Welcome to the North American Subaru Impreza Owners Club Friday August 1, 2014
Home Forums WikiNASIOC Products Store Modifications Upgrade Garage
NASIOC
Here you can view your subscribed threads, work with private messages and edit your profile and preferences Home Registration is free! Visit the NASIOC Store NASIOC Rules Search Find other members Frequently Asked Questions Calendar Archive NASIOC Upgrade Garage Logout
Go Back   NASIOC > NASIOC Technical > Engine Management & Tuning > Open Source Reflashes

Welcome to NASIOC - The world's largest online community for Subaru enthusiasts!
Welcome to the NASIOC.com Subaru forum.

You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our community, free of charge, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is free, fast and simple, so please join our community today!

If you have any problems with the registration process or your account login, please contact us.
* Registered users of the site do not see these ads.
Reply
 
Thread Tools Display Modes
Old 01-30-2008, 05:12 PM   #1
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default What is ECU execution frequency?

Does anyone know ECU execution frequency for USDM '02 WRX? Specifically interested in how often it executes boost control algo to calculate new wgdc.

If I do single variable log, I get ~90ms sample rate through OBDII interface. (thanks enginuity!) I imagine it's faster than that...

Related info that could help if anyone knows:
-Who makes subaru ecu's? Maybe manufacturer publishes specs.
-What do typical standalone EMS execute at?

TIA.
* Registered users of the site do not see these ads.
batchman is offline   Reply With Quote
Old 01-30-2008, 05:50 PM   #2
R4ND0M_AX3
Scooby Newbie
 
Member#: 67483
Join Date: Aug 2004
Location: Sacramento, CA
Vehicle:
2005 WRX STi

Default

I don't have an answer but I just want to ask, are you looking for the time portion of the Integral calc on the Boost control loop?
R4ND0M_AX3 is offline   Reply With Quote
Old 01-30-2008, 07:30 PM   #3
ExpO
*** Banned ***
 
Member#: 115819
Join Date: May 2006
Chapter/Region: Tri-State
Location: UDel, illidelphia
Vehicle:
2006 vf39 R0mRa1DeR
Darknessess Wagon

Default

Sounds like he's looking for the frequency.. not sure if this would be the cpu clock speed or a proportion of it... There are quite a few posts about the various ecu hardwares on Openecu.org forums, including some manufacturer spec sheets iirc.
ExpO is offline   Reply With Quote
Old 01-30-2008, 10:04 PM   #4
mickeyd2005
Scooby Newbie
 
Member#: 122595
Join Date: Aug 2006
Chapter/Region: NWIC
Vehicle:
2005 LGT Wagon
ABP

Default

I think turbo dynamics are updated every 70 to 80 ms on a 2005 LGT. However, it seems like the ecu should respond faster than that especially considering injector pulse width can be very low.
mickeyd2005 is offline   Reply With Quote
Old 01-30-2008, 11:29 PM   #5
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default

Quote:
Originally Posted by R4ND0M_AX3 View Post
are you looking for the time portion of the Integral calc ?
Basically. That's what initiated the question. How long between execution of integral algo.

Background: I did a little step test. Turn off wgdc compensation, set max wgdc limit at x% until 3700rpm then drop limit to (x-y)% at 3701 rpm. log mrp and you get a negative step test.

with my car and gmbcs 5% step in wgdc = ~1.5psi and took about 1200 ms to reach steady state. Can't remember time constant ~320ms? (Don't have file w/ me at home). I even was able to see a bit of dead time w/ obdII interface resolution. pretty neat if you're a geek.

Anyway, I need to know execution period of integral algo to finish my little exercise.

But now I'm also interested in base clock spd of our ecu. Assume all other algos will be executed as integer multiples of base.

I'll give openecu.org a search. Hadn't checked there.
batchman is offline   Reply With Quote
Old 01-30-2008, 11:35 PM   #6
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default

Quote:
Originally Posted by mickeyd2005 View Post
I think turbo dynamics are updated every 70 to 80 ms on a 2005 LGT. However, it seems like the ecu should respond faster than that especially considering injector pulse width can be very low.
Thanks. I'll use this for now and see what I get. Any background on source of your numbers?
batchman is offline   Reply With Quote
Old 01-31-2008, 12:25 AM   #7
mickeyd2005
Scooby Newbie
 
Member#: 122595
Join Date: Aug 2006
Chapter/Region: NWIC
Vehicle:
2005 LGT Wagon
ABP

