Injector Open Time

[Gokart]

Ms2 manual stated, "Injector Opening Time (ms) (InjOpen) is the amount of time required for the injector to go from a fully closed state to a fully opened state when a 13.2 volt signal is applied. Since fuel injectors are electro-mechanical devices with mass, they have latency between the time a signal is applied and the time they are in steady-state spraying mode. Typically, this value is very close to 1.0 milliseconds. " however, I can increase the Open Time till 25.5ms as allowed by TS/MT that is actually predefined in the ini.

My question is, can that such a high value be use in the actual tuning or should I restrict to only 1.3ms that was recomended in MT used with MS1? Or, since MS2 have different calculations, then using till 25.5ms Injector Open Time is a normal condition.

[jsmcortina]

If your injectors take 25.5ms to open something is seriously wrong.

Maybe up to 1.6ms is reasonable with series resistors.

James

[Gokart]

Thanks for the clarificaton. I'll edit my ini file to restrict the value.

[Peter Florance]

I found that changing from 1 to 3 squirts (6 cylinder) and adjusting the opening time until AFR's were equal gave me the best performance. Is it right at all PW? Probably not. But it works pretty well like that.

[muythaibxr]

That didn't work for me when I went to 4-squirts on the rx7... every time the PW went low, I'd go full lean... I found that any time I went below 1.3 ms it did that, so I switched to 1.3 ms for the opening time and retuned, and now I don't have that problem.

I think if I retuned my 2 squirts table with the new PW, the two tables would probably be pretty close.

Ken

[Peter Florance]

That didn't work for me when I went to 4-squirts on the rx7... every time the PW went low, I'd go full lean... I found that any time I went below 1.3 ms it did that, so I switched to 1.3 ms for the opening time and retuned, and now I don't have that problem.

I think if I retuned my 2 squirts table with the new PW, the two tables would probably be pretty close.

Ken

New PW?

[muythaibxr]

That didn't work for me when I went to 4-squirts on the rx7... every time the PW went low, I'd go full lean... I found that any time I went below 1.3 ms it did that, so I switched to 1.3 ms for the opening time and retuned, and now I don't have that problem.

I think if I retuned my 2 squirts table with the new PW, the two tables would probably be pretty close.

Ken

New PW?

I meant new opening time pulse width...

Ken

[darsondriss]

That didn't work for me when I went to 4-squirts on the rx7... every time the PW went low, I'd go full lean... I found that any time I went below 1.3 ms it did that, so I switched to 1.3 ms for the opening time and retuned, and now I don't have that problem.

I think if I retuned my 2 squirts table with the new PW, the two tables would probably be pretty close.

Ken

New PW?

I meant new opening time pulse width...
Ken

My opening time is 1.1 and average idle pw is 1.8 and i could hardly do the 4-squirts test becuz that brought PW down to like 1.3 lol.

Now im desparately trying to resist the need to lower req fuel for greater precision down at idle so i dont have to retune when ms3 comes out

[wes kiser]

The manual is flat out wrong when it says to find the actual open time and to enter that as an offset. Injectors take a finite amount of time to open and close. The correct offset is the "0" crossing of the linear curve fit applied to the dynamic opening curve. I have found this to be about .55ms @13v with delphi low-z's that take .9ms to open, .66ms with siemens ev6 that took 1.4 to open, and about 1ms for bosch ev1's that took 1.6ms to open. My measured curves for these injectors can be found here:
http://i88.photobucket.com/albums/k163/ ... _pic-1.jpg

http://i88.photobucket.com/albums/k163/ ... s60pic.jpg

http://i88.photobucket.com/albums/k163/ ... lowzpw.jpg

In all curves the red is the actual measured curve, while the black is a linear curve fit with the slope equation at the bottom. Data was obtained with a 10ms period, so 1 ms is 10% duty, 2 ms is 20% duty, and so forth. This was done @ 13v, with ms2 extra coil test mode. This will make the fuel equation always work, provided you stay above the minimum "full open" PW. While entering the "full open" pw as the offset will prevent you from operating in the "here be dragons" region, it will make all of your % corrections WAY off at low PW's. It is an adeuate work around for many, and ms2 extra provides enough modifiers that it can be made to work.

[jsmcortina]

Wes,
I totally agree and had been thinking the same after seeing your data.

The non-linear region at the bottom of the graph is the cause of confusion when 99.9% of us have not used a flow bench to test our injectors.

James

[racingmini_mtl]

As it has been mentioned before, the "opening time" used is a misnomer. It should be called dead time because it is the result of the fuel not being injected when the injector is opening minus the fuel being injected during the closing time.

I think this should be corrected at some point because it does add to the confusion of what it is and how to correct for it.

Jean

[svocapri]

The manual is flat out wrong when it says to find the actual open time and to enter that as an offset. Injectors take a finite amount of time to open and close. The correct offset is the "0" crossing of the linear curve fit applied to the dynamic opening curve. I have found this to be about .55ms @13v with delphi low-z's that take .9ms to open, .66ms with siemens ev6 that took 1.4 to open, and about 1ms for bosch ev1's that took 1.6ms to open. My measured curves for these injectors can be found here:
http://i88.photobucket.com/albums/k163/ ... _pic-1.jpg

http://i88.photobucket.com/albums/k163/ ... s60pic.jpg

http://i88.photobucket.com/albums/k163/ ... lowzpw.jpg

In all curves the red is the actual measured curve, while the black is a linear curve fit with the slope equation at the bottom. Data was obtained with a 10ms period, so 1 ms is 10% duty, 2 ms is 20% duty, and so forth. This was done @ 13v, with ms2 extra coil test mode. This will make the fuel equation always work, provided you stay above the minimum "full open" PW. While entering the "full open" pw as the offset will prevent you from operating in the "here be dragons" region, it will make all of your % corrections WAY off at low PW's. It is an adeuate work around for many, and ms2 extra provides enough modifiers that it can be made to work.

