ECU reprogramming
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
No biggie. I've started working on my own disassembly.
I haven't gotten very far yet, but far enough to say that we're pretty much on the right track. The initialization routine seems sensible, which means we've definitely got the right instruction set.
I think it's pretty much a matter or time until we crack this thing.
I haven't gotten very far yet, but far enough to say that we're pretty much on the right track. The initialization routine seems sensible, which means we've definitely got the right instruction set.
I think it's pretty much a matter or time until we crack this thing.
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
I'm having my doubts about the memory map posted on the 16paws site.
I just caught a memory access to 0x8000, which according to his map, there should be nothing there to access.
Looks like after I finish the analog board, I'll be doing schematics for the digital one.
I just caught a memory access to 0x8000, which according to his map, there should be nothing there to access.
Looks like after I finish the analog board, I'll be doing schematics for the digital one.
Senior Member
Joined: Jan 2004
Posts: 339
Likes: 0
From: So Cal.
****. you guys are fukin amazing. thats all i have to say. i have access to eeprom writers and know some stuff about programming but you guys are doing stuff way out of my league. GOD BLESS YOU!
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Ok, I'd like some opinions here.
Here's the way my test setup is going to go:
3.5" outlet cone air filter => Mustang MAF => 3" coupler => Spectre AFM adapter => stock AFM
I'm going to log the Sensor voltages against each other to build a conversion table.
What air filter should I use?
Here are a few I'm looking at:
#1
K&N
RX-4140
Endcap is also a filter. I already have an RX-4990.
http://www.knfilters.com/universal/X-stream.htm
#2
AEM
http://www.streetbeatcustoms.com/aem212049-sy.html
$71.96
#4
K&N
RC-4760
8" long
http://www.knfilters.com/universal/univrndtap.htm
Summit part #:
KNN-RC-4760
$48.69
#5
Apexi
Supposedly they're good, but I can't seem to find decent info on their website.
#6
Some other K&N filter with 3.5" i.d. outlet
Here's the way my test setup is going to go:
3.5" outlet cone air filter => Mustang MAF => 3" coupler => Spectre AFM adapter => stock AFM
I'm going to log the Sensor voltages against each other to build a conversion table.
What air filter should I use?
Here are a few I'm looking at:
#1
K&N
RX-4140
Endcap is also a filter. I already have an RX-4990.
http://www.knfilters.com/universal/X-stream.htm
#2
AEM
http://www.streetbeatcustoms.com/aem212049-sy.html
$71.96
#4
K&N
RC-4760
8" long
http://www.knfilters.com/universal/univrndtap.htm
Summit part #:
KNN-RC-4760
$48.69
#5
Apexi
Supposedly they're good, but I can't seem to find decent info on their website.
#6
Some other K&N filter with 3.5" i.d. outlet
Last edited by theloudroom; Jan 26, 2004 at 01:14 PM.
Junior Member
Joined: Aug 2001
Posts: 20
Likes: 0
From: Tampa, FL
with regards to what filter to put in front of your correlation rig, I would suggest you try and provide two or three points on the restriction curve.
If you test with a stock type filter (high restriction), a medium (K&N), and a low (think worn gym sock), we can see if the filter has a significant impact.
If it shifts the curve, thats an easy adjustment. If it changes the slope, that's VERY important. If one or more create shifts at certain flow levels... that's critical.
It may be that at Idle, the filter represents "no restriction" and as the rpms increase, a critical point is reached where the vacuum has to start increasing to achieve the flow rate. This creates a curve that changes slope at some critical value. All these things might make the chage of air filter more critical, depending on how much restriction difference, and its impact on airflow.
Porting may do the same thing. At low RPM, the pulsing of the air flow may also have some effect. At higher RPMs, the mass of airflowing probably damps those to where they have zero effect. But it is probably important to repeat these tests with other engines to see if wilder ports will work with the new\old setup using the generic adaptation.
Sorry, but as an engineer, I love more data. Just some thoughts.
