3P's TCCS Disassembly/Analysis

Nick M

Established 1994
Authorized Seller
Contributor
Joined
Sep 9, 2005
Messages
8,811
Likes
12
Location
U.S.
EUDM and JDM 2JZGTE are using MAP sensor.
They are not required to have the car run optimally under many different conditions. See logical fallacy of "appeal to authority" to understand.
 

mjwillia

New Member
Authorized Seller
Joined
May 13, 2013
Messages
2
Likes
0
Location
Perth
Hi Guys, is there a way to enable the obd output on 7mgte ecu?
So far as I can find the pin required ( Te2 ) to enable the output ( VF1 ) doesn't exist on any 7mgte ecu, however I did read you were able to turn on some diagnostics hiding in the code.

Thanks for any advice.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
There is no obd hiding in any 7M ecu I have looked at. The California ecus are the most advanced, with misfire detection in the later models, but the interface I am referring to is proprietary, and never intended to be a replacement for obd.
 

Nick M

Established 1994
Authorized Seller
Contributor
Joined
Sep 9, 2005
Messages
8,811
Likes
12
Location
U.S.
This thread is getting so big its hard to find stuff, but see here for timing computation. http://www.supramania.com/forums/sh...bly-Analysis&p=1712465&viewfull=1#post1712465

So, on a hot day you will see some impact to timing.

Another thing to consider is that the air intake temp goes through the roof on a hot day because the AFM is beside the exhaust manifold and gets heat soaked. Its not unusual to see intake temps above 60C on a hot day after 30 minutes of driving, even though the air intake for the factory filter is in a cool place. That will also reduce power.
This was hard to find but not impossible. What I notice is the lag time when at operating temps goes up and I assumed timing was being pulled a little bit. I am thinking of changing the intake (mine came with a FIPK) to get more fender air. I am not sure how just yet. I don't want to fabricate ducting, I would rather get the filter down where it started.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
I wouldn't bother with the ducting. My tests on the stock air cleaner show that the incoming air is pretty near ambient, the stock system pulls air from in front of the radiator and it works well. The problem is that the AFM get really hot. Put your hand on it after running a while. the air heats up as it goes through the hot AFM. Some sort of insulation for the AFM would be very helpful I suspect.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
I think the first thing to do is confirm that the high air temp readings are accurate. It could be that the hot AFM is making the sensor read hotter than the air temp actually is. I plan to measure the temp in the accordion tube with a thermocouple and see if it matches the AFM sensor.
 

Nick M

Established 1994
Authorized Seller
Contributor
Joined
Sep 9, 2005
Messages
8,811
Likes
12
Location
U.S.
I'm going to use Lex AFM with this ecu in few weeks (I need to finalize my fuel model tune), for the same reason, even if I was enjoying MAP for piping commodity.
Did you try a Lex meter yet? With the stock CT-26, my opinion is that the car showed noticeable improvement all around. Which is what was reported a long time ago.

I agree and I've never wrote that.
It was just to complete your information...
My bad.
 

nuttom

New Member
Authorized Seller
Joined
Jun 29, 2014
Messages
11
Likes
0
Location
SIGOLSHEIM, FRANCE
I decapped the 42-pin IC labelled "D 151802-2860" "572 130M" "9046 Z44". Supposed to be the knock MCU from a MR2 ECU
Nitric acid (60%) heated and a couple of hour !

Guess what ? It's a Fujitsu chip, the partnumber is also on the DIE: MB88572
There is also a year: 1987 !!!
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
Nice. It was known that the knock MCU was not the same architecture as the denso/toshiba MCU, but not much more than that has been made public. It would be nice if someone could figure out the pin functions for that chip, as it may be the same architecture as used for the auto trans MCU which is also Fujitsu. Are the bare die pads labelled?
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
Did you try a Lex meter yet? With the stock CT-26, my opinion is that the car showed noticeable improvement all around. Which is what was reported a long time ago.
I took the AFM calibration code from the Lexus ECU and compiled it into the Supra code. As a result, I am running the stock 440 injectors with the Lexus AFM and it behaves 100% like stock, you cannot tell the difference and the fuel trim is unchanged, as is the ignition timing.

By doing this, I can then scale the injector routine to support different size injectors, without altering timing or other maps inside the ecu. I have also added short pulse width adder table as used in GM ECUs to allow big injectors to be used and still get a decent idle. All of this is running out there on the street and pretty successfully I would say.

The Lex/550 mod has been used for years, but frankly its a hack, and the drivability is no way as good as the stock setup due mainly to the different behavior of the Lex AFM at low flow rates.

