Members Login
Username 
 
Password 
    Remember Me  
Post Info TOPIC: Turbos and stock ecu... controlling fuel based on boost...


Guru

Status: Offline
Posts: 2338
Date:
Turbos and stock ecu... controlling fuel based on boost...


While reading the code (for yoshbox and stuff) come to read the ambient pressure sensor stuff...

Basically ambient pressure is used for two things:
- While on vacuum (IAP maps) the ambient - manifold pressure sets the X-axis for map value, i.e. load value
- While on TPS map the ambient pressure is used for compensating any air pressure changes. E.g. including riding in mountains with less pressure means less fuel required.

Anyhow the ambient pressure also works other way round... if you have more pressure it will make the ECU to deliver more fuel also when on TPS maps.

So basically if you hook up the ambient pressure sensor to your airbox in the turbo engine you will detect the boost as ambient pressure increase and make the necessary fuel adjustments. As the there is no boost when TPS<11% the ambient - manifold calculation is not affected. The initial test on workbench reveal that up to +30% more injector pulsewidth can be delivered with higher ambient pressure.

The biggest limitation we have here is the stock pressure sensors ability to detect any significant amount of boost so to utilize this basically may require changing the ambient pressure sensor.

If and when anyone becomes interested about this concept I am more than happy to start looking more in to this. There is e.g. a couple of compensation maps for ambient pressure meaning that we may well be only talking about adjusting a couple of maps when changing the sensor and enabling a stock ecu being used with a turbo...

Anyway with one reservation - I am only talking about big singles (injectors) here. Dont know if the transistors would be good to drive secondaries... ? Usually its 11-14Ohm/injector, that means 13.5V/11Ohm= 1.23A additional load per driver transistor (MN611S). The spec says +-6A or +-10A in pulses so we are propably talking about a possibility to drive also secondaries from stock ecu drivers in parallel to originals.


__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 1247
Date:

Perhaps a different sensor could be used that will handle boost, and its output modified with resistors to narrow the voltage, thus making the new "ambient" smething closer to a lower value in the vacuum map, and making positive pressure be where current ambient or slightly higher is now.

After re-reading that, it is fairly difficult to read and inderstand what I am referring to...

On a side note, is there a way to change the bins in the maps? i.e. making the highest number in the map be 200 or 300 kpa or possily change the target "zero" vaccum number to be lower and add more compensation for a positive pressure?

Basically, if the lower two or three bins could be used for 100 kpa and vacuum and the remaining bins be used for boost, then another pressure transducer could be used (with bias resistors if the ecu input range cannot be modified) to make it a +/- pressure sensor.

John

__________________


Guru

Status: Offline
Posts: 2338
Date:

If we take any pressure transducer and replace the stock atmospheric sensor with that we can also reprogram the maps for that sensor so that when not at boos the ecu operates normally - but when at boost the sufficient amount of additional fuel will be delivered.

If you choose which pressure sensor you would like to run and have me to have a spec for that I can preconvert and test the map so that the stock ecu can handle all that is needed for calculating and delivering fuel to your turbo bike.

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 741
Date:

Petrik since ECU EDITOR now allows us to change the switching point between IAP and TPS MAPS we could make the IAP maps smaller so we could tune the small throttle opeing better, using the TPS maps when under boost.



__________________


Guru

Status: Offline
Posts: 1247
Date:

Here's a different take....what about switching the TPS with a pressure transducer and the pressure transducer with the TPS.

Then change the rpm ranges so that the new TPS correction (old IAP maps) work over the whole rev range.

Then use the 0-5V from the pressure transducer hooked into the TPS input to control the engine with pressure vs rpm rather than tpe vs rpm.

Basically the way some aftermarket ECU's and megasquirt works.

John


__________________


Guru

Status: Offline
Posts: 741
Date:

the guys that use two power commanders to controll boost does something similar by connection a GM 3BAR map sensor to the TPS input of the powercommander.  

__________________


Guru

Status: Offline
Posts: 2338
Date:

You dont really need to do anything else than replace the ambient air transducer with something like GM transducer which can detecth higher pressures than the stock transducer. Then you connect that transducer to your airbox so that it will detect boost on top of the throttle bodies.

