3P's TCCS Disassembly/Analysis

[3p141592654]

The max AFM frequency the ECU recognizes is 3333 Hz.

At 6000 rpm, 100% load is about 2000 Hz. I have never seen 100% load using a stock CT26 at 6000 rpm, you would need some mods to get there. So I believe the stock AFM has quite a bit of life left in it for a modified engine.

The maximum I have ever datalogged the stock AFM with a CT-26 is 1600Hz.

 

[3p141592654]

I got a PM about open-loop (OL) triggering, and thought others might be interested....

There are two classes of events, one results in instant triggering of OL, and the other is delayed by a timer.

The primary trigger is a table that sets a maximum load vs rpm for load. If load exceeds this value OL is indicated. Basically, at low RPM, the trigger point is 44%, and above that it is interpolated to 0% at 7600 rpm (redline is 6500). The effect is that it is easier to go OL at higher rpm.

Once the trigger is set, a countdown timer is started at 3s. When it hits 0 OL is activated. However, there are several conditions that bypass the timer. They are:
1) THW < 45C
2) RPM > 6000
3) throttle > 75% if rpm > 2000 or throttle > 45% if rpm < 2000
4) fuel enrichment > 43% (computed from 2D map)
5) engines stressed ( a calc based on ignition advance from high temps or knock)

So, the nice thing here is that the table can be tweaked easily to push OL sooner if you desire. I notice that the ECU fuel enrichment map is typically 0 below the trigger values, so there is some thought to matching the two. Also, mashing the throttle will force OL essentially immediately.

 

[Rollus]

That's an addition to what you already teach us ;-)

Actually its pretty straightforward. There is a map versus rpm that is compared against the current engine load. If the load is > than the map value it goes open loop:
2800 rpm 48%
7600 rpm 0 %

So for rpm <= 2800, if load exceed 48% it goes open loop. As the rpm increases, the threshold linearly drops, so at 5200 rpm the threshold is 22%, and so on. You can see that the faster the engine is spinning, the sooner it will drop into open loop.

There is also a throttle check, and if the the throttle is greater than a certain amount (48% to be exact) then it will go open loop regardless of rpm.

There are a bunch of other special cases that come into play, including rpm < 300, right after startup with no O2 sensor cross counts detected, derivative of the load variable, and some extreme odd behavior of the short term fuel trim variable, but those are just covering certain special cases.

 

[Nick M]

Recently I modified the base ECU software to raise fuel cut to 116% of the stock value. The reason I didn’t go higher was because my calculations showed that we would exceed 100% duty cycle above 5500 rpm, assuming VE is maintained at 116% at that rpm (5500 is where the stock auto 1-2 shift occurs). In practice, with the boost settings I have at the moment I get peak VE of about 108% at 4200 rpm, and it drops from there so it is probable that I am being too conservative.

Was the 19ms that you observed with the 440s?

 

[Nick M]

Nick, 3p ECU allows injectors and maf scaling, so he will confirm, but it allows you to make any combination you like.

I'm not sure to catch your MAF thing..

More than a decade ago, I asked Pro-M if they would build an airflow meter to replace the VAF (flapper door) and mimic the signal, and they said gladly.

 

[3p141592654]

Yes, 19ms would be with the stock 440s and AFM, but the available injector time is fixed regardless of injector size.

Going to larger injectors, like Reynald has with his setup, allows the injector time to be scaled back and avoid 100%. Then the challenge is at idle where the duration goes below 2ms. That's a nonlinear region for the injectors and requires some modeling changes to accommodate if you want a smooth idle. If you look at late model ECUs they add a few additional injector model parameters (for example Ford uses low (ALOSL) and high (AHISL) slope values and a FUEL_BKPT to switch between the two slopes).

 

[gaboonviper85]

So are you still offering this modification? Is it possible for me to send you my ecu?

 

[3p141592654]

So I recently found a Turbo-A ECU (GTTL trim) on Ebay and bought it to understand how they implemented speed density on the factory ECUs. I was hoping it would be all software, and we could modify standard 7M ECUs to support SD. Unfortunately,that is not the case. The ECU mainboard is different. It uses a different analog-to-digital converter (MP194) than other 7M ECUs. This makes it impossible to just swap over the code and go. Other than that though it looks very similar to a standard 7M. Same connectors (gray plug 26P 16P 22P).

The mainboard part number is 175031-2010. I am curious whether this board is used on other speed density 6 cylinder engines from 1990s, but do not have any real way to check. If anyone knows, please respond. If we could swap the mainboard going factory SD would be an option. Turbo-A ECUs are too rare to go that path, but maybe there is something out there that was made in more volume that we could use instead.