This is very interesting. I have been tuning a 4bbl holly TBI with staged injection and trying to get the injector open time setting correct so that when my secondaries open, my fuel curve doesn't go crazy. By swaping between the two squirts and four squirts and reducing the injector open time constant until they are about the same, I have come up with 4.9ms injector open time. I thought I was doing something wrong since my numbers were so far off from what I have come across elsewhere, but from your graphs it doesn't look to far fetched. These are the Delphi TBI injectors which I think are supposed to be even faster than the other lowZ injectors.

[wes kiser]

This is very interesting. I have been tuning a 4bbl holly TBI with staged injection and trying to get the injector open time setting correct so that when my secondaries open, my fuel curve doesn't go crazy. By swaping between the two squirts and four squirts and reducing the injector open time constant until they are about the same, I have come up with 4.9ms injector open time. I thought I was doing something wrong since my numbers were so far off from what I have come across elsewhere, but from your graphs it doesn't look to far fetched. These are the Delphi TBI injectors which I think are supposed to be even faster than the other lowZ injectors.

You do mean .49, and not 4.9ms, correct? 4.9 is unreasonably large.

[svocapri]

You do mean .49, and not 4.9ms, correct? 4.9 is unreasonably large.

I know this has been a long time but for clarifications sake, yes I do mean .49ms

[W113]

Wes Kiser,
thanks for a lucid and reasonable explanation. However, parts of it needs a bit more care for us nerds
QThis was done @ 13v, with ms2 extra coil test mode. This will make the fuel equation always work, provided you stay above the minimum "full open" PW UQ

This will work well and is reasonable at tunes where the minimum PW required is say 0,7 ms or more above the "opening time"

For the squirters that try to squeeze every little bit of controllability at low driving PW´s , things are a bit more complicated than that. The injector is a current controlled device , and to cater for the change in driving voltage MS has a correction factor " ms/V" . This helps, some. However:
The time constant of the injector is L/R. L in Henry is inherently stable , but R in ohms is not. R will increase some 20 % when the injector temperature goes from 20 C to 75 C. So the time constant will decrease as as will the opening time when the injector warms up towards equlibrium . At the same time the saturated current will reduce (U/R), which partly counteracts the reduction in L/R and opening time. The net effect is that a typical Bosch Hi Z has a 20 C opening time of 1,30 ms, which drops to just above 1,23 at 90 C( temperature increase measured by checking R, calculating the average copper wire temp)
So , waiting for several minutes and running low PWs before trying to tune for the last few tenths of ms´s is a good idea that saves a bit of frustration.

And sorry for being loZ ignorant, have not done any bench tests, in fact , I try to stay as far away from loZ injectors that I can!
An OT solution would be to use circuitry with current control to counteract the change in R, but not even I have even tried a breadboard on that!


Best regards

W113

[wes kiser]

All of my tests were conducted at 13v, as honestly that is the only point my setup is capable of holding repeatably.

I am also aware of many unaccounted for variables in my explanation. An offset vs. voltage curve has been the method used to account for this by most OEMs up until the late 90's (maybe later, but my knowledge of OE stuff stops there) and every standalone I have ever used. Ford interestingly used 2 injectors constants (one for the opening time, and one for the steady on time). With sequentially fired 1x/cylinder injection, OE computers using this (offset vs. voltage) strategy have no trouble linearly controlling modern 2000+cc injectors on sub .5l/cyl engines at low loads.

What I suggested isn't perfect, but it will get you relatively close (much closer than following the manual).

Something you don't mention, but is likely also significant, is I gave a flow rate curve, which is based on only 1 frequency. This is likely on truly valid for this frequency. I picked a midrange period (10 ms). I believe the "most right way" would be to calculate volume for a single pulse, at a given PW and map that.

[matthewb]

Is there by chance a good thorough write up on this injector testing using test mode on tunerstudio? I Think I am getting the basics but am unsure of the exact settings to use in test mode to get the different duty cycles. How do I know how many times I have squirted the injector? Do I always you the same test interval and just vary the pulse width? Do I need to do this test at multiple different voltages to be able to fill in the voltage correction plots?

Does anyone know a typical deadtime value or have this data for a ford 24 lb injector?

[chadr]

I know this post is old, but it has some excellent info. I've been retuning my ve table after increasing my fuel pressure from 43 to 58 psi. I turned off over run fuel cutoff and wanted to get the ve numbers in the lowest map region tuned correctly. I've had over run fuel cutoff turned on for 6 years and never have tuned the lowest map regions of the table reached during deceleration, but now I want to smooth the transition from over run to light throttle as much as possible. The afr was reading way lean when injector pw as shown in the log was right at 1.6ms. I could not get the afr to tune correctly in the lowest kpa region until I increased injector opening time to 1.5ms. Now I'm getting close to the afr target of 1 lambda.

Based on my results of trying various opening times up to 1.5 this graph of the 42 lb green top injectors (someone posted above) show the same 1.5 - 1.6ms dead time that I found. I'm running ms1 with msns high res code and my injector opening time has been set forever with about .8ms but now I can see that was quite far off.

Getting the opening time number correct has allowed the afr to be tuned just as expected. I may need to go to 1.6ms if 1.5ms isn't enough, but its very close to lamba right now when the injector pw is the smallest during very low kpa deceleration.