View Full Version : SL ECU Major Differences or Why NOT to Use an SL ECU
99 vr4
07-17-2011, 05:54 PM
SL ECU Major Differences or Why NOT to Use an SL ECU
As some of you know, I have been working with a team of people for almost a year trying to find a low cost solution for a flashable ECU on the 3000GT platform.
I have tried many different "donor" ECUs from other V6 platforms, Mitsubishi, Hyundai, and Mazda (yes Mitsu makes thier ECUs). I have had to reverese engineer circuits that are critical for the ability of the ECU to correctly manage the engine with out harm.
I know that some of you have bought or are buying SL ECUs thinking that this will work. Let me tell you that it is not a 100% solution. In fact, it is not even a 90% solution. It is a bad idea. Why?
Here are a few of the issues that have been identified, not in any particular order.
1. No stock boost gauge. Simple fix, replace your now non-functioning boost gauge with and aftermarket boost gauge.
2. No stock two speed fuel pump control - Fuel pump stuck on HI. No biggie you think, right? Wrong . . . see the next item.
3. No stock fuel pressure solenoid control. Oh Oh. The stock ECU controls the pressure of the fuel being delivered through a two stage system. First you have the two stage fuel pump (above). Then you have a Fuel Pressure Regulator on the end of the rail. And lastely you have the fuel pressure requlator controlled through a selonoid. This selonoid controls the amount of boost pressure that the FPR sees on the boost compensation port. If this is controlled incorrectly or not at all you can have either to high pressure (no big deal) or not enough pressure (big deal). How do you fix this issue? Install an aftermarket fuel pressure regulator. Solves both problems.
4. No wastegate control. Mitsubishi controls the amount of boost that the wastegate sees through a selonoid. The SL ECU does not have this circuitry on the board. How do you fix this issue? Install an aftermarket boost controller.
5. Knock Sensor! Ok, because of the design of the engine, Mitsubishi handles the knock sensor input differently. Although the knock sensor it self is used on many platforms, the twin turbo has a different hardware controlled circuit. There are significant hardware differences between the turbo and non-turbo ECUs which results in both "phantom" knock and the inability for the ECU to see "real" knock. What this means is when the ECU sees phantom knock, it pulls time reducing performance. When the ECU fails to see real knock, the knock continues and will slowly trash your bearings until such time the knock is loud enough that you can hear it and say "oh shit" or your bearings let you know what is up - then you join the spun bearing club. What can cause this knock? Lack of fuel caused by bad fuel pressure, or a bad tune are two of the many culprits. How do you fix it? You don't. The ECU needs to be modified.
Jesters Deadd
07-18-2011, 01:19 AM
SL ECU Major Differences or Why NOT to Use an SL ECU
As some of you know, I have been working with a team of people for almost a year trying to find a low cost solution for a flashable ECU on the 3000GT platform.
I have tried many different "donor" ECUs from other V6 platforms, Mitsubishi, Hyundai, and Mazda (yes Mitsu makes thier ECUs). I have had to reverese engineer circuits that are critical for the ability of the ECU to correctly manage the engine with out harm.
I know that some of you have bought or are buying SL ECUs thinking that this will work. Let me tell you that it is not a 100% solution. In fact, it is not even a 90% solution. It is a bad idea. Why?
Here are a few of the issues that have been identified, not in any particular order.
1. No stock boost gauge. Simple fix, replace your now non-functioning boost gauge with and aftermarket boost gauge.
2. No stock two speed fuel pump control - Fuel pump stuck on HI. No biggie you think, right? Wrong . . . see the next item.
3. No stock fuel pressure solenoid control. Oh Oh. The stock ECU controls the pressure of the fuel being delivered through a two stage system. First you have the two stage fuel pump (above). Then you have a Fuel Pressure Regulator on the end of the rail. And lastely you have the fuel pressure requlator controlled through a selonoid. This selonoid controls the amount of boost pressure that the FPR sees on the boost compensation port. If this is controlled incorrectly or not at all you can have either to high pressure (no big deal) or not enough pressure (big deal). How do you fix this issue? Install an aftermarket fuel pressure regulator. Solves both problems.
4. No wastegate control. Mitsubishi controls the amount of boost that the wastegate sees through a selonoid. The SL ECU does not have this circuitry on the board. How do you fix this issue? Install an aftermarket boost controller.
5. Knock Sensor! Ok, because of the design of the engine, Mitsubishi handles the knock sensor input differently. Although the knock sensor it self is used on many platforms, the twin turbo has a different hardware controlled circuit. There are significant hardware differences between the turbo and non-turbo ECUs which results in both "phantom" knock and the inability for the ECU to see "real" knock. What this means is when the ECU sees phantom knock, it pulls time reducing performance. When the ECU fails to see real knock, the knock continues and will slowly trash your bearings until such time the knock is loud enough that you can hear it and say "oh shit" or your bearings let you know what is up - then you join the spun bearing club. What can cause this knock? Lack of fuel caused by bad fuel pressure, or a bad tune are two of the many culprits. How do you fix it? You don't. The ECU needs to be modified.
What!??? A bad idea? really? Having an inexpensive completely tuneable ECU that is nearly plug and play with an adaptor harness is a bad idea?
I have tested it on my car and IT WORKS GREAT!
1. No stock boost gauge... the stock boost gauge doesn't even measure boost, you should have a real boost guage anyway!
2. No stock two speed fuel pump control... hotwired fuel pump. This is like stage one upgrade stuff.
3. No stock fuel pressure solenoid control...Install an aftermarket fuel pressure regulator and hook up the vacuum from the intake. The fuel pressure solenoid only cuts vacuum to the fuel pressure regulator during hot starts. Most people have already disconnected this with no noticable affect.
4. No wastegate control... install a manual boost controller or a better one truboost etc.. Again this is stage one upgrade stuff.
5. Knock Sensor!... I tested this on my VR4 and got the same amount of knock on the same program as my 99VR4 ecu at the same locations... it wasn't 100% scientific but it was very repeatable...
Basically have a boost gauge, hotwired fuel pump with an adjustable fuel pressure regulator and a boost controller and you are set with an 98/99 SL ECU. This is all stuff you will need for stage one upgrades. You will have all the tuning ability of the 98/99VR4 for much cheaper than a 99VR4 ECU. Just ask questions and know what you are doing before you do it.
99 vr4
07-18-2011, 02:34 AM
Trust me when I say the knock sensor circuits are COMPLETELY different. Obviously for an experienced tuner such as yourself this is a non-issue. But for the majority of DIY types that are not experienced tuners . . . you should not "play".
The best way to explain this in normal non technical terms is like this. The knock sensor is a piezo element which produces electrical "noise" in response to physical vibrations. Very much like a microphone produces electrical signals from a sound wave. The hardware on the circuit board consists of filter circuits and amplifier circuits. Very much like an audio amplifier. Now lets say you just bought a new 300 watt sub amp for you car audio. It can drive a subwoofer from 20 hz to 140 hz. But your subwoofer is only rated at 20 hz to 100 hz. So you put a band pass filter in line with the speaker. But you miscalculate and your band pass filter only allows 100 hz to 140 hz through. Will it work? Yes. Will it work well? No. Will you blow your speaker . . . time will tell.
Drummer
07-22-2011, 02:38 AM
So I can just replace the few circuits with the correct value resistors and cap and the knock will read correctly? I hope!
Jesters Deadd
07-22-2011, 11:38 AM
So I can just replace the few circuits with the correct value resistors and cap and the knock will read correctly? I hope!
That would be hard to do... you need the correct values and you would need to be able to solder surface mount components...
Jesters Deadd
07-22-2011, 11:47 AM
Just FYI to keep info together...
Jester, from what I can see is that the knock issue with the Diamante and Montero ECU's seems to affect some people more than others. Some people swear they don't get knock, and others get all sorts of issues with these ECUs. Knock itself is not always able to be duplicated. Lets say some guy in NY is getting knock. I could load his tune and not get any knock. There is so much more here than just the tune and fuel delivery.
You have to take into account fuel formulations. Not all gas is equal. And even if you buy Amoco Gold here in St. Louis you can bet it is not formulated the same as Amoco Gold in Florida or New York. Although the EPA mandates RFG vs CG in high polution areas, they also limit the Reed Vapor Pressure (basically the rate of boil off) so RFG will be formulated differently in the winter vs. summer vs. elevation for each major area. Frequent Questions | Reformulated Gasoline | US EPA (http://www.epa.gov/oms/rfg/faq.htm) Then you have to take into count air temperature, density, barometric pressure (elevation), and humidity. Throw all of this into the mix with your tune and thats how you might not get knock or you might get knock. :)
So although I am sure that your real world test yielded no additional knock, its not 100% because there are more variables. As far as the filter circuits are concerned, I am sure that Mitsubishi, if they could have cut corners (and cost) by using the same knock sensor circuits, they would have. Even though we are talking a small handful of parts, the engineers designed it, tested it, specified it, and signed off on it. It is there for a reason.
As far as working the math for the filter circuits, I hate laplace transforms and I don't see myself working out that math any time soon.
Interesting information per the 99 service manual...
"The knock sensor generates a voltage proportional to the magnitude of cylinder block vibration due to knocking and inputs it to the engine control module. Based on this signal, the engine control module provides retard control of the ignition timing."
So maybe it doesn't need Laplace transforms just a little V=IR.
