Jump to content


Welcome to Ultimate Subaru Message Board, my lurker friend!

Welcome to Ultimate Subaru Message Board, an unparalleled Subaru community full of the greatest Subaru gurus and modders on the planet! We offer technical information and discussion about all things Subaru, the best and most popular all wheel drive vehicles ever created.

We offer all this information for free to everyone, even lurkers like you! All we ask in return is that you sign up and give back some of what you get out - without our awesome registered users none of this would be possible! Plus, you get way more great stuff as a member! Lurk to lose, participate to WIN*!
  • Say hello and join the conversation
  • Subscribe to topics and forums to get automatic updates
  • Get your own profile and make new friends
  • Classifieds with all sorts of Subaru goodies
  • Photo hosting in our gallery
  • Meet other cool people with cool cars
Seriously, what are you waiting for? Make your life more fulfilling and join today! You and your Subaru won't regret it, we guarantee** it.

* The joy of participation and being generally awesome constitutes winning
** Not an actual guarantee, but seriously, you probably won't regret it!

Serving the Subaru Community since May 18th, 1998!

Guest Message by DevFuse
 

Photo
- - - - -

Frankenmotor and ECU hacking


  • Please log in to reply
25 replies to this topic

#1 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 07 July 2012 - 12:56 PM

I've finished the frankenmotor in my Vanagon, EJ25D shortblock with EJ22E heads and Delta 220 cams. I calculated comp ratio at 10.4:1. Things went pretty good, and I used everything else from my EJ22E, which is from a '92 Legacy.

It's been said that the cams cause idle problems, and the frankenmotor too. Each one can be handled by the ECU okay, but both together cause problems. Well not surprisingly my idle is pretty terrible, and the computer is not going closed loop AFR; I should mention it ran perfectly before. The ECU tries to correct AFR by +20%, but then gives up.

My plan is to remap the ECU. I ordered some custom PCBs to install a daughtercard into the ECU (design thanks to Alcyone.org.uk!) and I've figured out where the maps live in the ROM. This stuff is pretty much the same as the remapping I've done on my GC8 WRX motor, so I have some experience here.

Anyone with experience mapping these engines with aftermarket engine management? I know I need to richen the idle, but what else, and why? Do the cams allow unmetered air in from the exhaust during idle, is this why it's lean?

Blipping the throttle I can hear some pinging. I have 87 in the tank now, and I know I need to run 91. But I'd like to back off the timing a bit to be safe. So for a higher compression ratio, theoretically do I need to back off timing across the board or just at certain places (i.e. low RPM)?


Below is a pic of the motor, and the maps I downloaded from my "F9" ECU.


Posted Image
Posted Image

#2 renob123

renob123

    Subaru Nut

  • Members
  • 719 posts
  • Hillsburrito, OR

Posted 09 July 2012 - 12:58 PM

Huh. I hadn't realized you could use RR on a '92 ECU.

I have no idea about the tuning part. I ended up getting a pretty good idle on mine by adjusting the IACV. The upper adjustment didn't do much, but the adjustment on the bottom helped a lot. If you're running lean due to unmetered air, then I'd bet on the IACV over air sneaking through the exhaust valves.

Jacob

#3 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 09 July 2012 - 01:39 PM

Thanks for your reply!

Huh. I hadn't realized you could use RR on a '92 ECU.


You couldn't, until I found the map locations and created a new entry in the RR XML file. :cool:

I have no idea about the tuning part. I ended up getting a pretty good idle on mine by adjusting the IACV. The upper adjustment didn't do much, but the adjustment on the bottom helped a lot. If you're running lean due to unmetered air, then I'd bet on the IACV over air sneaking through the exhaust valves.

Jacob


What do you mean adjustment of the IACV on top and bottom? The top being the electrical part and the bottom being the valve itself? Correct me if I'm wrong, but the IACV is also metered through the MAF, right? So even if it's wrong it still should have a good AFR.

It's possible I didn't hook something up right, or introduced a vacuum leak somewhere, but I went over everything a couple times and it seems ok. I guess I can borrow a smoke machine sometime. GeneralDisorder mentioned a while back that frankenmotor+cams is too much for the OBD1 ECUs which seems to be the case for me.

I'm trying to get some insight into how the cams and frankenmotor change the effective AFR.

#4 renob123

renob123

    Subaru Nut

  • Members
  • 719 posts
  • Hillsburrito, OR

Posted 09 July 2012 - 02:08 PM

Thanks for your reply!



You couldn't, until I found the map locations and created a new entry in the RR XML file. :cool:


I may want to get in on that. Mostly out of curiosity, though.

What do you mean adjustment of the IACV on top and bottom? The top being the electrical part and the bottom being the valve itself? Correct me if I'm wrong, but the IACV is also metered through the MAF, right? So even if it's wrong it still should have a good AFR.