Sunrotor
If you test with a stock type filter (high restriction), a medium (K&N), and a low (think worn gym sock), we can see if the filter has a significant impact.
If it shifts the curve, thats an easy adjustment. If it changes the slope, that's VERY important. If one or more create shifts at certain flow levels... that's critical.
It may be that at Idle, the filter represents "no restriction" and as the rpms increase, a critical point is reached where the vacuum has to start increasing to achieve the flow rate. This creates a curve that changes slope at some critical value. All these things might make the chage of air filter more critical, depending on how much restriction difference, and its impact on airflow.
Porting may do the same thing. At low RPM, the pulsing of the air flow may also have some effect. At higher RPMs, the mass of airflowing probably damps those to where they have zero effect. But it is probably important to repeat these tests with other engines to see if wilder ports will work with the new\old setup using the generic adaptation.
Sorry, but as an engineer, I love more data. Just some thoughts.
Sunrotor
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Originally posted by sunrotor
If you test with a stock type filter (high restriction), a medium (K&N), and a low (think worn gym sock), we can see if the filter has a significant impact.
If you test with a stock type filter (high restriction), a medium (K&N), and a low (think worn gym sock), we can see if the filter has a significant impact.
What I'm figuring on doing is taking a bunch of sets of data points and averaging them and time aligning them as close as I can. Then I'll use Matlab's polyfit function to fit a low-order polynomial to the data to develop a conversion formula that I will use at any airflow readings higher than those reached in the tests.
Any luck finding info on those chips?
I'm still finishing up the schematics. I'll be setting up a website with them once I'm done. I have a MAF senor now that I can start playing with.
Full Member
Joined: Oct 2002
Posts: 89
Likes: 0
From: Canada
Originally posted by RETed
You are correct. Most ECU's are using the Motorola 68HC11 or a variant of it. The original 6811's clock-speed was around 10MHz, which would sound slow compared to your current Intel powered PC (>3.0GHz), but these microprocessors run a very slim RISC design. With the reduced language set and streamlined run processing, the 10MHz clock-speed is fast enough to handle engines revving up to 10k RPM with very little fuss. On top of that, these automobile uP's typically run a timing co-processor, which handles the ignition input duties; this lessens the responsibility of the main processor demands, making for a nicer running system.
The CPU chip isn't very quick, but it's "good enough" to handle running the engine.
NOW, it's a matter of what programming you're doing... Sometimes, just "changing the maps" doesn't always work, as there are lots of watchdog and range limitations coded in there typically. This means if you adjust something that the code doesn't like, it's either going to barf or revert back to a "backup map". Thus, "ROM tuning" does NOT change parameters in very extreme values.  Fuel maps are typically not adjusted past 20%, and ignition timing isn't adjusted past 5-degrees either way.
-Ted
You are correct. Most ECU's are using the Motorola 68HC11 or a variant of it. The original 6811's clock-speed was around 10MHz, which would sound slow compared to your current Intel powered PC (>3.0GHz), but these microprocessors run a very slim RISC design. With the reduced language set and streamlined run processing, the 10MHz clock-speed is fast enough to handle engines revving up to 10k RPM with very little fuss. On top of that, these automobile uP's typically run a timing co-processor, which handles the ignition input duties; this lessens the responsibility of the main processor demands, making for a nicer running system.
The CPU chip isn't very quick, but it's "good enough" to handle running the engine.
NOW, it's a matter of what programming you're doing... Sometimes, just "changing the maps" doesn't always work, as there are lots of watchdog and range limitations coded in there typically. This means if you adjust something that the code doesn't like, it's either going to barf or revert back to a "backup map". Thus, "ROM tuning" does NOT change parameters in very extreme values.  Fuel maps are typically not adjusted past 20%, and ignition timing isn't adjusted past 5-degrees either way.
-Ted
I don't know a lot about eproms and stuff, but I have an HC11 simulator and manual. I might look around for an ECU near me. Perhaps I can help.
Last edited by Canadian Rotary Man; Feb 1, 2004 at 02:30 AM.