Drummer
07-24-2011, 06:53 PM
That would be hard to do... you need the correct values and you would need to be able to solder surface mount components...
Thats fine. The values and locations are known, and surface mounted soldering is not a problem.
Drummer
07-24-2011, 07:29 PM
Trust me when I say the knock sensor circuits are COMPLETELY different. Obviously for an experienced tuner such as yourself this is a non-issue. But for the majority of DIY types that are not experienced tuners . . . you should not "play".
The best way to explain this in normal non technical terms is like this. The knock sensor is a piezo element which produces electrical "noise" in response to physical vibrations. Very much like a microphone produces electrical signals from a sound wave. The hardware on the circuit board consists of filter circuits and amplifier circuits. Very much like an audio amplifier. Now lets say you just bought a new 300 watt sub amp for you car audio. It can drive a subwoofer from 20 hz to 140 hz. But your subwoofer is only rated at 20 hz to 100 hz. So you put a band pass filter in line with the speaker. But you miscalculate and your band pass filter only allows 100 hz to 140 hz through. Will it work? Yes. Will it work well? No. Will you blow your speaker . . . time will tell.
That sub would last forever cause its getting higher frequencies and not working as hard as it could ;-) J/K lol But yea, someone was saying that the SL ecu has a slightly more sensetive knock circuit and could produce fantom knock. I have not opened up the 98 SL's I have yet, nor do I have a reference VR4 ECU.
green-lantern
07-24-2011, 07:39 PM
IDK, I wish this would get resolved because I've been told by a very reliable source that the circuitry is exactly the same.
Monster Stealth
07-24-2011, 09:02 PM
subscribed
mb3000
07-25-2011, 01:56 AM
Subscribed.
Drummer
07-25-2011, 02:51 AM
IDK, I wish this would get resolved because I've been told by a very reliable source that the circuitry is exactly the same.
He has seen many of these ecus open and has a reference ecu. I cant see a reason to say otherwise, cause its easy to take a look. I too have an sl ecu and can get access to a reference 99 ecu. Ill take his word that the sl may be a more sensetive circuit. If for some reason its not, ill see that later. The real work on this project is in the ecuflash file. Keep up the good work! Brett, when will the ecus be ready for sale?
99 vr4
07-25-2011, 01:07 PM
I have the knock issue resolved.. I also have the timing issue resolved.. When the ecu is from a distributor based car.. I am working on the o2 circuits..
Basically the 3000 dohc na and turbo use 4 o2 sensors.. All others use two, one front and one rear. Modifying the circuit to use two front is the last major hurdle.. Right now the montero and diamante ecus are only reading the front bank o2 and we all know that is not ideal.. I should have it soon..
99 vr4
07-25-2011, 01:09 PM
Also I have the reason why you see ghost o2 voltages when there are no sensors.. It's inerrant in the hard circuitry.. I will provide a more detailed explanation later
99 vr4
07-25-2011, 01:11 PM
He has seen many of these ecus open and has a reference ecu. I cant see a reason to say otherwise, cause its easy to take a look. I too have an sl ecu and can get access to a reference 99 ecu. Ill take his word that the sl may be a more sensetive circuit. If for some reason its not, ill see that later. The real work on this project is in the ecuflash file. Keep up the good work! Brett, when will the ecus be ready for sale?
If you have 98/99 sl ecus I can modify them for the knock circuitry..
Jesters Deadd
07-25-2011, 02:25 PM
I have the knock issue resolved.. I also have the timing issue resolved.. When the ecu is from a distributor based car.. I am working on the o2 circuits..
Basically the 3000 dohc na and turbo use 4 o2 sensors.. All others use two, one front and one rear. Modifying the circuit to use two front is the last major hurdle.. Right now the montero and diamante ecus are only reading the front bank o2 and we all know that is not ideal.. I should have it soon..
Yeah so I took the diamante ecu put it in my 99VR4 and tested the different O2 sensors readings. I was concerned about the missing O2 sensors on the diamante. Since I have two Lc1 wb o2s I can manipulate each of the 4 channels coming into the ECU. Well long story short the daimante ECu reads the front and rear upper (precat) o2 signals correctly. I couldn't really test the postcat o2 signals because I don't know the MUTIII readout for them.
also the 98/99 sohc 3000gt had four O2 sensors... so the 98/99 SOHC 3000gt would be better than the Diamante ECU. It would be the same as 98/99 SL ECU if the start timing issue is fixed.
green-lantern
07-25-2011, 09:17 PM
Knock circuit layout
1998 3000GT SL
MMC P/N MD347251
MFG P/N E2T61389 H3
PCB P/N JE331B157D
Con1-Pin 91 > R314 (TOP) > C45 (BOTTOM) > R315 (TOP) > IC2-Pin 78 < IC2-Pin 80 > IC1-Pin 92
R314 = 51 KOhm
C45 = 0.02 uF
R315 = 6.2 KOhm
IC2 = Mitsu Electronics E328 logic level converter
IC1 = Renesas H8/539F microcontroller, A/D input 6
1999 3000GT VR4
MMC P/N MD360163
MFG P/N E2T61495 H1
PCB P/N JE331B32804B
Con1-Pin91 > R71 (TOP) > C62 (BOTTOM) > R73 (TOP) > IC7-Pin 78 < IC7-Pin80 > IC9-Pin 92
R71 = 51 KOhm
C62 = 0.02 uF
R73 = 6.2 KOhm
IC7 = Mitsu Electronics E328 logic level converter
IC9 = Renesas H8/539F microcontroller, A/D input 6
CoopKill
07-25-2011, 11:44 PM
I have the knock issue resolved.. I also have the timing issue resolved.. When the ecu is from a distributor based car.. I am working on the o2 circuits..
Basically the 3000 dohc na and turbo use 4 o2 sensors.. All others use two, one front and one rear. Modifying the circuit to use two front is the last major hurdle.. Right now the montero and diamante ecus are only reading the front bank o2 and we all know that is not ideal.. I should have it soon..
How will this affect those of us that only have two o2's?
Jesters Deadd
07-26-2011, 02:08 PM
How will this affect those of us that only have two o2's?
Coop, I tested this. You can run it by just deselecting some of the peripheries in the code. Since I have two Lc1 wb o2s I can manipulate each of the 4 channels coming into the ECU. Well long story short the daimante ECu reads the front and rear upper (precat) o2 signals correctly. I couldn't really test the postcat o2 signals because I don't know the MUTIII readout for them. But it also depends how you set up the O2s. If you go front and then back bank it works fine. If you go pre cat and post cat there is a periph that changes the two banks to osilate on the same O2. but testing needs to confirm this still.
Jesters Deadd
07-26-2011, 02:09 PM
Knock circuit layout
1998 3000GT SL
MMC P/N MD347251
MFG P/N E2T61389 H3
PCB P/N JE331B157D
Con1-Pin 91 > R314 (TOP) > C45 (BOTTOM) > R315 (TOP) > IC2-Pin 78 < IC2-Pin 80 > IC1-Pin 92
R314 = 51 KOhm
C45 = 0.02 uF
R315 = 6.2 KOhm
IC2 = Mitsu Electronics E328 logic level converter
IC1 = Renesas H8/539F microcontroller, A/D input 6
1999 3000GT VR4
MMC P/N MD360163
MFG P/N E2T61495 H1
PCB P/N JE331B32804B
Con1-Pin91 > R71 (TOP) > C62 (BOTTOM) > R73 (TOP) > IC7-Pin 78 < IC7-Pin80 > IC9-Pin 92
R71 = 51 KOhm
C62 = 0.02 uF
R73 = 6.2 KOhm
IC7 = Mitsu Electronics E328 logic level converter
IC9 = Renesas H8/539F microcontroller, A/D input 6
Green, where did you get that info?
green-lantern
07-26-2011, 03:36 PM
Green, where did you get that info?
Bored At Work Crew, a very reliable source ;)
Drummer
07-26-2011, 05:14 PM
Hmm... Brett, what do you think of the info that green lantern has provided?
Jesters Deadd
07-26-2011, 05:28 PM
Bored At Work Crew, a very reliable source ;)
Do you have the difference info on the starting timing issue between the ditributor ECUs ie Diamante and SOHC ECUs and distributorless ones ie VR4 and DOHC NAs?
green-lantern
07-26-2011, 06:32 PM
Do you have the difference info on the starting timing issue between the ditributor ECUs ie Diamante and SOHC ECUs and distributorless ones ie VR4 and DOHC NAs?
Sorry I haven't been involved with that and isn't my info to share.
Drummer
07-27-2011, 04:32 PM
I was under the impression that you can disable the extra 2 02 sensors in the flash in order for the ECU to only read the 2 that are on my 94' stealth TT. Is this true?
Jesters Deadd
07-27-2011, 06:39 PM
I was under the impression that you can disable the extra 2 02 sensors in the flash in order for the ECU to only read the 2 that are on my 94' stealth TT. Is this true?
I believe so... there are some that disagree with me about a certain periphery... but this is what I got when i tested it.
You have options, depends on the periphs you flip.
You can have all four 02s, 2 precat and 2 post cat for each bank.
You can have both front and back precat O2s only and turn off the post cat o2s.
You can have both the front and back bank trim on one precat O2 and then have just one post cat O2.
You can have just one precat O2 and have the front and back oscillate off the same O2. (this is untested just my theory)
All this with no CEL.
