Open ECU?

[arghx7]

Attach the .csv files you have saved. I want to take a look at some of this stuff.

I am curious which load PID Evoscan is reading. It looks like it might be Calculated Load, not absolute load which is more useful. Calculated load is scaled so that in essence WOT is 100%. Absolute load is actually a measure of cylinder filling efficiency. It can exceed 100% in boosted applications.

I'm sure Evoscan is missing a lot of possible parameters on the bus. The latest list of universal parameters is attached. You need to log fuel rail pressure. The actual name is diesel fuel rail pressure but it's used for gas engines with GDI.

[jedibow]

Originally Posted by arghx7

Attach the .csv files you have saved. I want to take a look at some of this stuff.

I am curious which load PID Evoscan is reading. It looks like it might be Calculated Load, not absolute load which is more useful. Calculated load is scaled so that in essence WOT is 100%. Absolute load is actually a measure of cylinder filling efficiency. It can exceed 100% in boosted applications.

I'm sure Evoscan is missing a lot of possible parameters on the bus. The latest list of universal parameters is attached. You need to log fuel rail pressure. The actual name is diesel fuel rail pressure but it's used for gas engines with GDI.

Thank you, I will log tonight! You are correct it is Load Calc, not Absolute. I did see that I could log fuel pressure.

[Billy2224]

Originally Posted by JP

have you seen this?> http://www.romraider.com/forum/viewtopic.php?f=8&t=8475

Haltech has a car now for development, Motec's also doing an M1 plugin ECU.

awesome, my neighbour works somewhere high up in Haltech.. yay for my ECU flash!!

[Dustin@Dynosty]

We are very close with Haltech and they will be a little ways out before they have a replacement unit but in the mean time we are wiring up a Haltech Sport 1000 to start playing around with the car in stock form before we build a turbo kit.

The air fuel readings at least on our project car indicated that there may still be a little left on the table from Toyota/Scion/Subaru on the tuning side. We will know more once the car is up and running on the haltech.

http://www.dynosty.com/2012/06/scion...dyno-baseline/

[jedibow]

Originally Posted by arghx7

Attach the .csv files you have saved. I want to take a look at some of this stuff.

Per requested by arghx7

I attempted to attach the .csv, but it didn't work?

[jedibow]

Try this one!

[jedibow]

Originally Posted by Dustin@Dynosty

We are very close with Haltech and they will be a little ways out before they have a replacement unit but in the mean time we are wiring up a Haltech Sport 1000 to start playing around with the car in stock form before we build a turbo kit.

The air fuel readings at least on our project car indicated that there may still be a little left on the table from Toyota/Scion/Subaru on the tuning side. We will know more once the car is up and running on the haltech.

http://www.dynosty.com/2012/06/scion...dyno-baseline/

Based on your wideband AFR, and my log of the fuel map tracing it would seem that the MAF in our cars may not be calibrated exactly, however it may me injector scaling, or latency values also... Interesting

I know one of the first items while tuning wrx's is to calibrate the MAF sensors, so if this is true, then there will be power to be made with adjsuting the stock ignition map, and altering the MAF calibrations.

[arghx7]

This appears to be the data you generated last night, because it doesn't have fuel pressure, absolute load, or any of the other data I suggested you take. I will say that the sample rate is pretty quick considering it was taken over the OBD port.

In all graphs, X axis is time in seconds and AFR is commanded AFR from the ECU--it's not a measured value.

Below we can get an understanding of the enrichment strategy. Notice that as it goes WOT, there is a step change in the commanded AFR. Then it continues in this enriched region until around 4500 rpm, where it enriches again, most likely in open loop control. You didn't log estimated catalyst temp but I would expect the open loop enrichment to coincide with a catalyst temp of around 900-950C.




Also notice that timing is retarded at the same time the mixture begins to enrich. That is likely a knock-prone area or an area where valve timing changes significantly. Another way to look at it is that in the 3000-4500rpm area, valve overlap and scavenging creates reduced burned mass in the cylinder, allowing more ignition timing due to decreased knock sensitivity.

We need data read directly from the ECU to get a better understanding of what's going on. Below is TPS, calculated load, and some kind of MAF flow. Normally it's in grams/second but this is reading weird units. Also you can see that the shape of the calculated load curve roughly corresponds with the TPS percent, despite the offset. I'd like to see calculated and relative load data taken at the same time.