Then you can:
1) Adjust boost map by modifying the pressure transducer map (i.e. How much fuel is added per psi (or bar) in the pressure transducer map)
2) Tune by using the TPS map

IAP map is not affected by this. I may need to remove the fault code generation from the IAP sensor when it detects higher pressures to remove the need for a pressure relief valve from the intake air pressure sensor vacuum line.

And yes - regardless of the above you can make the IAP map smaller and just run TPS map - but dont see any benefits in doing that. The bike will run better with a properly tuned IAP map, even or should I say particularly if it is a turbo. In many cases the vacuum is less on turbos, so modifications are bigger.
Everything else is already built in to the bike. I tested it yesterday on the workbenc and seems to be working almost like the ecu would have been designed to be used with turbos too.

-- Edited by PetriK at 06:29, 2008-08-26

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Veteran Member

Status: Offline
Posts: 45
Date:

You can't allow the IAP sensor to see boost at all as it ruins the sensor.

__________________


Guru

Status: Offline
Posts: 2338
Date:

You mean physically ruins the sensor - maybe... but you maybe should change that to another type too. We are testing next weekend to replace the stock IAP with a GM sensor.

Logically the Advanced settings now includes fixing the IAP FI light issue - positive pressure should not anymore set the FI light.



__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Veteran Member

Status: Offline
Posts: 45
Date:

I remember reports on either SH or 200mph that constantly lighting the FI light due to not running enough check-valves soon led to wrecking the IAP sensor. I think that the IAP is non-linear in it's operation too so fitting a replacement might not be so easy.

__________________


Guru

Status: Offline
Posts: 1247
Date:

Sweet. I wasn't aware that the ambient pressure transducer had its own map. That is excellent news!

So it looks like basically all features of a PC3USB, Ignition Module, and hub should be able to be accomplished with ECU reflashing and a bit of hardware on the flapper / pair solenoid signals In addition to being able to re-map the vaccum maps for idle and low rpm low tps operation.



__________________


Guru

Status: Offline
Posts: 741
Date:

if replacing the stock ambient pressure sensor with a gm 3bar works that would be perfect. for the turbo application...all i would need now is some bigger primary injectors..

__________________


Guru

Status: Offline
Posts: 741
Date:

Geri_Pete wrote:

I remember reports on either SH or 200mph that constantly lighting the FI light due to not running enough check-valves soon led to wrecking the IAP sensor. I think that the IAP is non-linear in it's operation too so fitting a replacement might not be so easy.




I REMEMBER READING THAT AS WELL....(I COPIED THIS FROM GARY EVANS FORUM)

Most people know by now that the Busa DFI utilizes two methods for determining fuel delivery. For low load (0-5% throttle) the IAP sensor signal and rpm are employed and for high load (10-100% throttle) it switches to throttle position and rpm. Beyond that the IAP sensor has been pretty much a mystery except for the knowledge that high boost pressure will turn on an FI light. Hahn solved that problem early on by clamping the IAP output signal voltage at 4.3 volts by the use of a Zener Diode. Everyone else just installed check valves to vent boost pressure, which is actually the better method since high boost can damage the delicate sensor. I recently went through some diagnostics to try and find the cause of wandering A/F at constant low load throttle and in the process I learned several things about the IAP sensor.
The first being that it has an unusual pressure to voltage ratio (slope of gain) that prevents the use of an off the shelve map sensor which hurts allot if you must replace one since the full retail price is about $250. That a bunch when you compare it to a GM 1 bar sensor that can be had for $35. It also has a strange operational range which equals about 1 1/2 bar as show by this graph.

IAP%20Sensor%20Test.jpg.

For comparison the closet U.S. map sensor would be a GM 1 bar and the following graph show the output of both. The Suzuki sensor may actually measure as low as the GM but I only tested the useable range. You will note that at zero pressure the GM voltage is at 4.8v and the Suzuki is at 3v. This differance could perhaps be adjusted with diodes or otherwise but you would still be left with a differance in ramp rates. It may or may not be possible to map around that differance with a Power Commander but it clearly wouldn't be a plug and play swap. The only other map sensor that I know of which would work for sure would be the Split Second programmable map sensor that cost as such as the Suzuki OEM.

