Tuning The cp-e™ Standback uses a true PID (Proportional, Integral, Derivative) feedback loop in order to regulate boost pressure. Many customers have requested some information on PID tuning, and this guide is meant to point the needy in the right direction. What is a feedback loop? The best example (that I can think of) of a feedback loop is a toaster oven. The toaster oven is meant to warm up food, but too hot and the food burns, too cold and it’ll never cook. In order to target a specific temperature, the machine must have some feedback as to what the actual temperature is. As such, a thermocouple is placed in the toaster oven, and when a specified value is reached the heating element turns on or off. The thermocouple provides the feedback the oven needs in order to determine when to turn the heating elements on or off. How about a PID feedback loop? Let’s continue with the toaster oven example. In the case of the toaster, we used what is basically a thermal switch to turn the heating elements either on or off. Of course, this doesn’t give very fine control of the temperature, but then again fine control really isn’t needed in a toaster. But if we did need to very finely control the temperature in the oven? PID control uses the history of some process in order to make an “educated guess” as to the controllers next action. If we were to plot a normal toaster oven’s duty over time, it may look something like this: Notice that as we clearly overshoot our critical temperature of 400°F. This is because the heating element doesn’t shut off until we reach our set point. PID feedback control looks at the control process and uses math to guess when to shut off the heating element in order to prevent overshoot and precisely maintain our target temperature. The control loop uses three functions in order to determine which action to take: Where: kpe(t) = Instantaneous error kie(t) = Error summed over time k de(t) = Slope of process Each of these factors provide some feedback to the controller, and we can use these inputs in order to better regulate our process. Let’s take a look: kpe(t) You can think of this term as your instantaneous error, which is defined as your setpoint minus the process value (this is technically not correct, but works for our intents and purposes). As the magnitude (size of the number; positive or negative) of the error grows, so does this terms weight on the action taken by the controller. This term is basically the gain of the system. kie(t) If you could “add up” the area between our setpoint and the actual process over time, then this would be our ki term. It is meant to direct the process in the right direction when it is consistently above or below the target over time. k de(t) This value quantifies the instantaneous “slope” of the process. The greater the magnitude of the slope, the greater affect this term has on the controller. This term can help dampen out oscillations, or create them depending on the PID tuning. And what’s the result of good PID control on our hypothetical toaster oven? Notice that the overshoot is almost gone, and the oscillation above and below our setpoint is much smaller. Surprisingly, the jump from a toaster oven to boost control isn’t a very big leap. Instead of using heating elements to control temperature, we’re using a solenoid to control the change in pressure at the wastegate. So although the two processes are very similar, the response time and tolerance for error with the boost control is much much lower! Below is a datalog taken from a customer that didn’t have his boost well optimized, but we can use it to critique: 1. The first thing to notice is the overshoot. This is a perfect example of the need for either more kd, or less kp. What’s happening here is that the controller is closing the wastegate solenoid (which opens the wastegate) too late and the turbo doesn’t spin down soon enough. The result is a pressure spike. As you can imagine, slowing the process down would help, but most people want their turbo spooled up as quickly as possible. So a reasonable suggestion here would be to add more weight to kd. Since the slope is very high as we reach the process setpoint, a higher kd value will help reduce the spiking. 2. Notice here that despite our being set to 15psi we never really settle there. That’s an indication that our ki value is set too low. If we were to increase ki, then the summed error over time would have more of an effect on the system, which would open the wastegate slightly in order to bring boost back down to the 15psi setpoint. 3. This last marker is just to indicate the role of the kp value. Kp can be thought of as the gain of the system, and you can increase or decrease the response time by adjusting this parameter. However, keep in mind that more is not necessarily better! A Comment on (In)stability Any PID tuner must be aware that these systems are far from foolproof! If the weights of each coefficient is too high then you can create a condition where the process becomes unstable. The result is usually an oscillation in the boost control, but if it gets bad enough then you risk damaging something, so be very careful when making changes to the PID coefficients. Make small changes to begin with until you get an idea of how each coefficient affects the process. | |
The Following 3 Users Say Thank You to www.cp-e.com For This Useful Post: |
11-03-2008, 07:25 PM | #3 |
Join Date: Aug 2008 Location: jersey
Posts: 2,069
(View Stats)
iTrader: (3)
Rep Power: 609 Thanks: 388
Thanked 895 Times in 360 Posts
Groans: 11
Groaned at 42 Times in 35 Posts
| Not Ranked : 0 score subd |
02-15-2011, 11:46 AM | #4 |
Join Date: Mar 2010 Location: Manassas, VA
Posts: 3,314
(View Stats)
iTrader: (2)
Rep Power: 2313 Thanks: 4,791
Thanked 4,320 Times in 1,452 Posts
Groans: 35
Groaned at 20 Times in 15 Posts
| Not Ranked : 0 score I'm going to quote something from the first page in this thread....this should deff help us.