Here's what you should try to log if possible:

-- Fuel system status (Power take off status?) which indicates open loop
-- RPM
-- Ignition timing
-- Commanded AFR
-- Fuel pressure ("diesel" rail pressure)
-- absolute load
-- calculated load
-- accelerator pedal position
-- throttle position
-- catalyst temp
-- O2sensor Linear Lambda (attempting to log the stock Denso wideband)
-- Short term fuel trim
-- Long term fuel trim

there are multiple parameters to choose from for some of these. You may have to play around to see which ones are being reported, because every car is different.

[arghx7]

Originally Posted by jedibow

Based on your wideband AFR, and my log of the fuel map tracing it would seem that the MAF in our cars may not be calibrated exactly, however it may me injector scaling, or latency values also... Interesting

I know one of the first items while tuning wrx's is to calibrate the MAF sensors, so if this is true, then there will be power to be made with adjsuting the stock ignition map, and altering the MAF calibrations.

The MAF sensor calibration comes from the supplier (Denso in this case) and then is tweaked on an engine dyno early on in the development process. Normally you use lab-grade tolerances with the parts, so-called "mean spec" parts. So the MAF calibration is made with parts that are somewhere in the middle between production sensors that read high and production sensors that read low. The fuel trims take care of the rest of the variation in sensors.

Translation: adjusting the MAF scaling on a stock car... I guess if you have time on your hands and you are a fuel trim nazi you could do it, but if it's good enough to meet emissions it''ll be fine for you.

[jedibow]

You are correct these logs are still from last night. I'm still in the process fo determining which ROM addresses are located where, and the correct scaling factors, as I'm logging off the CAN connection, not the specific Subaru SSM's. Fuel pressure does infact log, I'm not sure whether the scale is correct though. Throttle percentage does not point to the correct location as it just stays at max value when logging no matter where the throttle is located, I have not attempted to log catalyst temps yet, but I can, and the O2 sensor is available however it appears to be 0-1V and a fifth degree scaling equation, so I'm still working on that. If you want raw unscaled data I can log it.

I believe that the above fuel command that I logged was not in Closed loop at all, as the Closed loop has a load threshold to keep it from staying in closed loop during WOT, atleast this is true with older subaru ECU's, this could be different. I could also log fuel trims in the RAW data form, as the scaling is also way off for them, but I will try.

Right now I'm attempting to get the scaling correct on the O2 sensors, to determine if they are infact wideband, or narrow band, and if I'm successful this should also determine the crossover point from closed to open loop fueling in the ECU.

[arghx7]

Did you use the info in the parameter list I posted? All the CAN stuff is standardized. If you can log fuel system status it will probably tell us if it's open loop or not.

Richened closed loop operation is a common CO emissions control strategy now. I suspect Subaru adopted it here.

[jedibow]

Originally Posted by arghx7

Did you use the info in the parameter list I posted? All the CAN stuff is standardized. If you can log fuel system status it will probably tell us if it's open loop or not.

Richened closed loop operation is a common CO emissions control strategy now. I suspect Subaru adopted it here.

I'm working on it, my nephews birthday was today, so I will be able to work more now that I'm home. I don't think the problem is with the parameters, it is with the evoscan definitions, but I'm using your parameters to hopefully set the correct scalings. Thank you!

[Ess Vee]

I'm completely ignorant on this kind of stuff, so forgive me ahead of time... but is there a way to tune a MPG efficient ECU, with no regards to performance?

Just out of curiosity...

[serialk11r]

Originally Posted by Ess Vee

I'm completely ignorant on this kind of stuff, so forgive me ahead of time... but is there a way to tune a MPG efficient ECU, with no regards to performance?

Just out of curiosity...

I think the primary reason the way things are done the way things are done is for emissions, so yes. You can crank the AFR very high ("lean burn") to reduce pumping losses, and this will improve your fuel economy at the cost of emitting a lot more NOx. We're talking like 20:1 or so, which is around what Honda had on their lean burn engines or something. Dunno how hard it is to do...

I don't think it should affect performance, because full load parameters can be the same. When you do physical modifications of course you can swing things towards fuel economy or performance.