Sensor%20Comparison.jpg

With the high OEM replacement cost it makes sense to protect the sensor as much as possible. Two things can easily damage it those being excess positive pressure and any liquids.
As mentioned previously the best solution for pressure is to employ a check valve but occasional even multiple check valves won't prevent an FI light. The reason is that sometimes even with the check valve/s open there is enough remaining pressure to trip an ECU error code. The reason can be seen from the first graph and the following test data on check valves. I used this set up for testing flow rates on several valves.

122-2213_IMG.JPG

I installed a restrictor right after the regulator that reduced flow to what you can expect from 4 throttle body ports (about .080 inch). Then applying a constant 15 psi and measured the remaining pressure with different check valves.
I found that 1 Suzuki OEM 1 valve only reduced the pressure to 8 psi and two only to 5 psi. I also tested several other plastic check valves from other manufactures with basically the same result. While it is possible to reduce the number of throttle body ports to cut back the air volume it is recommended to employ as many ports as possible to reduce pulsing. We know that somewhere between about 4.3 and 4.5 volts the ECU can produce an error code so it is obvious why sometimes even 2 check valves are not enough venting.
The best venting check valve I found was the large flapper valve shown in the picture which is made for fuel systems as has a big internal port. It kept the pressure to less than 1 psi even when the pressure in was increased to 20 psi. I just threaded the inlet for 1/8 npt x ¼ inch hose fitting and cut the AN threads off the other end.

The best way to keep liquid away from the sensor is to mount it above the throttle body ports with the sensor port pointed down. Any liquid that enters the hose will then be expelled through the check valve.

122-2214_IMG.JPG

The reason for protecting this sensor (beyond its high cost) is that unless you running a wide band 02 sensor and display you may not know that the IAP sensor has been damaged until drivability is effected and it is possible in time to run lean enough at cruise to burn a hole right through a piston crown.
I also suggest you don't buy a used IAP sensor that came from a turbo bike unless you feel lucky.



-- Edited by busa2001 at 00:45, 2008-08-27

__________________


Guru

Status: Offline
Posts: 2338
Date:

Thanks for the GM map, chaps - we can now remap the IAP valve to GM spec. Luckily for most sensors that ECU has there is a map for sensor information to make it ecu information.

Anyway regarding pressure sensors as the information is used where the ambient less manifold calculation is used and then maped to a value - it means that both sensors (ambient and manifold) should be the same type.


-- Edited by PetriK at 01:59, 2008-08-27

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 741
Date:

member of innovate motorsport forum send me this gm 3bar map sensor table. not 100% sure about it's accuracy.
 PSIg Volt   
-28.5  0.078   
-27.1  0.156   
-25.7  0.234   
-24.2  0.313   
-22.8  0.391   
-21.4  0.469   
-19.9  0.547   
-18.6  0.625   
-17.1  0.703   
-15.7  0.781   
-14.8  0.859   
-12.9  0.938   
-11.4  1.016   
-10     1.094   
-8.6    1.172   
-7.17  1.25   
-5.75  1.328   
-4.33  1.406   
-2.91  1.484   
-1.49  1.563   
-0.06  1.641   
0.67    1.719   
1.37    1.797   
2.06   1.875   
2.76   1.953   
3.46   2.031   
4.16   2.109   
4.86   2.188   
5.56   2.266   
6.25   2.344   
6.95   2.422   
7.65   2.5   
8.35   2.578   
9.05   2.656   
9.75   2.734   
10.4   2.813   
11.1   2.891   
11.8   2.969   
12.5   3.047   
13.2   3.125   
13.9   3.203   
14.6   3.281   
15.3   3.359   
16      3.438   
16.7   3.516   
17.4   3.594   
18.1   3.672   
18.8   3.75   
19.5   3.828   
20.2   3.906   
20.9   3.984   
21.6   4.063   
22.3   4.141   
23      4.219   
23.7   4.297   
24.4   4.375   
25.1   4.453   
25.8   4.531   
26.5   4.609   
27.2   4.688   
27.9   4.766   
28.6   4.844   
29      4.922   
30      5 

-- Edited by busa2001 at 02:46, 2008-08-27

-- Edited by busa2001 at 02:47, 2008-08-27

__________________


Guru

Status: Offline
Posts: 2338
Date:

Excellent - now we can compare the GM3bar with stock sensor... that curve should be suitable too if both ambient and manifold sensors are GM 3bar type sensors.