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Originally posted by Canadian Rotary Man
I'm glad to see you guys are trying this, it's something I've been interested in. I didn't know these ECU's used a variant of the HC11. Last year I learned how to code motorola HC11 chips in assembly language and machine code. Right now I'm in my 2nd processor design course at Simon Fraser University(3rd year physics)
I don't know a lot about eproms and stuff, but I have an HC11 simulator and manual. I might look around for an ECU near me. Perhaps I can help.
I'm glad to see you guys are trying this, it's something I've been interested in. I didn't know these ECU's used a variant of the HC11. Last year I learned how to code motorola HC11 chips in assembly language and machine code. Right now I'm in my 2nd processor design course at Simon Fraser University(3rd year physics)
I don't know a lot about eproms and stuff, but I have an HC11 simulator and manual. I might look around for an ECU near me. Perhaps I can help.
I'd watch the for sale section for someone parting out their car. It's pretty easy to find used ECU's on this board.
Junior Member
Joined: Aug 2001
Posts: 20
Likes: 0
From: Tampa, FL
loudroom,
It will be easier to figure out analog parts with some idea of circuitry. I think I have access to a N327 ecu myself, and will start looking a bit closer at it this week. if I have to, I'll just start pulling parts and running tests to see "what it is".
Is anyone considering building a simulator like the ecu wonder tester @ http://www.16paws.com/ECU/WonderTester/index.html
I'd think it would help figure out somethings if we could input various signals to simulate idle, part throttle, WOT, steady rpm, deceleration, sensor failures, etc...
while monitoring the memory addressing, etc... to make sure we had everything nailed as opposed to guessing. Well, since I also have a spare harness, I guess I should start on that part. ANy inputs are welcome. Take a look at the 16paws tester and see if you want more or less. I was thinking about making an in-car ecu monitor when I picked up these things a couple of years ago. If I build this right, it may work for both in-car and satndalone benchtop testing. The only difference would be cutting in dummy loads\voltages for benchtop use, which wouldn't be hard.
keep it up guys,
sunrotor
It will be easier to figure out analog parts with some idea of circuitry. I think I have access to a N327 ecu myself, and will start looking a bit closer at it this week. if I have to, I'll just start pulling parts and running tests to see "what it is".
Is anyone considering building a simulator like the ecu wonder tester @ http://www.16paws.com/ECU/WonderTester/index.html
I'd think it would help figure out somethings if we could input various signals to simulate idle, part throttle, WOT, steady rpm, deceleration, sensor failures, etc...
while monitoring the memory addressing, etc... to make sure we had everything nailed as opposed to guessing. Well, since I also have a spare harness, I guess I should start on that part. ANy inputs are welcome. Take a look at the 16paws tester and see if you want more or less. I was thinking about making an in-car ecu monitor when I picked up these things a couple of years ago. If I build this right, it may work for both in-car and satndalone benchtop testing. The only difference would be cutting in dummy loads\voltages for benchtop use, which wouldn't be hard.
keep it up guys,
sunrotor
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
I'll be posting schematics soon, but I'm still finishing them up. I know I keep saying that over and over, but just when I think I'm about done, I realize I'm not. (That and I had to replace the starter on my winter rat.) Maybe I'll just scan what I have while I'm at work tomorrow, so you can at least have something to look at. At this point I have every ECU pin traced and every connection to the ribbon cable traced, but I still need to ring out a few pins on various IC's. I also haven't worked on the 5V supply. I'm not really in a hurry to figure that one out since I already know what it should be doing.
As far as building an ECU tester goes, I've considered it, but I don't think it will be necessary just to determine what these components are. I should be able to just hook up 12V to the ECU and then only stimulate the particular circuit I'm looking at. If that won't work then we can just pull chips off the ECU and test them like you said.
When we start modifying software, it would be very nice to have a full-fledged rx-7 simulator. I should probably start looking for my own spare wiring harness that I can hack up.
