Injection timing in relation to valve events

[mdrew]

IMO you're better off characterizing your injectors to your ECU on your own, not using data provided by the manufacturers. The results will differ enough that it's worth doing.

I wouldn't even know where to begin.... I’m still learning this whole “tune by laptop” thing…. Might you have a recommendation for some reading literature that would outline the process to do that? Every book I’ve got talks about how important data is, but none speak to “characterizing injectors” as has been mentioned in this thread.

I am quite confident that the data I got with this last set of injectors is incorrect. I can dial in my VE table using a wide band, and get AFR error under 2% in every cell of the table, but can’t get MAF to cooperate. AFR error and / or short term fuel trims vary greatly in the same Hz air flow cell, but different rpm/load.

I’d love to know how to figure this out, without depending on someone else….

Sorry for the thread de-rail.. Please PM me if that is more appropriate.

[Belgian1979]

You need to make a test rig to mount your injectors and connect them to the fuel pump. Then you wire up the injectors you're going to test and have them spray in a measuring glas (preferably 200 ml or so). I used a remaining wire loom connected to the ecu to supply it with various set voltages. I ran the fuel pump of the battery that was connected to a charger.

You then use the injector test feature to do a predetermined amount of injections and based on the volume that ends up in the measuring glas you can calculate the dead time at the various voltages.

I have a spreadsheet that calculates all of that automatically and if you pm me your email, I will send it to you.

[MadBill]

Both Toyota and Ford Motor Co have written SAE white papers regarding the optimization of inector timing.
That was over 20 years ago.

The GM one I read in that time frame looked only at the emissions effect of injector timing. AIR, in general there was little difference except for being ~ 20% worse on an open valve.

[MadBill]

My injectors are (unfortunately) perpendicular to the airstream. So a lot of fuel ends up on the walls. Inevitably this means that the fuel will take some time to reach the valve/become a gas. For that reason I do think (not sure) that it will require the injection to end earlier.
The only factor that I think is hard to estimate is how much more fuel ends up in a gas depending on the airspeed moving through the runner.

Some programs include algorithms for 'wall-wetting', 'transport delay', etc. which if provided should help with your concerns...

Also, there are tons more variables affecting vaporization. E.g., fuel distillation curve, air,fuel, intake valve and manifold temperatures, manifold vacuum, etc. plus a number of in-cylinder factors.

[dieselgeek]

His megasquirt ECU comes with an advanced "X-Tau" model built in, if he wants to go that route. It works but it's really difficult to tune in my experience. Once it's calibrated properly it works great though. You can get very crisp throttle response this way.

[David Redszus]

After over 40 years of testing and reconditioning fuel injectors, I can honestly say we have made some progress
but much is left to be done.

The determination of fuel injected quantity is one area.
For a given size injector, given fuel pressure, and fuel viscosity, how much fuel is delivered during an injection period of 6ms? How much at 12ms? Does doubling the period double the delivery? If not why not?

Then there is the issue of fuel delivery as a function of fuel pressure. Does an increase in fuel pressure of 20% produce
20% more fuel? How does a change in rpm affect the delivery rate (per stroke)?

The next vexation is fuel spray quality.
How does spray cone formation change with pressure, fuel viscosity and rpm? How does droplet size vary with
fuel pressure, fuel viscosity and spray shape?

Next is a concern for uniformity.
How can we be sure each injector behaves like its cousins? Do each of the above parameters affect each injector
in the same manner? How much variance is there from injector to injector? Clean vs dirty?

Many answers to the above questions can be answered at your local diesel injection pump repair facility. Most have
injector cleaning machines that can also be used to evaluate the performance characteristics of injectors.
It then becomes possible to construct a "matched set" of injectors, with numerical pedigree for later reference.

[RednGold86Z]

His megasquirt ECU comes with an advanced "X-Tau" model built in, if he wants to go that route. It works but it's really difficult to tune in my experience. Once it's calibrated properly it works great though. You can get very crisp throttle response this way.

I've found real X-Tau to be a little too powerful - if you put in enough parameters to model the temperatures of the surfaces, you wouldn't even need cranking, afterstart, fuel cut return fuel, closed loop fuel, or almost anything. But, when you try to combine it with something that still has those parameters active, you have to find ways to pause and restart the X-Tau without losing place.

Also, X-Tau isn't inherently stable - at a cold enough engine, it could swing wildly but with the result of a steady lambda reading - as it tries to do the "impossible" task of responding to a change perfectly. It's also pretty hard to tune by watching the lambda reading - it's just not intuitive to see which number or combination of X and Tau is at fault.

So, I made a pseudo X Tau - i transposed it sort of. I'm not going to say it's perfect, but it's a lot easier to use. The result are exactly the same for a given transient, but without the possibility of being unstable, and with the flexibility to, say, just go rich in accel, and not hurt the decel transient during a very cold transient.

[Belgian1979]

The xtau is only active during accel and not anywhere else.

In fact the delay in itself is an xtau imo

As for Bill's comment : i think a lot of factors are already accounted for in the 8 ms delay

[RednGold86Z]

The xtau is only active during accel and not anywhere else.

In fact the delay in itself is an xtau imo

As for Bill's comment : i think a lot of factors are already accounted for in the 8 ms delay

Well, I guess it's not a complete X-Tau then.

I've read some Bosch documents and watched the actual injection timing on a stock Bosch system on a stock engine, and when it's in low RPM ranges that would be on the emission drive cycle, it's early injection that would reach the valve before it opens. At about 3900 RPM it switched to a later injection. The documents say it's there to improve VE, due to cooling (contraction) inside the chamber (the cooling may also be beneficial to the engine parts, if it makes it all the way to the walls).

The document also estimates the overall transport delay at 8ms, but in your case - almost all of the fuel is smacking into a wall, so most of the fuel entering the cylinder is either evaporated or puddled in. You might not be able to notice much difference in the injection timing.

[Belgian1979]

Well, I see a remarkable difference in engine running cold to warm, so in effect yes, vaporization due to heat will contribute a lot. However that vaporization takes time as well I would think.

[In-Tech]

xtau brown cow takes thousands of man hours to get correct, change injectors(even the same ones ) camshaft, port heads, etc etc and start over. I assume you aren't trying to pass a bag test at CARB It's too easy for guys like us to overthink things. Just give it what it wants when it wants it and tinker from there...

[dieselgeek]

xtau brown cow takes thousands of man hours to get correct, change injectors(even the same ones ) camshaft, port heads, etc etc and start over. I assume you aren't trying to pass a bag test at CARB It's too easy for guys like us to overthink things. Just give it what it wants when it wants it and tinker from there...

On the MS products I can have X-Tau working in 20-30 minutes work. A noob might spend twice that. I've calibrated X-Tau on a Woodward system and it was about the same amount of challenge. Not that hard.

But I wouldn't recommend it when the Accel Pump functions can make the engine run perfectly and are easier to set up. But definitely check out some of the consumer EFI systems iterations of X-tau. For example, Accel DFI *only* uses X-Tau for enrichments and their out-of-the-box config works so well I almost never have to adjust it.

[In-Tech]

Hi Scott, I think we are talking two different languages. I'll check into yours and see if I can post some examples of what I was thinking.

Bleh, tuning an early TBI 496 BBC today on the chassis dyno

[dieselgeek]

The MS guys are the only ones I know that took the time to write about how they implement it in plain english, good reading about X-Tau here:
http://www.megamanual.com/ms2/xtau.htm

[Belgian1979]

Well, the xtau is not something for right now. I was hoping to get a good starting point on injection angle.