GM3bar_sensor.jpg

-- Edited by PetriK at 02:52, 2008-08-27

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 2338
Date:

Have you guys heard of Toyota sensors for this purpose ? That seems to give same voltage at ambient pressure as stock busa sensor - says a local turbo guy who had used toyota map but got FI light which we now fixed with the latest ECUeditor version.

Based on this document the sensor bias may be resettable using a resistor:
http://www.autoshop101.com/forms/h35.pdf

http://gtfour.supras.org.nz/mapsensor.htm


-- Edited by PetriK at 07:32, 2008-08-27

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 2338
Date:

OK - having seen the specs for the Toyota MAP sensor it very well looks similiar to Busa sensors. Without having both on my desktop can not confirm but comparing graphs and known test results both fall within 0.2V of each other for all pressure settings and at ambient both are said to deliver same voltage. Based on spec this should be good for up to 1bar boost(14psi), particularly if both IAP and AP sensors are replaced.

Based on this information it certainly looks interesting - exept from one point of view. The Toyota MAP sensors are bit expensive here. I am talking about standard OEM Toyota Corolla MAP sensor - dont know the part number yet. But as talking about Toyota family car OEM replacement parts can not be expensive when proper source is found.

Is 1bar/14psi enough for most applications - or should we aim higher to start with ?







__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Veteran Member

Status: Offline
Posts: 45
Date:

14psi is good for 300-350hp depending on turbo size and with/without intercooler. That's also about the limit for stock rods, so what this means is that this would make an excellent and cheap upgrade for a FMU-based stage 1 turbo kit.

__________________


Guru

Status: Offline
Posts: 2338
Date:

So to have this for real pros, how much ? 1.5bar ?

That means different sensor and more work...



__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Veteran Member

Status: Offline
Posts: 45
Date:

Most popular is 3 bar sensor. That allows for up to 30psi of boost, not many people go over that.

__________________


Veteran Member

Status: Offline
Posts: 45
Date:

BTW a 1-bar sensor only does vacuum, not boost. the stock bike sensor is effectively a 1-bar because it's a measurement of absolute pressure. For 14psi of boost you would need a 2-bar sensor (1 bar of vacuum and 1 bar of boost).

__________________


Guru

Status: Offline
Posts: 2338
Date:

Yes, the Toyota is +-1bar sensor countin all vacuum areas and up to 1bar boost.

Must find a way for 3bar sensor on the ambient air...

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 741
Date:

1 bar won't be enough for most of us....smile i currently use a 3bar for my secondaries.



__________________


Guru

Status: Offline
Posts: 2338
Date:

3bar boost ???? no way !


__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Veteran Member

Status: Offline
Posts: 45
Date:

Lol no, 2 bar boost, 1 bar vacuum, same as I use :)

__________________


Guru

Status: Offline
Posts: 2338
Date:

Yep - even 2 bar is incredible amount of boost. Highest we have got around here is about 1.4-1.6bar. I recall that even Charlies 256mph busa was below 2bar boost when 636hp was dynoed - would need to check from my notes for the exact figure.

Anyway - would need to test how ECU works with GM sensor voltages then. Should be able to make that work too - even its not as easy and fast.

The first step would be just to hook up the OEM pressure sensor into the airbox and verify that the idea behind really works.






__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 1247
Date:

We run 27-29 psi (1.8-1.9 bar) on our turbo bike and "nly make about 40-460 hp. The thing is, turbo bikes can make big power on less boost psi which is of course very desirable, but with a smallish 50 trim turbo and only water injection rather than an intercooler, more boost per horsepower is required. I am seeing near 300 degree F intake temperatures AFTER the water injection.

If one had to pick a map sensor, the GM 3 bar would be the way to go, at least for in the states as it is very easy to get.

The issue will likely be injector size and being able to control very large injectors at low rpm / power levels. As most are aware, the S2000 injectors are a handful to tune to get a good idle and low RPM quality (although this may be much easier if you use the IAP mapping rather than just power commander tps modifications).