Kp = 27.5 Ki = 0.265 Kd = 225 2-15-2011.GIF Matt, More times than not I get the StandBack disconnected error when logging. It seems to happen less often when running on battery instead of AC power (power converter).
__________________ CP-e Atmosphere GTX3076r w1.06 Tial | JMF IM | NATOR D.C. FMIC | JBR Wide Path | Autotech Internals | HKS SSQV2 | Ebay DP | Denso ITV24 | CP-e RDM | AutoExe Strut Tower Bar | Whiteline RSB | Sonic Tuning Coils | DSS Driver Side Axle | JMF OCC BT - 395 awhp 345 awtq @ Moore Automotive 4/12/2014 on 93 octane Stock Turbo - 331 awhp 397 awtq @ Moore Automotive 11/10/2012 on ~e48 12.263 @ 111.75 mph 1.669 60' 11/17/2012 Last edited by RichieRichness; 02-16-2011 at 09:21 AM. | |
02-16-2011, 09:18 AM | #5 |
Join Date: Jul 2008 Location: Canton, Ohio
Posts: 698
(View Stats)
iTrader: (0)
Rep Power: 0 Thanks: 1,233
Thanked 434 Times in 195 Posts
Groans: 5
Groaned at 2 Times in 2 Posts
| Not Ranked : 0 score Richie, i would go back to the PID values you had where the boost was smooth but just a hair under target. Then just increase Ki to .3 and see how it runs and go from there. You had an excellent setup with that boost curve.
__________________ cp-e CAI, cp-e Turbo Inlet, cp-e catted DP, cp-e HPFP, cp-e flashes, cp-e Standback w/PNP, cp-e FMIC, TRZ RMM, TS Compact BOV (full re-circ), DDM 55W 6k HID's, Cobb RSB. HKS Hi-Power Exhaust, OCC |
The Following User Says Thank You to the_caruch For This Useful Post: | RichieRichness (02-16-2011) |
02-16-2011, 11:33 AM | #6 |
Join Date: Mar 2010 Location: Manassas, VA
Posts: 3,314
(View Stats)
iTrader: (2)
Rep Power: 2313 Thanks: 4,791
Thanked 4,320 Times in 1,452 Posts
Groans: 35
Groaned at 20 Times in 15 Posts
| Not Ranked : 0 score
P = 30 I = 0.3 D = 150 2-16-2011_3rd_gear.GIF
__________________ CP-e Atmosphere GTX3076r w1.06 Tial | JMF IM | NATOR D.C. FMIC | JBR Wide Path | Autotech Internals | HKS SSQV2 | Ebay DP | Denso ITV24 | CP-e RDM | AutoExe Strut Tower Bar | Whiteline RSB | Sonic Tuning Coils | DSS Driver Side Axle | JMF OCC BT - 395 awhp 345 awtq @ Moore Automotive 4/12/2014 on 93 octane Stock Turbo - 331 awhp 397 awtq @ Moore Automotive 11/10/2012 on ~e48 12.263 @ 111.75 mph 1.669 60' 11/17/2012 | |
The Following 2 Users Say Thank You to RichieRichness For This Useful Post: | MattJackson86 (02-16-2011), the_caruch (02-16-2011) |
02-16-2011, 02:48 PM | #7 |
Join Date: Jul 2008 Location: Canton, Ohio
Posts: 698
(View Stats)
iTrader: (0)
Rep Power: 0 Thanks: 1,233
Thanked 434 Times in 195 Posts
Groans: 5
Groaned at 2 Times in 2 Posts
| Not Ranked : 0 score
I would still raise Ki a little more. try .4 or .5 Yea not having the cp-e flash would make it hard to sustain that psi past 5500rpm where all sorts of power limiting shit comes into play. I assume that drop in psi is right at 5500. I dug up some older SB logs i had and my WGDC skyrockets when i go past 5500rpm. trying to make the boost level i targeted but just couldn't do it.