Simace
07-27-2011, 06:47 PM
Another option, if you don't have precats, is to use anti-plug foulers on the O2's just after the precats.
Drummer
07-28-2011, 03:03 AM
I will have no precats. I will have a complete turbo back exhaust, with the rear precat gutted. I have to get anti foulers for that your saying?
Jesters Deadd
07-28-2011, 11:45 AM
I will have no precats. I will have a complete turbo back exhaust, with the rear precat gutted. I have to get anti foulers for that your saying?
NO you don't need plug foulers... you can disable em in the code.
Personally I would get two LC1s, because they work awesome and are only like $150 each. Then you can see what each bank is doing.
green-lantern
07-28-2011, 12:35 PM
NO you don't need plug foulers... you can disable em in the code.
Personally I would get two LC1s, because they work awesome and are only like $150 each. Then you can see what each bank is doing.
:stupid:
Drummer
07-28-2011, 05:44 PM
I have an Innovated WB02 MTXL with guage ready to go. I was just going to refresh the NB02's and set the flash to see only those 2, and use the wideband to check up. The WB goes right after the F-pipe joins. What do you think of that? I will have 2 NB stock and 1 WB further down... ?
Your saying 2 WB's in the stock positions connected to the ECU or right to a logger? Forgive my lack of intelligence on this one...
green-lantern
07-28-2011, 06:23 PM
I have an Innovated WB02 MTXL with guage ready to go. I was just going to refresh the NB02's and set the flash to see only those 2, and use the wideband to check up. The WB goes right after the F-pipe joins. What do you think of that? I will have 2 NB stock and 1 WB further down... ?
Your saying 2 WB's in the stock positions connected to the ECU or right to a logger? Forgive my lack of intelligence on this one...
You will be fine with that. It's good to be able to have one in the front part of the down pipe and one and the rear before they join but really if you don't have the bungs already or the ability to just add another sensor then I wouldn't worry about it, I'm not.
Just log the one you have.
Jesters Deadd
07-29-2011, 11:54 AM
I have an Innovated WB02 MTXL with guage ready to go. I was just going to refresh the NB02's and set the flash to see only those 2, and use the wideband to check up. The WB goes right after the F-pipe joins. What do you think of that? I will have 2 NB stock and 1 WB further down... ?
Your saying 2 WB's in the stock positions connected to the ECU or right to a logger? Forgive my lack of intelligence on this one...
The two NBs, one for each bank are good and will work fine. You can just hook the WB into the BSLCDBC. Then you won't have redundant gauges... just saying.
Drummer
07-30-2011, 05:05 PM
The two NBs, one for each bank are good and will work fine. You can just hook the WB into the BSLCDBC. Then you won't have redundant gauges... just saying.
What is a BSLCBC? Thnx for the replies also! Cant wait to get my VR4 ECU....
Jesters Deadd
07-30-2011, 05:44 PM
What is a BSLCBC? Thnx for the replies also! Cant wait to get my VR4 ECU....
oh sorry, Black Stealth LCD Boost Controller
Drummer
07-30-2011, 07:39 PM
Ahh yes, just log it right to his BC. Sounds good.
prop7459
07-31-2011, 12:38 AM
How about intergating the 99sl knock sensor if your theory is right about false knock readings.
Drummer
07-31-2011, 03:49 PM
When you say integrate, do you mean alter a parameter in the flash in order to correctly read the SL circuit? Or modifying the circuit?
prop7459
07-31-2011, 06:18 PM
Passing the some parameters to mimic the knock function of a 98-98 VR4 car. I would like to avoid modifying circuits to prevent frying the ECU.
green-lantern
07-31-2011, 09:54 PM
Passing the some parameters to mimic the knock function of a 98-98 VR4 car. I would like to avoid modifying circuits to prevent frying the ECU.
You don't need to modify the circuits
Knock circuit layout
1998 3000GT SL
MMC P/N MD347251
MFG P/N E2T61389 H3
PCB P/N JE331B157D
Con1-Pin 91 > R314 (TOP) > C45 (BOTTOM) > R315 (TOP) > IC2-Pin 78 < IC2-Pin 80 > IC1-Pin 92
R314 = 51 KOhm
C45 = 0.02 uF
R315 = 6.2 KOhm
IC2 = Mitsu Electronics E328 logic level converter
IC1 = Renesas H8/539F microcontroller, A/D input 6
1999 3000GT VR4
MMC P/N MD360163
MFG P/N E2T61495 H1
PCB P/N JE331B32804B
Con1-Pin91 > R71 (TOP) > C62 (BOTTOM) > R73 (TOP) > IC7-Pin 78 < IC7-Pin80 > IC9-Pin 92
R71 = 51 KOhm
C62 = 0.02 uF
R73 = 6.2 KOhm
IC7 = Mitsu Electronics E328 logic level converter
IC9 = Renesas H8/539F microcontroller, A/D input 6
If the software is different on the SL it gets replaced when the VR4 stuff is flashed onto the ECU.
Drummer
08-01-2011, 12:44 AM
Well if green is right, then once the sl ecu is flashed to vr4, it will use the circuit like a vr4 should. But 99vr4 says there are differences. Need the final answer here!
Any idea what the ecu's will cost when ready for sale as a kit? Flashed for the correct 99 revision, and ready for install with harness?
Jesters Deadd
08-01-2011, 12:50 PM
Well if green is right, then once the sl ecu is flashed to vr4, it will use the circuit like a vr4 should. But 99vr4 says there are differences. Need the final answer here!
Any idea what the ecu's will cost when ready for sale as a kit? Flashed for the correct 99 revision, and ready for install with harness?
Worked fine for me... just saying...again. NO one else has said it hasn't worked for them.
95gto
08-01-2011, 01:51 PM
Well if green is right, then once the sl ecu is flashed to vr4, it will use the circuit like a vr4 should. But 99vr4 says there are differences. Need the final answer here!
Any idea what the ecu's will cost when ready for sale as a kit? Flashed for the correct 99 revision, and ready for install with harness?
There is a hardware difference in the knock circuits and the easiest way to think about the issue would be to think of it like having a non-matched gauge unit and sending unit.
So say you have an Apexi oil pressure gauge (round unit with the needle) but a Greddy oil pressure sending unit. The two parts may work well enough together that it is not an issue (ie it could read accurately enough) but I would personally prefer to use a matched gauge and sending unit.
The difference in the knock circuit is similar in that it may work well enough but its not correctly matched. Whether or not you want the matched set or the "should be okay" set is your call.
Drummer
08-01-2011, 02:03 PM
There is a hardware difference in the knock circuits and the easiest way to think about the issue would be to think of it like having a non-matched gauge unit and sending unit.
So say you have an Apexi oil pressure gauge (round unit with the needle) but a Greddy oil pressure sending unit. The two parts may work well enough together that it is not an issue (ie it could read accurately enough) but I would personally prefer to use a matched gauge and sending unit.
The difference in the knock circuit is similar in that it may work well enough but its not correctly matched. Whether or not you want the matched set or the "should be okay" set is your call.
Ok I get that, but what Im wondering about is how they differ. If greens information is correct, the sensor circuits are identical. Is there something else in the path that is not? Another part of the circuit? Im just trying to make sense of it. Perhaps the sensor itself is not identical in older models? I have not yet opened up my SL ECU, and I now own a VR4 ECU. So Im just asking ahead of time to save me the trouble. I will take a look personally in the end, as I have the correct values and locations.
95gto
08-01-2011, 02:46 PM
Ok I get that, but what Im wondering about is how they differ. If greens information is correct, the sensor circuits are identical. Is there something else in the path that is not? Another part of the circuit? Im just trying to make sense of it. Perhaps the sensor itself is not identical in older models? I have not yet opened up my SL ECU, and I now own a VR4 ECU. So Im just asking ahead of time to save me the trouble. I will take a look personally in the end, as I have the correct values and locations.
Based on a side by side comparison the knock circuits on the ecu boards, vr4 versus sl, are not identical. There are differences in the surface mount hardware and that circuity impacts the knock signal your ecu sees from the knock sensor. So going back to my metaphor, the ecu = the gauge and the sender is the knock circuit.
I'm not sure who Paul's very reliable source is but I know that Brett is first hand comparing a VR4 ecu to the others, not to mention that Paul's source may be some of the folks that Brett is already talking to.
green-lantern
08-01-2011, 03:45 PM
I'm not sure who Paul's very reliable source is but I know that Brett is first hand comparing a VR4 ecu to the others, not to mention that Paul's source may be some of the folks that Brett is already talking to.
It’s Jeff, it’s why I’m a bit baffled that I’m seeing conflicting information about it. It’s not like these guys haven’t talked to each other a lot about the ECU research.
prop7459
08-01-2011, 04:42 PM
If this is the case, then why not stick with the 99 SL Rom and use the 99 knock sensor to prevent false readings?
ChargerX3
08-01-2011, 05:08 PM
Interesting to see if this phantom knock will make itself known. Ive got a 98 Diamante ecu and a 99 vr4 ecu on the way. Will be glad to test this all out once i get hard wired in.
95gto
08-01-2011, 05:21 PM
It’s Jeff, it’s why I’m a bit baffled that I’m seeing conflicting information about it. It’s not like these guys haven’t talked to each other a lot about the ECU research.
That's what I figured so I will bug Brett again about it.