Large injector control at lower rpm / power as can be done with a Motec or other high end aftermarket unit, with stock reliability (and price!) woule be great. THe low pulse width of large injectors can be problematic for big horsepower (>350 hp) applications with single injectors.

__________________


Guru

Status: Offline
Posts: 741
Date:

with the range of adjustment we have with the stock ecu, we should be able to tune the low end much eaiser and better for larger injectors than in the past with the power commander. it just going to be time consuming. 

__________________


Guru

Status: Offline
Posts: 2338
Date:

OK - lets try the GM 3bar sensor ... I dont know if I can easily get one, but at least can test the ECU with the voltages from it.

How about this issue of IAP sensor failing under higher boost - is there any truth in it ?

I know racing ECUs also fairly very well - in our racing kawis we have KMS units... from my experience so far the stock busa ecu has been more versatile.

I dont understand the big injectors being problematic to tune. In my NA I have S2000 with HP fuelpump without any problems ? Please let me know why those are difficult to tune to understand if something can be done to the tuning process.

So you guys run water injectrion - same here with low output turbos. For bigger power (400+) they run intercooler.

Looks like the chaps over here are still 0.5bars behind. That is interesting... as it means propably some differences in exhaust and manifild tubing as well.

-- Edited by PetriK at 02:36, 2008-08-29

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 741
Date:

I have heard of more than one failing under boost a few years back, none since we started using check valves....

The problem with the larger injectors in the past was you could only tune by tps and rpm using the power commander, and you had to set the static pressure very low to tune them half way decent, it was alway a compromise.

Most of us are running water injection, it's cheaper and takes up less room and you can still lower the bike. i run an Aquamist 2c.   

__________________


Guru

Status: Offline
Posts: 2338
Date:

The intake air pressure sensor needs also be nominated, i.e. if stock can be used or something else.

We have had difficulties with one way valves with IAP - that messes up the measurements by IAP valve as the pressure from cylinders is not steady - it has peak vacuum but may well also have quite high pressure peaks while running. That means that the valves we have found do not work reliably so far.

Please clarify the IAP valve destroying stories. How ? What boos ? How was it detected ?

E.g. On AZTECH forums I have found so much confusing misinformation about IAP:s and other things that I could not read it throug without being very confused how someone can release so much misinformation.

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 741
Date:

any updates Petrik?

__________________


Guru

Status: Offline
Posts: 2338
Date:

I have some confirmation that stock IAP do not get bad when running normal boosts (0.9...1.4).

Additionally one turbo user measured that the stock toyotoa MAP sensor is outputting exactly same vacuum values as Suzuki IAP and Maniforld pressure sensors.

The software code with IAP turbo setting was released long ago.

So I am just waiting someone to remove the vacuum one direction valve and setting the IAP extension for turbo switch on using ECUeditor. Then I am just waiting for someone to report back that it improves the driveability and tuneability. Should happen here locally in a couple of weeks, but its getting autumn and water is pouring from the sky so not many of us are actively running the bikes for coming 6 mths.



-- Edited by PetriK at 07:54, 2008-09-07

__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Member

Status: Offline
Posts: 16
Date:

I've got a question regarding the ambient pressure sensor...

If we change the 2D-Table for the pressure sensor, how much weight do we have to start with?  meaning, if I ran this ecu on a 5Bar map sensor, (410% over amtosperic pressure).. given a semi-non-linear output from the sensor itself..
we could find ourselves in a position to where the ECU is having to calculate between breakpoints/points from the 2D table.

So if we've come thus far..  why not decompile the binary image & recode to use more room for the non-linear/linear 2D table(s)?  This is a pretty advanced ECU in the long view of things.  I've kept an eye on this (from an embedded developers point of view) thinking this would could be very slick on another application.

__________________


Guru

Status: Offline
Posts: 2338
Date:

The values used are weighted averages between two points (if I remember correctly) so dont really know if adding more points adds much value.

The calibration table is not included into ECUeditor - maybe in the future. Anyway I would start the calibration by measuring the pressure variable output using ECUeditor datastream and then convert the values to output the same numbers in that variable.