Default

32 bit ecus are a little different. I just set TD integral to 1% and then watched how long it took for TD integral to windup.
mickeyd2005 is offline   Reply With Quote
Old 01-31-2008, 09:41 AM   #8
gabedude
Scooby Guru
 
Member#: 10108
Join Date: Sep 2001
Chapter/Region: TXIC
Location: Lake Travis, Austin, TX
Vehicle:
07 STI
60Trim TQ monster

Default

All of that is only one algorithm running on one thread. I am having a hard time finding the specifics of the Denso ECU, but there is a spec sheet somewhere. My point is, you do not know the priority of each thread, so your tests would be limitted to that one function of the ECU. Bascially, each different algorithm will have a different frequency based on what is more important. Knowing the CPU frequency will not tell you much unless you wanted to use the Denso ECU hardware and rewrite all of their code to make a standalone. There are more powerfull boards out their that can be used to produce a better plug in replacement for cheaper if you wanted to write your own standalone EMS.

Gabe
gabedude is online now   Reply With Quote
Old 01-31-2008, 11:08 AM   #9
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default

Quote:
Originally Posted by mickeyd2005 View Post
32 bit ecus are a little different. I just set TD integral to 1% and then watched how long it took for TD integral to windup.
DOH Excellent idea. Thought hadn't occurred to me. Thanks! That will work for my purposes.
batchman is offline   Reply With Quote
Old 01-31-2008, 11:21 AM   #10
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default

Quote:
Originally Posted by gabedude View Post
All of that is only one algorithm running on one thread. My point is, you do not know the priority of each thread, so your tests would be limitted to that one function of the ECU. Bascially, each different algorithm will have a different frequency based on what is more important.
Gabe
That's the way I imagine it to work. Like a PLC, each function would have it's own execution frequency. If we knew base frequency, we'd at least know that frequency of a given function must be integer multiple of base.

You're right though, for my purposes consider this thread closed.

Quote:
Originally Posted by gabedude View Post
I am having a hard time finding the specifics of the Denso ECU, but there is a spec sheet somewhere.
Exactly. If anyone happens to know where to find this data and sees this thread before it gets pushed off front page, post it here. Maybe we could add specs to FAQ to so it would be easier to find in future.

Thanks for input folks!
batchman is offline   Reply With Quote
Old 01-31-2008, 08:06 PM   #11
Freon
Scooby Specialist
 
Member#: 88322
Join Date: Jun 2005
Chapter/Region: MWSOC
Location: Indianapolis, IN
Vehicle:
2009 BMW 135i

Default

There is a whole OS layer that handles execution priority and a very complex timer system. You can search around for "Hitachi Vehicle Operating System". Both the 16bit and 32bit ECU seem to use this OS.

One of the most important features of the specific chips our cars use is the counter systems they integrate. Lots of timers, and they are used to trigger interrupts to ensure certain things are run on a regular basis.

I suppose one could accurately test execution speed by inserting a counter into each code segment you'd want to test, then you can log it.

I suspect RPM and crank position are calculated every time the crank sensor detects a tooth. I.e., ridiculously fast.

Boost control is very slow, around 15hz.
Freon is offline   Reply With Quote
Old 01-31-2008, 10:10 PM   #12
mickeyd2005
Scooby Newbie
 
Member#: 122595
Join Date: Aug 2006
Chapter/Region: NWIC
Vehicle:
2005 LGT Wagon
ABP

Default

I'm not sure what OP's needs are, but the only two things that I have found a time interval requirement are boost control and tip-in. For boost control, we can calculate that time as mentioned above.

However, for tip-in, the delta throttle change needs to have a time reference. So far, all I have been doing is changing the slope of the tip-in curve (starting at the corner) to drop the lean spike down.
mickeyd2005 is offline   Reply With Quote
Old 02-01-2008, 04:16 PM   #13
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default The PhD Dilemma

I'm an engineer. <-- don't hold this against me. Among engineers, I find myself falling between the "neat to know/PhD" type pursuits and the "practical/B.S." type applications. Once in a while, the "neat to know" work leads to a great practical idea --- If I don't get too bored with the topic first.

So this thread is right up my alley

Re: my original purposes. I'm retuning (very slowly) my XPT stg2 gmbcs map. I took out a bit of timing to get rid of knock and now I'm learning about boost control. Along the way, I'm picking up info on neat subjects like rescaling MAF, how to start tuning from stock map, tip-in fueling, etc. This is the most fun hobby I've had in a while. It's applicable to my career to boot.