The Lexus AFM calibration is very different from the Supra, due to the way the bypass is designed. People who mess withe the Lexus screw in the AFM are making a big mistake. This has been modeled and confirmed (not by me so I cannot say more at the moment).
 

nuttom

New Member
Authorized Seller
Joined
Jun 29, 2014
Messages
11
Likes
0
Location
SIGOLSHEIM, FRANCE
The pinout fit with fujitsu databook for the mb885000 series (understand supply and crystal, also adc ref).

I contacted fujitsu, I hope they will answer with the datasheet.

Unfortunately my microscope is a cheap usb camera with a cheap lens, I'm investing in a low cost but true metallurgical microscope, then I will be able to answer. Right now it's too small to see anything I was lucky to find a marking on the die !

If you want I can decap your ic ! Maybe a partnumber can be read
 

nuttom

New Member
Authorized Seller
Joined
Jun 29, 2014
Messages
11
Likes
0
Location
SIGOLSHEIM, FRANCE
I desoldered with my hot air station both IC, but I'm unable to inspect the main MCU because I damage it. If someone own a broken ECU I would be happy to try again !
 

Rollus

Member
Authorized Seller
Joined
Jun 2, 2011
Messages
593
Likes
2
Location
Paris, France
I have also added short pulse width adder table as used in GM ECUs to allow big injectors to be used and still get a decent idle. All of this is running out there on the street and pretty successfully I would say.
I confirm this. I've drove 8812 miles with this ECU, My 960cc injectors have flew 667,62gal of E85. Injectors are scaled and with short pulse tuned.
I have yet to tune for cold start now, with some around 10-15°C with the fall comming.

The work on the Lex AFM seems good also, but I had no chance to install mine yet (I have a whole project with a FFIM)

3p your modded ecu is awsome, thank you and keep up with the good work ;)
 

turbotoy

Member
Authorized Seller
Joined
Apr 4, 2005
Messages
67
Likes
0
Location
NY
The Lex/550 mod has been used for years, but frankly its a hack, and the drivability is no way as good as the stock setup due mainly to the different behavior of the Lex AFM at low flow rates.

The Lexus AFM calibration is very different from the Supra, due to the way the bypass is designed. People who mess withe the Lexus screw in the AFM are making a big mistake. This has been modeled and confirmed (not by me so I cannot say more at the moment).
Being the one who performed the modeling work that 3p is referring to, I completely concur with these statements. I will complete the documentation of my modeling work and share that with the community shortly so people can understand the underlying reasons for why problems arise when using the Lexus AFM with the stock ECU code.

I have been using an ECU modified by 3p with the Lexus AFM calibration code and a Lexus AFM. I have also derived (by iteration) close to optimal injector parameters for the RC Engineering 550cc/min injectors that I currently have installed. I have experienced dramatic performance improvements with these changes; my car has never run better.

I could not be more impressed with the performance of this ECU and associated software, along with the outstanding support provided by 3p along the way!
 

Zazzn

OLD school
Authorized Seller
Joined
Apr 1, 2005
Messages
1,054
Likes
14
Location
Toronto / San Francisco
wow, keep up the good work been 6 years since this was started... Any real expectations when this will be a finished product. BTW all JDM JZA70's, Gen 1 / 2 Aristos, chasers ect are SD. When you said you got this working on the 2jzGE, how about a GTE ecu using SD instead of AFM? I have 5 mk4 ECU's that I would like to play with.

JZA80 USDM 6psd
Stock
TechTom

JZA80 JDM
1 6 spd modified (not sure by who or how, came with my JDM supra) has a very aggressive tune, and no speed cut, or boost cut and higher rev cut 7800

1991 airsto (I think JZA141?) auto
1 modified ecu (no boot cut, higher revs, no speed cut)
1 stock Aristo ECU

I'd love to dump these maps for you, or send you the ecu's to test (with the exception that I need some of them back eventually since i'm up in SF bay it's not hard for me to hand deliver as I work down in LA very often)

Finally, I don't understand why you are using the Lexus AFM and resealing the map to work with 440's? Why bother with the afm the first place? Isn't the goal to allow more air flow, while correcting the scaling so that the timing and all the other maps arn't all screwed up? At the end of the day, you could run larger injectors but whats the point if you can raise the FCO point when you need to gobble more air to take advantage of the injectors?