__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Member

Status: Offline
Posts: 16
Date:

Let me take this from a different point of view - has the source been decompiled/disassembled?
I've got access to IDA Pro Advanced which should support the processor. Looks like RR has done a good deal of reverse work - is this true?

I'm interested in opening up the ECU (source) and maybe opening the doors to what we could run this on.

__________________


Guru

Status: Offline
Posts: 963
Date:

Actually on the Busa things are kind of messed up. PetriK did a lot of the original work on it with IDA then he sent it to me. I redid some of his stuff and did some new stuff. Meanwhile Petrik did a lot more work with his as I moved on to other bikes.

It is hard to say at this point who has the more complete example. The best version would be a combination of the two but at this point it would be a real mess to integrate. It would contain

Original Petrik
Original PetriK modified by RR
new PetriK
new RR
overlapping new

__________________


Member

Status: Offline
Posts: 16
Date:

lets get something going in terms of a single version.. ?? :)

__________________


Member

Status: Offline
Posts: 16
Date:

ohh for whats its worth.. you can snag most sensor values/ranges etc from AEMPro software.

Personally I've done it.. for o2 sensors & amps, map sensors, coolant,etc. its a PNP ecu with tons of base maps so you'll have plenty of sensor-data to collect from.

__________________


Guru

Status: Offline
Posts: 2338
Date:

It may be more beneficial to enable e.g. the temperature/boost correction and other maps which are not utilized in stock configuration rather than just go for bigger one map. Everything indicates that busa stock ecu is turbo enabled to start with.

For a map extenstion I would not expect much more being needed than to rewrite the map size word and locate the map address to an unused area. No disassembly is needed. But still dont understand the benefits of that ?

We have also started to write plugins to the ecu code, i.e. using the Renesas HEW we are compiling C-programs that then get executed with the ECU itsef. This is our strategy to look into the frequency controlled solenoid issue - e.g. nitrous or boost control solenoid.



__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Member

Status: Offline
Posts: 16
Date:

So what you're saying is you're recompiling & debugging your recompiled binaries?

you're surely not compiling C-code right? if you were then you'd have the disasm code back into embedded C? maybe i misread you.

the benefits of higher map resolution is simple. I know the euc works on a much larger scale than 12x12 but assume it was 12x12 and your RPM range was 100-rpms to 9000-rpms... that gives you a 750RPM incremental step.. so every 750 rpms you've got a breakpoint/cell to edit for RPM axis. the same goes for the Y axis.

if you're at 32x48.. at 32 RPM points on a 100-9000 scale, thats about every 280-300 rpms. now the higher count of rows and columns, the better off you are. you have more detail to the map;

Keep in mind, my goal isn't to run a motorcycle, i'm using the busa engine in a custom setup (think Micro Sprint car) with a med-sized turbo setup. so my needs are a little bit different than yours. I can't use alpha-n method as I dont 100% argee with it. using alpha-n assumes your barometric pressure stays about the same (or atmospheric pressure is static). on top of that, alpha-n .vs speed density.. SD wins only because if you couple that SD setup with a baro-sensor.. you'll never have to return your maps again (given you can add code to compensate based on the baro input).

another thing would be injectors.. are you in a position to run non-factory injectors (low resistance or high resistance.. doesn't matter) by modding the source? say if i ran 550cc injectors on the ECU, how do I change the PW and dead times? I think this is a great project to be a part of..

What would be really cool is to setup a simple interface / GUI that lets you `build` the coolant sensor calibration, crank ign tooth setup, etc. or even change the dwell times on the coils. or reconfiure the source to maybe fire sequential .vs wasted spark?

...and just MAYBE we could get to a point to say build a drag application ECU to where any unused inputs & outputs could be used for aux injection or something along those lines.

could you imagine an ionic-ignition capable platform?

__________________


Guru

Status: Offline
Posts: 2338
Date:

Nope - recompiling of the binaries was discussed some time ago. RR has done that, to me it sounds too risky as it adds a point of failure without adding understanding of the controlling logic.

I modify the code directly with either parametrizising or writing assembler instructions into the binary. The add on modules we write with C that we compile and then relocate the binary to an unused space. The first module is kill shift module (shifter) which you can install using the GUI that was named as Hayabusa K2-K7 ECUeditor.

