Interesting graph about nonlinear injector behavior

[jsmcortina]

Wes,
good data! Shame all these "informational site" get so clouded with their marketing machine.

Do you have side by side comparative data for your injectors for all of the set? That's what I'd really like to know.
Say on my engine, I have 8 injectors - what will be different between each of them? how to trim that to make them "perfect" ?

James

[wes kiser]

I just mapped one for the entire curve, I mapped the other 3 to 1.1ms and 9.0 ms. I figured if they were within x% at high and low, that was good enough. Both numbers were with 3% across the set. I will be mapping my 60lb highz siemens soon, and I will post that. My biggest stumbling block right now is voltage, I can only do one stable voltage at the moment, adjustable power supplies capable of sinking any current are surprisingly expensive.

[muythaibxr]

I've been doing some less scientific stuff with my rx7 lately... I recently swiched from 2 squirts simultaneous to 4 to get better throttle response, but that killed low-PW tunability... I was finding that below 1.3ms I got no fuel injected! So I set the opening time to 1.3 ms and now I don't have that problem anymore, but I'm noticing at the bottom of my table when the PW gets below 2 ms, the VE table has to be dramatically increased to get the AFR I am targeting, so it looks like between 1 and 2 ms, these old injectors are VERY nonlinear.

I'm going to try out James' non-linear injection code to see if I can smooth that out.

Not nearly as useful as having the actual curves, but using this method I should be able to tune it reasonably close using data that your average DIY tuner can get from datalogs without having a good injector test bench.

Ken

[wes kiser]

If you have an MS and a functional car, you almost have the test bench. I just made a baby fuel rail, and feed it from one of my ports on my regulator. I then unplug my in car injectors, and plug up the one in the baby fuel rail. Then just spray into a graduated cylinder. I make sure I have a friend handy with a fire extinguisher when I do it. I hood up the battery charge to get consistant (although not variable) voltage.

I ultimately am going to make an actual test bench, but this works well for almost no money.

[muythaibxr]

That would work great if I could get my injectors out, and/or they weren't weird.

The rx7's primaries are under the intake manifold, you have to take out the manifold to get to them.

The Corolla has side-feed injectors, which are a bit strange.

Ken

[muythaibxr]

In other words I could do it, but it's a pain.

Ken

[wes kiser]

Here are bosch 42 lb (high z, old school ford racing) and siemens 60 lb (high z, new ev6).





Notice the siemens ev6 have an offset similar to the low z delphi's, but are a high z injector. The equations at the bottom right are the slope of the linear fit (and can be used to calculate offset as well as flow at 10ms, which is 100% duty with the period I chose). Both are at 45 psi. FYI, the siemens were 66 lb at that pressure, and the green tops were 48 lb at that pressure.

[jsmcortina]

Again, very interesting information. The steep rise at low pulsewidths is similar enough between the two injectors that I'm inclined to try modelling that by default, with the zero crossing point set as the opening/dead time.

James