The real question here to me is if the resolution of the AFM is all the way to 3000 why can't we allow it to process up to that and move the cut point to 2950 or something like that? (lets assume you have fueling in place). Furthermore, is it possible to manipulate redline? Also when do you believe you guys will be selling these?

Realistically, my application is all 2jz based since my mk3 is 2J right now.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
As you note there is no real point in installing the Lexus AFM and then running 440s as you will be fuel limited. However, fuel cut is manipulated in these ECUs such that it is about 24% higher than stock, so there is about the same improvement as with the original Lex/550 mod.

This is just an interim test. I wanted to make sure we could run the Lexus AFM with no drivability concerns, after the modeling mentioned above showed it could be an issue. The next step is to double the fuel cut limit. That will require some surgery on internal maps, and there are a lot of them.

Rev limits and such are easily manipulated, along with lots of other things like fuel pump speed control, open loop transitions, fuel enrichment, and so on. My beta testers have been a huge help sorting this stuff out, top notch bunch of guys.

2JZ-GTE dual processor ecu is another can of worms. Jon Sole has a solution for the 3S-GTE that in principal works for the 2JZ, but because space it tight and things are situated differently in the 2JZ ecu, his solution doesn't quite fit on some versions. Also, later ECUs use a different chip package that we don't support either. So 2JZ is not a focus at the moment.
 

Zazzn

OLD school
Authorized Seller
Joined
Apr 1, 2005
Messages
1,054
Likes
14
Location
Toronto / San Francisco
Help me understand what is going on here.

You stated that there is a VE (load) table that FCO is based off. If you hit > 100% it will trigger FCO. You have the ability to go to 114% (with some tweaks you said in the previous pages) If you were to go tot 200% you basically would need re scale all the maps and match the injectors. If this logic is right could you not simply design the software to interpolate and export to the rom custom new maps like a standalone? Meaning, I install say 1200 CC injectors, I set the injector size in the software (like emanage or AEM) and it rescales the injector map to represent the duration of the size injectors i just installed. After that there would need to be fine tuning as I'm sure the interpolation would be off a bit (just like emanage and aem). Realistically, I dropped the 7m may years ago because the 2jz was just more reliable and easier to make power on however, I've always preferred not to surpass the power threshold of stock computer + emanage, because while there are good tuners, I don't believe any of them come near the complexity and amazing tune of the factory computer.

I realize we are looking in to the 7m ECU, but I though you said similarly the other party looking at the 3s ECU was similar to the 2jz, wouldn't the software and hardware end be very easy to combined to allow a multi ECU solution?

(out of scope for this project but non the less curious question) I believe one of the main things here is that the 2j is very abundant now, and is virtually swapped into every platform almost. Providing a project like this after this one completes to 2jz would generate an insane amount buzz (as this one has).

I wish I knew how to help you or even where to start to do this for the 2jz side of things.


I'm trying to understand this in terms of how you are doing this vs say a standalone AEM which will do the following.

I can adjust Timing/Injection maps completely independent of each other. Meaning if I add fuel, I still have ignition maps which are specified. Wouldn't it be easier to adjust the single map than the whole Load/VE map? I understand adjusting the VE makes it really easy to tune and use what toyota has put in, but using toyota as a base line and building from there is 3/4 of the way there and most supra owners I feel can do this. The VE / Load adjustment could be done by every supra owner I agree, but that will take much more work for you guys and is much less flexable for the guys that want to run more power say in the range of 500-600 HP. I can see this all making sense, if I could disjoin (via software) injection/timing maps and then the software puts it all back together when writing to the rom.

Meaning I setup in the AEM/emanage software the type of injectors / size. I can set some basic dwell settings or (you tell us what injectors we need ranging from the common upgrades) RC550/680/880/1200 we input what we have, and then the maps (don't have to be filled out by you, we could scale them and fine tune them as needed cell by cell like any standalone with interpolation features so that I can say well i had 440's now I have 1200's so here's the % of difference lets reduce our injection map by that amount. Then the end user can clean up the map to fine tune if needed.

I realize that this is focused on the 7m
 

Rollus

Member
Authorized Seller
Joined
Jun 2, 2011
Messages
593
Likes
2
Location
Paris, France
Actually, we can edit independently these maps (incomplete list, I've selected main maps only)
  • Max Knock Retard Map
  • Ignition Timing
  • Fuel Enrichment
  • Short PW Adder Cal
  • Idle Speed
  • Injector Latency vs Vbat
  • VE
  • Fuel Cut
  • Injector Scale Factor
etc...
 

