dead time calculation

[subwoofer]

That was more like what we expect to see. I am still curious as to why the Green Giants behaved like they did, was the fuel pressure dropping as the flow increased, is there some electronics fault?

Again, how do you drive them? What happens if you swap polarity on them? For the record: I haven't studied injector tech thoroughly, could be that they are polarity agnostic for all I know. Or even that some are and some aren't - that could explain the difference.

[hoveringuy]

They were being driven directly off MS with the full fury of fuel pressure from the car. Polarity was identical because I was using the same EV1 plug.

All I can think is that they were new or defective.

[vw_chuck]

So I have a question.
Now you ran an injector using the drip method and saw that fuel does not flow until 1.1 ms actually driving the injector.
Now you input your new deadtime that seems reasonable at .560ms into the MS
Now what will happen if the MS calls for say .4 ms pulsewidth at idle? You will only be driving the injector for .960ms total so no fuel will be coming from the injector... right.
I still don't see how the .560 is correct.
Let us know how she runs.

[jsmcortina]

If you injectors are too large for your engine you can run into this. The non-linear small pulsewith curve can mitigate it somewhat too.

The new deadtime figure looks good.

James

[hoveringuy]

So I have a question.
Now you ran an injector using the drip method and saw that fuel does not flow until 1.1 ms actually driving the injector.
Now you input your new deadtime that seems reasonable at .560ms into the MS
Now what will happen if the MS calls for say .4 ms pulsewidth at idle? You will only be driving the injector for .960ms total so no fuel will be coming from the injector... right.
I still don't see how the .560 is correct.
Let us know how she runs.

Yes, you're right. This is a 3.0litre, 6 cyl. With the 42# injectors it was idling at 2-3 ms. With the 24# it's closer to 5ms. I wasn't enamored with running the larger injectors but some of the Bosch Style 3 injectors have the groove in the top for the fuel rail retaining clip and some don't. The 42# ones do.

I should still be Ok with maxing out the duty cycle, and I'll never worry about getting into the small PW region.

The .560 figure is correct because that figure makes the entire plot linear and scale correctly. My raw data is that I had 17.36 ml/1000 injections at 4ms and 37.5 at 8ms. Those flow figures are not a factor of 2, but when you subtract the .56 from each you get 3.44ms and 7.44ms. That's a ratio of 2.16, and 37.5 is 2.16 times 17.36. It works all the way up the plot. My 20ms data is 98ml/1000 injections. Do the math.. it works!

So, if I tested for PW of 4.56 and 8.56mS, I would expect the volume to be exactly double. 16.56mS would result in 4x that of 4.56mS.

In my opinion, the dead time has little to do with the opening time of the injector and everything to do with scaling the response. It ensures that a 100% duty cycle flows 24# and a 10% duty cycle flows 2.4#.

[vw_chuck]

It almost seems like you need two deadtimes. One for the non linear region and one for the linear region. I think this would help with idle quality when running big injectors.

[subwoofer]

There's already a function for that, but the best solution is to run correctly sized, sequential (and possibly staged) injection to stay out of that problem area.

[vw_chuck]

There is a dual deadtime calibration in the strategy? Where?
Or are you talking about the small pulsewidths calibration?

[jsmcortina]

He's talking about small pulsewidths.

James

[vw_chuck]

James there is a flaw in the small pulsewidths code. It says that the ends of the graph must be the same number so 0ms is 0ms and 2ms is 2ms. The problem is that your 0ms point needs to be the largest number on the graph because you are farthest away from the linear injector line.
So lets say you are at .5 ms for deadtime but your injector doesn't start to flow fuel until 1.1 ms. Then your 0 ms point needs to be .6ms but the code limits it to 0ms. How could this work? or am I confused about this calibration?

[vw_chuck]

Tumbleweeds....

[subwoofer]

I don't understand where you are trying to get at, but I think the flaw is mainly in you understanding of the problem? For most installs, unless you either run one channel batch fire with many cylinders or have ridiculously large injectors, the non-linear part of the curve can safely be ignored. There is no point in making life more difficult for yourself than necessary.

[vw_chuck]

Dude some of us like to make big power on 4 cylinders and methanol. Not everyone is status quo 2hp/ci..
It seems that MS has a lot of issues related to idle quality as per the forum topics.
If you make the strategy more robust in this area then you won't have these issues.
Just trying to point out an issue that I see. It can't work properly the way the strategy is written.

[racingmini_mtl]

Dude some of us like to make big power on 4 cylinders and methanol. Not everyone is status quo 2hp/ci..
It seems that MS has a lot of issues related to idle quality as per the forum topics.
If you make the strategy more robust in this area then you won't have these issues.
Just trying to point out an issue that I see. It can't work properly the way the strategy is written.

The problem is not the code but your understanding of injector characterization. You don't seem to understand the difference between dead time and a minimum effective pulse width and how the non-linear behaviour of the injector at low pulse widths affects all that.

It may not be simple and require some careful testing, measurements and characterization but what is provided in the MS3 settings does allow you to specify the injector behaviour so that you get a correct fueling throughout the entire operational range. Of course, that assumes you understand all the variables involved, that you can perform the characterization adequately and that you don't operate the injectors outside their operational range. It also assumes you have good injectors that give repeatable results in the range you operate them.

By the way, I want to make clear that my intention here is not to insult you or shut you down but just to point out that you don't have all the information needed to make a good assessment of the situation, or at least your posts tend to indicate that. The information is available on the web so I would propose you search for more information on injector characterization.

Jean

[Paul_VR6]

If you make "big power" use id1k and use their ms specific settings. They work perfectly, 800rpm idle on 500cc/cyl, on gasoline.

[Marek]

The main problem with MS is that people either do things on the cheap or shoehorn the wrong solution onto a non-ideal, ill thought through starting point.

It isn't the intended modus operandi, but you can have two separate deadtimes for one set of injectors by using the table switch parameters to your advantage.

The most sensible way to make the car work is to have properly sized injectors with the correct deadtime dialed in, being operated in their predictable linear region.

kind regards
Marek

[vw_chuck]

Yea I am sure there are lots of band aids but all the time spent trying to figure out a work around you could have fixed the problem 10 fold.

[subwoofer]

I don't quite see what is broken? If you run full sequential and have ginormous injectors so you touch into the non-linear area in actual operation, there is already a tool to linearize the operation in the trouble area. May not be the easiest to understand since it remaps pulsewidths into linearized space, but then again if you find yourself in the situation you have to run into non-linear response PWs you should be able to figure out how to map this.

Also make sure you understand the difference between calculated PW (from REQ_FUEL) and actual commanded PW (which includes dead time). If not running full sequential, rework the installation to do so since it effectively doubles (or more) the calculated PWs.

[racingmini_mtl]

Yea I am sure there are lots of band aids but all the time spent trying to figure out a work around you could have fixed the problem 10 fold.

You're the one asking for a band-aid solution. What is there is the correct way to characterize the injectors. If you don't know how to do that and/or don't want to, that's your problem. As I said before, you don't understand how an injector works.

Jean

[vw_chuck]

0ms cant be 0ms in small pulsewidths. So no it isn't correct. Lets say you figure out a deadtime of .5ms but your injector doesn't flow until 1.1 ms.
Your not stupid so tell me how could the 0 ms be 0ms in small pulsewidths??? It needs to be .6ms.