It's possible I didn't hook something up right, or introduced a vacuum leak somewhere, but I went over everything a couple times and it seems ok. I guess I can borrow a smoke machine sometime. GeneralDisorder mentioned a while back that frankenmotor+cams is too much for the OBD1 ECUs which seems to be the case for me.

I'm trying to get some insight into how the cams and frankenmotor change the effective AFR.


I'm happy with my Frankenmotor+cams and OBDI ECU. It took a long time to figure out the idle, though. I guess it has something to do with the overlap reducing dynamic compression which reduces the vacuum. We're dealing with a limited sample size, so I think it's too early to tell whether you're unlucky, or I'm lucky, or if there's some magic step to take that will help everyone. To me, my Frankenmotor just feels like an engine that has cams. If you've been to a "hot rod" show (aka: cars that just sit there show), then you'll know how a car with cams will behave.

For the IACV, yes, I'm talking about the bottom one. The top is the electric thing with an electric plug on it. The bottom will have a hex on it so you can use a socket wrench on it. The electric thing didn't help. Turning the hex on the bottom helped. Yes, the air has been metered. However, my understanding is that in a MAF car, under some conditions, the time between when the air is metered and when it is used in combustion matters. If a ton of air gets metered, then the computer expects to dump a ton of fuel some time down the line. Lengthen or shorten that time, and you'll get lean/rich conditions. That's why STIs could use a retune after the addition of most aftermarket intakes.

Does that sound reasonable?

Jacob

#5 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 09 July 2012 - 05:00 PM

I may want to get in on that. Mostly out of curiosity, though.


Sure thing. I'll also be releasing my modified ROM. What kind of ECU do you have?

I'm happy with my Frankenmotor+cams and OBDI ECU. It took a long time to figure out the idle, though. I guess it has something to do with the overlap reducing dynamic compression which reduces the vacuum. We're dealing with a limited sample size, so I think it's too early to tell whether you're unlucky, or I'm lucky, or if there's some magic step to take that will help everyone. To me, my Frankenmotor just feels like an engine that has cams. If you've been to a "hot rod" show (aka: cars that just sit there show), then you'll know how a car with cams will behave.


I'm glad to hear you have it working ok. I'll ponder that dynamic compression ratio thing... I'm not driving the motor yet until I can get the map closer to normal, but maybe after some driving it will eventually learn the AFR. But right now at idle it's running very lean.

For the IACV, yes, I'm talking about the bottom one. The top is the electric thing with an electric plug on it. The bottom will have a hex on it so you can use a socket wrench on it. The electric thing didn't help. Turning the hex on the bottom helped. Yes, the air has been metered.


I can up the idle by turning the top electric adjustment, but then the computer can't close the loop at idle. Does yours still control the idle? What is your idle speed? You can see in the maps above that warmed-up idle for my ECU is 700 RPM.

However, my understanding is that in a MAF car, under some conditions, the time between when the air is metered and when it is used in combustion matters. If a ton of air gets metered, then the computer expects to dump a ton of fuel some time down the line. Lengthen or shorten that time, and you'll get lean/rich conditions. That's why STIs could use a retune after the addition of most aftermarket intakes.

Does that sound reasonable?

Jacob


I understand what you're saying, but I'm not sure the difference in time between the throttle butterfly opening and the IACV opening is that much different, something more to ponder, hmm.

Thanks for your comments!

#6 renob123

renob123

    Subaru Nut

  • Members
  • 719 posts
  • Hillsburrito, OR

Posted 09 July 2012 - 07:11 PM

Sure thing. I'll also be releasing my modified ROM. What kind of ECU do you have?


Cool. I have an ECU from the '91-'93-ish era. I've always wondered what was going on inside of it.

I'm glad to hear you have it working ok. I'll ponder that dynamic compression ratio thing... I'm not driving the motor yet until I can get the map closer to normal, but maybe after some driving it will eventually learn the AFR. But right now at idle it's running very lean.


How does it look when it's not idling? I'd guess it's not lean at higher RPMs/loads. Have you considered anything else to be your problem? Bad MAF/O2/Injectors/Plugs?

I can up the idle by turning the top electric adjustment, but then the computer can't close the loop at idle. Does yours still control the idle? What is your idle speed? You can see in the maps above that warmed-up idle for my ECU is 700 RPM.


Don't bother with the top thing. I tried moving it all over the place, and it didn't help. Using the bottom adjustment helped, though. I'd phrase it this way. Right now, your IACV isn't controlling the idle. Mine is. Messing with the thing on the bottom gives the ECU/IACV some ability to adjust the idle. Right now yours could very well be pegged, with the conditions that are way outside of what the ECU is willing to do. I can't remember what my warm idle is. The car is in the shop for some work.

