SVO red skinny injector data issue

[natesully]

I bought a set of those skinny all-red 36# injectors, aka "SVO injectors" for my LS1 Miata in preparation for future upgrades. It's a stock engine from a 2000 Camaro. For some reason, the injectors just don't seem to flow what they should, I had to drop the IFR table a bit to get anything other than +25% fuel trims. Stock injectors ran fine, other than the fact that one of them was leaking.

I used the spreadsheet here for the data, double checking the FRPP data matches what was entered already:
http://www.hptuners.com/forum/showth...on-spreadsheet

Here is my current tune, with a "fudged" IFR table, offset/short pulse from that spreadsheet and stock everything else:
fudgingIFR.hpt

And a datalog running on that tune, with rich-but-not-insane LTFTs:
drivefromwork.hpl

I checked fuel pressure, and it measured "60 PSI" on an autozone gauge. Close enough I would think. Other than the data just being plain wrong, what else could cause flow to not match what it should be?

[natesully]

Well, I don't get it. Took a drive with the fuel pressure gauge hooked up - it bounces a few PSI when I floor it, but it's pretty stable. Battery voltage is OK, MAF matches dynamic airflow, LTFTs are +/- a few percent in the range I usually drive with my made-up IFR table... But these injectors don't seem to be flowing 36# at 58 PSI like everyone says they should. Seems more like 32.3# at 58 PSI, at least that's what's in the 0 bin of my IFR table right now.

Of course, the car runs extremely well right now, at least from what I can tell with my narrowbands and butt dyno.

[natesully]

I think I figured it out - the spreadsheet in the linked thread seems to be wrong, at least on offset. The Ford cal data shows an offset compensator thing of 1.1639 at 60.03 PSI, and the offset at 12V is 1.083. So, at 12V and 0 vacuum on the LS1, the offset should be 1.26 ms or so. The spreadsheet has the SVO reds having an offset of 0.987 at 12V and 0 vacuum - something looks wrong there.

So, I went back to the IFR from the spreadsheet, and added 0.27 to the offset table as a temporary hack just for testing, reset trims and went for a short drive. LTFTs are -4 -> 7 or so like that, which is WAY better than +25%. Hmmm...

Time to make my own spreadsheet I guess. Seems straightforward enough, though the ford data only goes to 60 PSI - at full vacuum, I'll have like nearly 70 PSI across the injector, so I suppose you just have to curve fit and hope for the best up there.

[DSteck]

Just FYI, Ford offset and GM offset are two VERY different values. That's why the GM and Ford values don't match up.

That spreadsheet is not a run of the mill easy conversion. You need to look into where the Ford data and where the GM data come from.

[natesully]

Yeah, I see. From the ID site, I gather that Ford offset is basically "how long it took to open", while GM offset is some crazy "time between ideal and actual slope line" type thing, and the short pulse adder brings it in the rest of the way. That explains why the Ford offset is bigger, but I still have no idea why the data I originally tried led to +25% LTFTs.

That said, if you have sufficiently long pulses, isn't the GM pcm relying on the offset table alone to correct things? Seems like at that point, you want offset to be how long it took the injector to open all the way, like the Ford offset. Or I just don't understand how the short pulse table works, since I have yet to find a better writeup than the ID site about it.

[DSteck]

I think you're not fully understanding how the data is used. The offset doesn't change based on pulsewidth, so whether or not you're in the short pulse region doesn't change how it should be calculated.

What you're seeing is the loss of accuracy going from format to format and manufacturer data variation.

[natesully]

Are you saying my awful cal is just a fact of life with data from the original spreadsheet?

[DSteck]

Rescale the VE and MAF, and move on with life is what I'd suggest.

[natesully]

Fair enough, I'm probably overthinking this. Thanks for the advice!

[DSteck]

Like I said in the other thread, it's imperfect data, but better than nothing. If a factory GM injector can't handle whatever I'm doing, I go right to IDs.