Busa stock ecu has both fuel calculation methods, speed density and alpha-n already built in. We can set the fuel calculation method what I call valuum vs tps can be already set by a parameter called map switching point using ECUeditor. The stock ecu runs cruising based on vacuum pressure and higher rpms based on tps positions. The tps calculations always include also barometric correction, which means that if you relocate your pressure sensor you can actually use the barometric correction as your boost signal.

The injector size can be set with fuel pressure/injector size value and fuel maps can be modified to change your pw times using the ECUeditor.

All this can be modified using a graphical user interface already today, as you must have seen on this board. Dwell changing will be added to the GUI as soon as we find the right way to representing it on the Gui.

With the existing code the crank sensor is something that is a bit of a work to change. That part of the code I have not yet completely grasped.

Dont really see any benefits of changing the fuel maps to any bigger resolution than current 21x43. Its already now a beast to tune on the dyno. Anyway you can use the GUI to modify those.

For your car setup you can use the GUI to have a small screen installed to your ****pit and view any of the engines 74 parameters on line using the data monitoring screen - including error codes if you dont have busa std gauge installed.

In addition to the above the stock busa has maps e.g. to calculate ignition advance based on intake temperature and pressure - to me its something that is very useful for turbo engines. Anyhow as stock those are not used, but you can take em. in use by just changing one parameter value. The biggest hurdle to understand the programming logic fully, to understand how changing one thing affects another.

All that is available already today smile.gif - what more do you need ? Of course some of us are pioneers who want to reinvent everything and in that case the world is not enough.





__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Member

Status: Offline
Posts: 16
Date:

Everything is a point of failure.  Seriously. 
Seems like you're very happy with what you're doing.

I hope I dont offend you here but, seems like you're just creating a bunch of work-arounds.  I dont know why you're really bothering with modofiying the binaries when the changes you are making are do'able via hex-editor or OpenECU-related.  What I mean by that is this: I dont understand the efforts put forth when you're not after the whole cake, only a single slice.

Just follow me here.. If I handed you a fully documented source with the ablity to compile the code, giving you an extact duplicate of the factory binary.. where is the failure?  With documentation (as you're already doing), I could hand you the code, let you read it, understand it then make changes based on your needs.  Hell you could even wire in better sensors (IAT, Coolant, etc) as long as you knew where the logic & lookup tables were.

So (to me) it seems most logical to completely understand what you're working with before saying "its only a matter of changing PW for other injectors".. which isn't true.  since the ECU has injector drivers built in for high resistance injectors, you can use injectors that draw a lot of power... BUT high resistance injectors aren't best choice.. I wont go off on a rant here but there is more possible here than whats surfaced.

I take it you're more the programmer than a tuner, right?  I'm really not trying to pick on you at all, in fact I followed you here from another forum.  I am maybe too blunt and sometimes come off harsh.. sorry if i have.

__________________


Member

Status: Offline
Posts: 16
Date:

btw, you really shouldn't replace the Baro sensor on any ECU.
its there as a reference - if you tune your bike at sea level then move up to the mountains... it wont run correctly as air up there is too thin, etc etc... it'll run pig rich. so you'll need some type of ambient air pressure, etc. ...just my 2cents

__________________


Guru

Status: Offline
Posts: 1247
Date:

I believe he was working on replacing the baro sensor and plumbing it into the airbox. This will still take a baro reading on startup and will still read moderatel close in off boost sitations as long as it is not below the throttle blades. Some aftermarket ECU's use this method.

I believe the whole point of the ecu hacking work is to tune a hayabusa engine while in a hayabusa. Sure it can be done in another environment and ceratinly it can be done differently, but it seems to work pretty damn good for its intended purpose. Many areas of the code are still unknown and are being looked at piece by piece. I am sure other features will be looked at that will make it more tunable outside the standard Busa environment. Making the ECU drive low Z injectors may require a hardware change rather than just software. If you are looking to use low Z injectors due to the inherent benefits relating to opening adn closing time or other benefits, you should probably invest in a high end aftermarket ecu and look at multiple pulse and multiple spark and higher resolution tuning anyway.

__________________


Guru

Status: Offline
Posts: 2338
Date:

hwnd, dont worry - not offended, any fresh views will generate thinking processes.