Does anyone know of any A70 engines that were SD other than the Turbo-A?

 

[Rollus]

The ECU mainboard is different. It uses a different analog-to-digital converter (MP194) than other 7M ECUs. This makes it impossible to just swap over the code and go. Other than that though it looks very similar to a standard 7M. Same connectors (gray plug 26P 16P 22P).

Why not use a 0-5V to 20-2000Hz (from memory) oscillator circuitry to emulate the AFM from a MAP sensor?

The mainboard part number is 175031-2010. I am curious whether this board is used on other speed density 6 cylinder engines from 1990s, but do not have any real way to check. If anyone knows, please respond. If we could swap the mainboard going factory SD would be an option. Turbo-A ECUs are too rare to go that path, but maybe there is something out there that was made in more volume that we could use instead.

Does anyone know of any A70 engines that were SD other than the Turbo-A?

The Map sensor and Ecu of the Turbo-A are very specific to the Turbo-A Supra, here are the cross-reference results in the new Live EPC:

As Ecu is not detailled, the mainboard reference shows nothing.

 

[LordDigital]

The mainboard part number is 175031-2010. I am curious whether this board is used on other speed density 6 cylinder engines from 1990s, but do not have any real way to check. If anyone knows, please respond. If we could swap the mainboard going factory SD would be an option. Turbo-A ECUs are too rare to go that path, but maybe there is something out there that was made in more volume that we could use instead.

Does anyone know of any A70 engines that were SD other than the Turbo-A?

Finding other 6 cylinder engines that ran SD will be difficult as at was too early for turbocharged MAP based EMS systems in the Turbo A era ,if you look at the R32 Group A Homologation model (NISMO R32 GTR) they maintained the MAF sensor for the 530 car run from the 5000+ production model.

Luckily the JZA70 with 1JZ-GTE has a SD based ECU ,there is a chance that the converter may be the same as I have a feeling that Toyota used the 250 or so Turbo A cars for staging for further SD based turbo engines.

 

[3p141592654]

Here is my guess. The SD code was pulled from an existing ECU, and replaced the AFM code. That would save a lot of development time. I am trying to track down likely donor ecus they might have used. Would also explain why they had to change the hardware from the standard 7M design.

I did not realize the IJZ-GTE was SD, or is it only some variants of that family?

 

[Backlash2032]

I thought all JZs were SD? Could also look to like Camry v6s. Not sure if those are SD or not. Pretty sure all the trucks ran MAF.

 

[Backlash2032]

The 3vz-fe is SD. 1mz-fe is MAF.

Edit... I misread. the 3vz is also maf.

 

[3p141592654]

What would be best is something SD before the JZ/VZ/UZ family, as these are later engines that use newer ECU technology than the 7M.

I do have the 1JZ ecu code, so I'll poke into it.

 

[JonS]

The 4A-G(Z)E was produced in both speed density and air-flow versions, later versions used the same 64-pin Denso CPU.

 

[LordDigital]

The 4A-G(Z)E was produced in both speed density and air-flow versions, later versions used the same 64-pin Denso CPU.

I was also going to say SD versions of the 4A-GE ECUs for emission unfriendly markets (JDM ,Thailand ,etc) ,also IIRC a the 128PS 4A-GE ECU was used by Toyota GB as ECM for their Group A race Supras in he mid 80s , it was modified by GEMS to work with the 6 cylinder engine and to be eprom programmable.

 

[Toyoloog]

I've got several different 4-GE ECU's laying around but the 16V N/A versions are all 40 & 42 pins. I still havent found a way to fabricate an 42 - 64 pins adapter for the ECU code readerboard.

 

[Nick M]

I thought all JZs were SD? Could also look to like Camry v6s. Not sure if those are SD or not. Pretty sure all the trucks ran MAF.

The MAP is bottom of the barrel. It also goes on Tercels of the era and 5SFE Camrys. The 2JZGE uses the same high dollar Karman Vortex volume airflow meter. The 2JZGTE uses the hot wire mass air flow sensor. I am only referring to Toyota Motor Sales which is under Toyota Motor Corp.

 

[Suprapowaz!(2)]

What is the most horsepower a stock afm has seen? When does it become a limitation?

BTW, I'm not trying to prove anybody right or wrong here. I agree with Nick, and don't like MAP based tuning. I'd prefer my stock AFM to do all of the calculations. This is why I'm very interested in the completion of this project.

 

[Rollus]

The 2JZGTE uses the hot wire mass air flow sensor. I am only referring to Toyota Motor Sales which is under Toyota Motor Corp.

EUDM and JDM 2JZGTE are using MAP sensor.