95gto
08-01-2011, 05:23 PM
If this is the case, then why not stick with the 99 SL Rom and use the 99 knock sensor to prevent false readings?
The issue isn't the sensor, the issue is the circuit on the ecu board. The knock sensor output does not directly go to a chip in the ecu there is other hardware in the circuit that the signal has to go through first.
prop7459
08-01-2011, 05:51 PM
The issue isn't the sensor, the issue is the circuit on the ecu board. The knock sensor output does not directly go to a chip in the ecu there is other hardware in the circuit that the signal has to go through first.
So you are telling if I used the 99 SL ecu with a 99 SL Rom(reflashed to my needs) with the 99SL Knock sensor that they will be still an issue with false knock readings. Have you guys tested a TT car with a 99 SL ECU and 99 SL knock sensor to confirm this? What other hardware you are talking about that is part of the knock system?
prop7459
08-01-2011, 05:54 PM
Interesting to see if this phantom knock will make itself known. Ive got a 98 Diamante ecu and a 99 vr4 ecu on the way. Will be glad to test this all out once i get hard wired in.
Lucky you, my stealth bite the dust after doing a compression test with three dead cylinders. Are you going to make your patch cable or have boomslang make you one?
ChargerX3
08-01-2011, 06:01 PM
Lucky you, my stealth bite the dust after doing a compression test with three dead cylinders. Are you going to make your patch cable or have boomslang make you one?
I plan on doing my own. Thanks to some pin out info on here it should be pretty much straight forward.
green-lantern
08-01-2011, 07:00 PM
That's what I figured so I will bug Brett again about it.
Thanks, I respect Brett a lot and I consider Jeff a pretty good friend. I talk to Jeff pretty much every day on BAWC so he happens to answer a lot of my questions I have about this. I have an interest because I plan to use an SL ECU and I don't want this site to have the wrong info. This place needs all the help it can get.
mb3000
08-01-2011, 07:07 PM
Part #'s on ASA are the same for both sensors.
MD159216
Drummer
08-01-2011, 09:41 PM
Well now we know that the sensor itself is the same. It makes sense that the 2 engines are read differently on the ECU end. I can imagine that noise readings from an SL will vary from the readings on a TT. Considering a fair bit more "noise" is coming from a TT, the SL circuitry could be a little more sensitive, and understandably so.
I realize that someone is very sure that the circuit is different. They probably do not want to post that info because this kit can be a money maker. Again, thats understandable. All of the ECU's that are not an original vr4 have their issues. Id like to see the price of a kit soon though. Harness+Flash cable+ECU+disc with all the software, logger too. That would be a cool kit!
95gto
08-02-2011, 10:47 AM
So you are telling if I used the 99 SL ecu with a 99 SL Rom(reflashed to my needs) with the 99SL Knock sensor that they will be still an issue with false knock readings. Have you guys tested a TT car with a 99 SL ECU and 99 SL knock sensor to confirm this? What other hardware you are talking about that is part of the knock system?
If you have an SL and you use a 99' SL ecu your knock will read correctly. If you have a TT car and use a 99' SL ecu that is unmodified you will not get accurate knock readings, you would get phantom knock.
Thanks, I respect Brett a lot and I consider Jeff a pretty good friend. I talk to Jeff pretty much every day on BAWC so he happens to answer a lot of my questions I have about this. I have an interest because I plan to use an SL ECU and I don't want this site to have the wrong info. This place needs all the help it can get.
I called Brett yesterday and told him to talk to Jeff so that we could get a consistent answer. I would agree that we need to provide as accurate of information as possible.
Well now we know that the sensor itself is the same. It makes sense that the 2 engines are read differently on the ECU end. I can imagine that noise readings from an SL will vary from the readings on a TT. Considering a fair bit more "noise" is coming from a TT, the SL circuitry could be a little more sensitive, and understandably so.
I realize that someone is very sure that the circuit is different. They probably do not want to post that info because this kit can be a money maker. Again, thats understandable. All of the ECU's that are not an original vr4 have their issues. Id like to see the price of a kit soon though. Harness+Flash cable+ECU+disc with all the software, logger too. That would be a cool kit!
As of right now the harnesses are figured out and the bulk of the flashing is figured out. The main issue is getting all of the hardware differences worked out between a VR4 ecu and viable donor ecu. My understanding is that once some dyno testing is completed kits containing harnesses, ecus, and basic rom flashes will be sold. I do not think they will try to bundle in the evoscan or tactrix cable since that would just inflate the price.
Jesters Deadd
08-02-2011, 12:24 PM
Thanks, I respect Brett a lot and I consider Jeff a pretty good friend. I talk to Jeff pretty much every day on BAWC so he happens to answer a lot of my questions I have about this. I have an interest because I plan to use an SL ECU and I don't want this site to have the wrong info. This place needs all the help it can get.
Agree!...
prop7459
08-04-2011, 03:16 PM
So does anyone has a circuit diagram to our knock system?:DevilSmile: I will check the electrical system in my 3s manual to see if I can get some information rolling on the differences between the TT's and NA's knock circuitry. Any EE guys out there that want to help would be great.
95gto
08-04-2011, 05:40 PM
So does anyone has a circuit diagram to our knock system?:DevilSmile: I will check the electrical system in my 3s manual to see if I can get some information rolling on the differences between the TT's and NA's knock circuitry. Any EE guys out there that want to help would be great.
Awaiting confirmation from Brett and Jeff. Brett has mapped out the TT circuit and the circuit on some na ecu's. He is supposed to talk to Jeff to figure out where Jeff pulled the info that the SL and TT circuit are the same.
So pending that conversation we should have a valid answer. If we are lucky they will be the same and make life very easy on people.
Drummer
08-05-2011, 01:50 AM
Well what can you do LOL Im still waiting for my vr4 ECU to arrive :)
Anyone know if you can delete the EGR valve in the flash after taking it off the engine to avoid any CEL?
Jesters Deadd
08-05-2011, 11:17 AM
Well what can you do LOL Im still waiting for my vr4 ECU to arrive :)
Anyone know if you can delete the EGR valve in the flash after taking it off the engine to avoid any CEL?
yep... had it for awhile like this.
Jesters Deadd
08-11-2011, 12:18 PM
Soooo.... are you going to say if there was or was not a difference?
Drummer
08-11-2011, 09:53 PM
You mean me? If so, it will be a while until I get to put my car together :)
green-lantern
08-11-2011, 10:07 PM
You mean me? If so, it will be a while until I get to put my car together :)
No, he doesn't mean you
Jesters Deadd
08-12-2011, 03:04 PM
You mean me? If so, it will be a while until I get to put my car together :)
Didn't mean you... :)
95gto
08-12-2011, 03:12 PM
No, he doesn't mean you
Didn't mean you... :)
I'm guessing the question is pointed to me, and I have not heard back yet. I will send Brett a follow up to see if we can get this put to bed.
Drummer
08-15-2011, 01:30 PM
Please do.
95gto
08-19-2011, 03:31 PM
Soooo.... are you going to say if there was or was not a difference?
Please do.
So in talking with Brett the answer is that they can't confirm if it is the same or if it is different.
The issue is that neither he nor Jeff have actually inspected a 98'-99' SL ecu to validate that the circuits are the same. Brett has gotten ahold of a SOHC ecu but not an SL. To further complicate the situation the knock circuit has not been fully mapped on the VR4 ecu, there are blind via's on the board that have not been properly traced out. Which means even if an SL ecu has been inspected they wouldn't have known if the whole circuit matched or not since they don't actually know the whole circuit. This same issue persists for the Montero and Diamante ecu's as well.
Jesters Deadd - Brett suggested that you two should talk, he thinks getting you involved in the reverse engineering would be beneficial. If you are interested pm me and I will give you his contact number.
Jesters Deadd
08-19-2011, 04:00 PM
Interested... I have an 98 SL Ecu here and a DiaECu and a montero ECu.
Jeff V.
08-23-2011, 02:36 PM
I didn't intend to sign up for this board, but apparently passing info to Paul and calling Brett to correct him isn't working.
For starters..
The issue is that neither he nor Jeff have actually inspected a 98'-99' SL ecu to validate that the circuits are the same.
I've lost count of the times I've told Brett that I have had both a 98 SL ECU (on loan from Hans at GZP) and a 99 VR4 ECU in my possession for a couple YEARS now. I've posted about it on 3STech and to Brett personally via email and on the phone. I have examples of damn near every V6 Mitsu ECU made from 1990 to 2003 (some of them I got from Brett actually). I can take pictures if anyone would like to continue arguing this.
Anyway, as I posted in the info I gave to Paul when he asked me on BAWC, the SL and VR4 knock circuits are IDENTICAL. I even found another portion of the circuit that's not on the text file in the original post, and that portion is identical as well. Again, I can post pictures and notes if anyone wants to keep arguing this.
The Diamante (except for maybe that rare VR-X unit) and the Montero don't even HAVE knock circuits. The boards have the paths and the pads for the knock circuit, but are missing a couple little pieces. This info has been public knowledge ever since Aaron posted that we got his car working on a Montero ECU back in October of last year.
Reference: 99 Montero ECU, 92 Stealth TT - 3000GT/Stealth International Message Center (http://www.3si.org/forum/f1/99-montero-ecu-92-stealth-tt-507402/)
Now, back to the origin of this thread.