__________________ cp-e CAI, cp-e Turbo Inlet, cp-e catted DP, cp-e HPFP, cp-e flashes, cp-e Standback w/PNP, cp-e FMIC, TRZ RMM, TS Compact BOV (full re-circ), DDM 55W 6k HID's, Cobb RSB. HKS Hi-Power Exhaust, OCC | |
03-04-2011, 10:04 AM | #8 |
Join Date: Dec 2008 Location: Los Angeles, CA
Posts: 1,407
(View Stats)
iTrader: (6)
Rep Power: 0 Thanks: 541
Thanked 681 Times in 387 Posts
Groans: 3
Groaned at 15 Times in 10 Posts
| Not Ranked : 0 score So I did end up trying these for a few days, but my boost spiked really badly. I was spiking to 24psi, where as on my PID settings (29, .075, 250) I spike to maybe 20. Your graph looks way smoother than mine did for sure, but I couldnt take that spike. I tried a P of 29 and also an I of .35. No major change. One I install a bunch of mods in the coming weeks I will be coming back to my tune and finishing the PID stuff. I did already change my timing and have that under control. The timing should look a lot better now as I only pull 1.5 degrees around 5200RPM and not the 5 degrees like before. Car definitely feels stronger.
__________________ GTX3071 - SURE 3.25" Full3 Aeros - CP-E FMIC - CP-E Catted Downpipe - CNT Catback Exhaust - HKS SSQV BPV - James Barone 88 Trilogy - KMD CDFP Internals - CP-E Injector Seals - Cobb EBCS - ITV22 at .025 - 3 Bar MAP Bitwise Motorsports http://www.bitwise-motorsports.com - Mazdaspeed 3 BAR MAP Harness, Torque Plugins for Android, and 3-Port EBCS (coming soon!) |
The Following User Says Thank You to MattJackson86 For This Useful Post: | RichieRichness (03-04-2011) |
03-09-2011, 02:33 PM | #9 |
Join Date: Jul 2008 Location: Canton, Ohio
Posts: 698
(View Stats)
iTrader: (0)
Rep Power: 0 Thanks: 1,233
Thanked 434 Times in 195 Posts
Groans: 5
Groaned at 2 Times in 2 Posts
| Not Ranked : 0 score
__________________ cp-e CAI, cp-e Turbo Inlet, cp-e catted DP, cp-e HPFP, cp-e flashes, cp-e Standback w/PNP, cp-e FMIC, TRZ RMM, TS Compact BOV (full re-circ), DDM 55W 6k HID's, Cobb RSB. HKS Hi-Power Exhaust, OCC | |
02-16-2011, 09:51 AM | #10 |
Join Date: Dec 2008 Location: Los Angeles, CA
Posts: 1,407
(View Stats)
iTrader: (6)
Rep Power: 0 Thanks: 541
Thanked 681 Times in 387 Posts
Groans: 3
Groaned at 15 Times in 10 Posts
| Not Ranked : 0 score
"Standback Tuning For Dummies Guide" The Official Q&A...