The setup on the 16paws site looks decent, with the exception that it's using a real CAS. It would probably be a good idea to design a little circuit to mimic the CAS output at various RPM's (controlled via a potentiometer). This would allow us to use an oscilloscope to verify things like the fuel injector and spark signals are functioning properly before sticking a modified ECU in a real RX-7. I like how he had pots for all the analog sensors, but that may end up being overkill.
As far as building an ECU tester goes, I've considered it, but I don't think it will be necessary just to determine what these components are. I should be able to just hook up 12V to the ECU and then only stimulate the particular circuit I'm looking at. If that won't work then we can just pull chips off the ECU and test them like you said.
When we start modifying software, it would be very nice to have a full-fledged rx-7 simulator. I should probably start looking for my own spare wiring harness that I can hack up.
The setup on the 16paws site looks decent, with the exception that it's using a real CAS. It would probably be a good idea to design a little circuit to mimic the CAS output at various RPM's (controlled via a potentiometer). This would allow us to use an oscilloscope to verify things like the fuel injector and spark signals are functioning properly before sticking a modified ECU in a real RX-7. I like how he had pots for all the analog sensors, but that may end up being overkill.
Last edited by theloudroom; Feb 2, 2004 at 12:25 AM.
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Schematics, come 'n get 'um!
Schematics here:
http://24.175.236.46/mark/Binder2.pdf
The details:
These are still incomplete, but complete enough to answer many questions.
I still need to add a few pages showing miscellaneous IC connections and the power supply.
The fine print:
These schematics are copyright me. I will release them under a copyleft liscense when I find the right one (the GPL doesn't seem right for schematics), but for now "all rights reserved"
Conventions:
CONN1, means wire 1 on the internal ribbon cable.
2B, means connector pin 2B, as shown in the FSM.
Gotchas:
One of the pages is upside down
My bad.
Schematics here:
http://24.175.236.46/mark/Binder2.pdf
The details:
These are still incomplete, but complete enough to answer many questions.
I still need to add a few pages showing miscellaneous IC connections and the power supply.
The fine print:
These schematics are copyright me. I will release them under a copyleft liscense when I find the right one (the GPL doesn't seem right for schematics), but for now "all rights reserved"
Conventions:
CONN1, means wire 1 on the internal ribbon cable.
2B, means connector pin 2B, as shown in the FSM.
Gotchas:
One of the pages is upside down
My bad.
Senior Member
Joined: Jun 2003
Posts: 275
Likes: 0
From: Colorado
just want to post I made a forum just for this on my New website www.rotorygod.com mabe start posting there? and get this going some more!
Daily Domestic Killer
Joined: Mar 2001
Posts: 2,425
Likes: 0
From: San Antonio, Tx, USA
Okay Well I am able to get my hands on an e-prom reader from school, so I can pull the code off my supposedly "aftermarket" chip that I bought off ebay
my Question is in what format should i pull it off
binary or something else? sooner I get answers the sooner I post the data
my Question is in what format should i pull it off
binary or something else? sooner I get answers the sooner I post the data
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Originally posted by BlackRx7
Okay Well I am able to get my hands on an e-prom reader from school, so I can pull the code off my supposedly "aftermarket" chip that I bought off ebay
my Question is in what format should i pull it off
binary or something else? sooner I get answers the sooner I post the data
Okay Well I am able to get my hands on an e-prom reader from school, so I can pull the code off my supposedly "aftermarket" chip that I bought off ebay
my Question is in what format should i pull it off
binary or something else? sooner I get answers the sooner I post the data
Senior Member
Joined: Sep 2003
Posts: 488
Likes: 0
From: Syracuse, New York
Schematics for the analog board are now done. It turns out there were a couple errors in the partile schematics I've posted.
I'll have to wait 'til I get to a scanner to post them though.
My next task is a full set of schematics for the digital board. There are some refrences (in the ECU code) to memory addresses that aren't mapped on the 16paws memory map. I'm going to see if he missed something.
I'll have to wait 'til I get to a scanner to post them though.
My next task is a full set of schematics for the digital board. There are some refrences (in the ECU code) to memory addresses that aren't mapped on the 16paws memory map. I'm going to see if he missed something.