I understand what you're saying, but I'm not sure the difference in time between the throttle butterfly opening and the IACV opening is that much different, something more to ponder, hmm.


Assuming the air travels at the same speed, then yes, I'd guess the difference in time is minimal. I don't know anything about fluid dynamics, but I'm betting the air takes less time to get to the cylinder when it has 2 places to go vs. 1 place to go, at least when both places are summoning the air with vacuum. So sometimes the air gets to the cylinders quickly, and the injectors need to fire soon, and sometimes the air takes a bit longer. That is totally my rudimentary understanding, though. Cobb would probably have a way better explanation. Maybe williaty on NABISCO would as well.

Jacob

#7 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 12 July 2012 - 09:36 AM

Cool. I have an ECU from the '91-'93-ish era. I've always wondered what was going on inside of it.


If you have an "F9" ECU then my modified ROM will work for you, assuming I can get the idle sorted and running smooth.

How does it look when it's not idling? I'd guess it's not lean at higher RPMs/loads. Have you considered anything else to be your problem? Bad MAF/O2/Injectors/Plugs?


I ran it no-load at high RPM and it was smooth, but still running lean. I am going to remap the ECU to get the mixture in the ballpark before driving it. The engine (and all the EJ22 parts) were working perfectly a week before doing the cams/shortblock so unless I messed something up during the engine work they should all be ok.

Don't bother with the top thing. I tried moving it all over the place, and it didn't help. Using the bottom adjustment helped, though. I'd phrase it this way. Right now, your IACV isn't controlling the idle. Mine is. Messing with the thing on the bottom gives the ECU/IACV some ability to adjust the idle. Right now yours could very well be pegged, with the conditions that are way outside of what the ECU is willing to do. I can't remember what my warm idle is. The car is in the shop for some work.


I could get it to idle fast with the top adjustment, but the ECU was obviously not in control. And I could also get it to idle slow and it would be correct at around 700 RPM (also monitoring IACV duty cycle with SSM), but it was running like poo, just shaking a lot.


No comments from anyone else? :-p Not even on my choice of alternator? :brow:

#8 ivans imports

ivans imports

    1000+ Super USER!

  • Members
  • 2,954 posts
  • lumby bc canada

Posted 12 July 2012 - 10:28 AM

not everyone needs a wleder on there volksy van what 200 amps or more bet your lights are good

#9 GeneralDisorder

GeneralDisorder

    Elite Master of the Subaru

  • Members
  • 20,281 posts
  • Portland

Posted 12 July 2012 - 10:47 AM

Since you have gone this far.... The community (and especially myself) would be forever grateful if you could do some similar work on the EJ22T ECU which is almost identical to the 92 ECU you are working on - with the addition of boost control.

If we had access to the maps and the boost control it would make tuning them a viable option and I would very much like to see that happen. As a former software engineer I wouldn't mind setting up to do this... If you could point me in the direction I need to go for aquireing hardware and setting up to interface with the ROM.....

NICE WORK btw!

GD

#10 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 12 July 2012 - 11:41 AM

not everyone needs a wleder on there volksy van what 200 amps or more bet your lights are good


Heehee, yes good eye; it is 200 amps and 15V. Barely fits, can just fit a credit card between the alt and the intake. I use it to charge my Odyssey house batteries.

Since you have gone this far.... The community (and especially myself) would be forever grateful if you could do some similar work on the EJ22T ECU which is almost identical to the 92 ECU you are working on - with the addition of boost control.

If we had access to the maps and the boost control it would make tuning them a viable option and I would very much like to see that happen. As a former software engineer I wouldn't mind setting up to do this... If you could point me in the direction I need to go for aquireing hardware and setting up to interface with the ROM.....

NICE WORK btw!

GD


Thanks GD!

I started a new thread just for ECU hacking here:
http://www.ultimates...d.php?p=1122184

#11 renob123

renob123

    Subaru Nut

  • Members
  • 719 posts
  • Hillsburrito, OR

Posted 12 July 2012 - 11:56 AM

I ran it no-load at high RPM and it was smooth, but still running lean. I am going to remap the ECU to get the mixture in the ballpark before driving it. The engine (and all the EJ22 parts) were working perfectly a week before doing the cams/shortblock so unless I messed something up during the engine work they should all be ok.


Where are your IDCs?

Jacob

#12 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 12 July 2012 - 09:01 PM

Where are your IDCs?

Jacob


The IACV duty cycle? It was anywhere from 25-50%, depending on where I turned the motor.

#13 GeneralDisorder

GeneralDisorder

    Elite Master of the Subaru

  • Members
  • 20,281 posts
  • Portland

Posted 12 July 2012 - 09:04 PM

I think he meant Injector Duty Cycle.