__________________ GTX3071 - SURE 3.25" Full3 Aeros - CP-E FMIC - CP-E Catted Downpipe - CNT Catback Exhaust - HKS SSQV BPV - James Barone 88 Trilogy - KMD CDFP Internals - CP-E Injector Seals - Cobb EBCS - ITV22 at .025 - 3 Bar MAP Bitwise Motorsports http://www.bitwise-motorsports.com - Mazdaspeed 3 BAR MAP Harness, Torque Plugins for Android, and 3-Port EBCS (coming soon!) | |
The Following User Says Thank You to MattJackson86 For This Useful Post: | RichieRichness (02-16-2011) |
11-04-2008, 11:27 AM | #11 |
Approved MSF Vendor Join Date: May 2008 Location: Beltsville, Maryland
Posts: 1,728
(View Stats)
iTrader: (0)
Rep Power: 2499 Thanks: 425
Thanked 4,675 Times in 997 Posts
Groans: 10
Groaned at 115 Times in 51 Posts
| Not Ranked : 0 score
First, yes, it does take the car a long time to transition to a rich(er) mixture when you first start generating boost, and this is part of the reason why I pull timing in that region. As you'll notice the car already pulls a lot timing there, but if we're upping boost beyond factory levels then it isn't a bad idea to pull more timing out. I usually yank between one and two degrees from that area if boost is above ~15psi. How do you correct for this lean spot? Well, we're working on a flash that changes the ECU's target mixture at lower RPM's. Mazda likely kept the mixture lean at least initially because it isn't uncommon for these small turbo cars to generate boost just puttering around town. As you can imagine, if the ECU richened the mixture up every time the driver got into some sort of boost, fuel mileage would drop and emissions (by and large) would increase. This chart below illustrates which pollutants are affected by a non-stoich mixture (notice carbon monoxide and hydrocarbons both decrease as the mixture leans out): Until we get the flash ready there are some options. Some folks add fuel in that area using the MAF tables. Now, this works to some degree, and basically what you're doing at that point is capitalizing on the ECU's response time to fueling changes. We're essentially jacking up the MAF voltage which increases fuel demand. The ECU will temporarily add fuel until it sees that it's straying from it's intended target, at which point it will begin pulling fuel out to compensate. I don't prefer this approach though to be candid. The best fix that we've found that's currently available is water/meth. Our water/meth customers are tuning their srpay to come on with boost pressure, and this drops the AFR immediately, and the ECU can't trim fuel out fast enough, so you get the mixture you intend to see. I have another option that I'd like to try, and it has to do with manipulating fuel pressure with the Standback programmer. I've got a map and no one to try it out for me, lol! But what I've done is commanded fuel pressure to rise when boost is between 3-16psi in the ranges of 2000-4000rpm. The fuel pump has a feedback loop attached to it just as the air/fuel control does, but I don't think the pump will respond as quickly as the closed-loop fuel control does. So my thinking is that we can bump pressure up gradually as boost pressure rises. I don't know if this will be a proper solution, but I'd like to give it a try sometime. I can post a picture of the map I whipped up if anyone else wants to try it too. The KR seems pretty consistent, which is convenient. I'd slowly start removing timing from that region and see if that helps quell the knock. Remember, if your air/fuel ratio is solid, then increasing fuel is not a smart way to quell knock. Gerald, you also commented that your boost isn't holding very well with the factory PID parameters, so let's see if we can tweak them to improve how well you hold your boost target. You said in your email that you have the controller set to 18psi, and it gets up there, but then beings to taper down almost immediately. If you look back at the PID explanation above, it sounds like you need to increase the ki term. I know this because you're consistently below your boost target, and that means the i-term doesn't carry enough weight to close the wastegate under those conditions. So you might try raising your ki term from 0.09 to maybe 0.2. The ki term will also add to the "urgency" when yopur boost is building so you might need to compensate by adding a little more kd. So you might leave your kp term at 30, and then bring your kd up to maybe 150. Try that and see if you can hold boost better.
Thanks! That's actually how I got introduced to PID controllers. I was a guest researcher at the National Institute of Standards and Technology, and I worked in the thermometer calibration lab. All of the calibration baths (I worked in low-temperature measurement) has PID loops controlling their temperature, and I had no idea that a few years later I'd be controlling turbos using the same technology. How cool is that?! I'm not sure if you ever received my email, but I'd like to work directly with you in tuning your application remotely. If you'd like to talk about getting your car dialed in you're welcome to call (301-576-6142) or email me (jgartenhaus@cp-e.com) anytime. | ||
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Access Tuner Race Tuning Guide | Haltech | Cobb ATR Support | 6 | 04-06-2010 12:40 AM |
vta tuning on the standback | bykeryder4life | MazdaSpeed 3/6 - ECU Computer Tuning | 14 | 09-18-2009 08:30 PM |
Standback Tuning for Dummies thread?? | vbt101 | MazdaSpeed 3/6 - ECU Computer Tuning | 12 | 09-02-2009 09:13 AM |
FFS and LC in the new tuning guide | twisted99 | Cobb ATR Support | 4 | 07-23-2009 12:25 AM |
Cp-e standback tuning | Wisedom | MazdaSpeed 6 - How-To's | 10 | 03-31-2009 07:17 PM |