Zazzn

OLD school
Authorized Seller
Joined
Apr 1, 2005
Messages
1,054
Likes
14
Location
Toronto / San Francisco
So, when do you think a full product will be available? Sounds like it's a standalone now pretty much! I mean end of the day i realize what FCO does, but if you know what you are doing and tuning, you can by pass this like most do with FCD, and tune away extending the maps!

God i wish this work was done on the 2jz ecu's...

I hope you guys tackle that next.
 

GC89

BoostRodeo
Authorized Seller
Contributor
Joined
Jun 13, 2007
Messages
1,035
Likes
22
Location
Spokane, WA
3p how does one become one of the lucky few beta testers. I have been following your assembla project for a while now, but unfortunately had to step away from the car for a bit until recently.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
There is at this point no difference between what we can do to the 7M ecu and a standalone, except that the code comes with a factory tune that was optimized only for the 7M. Anything and everything can be changed or modified.

The caveat is that the whole thing is written in 8 bit assembly language, and the math is highly optimized to the factory expectations for the engine.

To give an example, the load variable is 16 bits. The factory factory fuel cut occurs at C800h. The variable overflows at FFFFh (128% of factory fuel cut), so that is the upper bound for simple mods to this variable. If you want to handle more than 28% you need to scale this variable, and since it is an index to most of the tables in the ecu, they also need to be scaled. Its not like its an impossible task, just a lot of work to get it right, and because its assembly language, its not easy to automate it so room for stupid errors.

That is why we are proceeding in steps. First step was to support injector scaling and SPA so that we can use bigger injectors. Then integrate the Lex AFM and maintain factory drivability. Both are done now. Next step is to support loads above 128%.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
You stated that there is a VE (load) table that FCO is based off. If you hit > 100% it will trigger FCO. You have the ability to go to 114% (with some tweaks you said in the previous pages) If you were to go tot 200% you basically would need re scale all the maps and match the injectors. If this logic is right could you not simply design the software to interpolate and export to the rom custom new maps like a standalone? Meaning, I install say 1200 CC injectors, I set the injector size in the software (like emanage or AEM) and it rescales the injector map to represent the duration of the size injectors i just installed. After that there would need to be fine tuning as I'm sure the interpolation would be off a bit (just like emanage and aem). Realistically, I dropped the 7m may years ago because the 2jz was just more reliable and easier to make power on however, I've always preferred not to surpass the power threshold of stock computer + emanage, because while there are good tuners, I don't believe any of them come near the complexity and amazing tune of the factory computer.
Load is a calculation, based on air flow, air temp, barometer, and rpm mainly. Its not a map, but is used as an index to many maps because it is so closely tied to fuel flow. The injector pulse duration calculation is directly based on this load variable as well. Load and rpm are the main variables. The next most important after those two is air and coolant temp. I am not saying that higher loads cannot be supported, just that the way the code is written in assembly language makes these changes challenging. The code was not written to make changes easy, rather the biggest objective of the code was to make everything as small as possible so that it would fit in a 1980's chip technology that was extremely memory and processor power limited. What they achieved is remarkable, but the cost is its hard to modify later. Hard, but not impossible.

I realize we are looking in to the 7m ECU, but I though you said similarly the other party looking at the 3s ECU was similar to the 2jz, wouldn't the software and hardware end be very easy to combined to allow a multi ECU solution?
The 3S and 2JZ ecus are based on the same technology, and that technology is based on the 7M, since there is a lot of code reuse between all of them. The hardware is different though. A six cylinder has different needs than a 4, and the connectors and circuitry aren't compatible. There will never be one ecu that works on everything, because then you are throwing out the reason for doing this in the first place, which is to start with a factory optimized ecu and make just the changes needed to support your new goals. A single ecu solution is what you get when you buy a generic standalone from AEM. That market is already well covered.

The 3S and 2JZ solutions are very close. The problem is that the internal circuit board layouts are different enough that the 3S extender card bangs into a wall or an adjacent tall component when used in the 2JZ application. The fix is a reworked extender board that fits in a 2JZ ecu chassis. Such changes take time and money, and have not be done yet.


(out of scope for this project but non the less curious question) I believe one of the main things here is that the 2j is very abundant now, and is virtually swapped into every platform almost. Providing a project like this after this one completes to 2jz would generate an insane amount buzz (as this one has).

I wish I knew how to help you or even where to start to do this for the 2jz side of things.
Agreed. We have had queries on 2JZ applications for a long time.

I'm trying to understand this in terms of how you are doing this vs say a standalone AEM which will do the following.