1. No stock boost gauge. Simple fix, replace your now non-functioning boost gauge with and aftermarket boost gauge.
True. Who cares? The stock gauge is worthless anyway. But if you really want to enable it, it's simple enough to add the transistor.
2. No stock two speed fuel pump control - Fuel pump stuck on HI. No biggie you think, right? Wrong . . . see the next item.
True, but the conclusion is wrong. Because...
3. No stock fuel pressure solenoid control. Oh Oh.
Also true, but again, the conclusion is wrong. The stock FPR solenoid is only used to bump up the pressure during hot restarts where there may be vapor lock conditions. The solenoid is disabled 2 minutes after the engine enters a running condition, and is not re-enabled until a key-off followed by another hot restart / vapor lock condition.
It has no effect on an engine that has been running for more than 2 minutes.
This info doesn't even require you to read the code or know how the ECU works at all. It's spelled out in the service and tech info manuals.
4. No wastegate control.
Also true. As Brett said, easily fixed with an aftermarket boost controller, or by adding the necessary bits to the board.
5. Knock Sensor! Ok, because of the design of the engine, Mitsubishi handles the knock sensor input differently.
And this is why I'm here. See what I posted above. The SL and VR4 knock circuits are IDENTICAL. All of the other ECUs need to have it added, and this was well established 10 months ago. It's about $0.60 worth of parts and some time from someone who can do surface mount soldering.
I tried to solve this over the phone. I've tried to solve this by forwarding info to someone contemplating this mod. This could have been solved via email or on 3STech.
That didn't work, so now we'll solve it here. If anyone wants to dispute anything I've said, bring some facts to the table and we'll hash it out in public now.
If anyone would like to discuss this outside the scope of what was covered in this thread, I suggest you sign up at BAWCforums.org and discuss it there. I do not plan on maintaining a presence here.
ChargerX3
08-23-2011, 02:57 PM
I would be interested in what/where the surface mounts need to be soldered in on my Diamante ECU. Care to show a picture?
Jeff V.
08-23-2011, 03:04 PM
What's your circuit board number? It'll be on the edge of the board and say something like 'JE331Bxxxx'.
Pretty much everything you need is in the text file Paul (Green Lantern) posted earlier. As far as I know right now, there is only one V6 circuit board each for 98 and 99. The only difference between models is that some pads are left unpopulated (like the knock stuff). Match your board number up to Paul's post and see if that helps.
If you're still unsure of what to do after that, post your concerns here.
ChargerX3
08-23-2011, 03:18 PM
JE331B157D, OR 0 at the end. I cant really make it out in my pic atm. Here is a pic of it.
http://s242.photobucket.com/albums/ff27/chargerx3/Random Pics/Fall2010006.jpg
Jeff V.
08-23-2011, 03:28 PM
Standard 98 board. Match that up to the numbers in Paul's post and look for the silk screen markings listed in his post. That'll tell you what you need.
I'll warn you though, this kind of soldering isn't like replacing caps on the older ECUs. If you can't figure out which parts to order and haven't done this kind of work before, I'd seriously suggest sending this out to someone who has. You run the risk of ruining the ECU if you do it wrong.
ChargerX3
08-23-2011, 04:39 PM
Knock circuit layout
1998 3000GT SL
MMC P/N MD347251
MFG P/N E2T61389 H3
PCB P/N JE331B157D
Con1-Pin 91 > R314 (TOP) > C45 (BOTTOM) > R315 (TOP) > IC2-Pin 78 < IC2-Pin 80 > IC1-Pin 92
R314 = 51 KOhm
C45 = 0.02 uF
R315 = 6.2 KOhm
IC2 = Mitsu Electronics E328 logic level converter
IC1 = Renesas H8/539F microcontroller, A/D input 6
1999 3000GT VR4
MMC P/N MD360163
MFG P/N E2T61495 H1
PCB P/N JE331B32804B
Con1-Pin91 > R71 (TOP) > C62 (BOTTOM) > R73 (TOP) > IC7-Pin 78 < IC7-Pin80 > IC9-Pin 92
R71 = 51 KOhm
C62 = 0.02 uF
R73 = 6.2 KOhm
IC7 = Mitsu Electronics E328 logic level converter
IC9 = Renesas H8/539F microcontroller, A/D input 6
You mean this? Sorry I i dont follow.
Jeff V.
08-23-2011, 05:06 PM
This will probably piss you off, but oh well.
If you can't figure out what you need to do based on this info, then you probably don't have the skills to finish the job correctly. I'd hand it off to someone more qualified.
I'm really not here to give people how-tos anyway. If you want that kind of discussion, head over to BAWC. This is more about correcting a pile of bad info.
Drummer
08-24-2011, 11:08 AM
Surface mount soldering :
http://www.youtube.com/watch?v=3NN7UGWYmBY
And, thank you Jeff for coming and clearing this up!
Jesters Deadd
08-24-2011, 05:11 PM
The Diamante (except for maybe that rare VR-X unit) and the Montero don't even HAVE knock circuits. The boards have the paths and the pads for the knock circuit, but are missing a couple little pieces. This info has been public knowledge ever since Aaron posted that we got his car working on a Montero ECU back in October of last year.
Reference: 99 Montero ECU, 92 Stealth TT - 3000GT/Stealth International Message Center (http://www.3si.org/forum/f1/99-montero-ecu-92-stealth-tt-507402/)
Jeff you are correct about the montero ECU not having a knock circuit populated. However you are wrong about the Diamante ECU not having a knock circuit populated.
Jesters Deadd
08-24-2011, 05:22 PM
I didn't intend to sign up for this board, but apparently passing info to Paul and calling Brett to correct him isn't working.
For starters..
I've lost count of the times I've told Brett that I have had both a 98 SL ECU (on loan from Hans at GZP) and a 99 VR4 ECU in my possession for a couple YEARS now. I've posted about it on 3STech and to Brett personally via email and on the phone. I have examples of damn near every V6 Mitsu ECU made from 1990 to 2003 (some of them I got from Brett actually). I can take pictures if anyone would like to continue arguing this.
Anyway, as I posted in the info I gave to Paul when he asked me on BAWC, the SL and VR4 knock circuits are IDENTICAL. I even found another portion of the circuit that's not on the text file in the original post, and that portion is identical as well. Again, I can post pictures and notes if anyone wants to keep arguing this.
The Diamante (except for maybe that rare VR-X unit) and the Montero don't even HAVE knock circuits. The boards have the paths and the pads for the knock circuit, but are missing a couple little pieces. This info has been public knowledge ever since Aaron posted that we got his car working on a Montero ECU back in October of last year.
Reference: 99 Montero ECU, 92 Stealth TT - 3000GT/Stealth International Message Center (http://www.3si.org/forum/f1/99-montero-ecu-92-stealth-tt-507402/)
Now, back to the origin of this thread.
True. Who cares? The stock gauge is worthless anyway. But if you really want to enable it, it's simple enough to add the transistor.
True, but the conclusion is wrong. Because...
Also true, but again, the conclusion is wrong. The stock FPR solenoid is only used to bump up the pressure during hot restarts where there may be vapor lock conditions. The solenoid is disabled 2 minutes after the engine enters a running condition, and is not re-enabled until a key-off followed by another hot restart / vapor lock condition.
It has no effect on an engine that has been running for more than 2 minutes.
This info doesn't even require you to read the code or know how the ECU works at all. It's spelled out in the service and tech info manuals.
Also true. As Brett said, easily fixed with an aftermarket boost controller, or by adding the necessary bits to the board.
And this is why I'm here. See what I posted above. The SL and VR4 knock circuits are IDENTICAL. All of the other ECUs need to have it added, and this was well established 10 months ago. It's about $0.60 worth of parts and some time from someone who can do surface mount soldering.
I tried to solve this over the phone. I've tried to solve this by forwarding info to someone contemplating this mod. This could have been solved via email or on 3STech.
That didn't work, so now we'll solve it here. If anyone wants to dispute anything I've said, bring some facts to the table and we'll hash it out in public now.
If anyone would like to discuss this outside the scope of what was covered in this thread, I suggest you sign up at BAWCforums.org and discuss it there. I do not plan on maintaining a presence here.
I said most of this in the second post of this thread, except for the incorrect statement about the ECU not having a populated knock circuit. Thanks for agreeing with me.
Jesters Deadd
08-24-2011, 07:21 PM
JE331B157D, OR 0 at the end. I cant really make it out in my pic atm. Here is a pic of it.
http://s242.photobucket.com/albums/ff27/chargerx3/Random Pics/Fall2010006.jpg
You can clearly see the circuit is populated here.
Jesters Deadd
08-24-2011, 07:26 PM
30853086
Jeff you are correct about the montero ECU not having a knock circuit populated. However you are wrong about the Diamante ECU not having a knock circuit populated.
green-lantern
08-24-2011, 08:57 PM
I said most of this in the second post of this thread, except for the incorrect statement about the ECU not having a populated knock circuit. Thanks for agreeing with me.
So what? you want recognition for something Jeff has known for a long time? Hell even I knew your post was correct. "Thanks for agreeing with me"? This seems very petty, I'm hoping this is a lost in translation thing. I've watched Jeff pave the way for this for years. It would not be where it is without him.
Drummer
08-24-2011, 11:10 PM
Again, thnx to those who made this possible.