GD

#14 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 12 July 2012 - 09:12 PM

I think he meant Injector Duty Cycle.

GD


Well, the SSM interface does not show it as duty cycle (time on vs overall time), but as just the pulse width (time on). At idle it was showing around 2ms injector pulse width, which is the same as my EJ20G. About 1ms of that is the injector latency anyways. However my EJ20G has 550cc/min injectors and the frankenmotor has 280cc/min ones...

But it's not wise to put too much faith in the numbers from the SSM software, as a lot of the conversion constants are wrong.

Edited by presslab, 12 July 2012 - 09:17 PM.


#15 nncoolg

nncoolg

    USMB Regular

  • Members
  • 144 posts
  • Gold Coast, Australia

Posted 13 July 2012 - 09:49 PM

the bottom part of those EJ IAC valves are prone to corrode and seize if the coolant hasnt been cared for, it may be jammed, take the electric part off and see if the spigot turns.

#16 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 17 July 2012 - 10:30 AM

Quick update. EEPROM board is in:

Posted Image


Mixture is only wrong at idle, otherwise it's pretty spot on. I guess it's due to cam overlap.

I'm having to pull A LOT of timing, but it's still running very strong. For sure it's due to higher compression, but I'm sure the piston shape, cam, et cetera play into this.

Modified the ECU code to flicker the CEL on knock, a very useful tool. Found the idle timing map (separate from base map) and experimenting with retarded timing at idle to smooth out the cam lope. Seems to work. Set idle speed to 1000 RPM and it's decently smooth.

#17 GeneralDisorder

GeneralDisorder

    Elite Master of the Subaru

  • Members
  • 20,281 posts
  • Portland

Posted 17 July 2012 - 11:45 AM

Very cool. When I get a spare minute I'm going to pull my ECU down and get some pics....

GD

#18 grossgary

grossgary

    Elite Master of the Subaru

  • Members
  • 20,093 posts
  • WV

Posted 17 July 2012 - 10:22 PM

wow.
well.
done.
:clap:

i would love to be able to dive into that too, that's just too cool!

#19 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 18 July 2012 - 12:08 PM

Thanks! :drunk:

Here is a quick video I took. Remember this is a 4800 lb Vanagon Westfalia.

https://picasaweb.go...feat=directlink

#20 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 23 July 2012 - 05:42 PM

Here's links to my modified ROM and the XML file for RomRaider.
https://docs.google....YmExVDBhOUFRZjg
https://docs.google....aDRQbzBCbHhydW8

#21 nncoolg

nncoolg

    USMB Regular

  • Members
  • 144 posts
  • Gold Coast, Australia

Posted 24 July 2012 - 07:40 AM

Would it be possible to make the CEL light at high temp seen by the CTS? May be a useful fail safe warning to draw attention to the dash...

#22 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 24 July 2012 - 12:54 PM

Would it be possible to make the CEL light at high temp seen by the CTS? May be a useful fail safe warning to draw attention to the dash...


Yes that's possible. And in fact, the ROM from pmugabi for the EJ20G does this along with limiting the RPM to ~3000.

The issue with the "F9" ECU is that all the ROM space is used. The CEL on knock I was able to fit in, but it was very small. To get more ROM space some OBD stuff could be chucked.

#23 nncoolg

nncoolg

    USMB Regular

  • Members
  • 144 posts
  • Gold Coast, Australia

Posted 25 July 2012 - 06:26 AM

That is very cool. I imagine that would be a very useful feature.
So, can you update your tune while the ECU is in the car, or do you have remove it and put the chip in the eeprom burner?

#24 presslab

presslab

    Subaru Nut

  • Members
  • 867 posts
  • N. California

Posted 25 July 2012 - 01:42 PM

I leave the ECU installed, but I do need to remove the EEPROM ICs and program them. I use a ZIF socket on the ECU daughterboard to make this easier.

#25 86 Wonder Wedge

86 Wonder Wedge

    USMB Regular

  • Members
  • 374 posts
  • Akron, OH

Posted 31 July 2012 - 02:07 AM

With the daughterboard installed, could you use an emulator to run/tune the maps in realtime instead of having to burn/reburn EEPROMs? Run/Record with the emulator, fine tune on the fly, then burn one EEPROM or keep the emulator in place?

I was able to use my Emu emulator on an old 3rd gen F body ECM in real time to run my "warmed" :D 80 Turbo TA. It would load an "EPROM" to the board and allowed the laptop to record and modify the timing/tables w/o having to remove the chip.. I just left the Emu plugged in so I could have USB connectivity whenever and change maps/timing to suit the weather or driving style (track/street/strip)...




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users