Re: Boost tuning. Not many people talk TDB. Is the TDB thread triggered by change in TPS or change in Boost Target? And how often is trigger evaluated? Same frequency as TDI (I got ~84ms for my '02)?

With mickeyd's TDI experiment as inspiration, I am thinking of a way to find out about TDB. We'll see.
batchman is offline   Reply With Quote
Old 02-01-2008, 04:41 PM   #14
Tea cups
Scooby Specialist
 
Member#: 103136
Join Date: Dec 2005
Default

Tea cups is offline   Reply With Quote
Old 02-01-2008, 04:41 PM   #15
gabedude
Scooby Guru
 
Member#: 10108
Join Date: Sep 2001
Chapter/Region: TXIC
Location: Lake Travis, Austin, TX
Vehicle:
07 STI
60Trim TQ monster

Default

Quote:
Originally Posted by Freon View Post
There is a whole OS layer that handles execution priority and a very complex timer system. You can search around for "Hitachi Vehicle Operating System". Both the 16bit and 32bit ECU seem to use this OS.

One of the most important features of the specific chips our cars use is the counter systems they integrate. Lots of timers, and they are used to trigger interrupts to ensure certain things are run on a regular basis.

I suppose one could accurately test execution speed by inserting a counter into each code segment you'd want to test, then you can log it.

I suspect RPM and crank position are calculated every time the crank sensor detects a tooth. I.e., ridiculously fast.

Boost control is very slow, around 15hz.
Nice info. A quick search on google brought up the datasheet for the SH2 RTOS.
gabedude is online now   Reply With Quote
Old 02-01-2008, 05:39 PM   #16
batchman
Scooby Newbie
 
Member#: 94656
Join Date: Aug 2005
Chapter/Region: TXIC
Location: SE Texas
Vehicle:
02 WRX Quicksilver
hums Indiana Jones theme

Default

Quote:
Originally Posted by Tea cups View Post
http://forums.nasioc.com/forums/show....php?t=1433698

The 'Turbo Dynamics Burst' table is the least used of the two correction tables. It is only active when boost error swings extremely quickly from negative to positive or vice versa (outside the range determined by the activation thresholds). Otherwise, it does not come into play.
Had read before.. read it again to see what I missed.

So the TDB action is somewhat of a "one-shot" correction, only applied when boost error changes signs w/ value outside activation window? OR is there more implied by the "extremely quickly"?

Reading about the 32bit approach to boost control makes me want to give up on my 16bit control and go w/ hybrid mbc setup as some have suggested. Only concern I need to squash is mbc unable to taper boost near redline. I don't fully understand what ill-effects this might have. Slightly less power due to operating at lower efficiency range?

Quote:
Quote:
Originally Posted by Freon
There is a whole OS layer that handles execution priority and a very complex timer system. You can search around for "Hitachi Vehicle Operating System". Both the 16bit and 32bit ECU seem to use this OS.


Nice info. A quick search on google brought up the datasheet for the SH2 RTOS.
No joke! Good lead. Lots of good reading material

On a side note, I'm currently perusing other job options and did a search for "powertrain engineers".. For those future ME's on the board, might be worth considering. Too late for me. I went down the chemistry side. DOH! I should have been a ME.
batchman is offline   Reply With Quote
Old 02-01-2008, 06:12 PM   #17
Tea cups
Scooby Specialist
 
Member#: 103136
Join Date: Dec 2005
Default

Quote:
Originally Posted by batchman View Post
So the TDB action is somewhat of a "one-shot" correction, only applied when boost error changes signs w/ value outside activation window? OR is there more implied by the "extremely quickly"?
boost error change as compared to the last execution.
Tea cups is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Why doth thou Executive Admin Assistant thinketh she is an executive? -=C=- Off-Topic 34 11-20-2009 02:11 PM
RE: ECU sampling frequency? verc Engine Management & Tuning 2 08-21-2004 01:48 AM
Anyone know what year ecu this is? Avenger Normally Aspirated Powertrain 4 05-22-2003 08:46 PM
What is stock ECU max boost per gear? WRXTom Factory 2.0L Turbo Powertrain 11 09-13-2002 10:42 AM
Group N ecu, what is it? skidplatez Motorsports 5 06-05-2001 10:06 PM


All times are GMT -4. The time now is 10:49 PM.


Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Powered by Searchlight © 2014 Axivo Inc.
Copyright ©1999 - 2014, North American Subaru Impreza Owners Club, Inc.