Jesters Deadd
08-25-2011, 02:17 AM
So what? you want recognition for something Jeff has known for a long time? Hell even I knew your post was correct. "Thanks for agreeing with me"? This seems very petty, I'm hoping this is a lost in translation thing. I've watched Jeff pave the way for this for years. It would not be where it is without him.
Sorry that wasn't ment to be petty toward anyone, especially Jeff or you. It's frustrating to everyone when people make statements then threads then more statements mislead people then in the end are incorrect. Also Jeff didn't tell me any of this info. So I wasn't sharing his information. I was pointing out he was reaffirming what I said earlier of his own accord.
Jesters Deadd
08-25-2011, 02:30 AM
Also I think I know what the confusion is in Brett's and Jeff's info and where the general confusing statement in this thread came from. The SOHC 99 3000gt has a knock sensor with a different resistor value in it. So maybe he thought he had a SL ECU when he had a 99 SOHC 3000gt ECU. The Diamanté ECU has the same knock circuit as the 99 SOHC 3000gt ECU.
green-lantern
08-25-2011, 06:22 AM
Sorry that wasn't ment to be petty toward anyone, especially Jeff or you. It's frustrating to everyone when people make statements then threads then more statements mislead people then in the end are incorrect. Also Jeff didn't tell me any of this info. So I wasn't sharing his information. I was pointing out he was reaffirming what I said earlier of his own accord.
It's cool, I must be reading into your post too much. The reason I asked Jeff in the 1st place is I don't want any wrong info on this site and of course I'm doing this mod.
I know Jeff didn't directly tell you this info but most of this stuff came from him in the 1st place. People wouldn't be doing this yet if it wasn't for him. He just isn't into it that much anymore so guys like Arron, Greg and Brett have taken the lead (you too). He deserves a lot of credit for this and I'd like to encourage him to post here more.
95gto
08-25-2011, 12:55 PM
Also I think I know what the confusion is in Brett's and Jeff's info and where the general confusing statement in this thread came from. The SOHC 99 3000gt has a knock sensor with a different resistor value in it. So maybe he thought he had a SL ECU when he had a 99 SOHC 3000gt ECU. The Diamanté ECU has the same knock circuit as the 99 SOHC 3000gt ECU.
Brett's SOHC ecu has only one resistor in the knock circuit and it is tied to ground, not the path noted in the file that Paul listed. That path has empty pads, I have seen that first hand.
Sorry that wasn't ment to be petty toward anyone, especially Jeff or you. It's frustrating to everyone when people make statements then threads then more statements mislead people then in the end are incorrect. Also Jeff didn't tell me any of this info. So I wasn't sharing his information. I was pointing out he was reaffirming what I said earlier of his own accord.
It isn't confirmed if Brett is incorrect or not, so I would hold off on the verdict just yet. I have started dialogue elsewhere to address the issue in detail but essentially we have potential dissagreement with the absence of knock circuits on ecu's in conjunction with motors actually having knock sensors. So the issue is being investigated.
What I want to make emphatically clear is that this thread and the posts associated with supporting it are being made from a cautious perspective and NOT a misleading one. Brett started the thread in an effort to prevent people from potentially grenading their motors, not handicapping their ability to switch to a flashable ecu.
In the long run I hope the issue turns out to be moot since it would mean greater flexibility for the platform with less re-work.
99 vr4
08-25-2011, 01:03 PM
Like Jeff, I really don't spend a lot of time on this board as I have way too much going on in my life that actually matters than to follow the drama here. But never the less, I have been getting phone call after phone call about this issue.
But first let me address an earlier issue before I settle this knock sensor issue.
Jester, when someone such as Jeff or I makes a post, it is considered very rude to knock the post down bit by bit based upon your beliefs, and you can certainly not expect a timely response. I don’t have time for drama. For example, I posted both the problem and the solution:
“1. No stock boost gauge. Simple fix, replace your now non-functioning boost gauge with and aftermarket boost gauge.”
Your reply, however immature, is certainly not winning any points in the “he knows what he is talking about category”:
“1. No stock boost gauge... the stock boost gauge doesn't even measure boost, you should have a real boost guage anyway!”
Really? Didn’t I already say that? Your antagonistic, juvenile response is why you have not seen me post anything else up here because I don’t have time for the pettiness or the drama.
Now on too way more important issues, KNOCK SENSOR TIME:
99 vr4
08-25-2011, 01:30 PM
Now on too way more important issues, KNOCK SENSOR TIME:
Jeff and I have spoken at length about this issue. Jeff did clear some things up, particularly in regards to the knock sensor and which cars actually have a knock sensor.
I have been working on the assumption (yes I know, bad) that ALL cars had a fully functional knock sensor. This was actually, and falsely, backed by the fact that Mitsubishi ASA was actually listing a knock sensor as a valid part for the Montero and Diamante platform when looking up certain models by model code. The classification and sometimes the VIN makes the difference.
But its way more complicated than that (hey it’s Mitsubishi). Let’s look at the Montero. Apparently 3.0L and 3.5L Montero Sports do NOT have a knock sensor (Code K86W & K89W). BUT SOME Montero 3.5L Long Wagons (Code V45W) DO have knock sensors as well as the newer 3.5L and 3.8L Long Wagons (Codes V75W & V77W). BUT NOT ALL – This is where Mitsubishi gets freaky.
BEGIN EDIT
I have a MD361885 which I know for a fact came out of a 1999 Montero Long Wagon 3.5L Code V45W. This is a FED ECU, and it does NOT have the completed knock sensor parts on the circuit board and it turns out it does not have a knock sensor on the engine (more on this later). While I have actual VIN numbers of the donor cars for most of my ECU's, this one ECU did not have a VIN with it. When I crossed the part number it was determined that it came from a V45W GRXVL2M, a 1999 FED Montero. When you search for "knock" in ASA it returns a valid part number code which then returns a valid part number for a "SENSOR, ENG CONTROL KNOCK". So I thought, ok, there is a knock sensor, and the parts are missing on the ECU circuit board.
When reverse engineering the Knock sensor circuit, since there were critical parts missing between pin 91 on the ECU main plug and pin 78 of the E328 chip, this indicated that there was a blind trace between pin 91 on the ECU main plug and the E328 chip. Well I was partially wrong. (More on this later, as there still remains a blind trace.)
Back to my 1999 Montero Long Wagon 3.5L and its missing knock sensor. If you dig deeper in ASA, and try to find a diagram, you won't. If you change your model and classification to V45W GRXML2M (FED) or GRXML7M (CAL) and search for a knock sensor you can find the diagram indicating that there is a knock sensor. So in later years, they did away with the knock sensor. While ASA is smart enough to filter the diagrams based upon model classification, it apparently is not smart enough to filter the search results the same way. Damn! This lack of search filtering has caused a LOT of problems.
END EDIT
As any engineer can tell you, when in doubt, you error on the side of caution until empirically proven wrong. And I am sorry; the statement that it works in my car is not really proof enough. Replicating the results on a dyno, side by side, in a controlled environment, with external testing and measurement, well that would be what I would consider proven.
I have completely reverse engineered the O2 sensor circuits and successfully converted Montero and Diamante ECUs. I have completely reverse engineered the ignition circuits and successfully converted SOHC distributor based ECU’s. I have not been able to completely reverse engineer the knock sensor circuit.
Here is why:
There is a BLIND via off of pin 78 and I have no idea where it goes. I know it does NOT go to the ground bus or the voltage bus. So where does it go? Until we know where it goes, there is a portion of the knock senor circuit that we don’t know about. This could be a benign input to a counter, or it could be more filtering which will directly affect knock voltage and knock sum.
EDIT:
I asked Jeff the other day to please if he has time look at the dual chip ECU's that I sent him (with integrated auto trans controllers) and see if this blind via off of pin 78 is still a blind via. There is some likely hood that the circuit is traceable due to a difference in circuit board design and we will be able to completely map out the knock circuit sensor and say yes it is 100%.
Again, I error on the side of caution. This entire post originally was not meant to dissuade any form of ECU swap, but to inform the general uninformed public that swapping with the SL ECU is not a 100% compatible solution, and there are inherent risks that you are taking. Some of these risks include blowing your motor. And unlike some of the upgrades that are DIY, like boost control, where most people know that over boost without proper fuel compensation will result in a blown motor, ECU swaps for most people are in the realm of electronic witchcraft. They know it does something, but cannot explain how or why because they don’t know the how or why. And if you don’t know the how or why, you run the risk of damaging your motor.
Only when people such as Jeff and myself reverse engineering these circuits to determine the how and why will anyone be able to say with 100% certainty that this is a non issue. Drivability tests with your butt dyno and data logging is not the same as side by side dyno tests in a controlled environment.
And with the cost of these motors, I will not give my stamp of approval on something that I cannot diagram out and say, “Yes it is the same”.
EDIT:
ALSO PLEASE NOTE:
On the schematic, "Not Applicable" is probably misleading. It merely means that the part is missing on the circuit board. It is VERY applicable and needs to be added to the circuit board in order to complete the knock sensor circuit.
99 vr4
08-25-2011, 05:20 PM
Before I forget . . . someone asked about the O2 circuit . . . if you have a 3000GT board (and some CAL spec boards I think) it will have all four O2 circuit populated. If your unlikely and you have a Diamante or Montero board, then only the upstream (pre cat) front bank and downstream (post cat) front bank will be populated. This means no "REAR" O2 sensor for our cars. The fix is to modify the ECU eliminating the downstream and creating TWO upstream (pre cat) circuits for the front and rear O2. You simply can't change the wiring harness because the downstream circuit is processed differently both by hardware and by code. You MUST modify the ECU. Then you turn off the downstream (post cat) in the periphery bits and your good to go.