Let me give you some background of my involvement on this project.

I started to work on busa ecu as i just wanted to make my bike run faster. This summer went 208mph on a mile, could have run more but there was not enough fuel from the pump so did not want to push it. My interest was sparkled because of realizing the limitations and failure of a powercommander. Then having history with disassembling some car ecus i wanted to pursuit this route.

Initially I used enginyity and alike mapping tools in dyno, but those are very clumsy on dyno. So that lead me to write a software that can be used in dyno to simplify fuel and ignition mapping. You will understand what i mean if you ever try to tune busa ecu using other tools than ecueditor.

Saying all the above, please note that even worked with it for 20+ years, i am not a programmer - rather i pursuit excellence in engine building. This hobby includes my own engine building and being a mech in this years winner pro street dragrace team.

Recompiling is a word that is technically appealing, but involves unnecessary steps for tuning process. In real tuning world i am yet to see a builder who could recompile a source. Mapping is what they/we understand. Therefore we have a GUI for mapping that has been developed running it dyno for over 1000 runs.

The key here is to understand how busa ecu works and then making a tool that anyone can use in dyno environment. Recompiling does not add value in understanding the functionality of stock ecu.

Sounds that you share some views and desires with RR, please talk with him about this idea. To me its not worth while thinking more as there is very little practical value for most of the busa owners. If i wanted to get a custom ecu i rather would choose kms or maybe megasquirt - those i know well having worked with both. If i wanted to start recompiling my own ecu source code, would go for megasquirt.

E.g. About the baro sensor - if and when you know more about the logic how ecu works you come to realize that of course it can be changed. Then when you test it on your workbenc you can make a proof of concept. Then it can be put in practise.



__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Guru

Status: Offline
Posts: 2338
Date:

oh -let me add one thing here which may not become clear otherwise. We do have a fully disassembled and commented source we are using to understand the programming logic. it would be very easy to output that from disassembler to assembler and regenerate the code if one has a compiler environment up and running.

Anyhow based on this disassembly work i can say that the stock busa ecu code is extreamly versatile. Many modules are taken into use by e.g. just converting one 0x00 byte to 0x01 or 0xff. All (most) the sensors use a conversion map so changing any sensor is easy. Even map size is parametrized by map definition longword - so changing map size is just a matter of changing the longword and replace pointer addresses with the new tablle.

Based on this and previous experience with denso car ecus I believe that denso is using same basic program for many engine manufacturers and configure the code just by parametrization and only change few things. Many of the subroutines are same or similiar as used widely in the industry - e.g. calculating the ignition using timers and estimated crankwheel position.

But back to the business. The turbobusa (215mph/one mile) is back in the dyno as we speak. It looks very promising. Finally the powercurve goes steadily up until hitting the limiter. Previously the peak power was around 9krpm. The tuning is all done with ecueditor. Now with 0.8bar we already got 350rwhp. Next its time to up the pressure and go beyond 400rwhp. I believe the previous problem was because of both too large and misadjusted cams for turbo use. Now its 2 x stock intake and its way better.

We will be having two maps on this bike, one for 0.8bar other for 1.4bar and use the stock ecu map switching feature.


__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom



Member

Status: Offline
Posts: 16
Date:

Sounds great! The disassembly you speak of, is that available online?

__________________


Guru

Status: Offline
Posts: 2338
Date:

Early versions may be floating around.

I stopped publishing information as after one bad experience. Shared everything with someone just to see it being used in a software without any thanksm not even reference to this board which really is the mother of motorcycle ecu hacking. Then when complained I was promised to receive a test copy of the software, never received that. Anyway got an opportunity to try tuning using it and its really crappy so I understand why things went only one way.

Since that have been just publishing free tools for end users.

If there is someone honestly contributing to this information board then all the doors are open, but oneway street - done once, not anymore.




__________________

When asking a question, you can also consider posting it to facebook:

http://www.facebook.com/ecueditorcom

1 2 3 4  >  Last»  | Page of 4  sorted by
 
Quick Reply

Please log in to post quick replies.

Tweet this page Post to Digg Post to Del.icio.us


Create your own FREE Forum
Report Abuse
Powered by ActiveBoard