How is injector duration (ms) calculated?

Vetal · Post by **Vetal** » Wed Jun 27, 2007 6:06 pm

Not a high priority, cause I don't really care about inj.time (until injectors are at their max, then it would be good to see that), just let you know the problem

Matt · Post by **Matt** » Thu Jun 28, 2007 3:13 am

If you can send through a log of that I can play it back and have a look

Got another CA18 ECU here now with knock detection board working on it... so can have a look at knocking problems too

The Injection value, I have used reverse engineering and routine comparison against the R32 skyline again for this value. It is what the ECU is reporting for this value, and we do no additional calculation on it

What I can do (raised a CR on this) is look at the R32 and see if the pulsewidth on my scope matches that of the reported value. Then plug the CA18 ECU in and see if it does the same. Its the best way to check and measure this stuff

Also with the % adjustment.... I was thinking I can do this... using the current K value, mulitply by 1/ X% will give you a new K value which you can up/down adjust to and make changes in realtime so you can hold a button down and hear the engine change. item added for upcoming version

Vetal · Post by **Vetal** » Thu Jun 28, 2007 7:38 pm

Here is a log, you will see that ~6 times during the 2nd gear run injection time changes from ~10ms to ~5ms and back

Vetal · Post by **Vetal** » Fri Jul 06, 2007 9:16 pm

Any ideas on that?

Fusion Ed · Post by **Fusion Ed** » Fri Jul 06, 2007 10:44 pm

Interesting, You need a bigger airflow meter too.

Ed

Vetal · Post by **Vetal** » Sat Jul 07, 2007 2:38 am

Will do

Matt · Post by **Matt** » Sat Jul 07, 2007 11:33 pm

trying to look for correlation. seems to follow TP but just goes really rich some of the times (like spikes)

perhaps O2 sensor feedback related? i also draw O2 against it and was looking for correlation. something is making it richer

this log has some extra empty columns in it. had to remove those from the CSV file to get it working. fixed in latest NIStune version when you record now

Matt · Post by **Matt** » Sun Jul 08, 2007 1:33 am

E-manage ultimate has the nicest way about it. Software is freely available here: http://www.mohdparts.com/emanage/pc/ult ... 9_full.zip

this requires a serial number to install...

Vetal · Post by **Vetal** » Sun Jul 08, 2007 2:13 am

Matt wrote:trying to look for correlation. seems to follow TP but just goes really rich some of the times (like spikes)

perhaps O2 sensor feedback related? i also draw O2 against it and was looking for correlation. something is making it richer

this log has some extra empty columns in it. had to remove those from the CSV file to get it working. fixed in latest NIStune version when you record now

O2 area is the left part, but here all action happens in the 1-2 right columns

I just compared logs from 2nd gear and 4th gear pulls. Resume:
1st switchover appears to be dependent on AFM reaching 4.9-5V, because it happens at differerent revs and TP values
2nd switchover happens at ~6000 at both cases, but MAF is maxed out, respectively TP values are the same at this point, so can't really tell what is the reason.

Matt · Post by **Matt** » Wed Dec 19, 2007 1:50 am

Investigate item for CA18 injection resolved

10393 CA18 Injection not reflecting actual injection?

Current value X004D change to X1050 (LHS_INJECT_TIME_MSB)

This will require a firmware update on the boards (ie send here for reprogramming)

Before I do that ... see if there is a way I can make the consult table updateable in the firmware for future changes

Matt · Post by **Matt** » Wed Dec 19, 2007 1:51 am

btw this was verified with a scope on the injector line and adjusting K constant / AFM / RPM and monitoring the pulsewidth value against the display on the screen. matches perfectly

streetbike · Post by **streetbike** » Wed Mar 12, 2008 10:22 pm

Hi all ! have same problem on my CA18DET

AFM Z32, 550cc injectors GT2871R turbo

at 1.1bar of boost TP calculated is 88-90

the AFR expected is 11.5, my innocate mesure 11.5, so map is ok.

but nistune calculate a pulswidth of about 14ms and also a duty cycle at 7400rpm of 82% wich is surprisely high....

but consult mesure a pulsewidth of 8ms ( about 50% of duty at 7500rpm )
off course I prefer what is reported by the consult, but I would like to know wich one is the accurate value.

Is it possible to correct the values in the map if they are false ?

streetbike · Post by **streetbike** » Sat Mar 22, 2008 7:53 am

anyone to explain me ?

is the consult trace accurate or must I trust the calculated pulsewidth ??

because it is allmost twice as the consult trace....

Matt · Post by **Matt** » Sat Mar 22, 2008 9:57 am

The fuel map in its raw form contains values 0 - 255 these translate to:

O2 FLAG OFF (< 128)
0 = 14.7:1 AFR .... 127 = 7.38:1 AFR

O2 FLAG ON (> 128)
128 = 29.4 AFR .... 191 = 14.82 AFR
192 = 14.7:1 AFR ... 255 = 9.55 AFR

Now these are target values for the ECU to inject based on current RPM/TP. We translate them to target AFR, duty cycle and pulsewidth

They are based around the following calculation
target pulsewidth = base pulsewidth (16.5ms max) * (filtered raw value/128) * TP column value

Changing scalers, Injection multiplier (K) will affect end calculations

Use the consult reported injection amount since I measure this out on the oscilloscope for the injector output to verify it. Latest NIStune versions will fix a problem the older CA18 boards which reported the wrong injection pulsewidth address (hence wrong values displaying)

Matt · Post by **Matt** » Sat Mar 22, 2008 10:06 am

i can make the base pulsewidth adjustable which will let you change the map to be more reflective of the output. action item raised