And of course there is the question of SOHC distributor based vs. DOHC. In order to convert a SOHC Distributor based to a DOHC coil based there are parts that must be removed, parts that must be added, and finally parts that must be changed. Again, this is a HARDWARE modification on the ECU itself. Sorry guys. And again, these are all very tiny little surface mount parts. I utilize a hot air preheat system in order to minimize thermal damage to the board and components. Without this, you will likely lift a trace permanently damaging your ECU. Remember, these ECUs have a conformal acrylic coating on them to keep moisture from the components. The video did not cover that aspect. Although it is considered “solder through” meaning you don’t have to strip it away, parts will act as if they are glued on because of it and solder sticks to it like crazy. You need to strip the coating off with the appropriate chemicals when making repairs and changes. As you can see from that video that was posted, surface mount is not your normal soldering . . .
99 vr4
08-25-2011, 05:33 PM
Before I forget . . . someone asked about the O2 circuit . . . if you have a 3000GT board (and some CAL spec boards I think) it will have all four O2 circuit populated. If your unlikely and you have a Diamante or Montero board, then only the upstream (pre cat) front bank and downstream (post cat) front bank will be populated. This means no "REAR" O2 sensor for our cars. The fix is to modify the ECU eliminating the downstream and creating TWO upstream (pre cat) circuits for the front and rear O2. You simply can't change the wiring harness because the downstream circuit is processed differently both by hardware and by code. You MUST modify the ECU. Then you turn off the downstream (post cat) in the periphery bits and your good to go.
And of course there is the question of SOHC distributor based vs. DOHC. In order to convert a SOHC Distributor based to a DOHC coil based there are parts that must be removed, parts that must be added, and finally parts that must be changed. Again, this is a HARDWARE modification on the ECU itself. Sorry guys. And again, these are all very tiny little surface mount parts. I utilize a hot air preheat system in order to minimize thermal damage to the board and components. Without this, you will likely lift a trace permanently damaging your ECU. Remember, these ECUs have a conformal acrylic coating on them to keep moisture from the components. The video did not cover that aspect. Although it is considered “solder through” meaning you don’t have to strip it away, parts will act as if they are glued on because of it and solder sticks to it like crazy. You need to strip the coating off with the appropriate chemicals when making repairs and changes. As you can see from that video that was posted, surface mount is not your normal soldering . . .
Jesters Deadd
08-29-2011, 07:30 PM
Like Jeff, I really don't spend a lot of time on this board as I have way too much going on in my life that actually matters than to follow the drama here. But never the less, I have been getting phone call after phone call about this issue.
But first let me address an earlier issue before I settle this knock sensor issue.
Jester, when someone such as Jeff or I makes a post, it is considered very rude to knock the post down bit by bit based upon your beliefs, and you can certainly not expect a timely response. I don’t have time for drama. For example, I posted both the problem and the solution:
“1. No stock boost gauge. Simple fix, replace your now non-functioning boost gauge with and aftermarket boost gauge.”
Your reply, however immature, is certainly not winning any points in the “he knows what he is talking about category”:
“1. No stock boost gauge... the stock boost gauge doesn't even measure boost, you should have a real boost guage anyway!”
Really? Didn’t I already say that? Your antagonistic, juvenile response is why you have not seen me post anything else up here because I don’t have time for the pettiness or the drama.
Now on too way more important issues, KNOCK SENSOR TIME:
No drama, I just want correct information in the thread. That particular statement and emphasis was toward anyone who doesn't have a real boost gauge installed already, not you. Also I really don't care since I am not selling these things. I just have a couple I took apart for fun and tested in my car.
Who is all on your team to sell these things anyway?
As some of you know, I have been working with a team of people for almost a year trying to find a low cost solution for a flashable ECU on the 3000GT platform.
99 vr4
08-30-2011, 01:03 PM
@jeastersdead
The team that I have been working with includes both hardware and software hackers as well as testers. I have worked closely with jeff, paul, aaron, levi, and others who I won't mention by name.
A lot of this will be released public domain but some of it is reservd.
I normally won't share information here because of the technical nature.
if you can find out where the blind via goes that would be great!
Brett
Jesters Deadd
09-09-2011, 10:53 AM
From my four plug Montero Sport ECU.
3195
Jesters Deadd
09-09-2011, 11:01 AM
98 Diamante ECU
3192
Jesters Deadd
09-09-2011, 04:56 PM
These are 99vr4's montero ECU diagrams. I just wanted them all near the same place.
Jesters Deadd
10-08-2011, 10:03 PM
Edited
prop7459
10-11-2011, 01:37 AM
With the all those ecu listed, which ones have you tested your TT? Have you confirmed that the knock circuitry of those ecu work correctly with our knock system?
Jesters Deadd
10-13-2011, 01:46 PM
With the all those ecu listed, which ones have you tested your TT? Have you confirmed that the knock circuitry of those ecu work correctly with our knock system?
Several of those ECUs need circuit modification to work correctly.
EvilNCarnate
01-20-2012, 09:10 PM
Subscribed.... seems to have fallen off the edge of the world.
green-lantern
01-20-2012, 09:31 PM
Subscribed.... seems to have fallen off the edge of the world.
Chris, you might want this thread.
http://www.3sgto.org/f2/so-who-wants-their-ecus-chromed-7175.html
99 vr4
02-07-2012, 09:38 AM
Not really fallen.. Its just that there are some technical issues which no one seems to have been able to verify or work out related to the knock sensor circuit.
jazz9
02-07-2012, 10:14 AM
Anyone want me to take photos or anything from my JDM 94 GTO ECU? I'm pulling it tomorrow.
Drummer
02-14-2012, 09:55 PM
Yea, its not exactly totally refined yet... Im hoping that when it is, it will be a simple, very elegant plug and play solution!
Jesters Deadd
02-15-2012, 01:54 PM
This thread should be deleted. It confuses people.
99 vr4
02-16-2012, 08:05 AM
This thread should be deleted. It confuses people.
Not at all, there is some very good information in this thread including schematics of circuits related to the ecu.
futurevr4man
03-29-2012, 12:19 PM
there is also a lot of information that has been found out/changed in this thread over the last few years. a new, all-inclusive, thread needs to be made so the people who are coming into this fresh can learn the correct information off the bat
CoopKill
03-29-2012, 01:58 PM
there is also a lot of information that has been found out/changed in this thread over the last few years. a new, all-inclusive, thread needs to be made so the people who are coming into this fresh can learn the correct information off the bat
Yes please!!!
99 vr4
04-04-2012, 06:13 PM
there is also a lot of information that has been found out/changed in this thread over the last few years. a new, all-inclusive, thread needs to be made so the people who are coming into this fresh can learn the correct information off the bat
Last few years? Really? Actually this thread is only 15 months old, and ironically, the information in the VERY FIRST POST is still the same. The most important information is that the knock sensor circuits still have not been 100% fully reverse engineered due to buried traces and blind vias in the actual printed circuit board. There are some un-knowns, although some will swear it does not matter. I prefer the scientific methods that rely on schematics, circuit reverse engineering, etc. The idea of "it works for me" is not good enough for some of us to 100% endorse. But this will be changed soon enough with the release of new information. I just have to wait until my findings are cross checked.
ChargerX3
04-04-2012, 10:16 PM
Last few years? Really? Actually this thread is only 15 months old, and ironically, the information in the VERY FIRST POST is still the same. The most important information is that the knock sensor circuits still have not been 100% fully reverse engineered due to buried traces and blind vias in the actual printed circuit board. There are some un-knowns, although some will swear it does not matter. I prefer the scientific methods that rely on schematics, circuit reverse engineering, etc. The idea of "it works for me" is not good enough for some of us to 100% endorse. But this will be changed soon enough with the release of new information. I just have to wait until my findings are cross checked.
Well put. This is a process and very little has been proven yet. We are all still learning more about this, so any conclusive thread atm would be foolish.
BTW what new info do you have? I would like to be privy to this. ;)
futurevr4man
04-05-2012, 02:02 AM
Last few years? Really? Actually this thread is only 15 months old, and ironically, the information in the VERY FIRST POST is still the same. The most important information is that the knock sensor circuits still have not been 100% fully reverse engineered due to buried traces and blind vias in the actual printed circuit board. There are some un-knowns, although some will swear it does not matter. I prefer the scientific methods that rely on schematics, circuit reverse engineering, etc. The idea of "it works for me" is not good enough for some of us to 100% endorse. But this will be changed soon enough with the release of new information. I just have to wait until my findings are cross checked.
i mean, im not trying to ruffle your feathers, pumpkin. i am just saying that as someone who doesnt know anything about this stuff, and trying to learn, there has been mis-information and changes in information as you read through this thread. a lot of the more obvious mistakes are gone now because i pm'd certain people and asked them if it was true or not, and they edited their posts accordingly. for an example of something that is mis-information (AND/OR debatable)... you say in your first post "There are significant hardware differences between the turbo and non-turbo ECUs" but i have read on a few threads, from a few different people, that the knock circuitry was exactly the same. inconsistent, and i dont know who/what to believe in. i can tell you that jeff v. is supposed to be one of "THE" guys on this topic and he said the circuitry was exactly the same (post 71 for reference)... so you tell me (and im aware of the blind via no one has traced). even the thread title itself is enough to make people think twice before getting an SL ecu, when in reality, they are just fine...
also, the 'chromed' thread is very helpful, but ironically, its not in this section (ecu/tuning section) so i hadnt read through that yet when i posted this here.
im not trying to hate, or be a jerk. i just dont like mis-information or even conflicting information on something as 'revolutionary' (for lack of a better word) as this. especially for those of us who are thinking/acting on integrating this into our cars :suspect:
thank you for your help, as well as everyone else who has put time into this project
Jeff V.