I can adjust Timing/Injection maps completely independent of each other. Meaning if I add fuel, I still have ignition maps which are specified. Wouldn't it be easier to adjust the single map than the whole Load/VE map? I understand adjusting the VE makes it really easy to tune and use what toyota has put in, but using toyota as a base line and building from there is 3/4 of the way there and most supra owners I feel can do this. The VE / Load adjustment could be done by every supra owner I agree, but that will take much more work for you guys and is much less flexable for the guys that want to run more power say in the range of 500-600 HP. I can see this all making sense, if I could disjoin (via software) injection/timing maps and then the software puts it all back together when writing to the rom.
It is the same here. Just some extra limitations that comes from a design that uses only integer math, and limitations on variable sizes and such due to 8-bit 1980s technology.
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
3p how does one become one of the lucky few beta testers. I have been following your assembla project for a while now, but unfortunately had to step away from the car for a bit until recently.
Not much glory in being a beta tester. You have to put up with my mistakes that can crash the ecu when you try to adjust a variable. An LRIP run is in the works.
 

GC89

BoostRodeo
Authorized Seller
Contributor
Joined
Jun 13, 2007
Messages
1,035
Likes
22
Location
Spokane, WA
Understood, please keep me posted on the limited run or if you would like another tester. I am currently on 440s but have a set of P&P 680s I would like to swap in eventually. Have you tested injector scaling with anything bigger than 550s yet?

I am happy to deal with R&D hiccups, they have been the story of my cars life :)
 

3p141592654

90T
Authorized Seller
Contributor
Joined
Oct 11, 2005
Messages
4,094
Likes
66
Location
Thousand Oaks, CA
Reynald is running 960cc injectors with his E85 setup, so yes we are running big injectors. It is a big help to have good injector calibration data, unless you have the patience to work out the parameters on your own. Injector Dynamics and Deatschwerks are the only companies that seem to have a good handle on this.
 

Zazzn

OLD school
Authorized Seller
Joined
Apr 1, 2005
Messages
1,054
Likes
14
Location
Toronto / San Francisco
Hi 3p,

What if I didn't care about the ECU closing up nicely with daughter board, and I made a jumper cable or sorts to allow it to run outside the 2jz ecu housing. I would certainly beta test if you want to go down that road. I have no fear of picking up a solider iron. I have a very nice station, just gets tricky on modern solders which are very very small and I just don't have stable enough hands to make those connections well ;). Them 90's boards though are no problem, like doing a chip in a playstation 1 back in the day.
 

GC89

BoostRodeo
Authorized Seller
Contributor
Joined
Jun 13, 2007
Messages
1,035
Likes
22
Location
Spokane, WA
Reynald is running 960cc injectors with his E85 setup, so yes we are running big injectors. It is a big help to have good injector calibration data, unless you have the patience to work out the parameters on your own. Injector Dynamics and Deatschwerks are the only companies that seem to have a good handle on this.
The 680's I have are Denso's so I would have to do some analysis on them. I have done all of my own tuning to-date on all of my various vehicles so I am aware of the time it takes to get it right. The tccs is more complicated than most standalone so It will take much more time than an off the shelf box but that is also what appeals to me about it. Cant wait to see what your planning on releasing.
 

Rollus

Member
Authorized Seller
Joined
Jun 2, 2011
Messages
593
Likes
2
Location
Paris, France
The tccs is more complicated than most standalone
No it's not.

My main concern for injector tuning was to not have data available, so I've tuned with an empirical method. If I ever change my injectors, I will build a flow bench, that's easy with fuel system left over.

What may be easier with a standalone is maybe copy/pasting data from others because it's more available, but 3p gave me basemaps from Ford engine for injectors function.

++

Reynald
 

nuttom

New Member
Authorized Seller
Joined
Jun 29, 2014
Messages
11
Likes
0
Location
SIGOLSHEIM, FRANCE
I have a metallurgical micrscope that I use for DIE inspection, as said a couple of week ago I was able to read the partnumber and manufacturer's logo for the knock MCU :

The IC is labelled:
D151802-4800
74331380
1B2524

On the DIE is a Fujitsu logo and also "FUJITSU", "87" maybe a year, and "MB88572" supposed to be the partnumber of the IC.

I made the same job for the main MCU labelled as follow:
D151802-4800
7433-1380
1B2524

unfortunattelly I damaged the DIE and was only able to inspect 3/4 of the surface, this is what I found :


Is it possible that the chip came from SGS ? (back in time SGS and Thompson merged and ST micro was created)