04-05-2012, 02:13 PM
How many people are obsessing over a knock circuit that apparently works fine, while still using a 15 year old sensor connected to technology developed and patented in the 80s?
That blind via Brett is so obsessed with is probably just a test point.
It's possible to log all of the values that go into building the knock profile for the engine. You don't need an engine dyno, you don't need a chassis dyno, you don't need lab grade ion sensing knock detection equipment, and you don't need a masters in electronics or chemistry or engineering. You also don't need to intentionally knock your engine to generate the data.
Set up the parameters in EvoScan, and set up some test drives in consistent conditions. A chassis dyno would help with this, but it's not necessary.
Log the parameters on an ECU you're sure about (legit 98 or 99 VR4 ECU) and then again on one of the suspect ECUs (98 SL, converted Montero, etc)
See what your margin of error is. Throw the results in Excel and post them up.
We should not still be arguing about this after NINE MONTHS.
anyonebutme
04-05-2012, 02:28 PM
It shouldnt have been argued in the first place....not like there wasnt like 3 guys willing to test it on a chassis dyno over 9 months ago...just sayin'.
Jesters Deadd
04-05-2012, 03:53 PM
I have an extra board I am gonna cut open and see if that dang via goes anywhere!
95gto
04-05-2012, 04:58 PM
I have an extra board I am gonna cut open and see if that dang via goes anywhere!
There are plans to perform the very testing that Jeff has described, so I would suggest not disecting an ecu.
Greg E
04-05-2012, 07:38 PM
Brett, have you read my thread on BAWC about how the knock sensor filtering works? I included the code and a patch for EVOScan to log knock raw. My hopes were that someone would try back to.back runs with an SL & VR4 ECU.
If someone would let me.borrow their SL ECU I will test.... This debate needs to be resolved. No one had done anything about it for too long.
99 vr4
04-15-2012, 11:39 AM
Yes jeff you could say I am obsessed. I have a "test" ecu that I routinely remove components for "out of circuit testing", especially capacitors, diodes, and transistors.
And we should not be arguing about this. Some of what you say is true. Old technology. But old technology doesn't mean you can throw caution to the wild.
Testing will be done soon enough.
Jeff V.
04-15-2012, 06:43 PM
Nobody is throwing caution to the wind. There's reasonable caution, and then there's what you're trying to do.
Unless your ECU packages are going to include new knock sensors and replacements for the all too common frayed and corroded connectors found on these 13-22 year old cars, you're obsessing over nothing.
We've got results showing that it works, and nobody has yet taken me up on my suggestion of logging the raw component variables to confirm or deny these findings. You haven't presented one single reasonable scenario showing WHY we're supposedly wrong, other than that one single test point / via you can't trace.
Are you going to trace back every single circuit on the conversion ECUs to make sure they match too?
whitedragon
04-15-2012, 08:57 PM
It shouldnt have been argued in the first place....not like there wasnt like 3 guys willing to test it on a chassis dyno over 9 months ago...just sayin'.
If you're in ECU Test Club, you don't talk about ECU Test Club!
99 vr4
04-17-2012, 04:04 AM
Brett, have you read my thread on BAWC about how the knock sensor filtering works? I included the code and a patch for EVOScan to log knock raw. My hopes were that someone would try back to.back runs with an SL & VR4 ECU.
If someone would let me.borrow their SL ECU I will test.... This debate needs to be resolved. No one had done anything about it for too long.
Yeah I already had that . . . I also had the actual knock voltage - see my post on bawc
Jesters Deadd
04-17-2012, 12:04 PM
Just so this information is here too. I dissasembled a board and it had 4 layers. The "blind via" doesn't attach to anything. I think this is conclusive.
99 vr4
04-18-2012, 02:46 AM
Adam, burning through the board with a propane torch is hardly conclusive. The 1/2 mm traces would burn off before you even noticed them.
But I am working on the problem with several others and should have verifiable conclusive test results when I get back from the KC area.
On another note, fun with an o-scope chasing down the knock sensor circuit...
https://www.youtube.com/watch?v=pGrcamhLg10&feature=youtube_gdata_player
Jesters Deadd
04-18-2012, 05:39 PM
40104013401140124014
Adam, burning through the board with a propane torch is hardly conclusive. The 1/2 mm traces would burn off before you even noticed them.
But I am working on the problem with several others and should have verifiable conclusive test results when I get back from the KC area.
On another note, fun with an o-scope chasing down the knock sensor circuit...
https://www.youtube.com/watch?v=pGrcamhLg10&feature=youtube_gdata_player
It was butane. And there are no 1/2 mm traces in the inner layer. I was worried the inner layers would be as complex as the two outside layers. However I was pleasantly surprised that the inner layers were pretty simple. The ground layer, we will call it layer 3, was like one big block. The other layer, we will call it layer 2, had large blocks that would connect certain pins. I found pins that connected to the blocks the blind via went through and checked for continuity with the blind via. I conclusivley can say it doesn't attach to an inner layer. In fact there are many that go through the board and don't attach to an inner layer.
Jeff V.
04-19-2012, 10:28 AM
And there are no 1/2 mm traces in the inner layer.
What he's saying is that the via may attach to those big blocky inner layer pads via tiny traces. Those tiny traces could burn off before you ever see them.
I've seen those traces on the older PCBs. They radiate out from the via like a small X. They're bigger and more visible on the older ECUs.
Jimvr4
04-19-2012, 02:19 PM
What he's saying is that the via may attach to those big blocky inner layer pads via tiny traces. Those tiny traces could burn off before you ever see them.
I've seen those traces on the older PCBs. They radiate out from the via like a small X. They're bigger and more visible on the older ECUs.
That's typically done to connect vias to inner power or ground planes while maintaining some thermal isolation to improve soldering flow
Jesters Deadd
04-19-2012, 03:39 PM
What he's saying is that the via may attach to those big blocky inner layer pads via tiny traces. Those tiny traces could burn off before you ever see them.
I've seen those traces on the older PCBs. They radiate out from the via like a small X. They're bigger and more visible on the older ECUs.
oh I gotcha. You can actually see some of the vias that go through the board and attach to the blocks if you look in my pictures. The knock ground pin is a perfect exacmple. Also I found pins that connected to the blocks that the blind via goes through.
On an unmolested board I checked for continuity with the blind via with at least two other pins. Those pins I know connect to the block the blind via passes through.
Am I clear about what I did? I am sometimes not the best at explaining my madness. But there is a method to it.
99 vr4
05-05-2012, 05:03 PM
oh I gotcha. You can actually see some of the vias that go through the board and attach to the blocks if you look in my pictures. The knock ground pin is a perfect exacmple. Also I found pins that connected to the blocks that the blind via goes through.
On an unmolested board I checked for continuity with the blind via with at least two other pins. Those pins I know connect to the block the blind via passes through.
Am I clear about what I did? I am sometimes not the best at explaining my madness. But there is a method to it.
Yes but you sure did not prove anything except you can take a butane torch to a ecu and burn it up real good.
Jesters Deadd
05-05-2012, 07:01 PM
I guess you just don't get it...
mojo3120
12-04-2014, 08:11 PM
looks like you lifted a pad or two with your torch :highly_amused:
a little late to the party, but that's not a blind via. You rarely see blind (or buried) via's used in any pcb bc it vastly increases costs, and rarely necessary.
On a 4 layer PCB, the inner 2 layers are typically reserved for GND and VCC planes. You can see that via runs straight through a plane on both inner layers. If it connected to either you'd be able to see it with a flashlight & loupe (it wouldn't be a 6 mil trace, it'd look similar to the via on the resistors above it (R300 / R315)). Also, easy to meter it against both planes to make sure like Jester explained.
You guys posting schematics, care to post your eagle file? I'm curious to see the schematic and how much of it is mapped.
99 vr4
12-05-2014, 09:59 AM
Wow way to digg up an old post.
So I have not found where the via we are talking about goes to. I have metered out quite a bit of that board. My guess is that it pops up under a chip somewhere. Adam said he confirmed where it goes but he has never posted the info. . so it remains unverified.
BTW, I have tried the loop and light method. .. One of the planes makes it hard to see.
Other differences exist in the O2 circuit in the form of a resistor network to divide the input into a mux. The SL is not as sensitive to the lower O2 voltages. An issue Greg says we can ignore... but again hardware changes are expensive and the Mitsu engineers would not have made the change just for the VR if they did not feel it was required.
I have some of this built out in eagle but i am not ready to share this as yet. I eventually may open source the entire schematic and board design but would like a "qualified" EE to work with with experience in MCU design.
Powered by vBulletin® Version 4.2.3 Copyright © 2026 vBulletin Solutions, Inc. All rights reserved.