if its a flow rate modifier cant you just scale the injector constant to bias the table?
if its a flow rate modifier cant you just scale the injector constant to bias the table?
I was thinking more about this. NO injector out there is going to have a Negative Short Pulse Adder used in a GM vehicle even though GM may have calibrated some of the later vehicles with negative numbers. The negative numbers in that situation are a skewed value to correct something else.. The way GM uses that table in the operating systems I have physically looked at is to correct the delivered fuel charge of the theoretical whole. Same with the offset table. While it is true that injectors typically flow more in the early non linear area, to make the actual injector delivery linear with pulsewidth, it has to have a positive value. In the Injector Dynamics diagram, you can clearly see the theoretical white line and the actual fuel delivered in the red line. The red line will always be lower than the white.Originally Posted by steveo
http://injectordynamics.com/articles...racterization/
Last edited by Fast355; 02-07-2023 at 10:44 AM.
from the standpoint of actual tuning output -- i have always just left the table stock and tuned TONS of lt1s without issues
here's some reasoning for you
with a reasonably sized injector the threshold for non linearity is low enough where at worst it's basically at idle and transition off-idle, or at best your engine would damn near be stalling before you got there because it's way below normal operating range.
look at the stock table. it starts to build a curve at 1.7msec but barely, in effect it doesn't do much until below 1.0msec
i don't know what you guys are finding for pulsewidths at idle, but i usually see 1.5-2.5msec hot. i can't ever remember seeing an idle below 1.5msec pulsewidth.
if you DO cross that threshold, it's because you probably have a gigantic engine with a big cam and really big injectors, and lets not kid ourselves, idle fueling on engines like that is a joke. you're basically pissing into a river, there's no real target, you just play with it till it runs well.
so i'd stick with either zeroing the table or just ignoring it and just botching your airflow (ve or maf) in that region to get your fueling in the ballpark, can't imagine it making a gigantic difference, unless you run gigantic injectors and the 'knee' or whatever you'd call it suddenly encompasses a real operating range
Well, it makes sense
Just a quick question along those lines steveo - do you know, is the pulsewidth reported in the $EE datastream before of after this adder?
Your point is valid however - the operating range we're talking about here is going to be very rarely encountered for reasonably sized injectors.
No, you'd need to skew the entire offset table. The slope of the injector is assumed by the fueling equations to be a linear constant (even though it rarely is).
Edit: also, it's not a flow rate modifier. It's correcting delivered fuel mass for a region of the injectors flow curve that doesn't match the injector flow constant slope.
Though I'm scarcely qualified to teach a remedial course in fuel injection flow characterization, if I get bored enough I'll try to draw up a couple graphs to illustrate the difference between older type 1 / 2 knees and what a type 3 injector looks like. The graphs Fast355 referenced on injectordynamics are of type 3 injectors.
The topic I'm more interested in is how would you advise yoheer to proceed given the injector characterization posted here?
Code:ACCEL 150136 - 36# Hr Fuel Injector Specifications Drive Circuit: "Saturated", R-C Voltage Suppression, 14.0 VDC Coil Resistance: 14.5 ohms Fuel Compatibility: Standard Gasolines and Ethanol Flex Fuels Fuel Pressure: 300 kPa (43.5 psi) Static Flow Rate: 252 gm/min n-Heptane Dynamic Flow Rate: 7.56 mg/pulse at 2.5 ms, 10 ms period (100 Hz) Approximated Time Offset: 0.70 ms Approximated Slope: 4.20 mg/ms Minimum Linear PW: 1.38 ms Linear Flow Range (SAE): 15.3 Opening Time: 1.35 ms Closing Time: 0.67 ms SMOV: 4.23 volts
Last edited by spfautsch; 02-07-2023 at 07:41 PM.
The point you are missing is that the angle of the knee does not matter. It still going to deliver less than the theoretical delivery base off pulsewidth thus the numbers will still need to be a positive number. The injector that has more initial delivery will just have lower values in the adder table. The injector that has less flow initially will have larger numbers. Either way you still have to add to the total pulse width to get the full fuel charge.
I spent some time messing around getting the injector data correct on the 42# Ford 6.2L "Raptor" injectors in my P59 run 383 with a L31 Marine intake. Using the data in the Ford Raptor PCM, at a differential pressure of 58 psi, the break point corresponds to 1.667 msec. Over 1.6 msec the injector flow is characterized by the high slope values, under 1.6 msec it is the low slope. The low slope value has a higher flow rate, but the values are 0.121% different. I used an exponential curve to mathematically approximate the difference. At 0.06 msec commanded pulse width the injector in theory needs 0.2 msec more to get that 0.06 msec desired flow amount from that point on the difference reduces until it disappears into a number the P59 will not even save at 1.64 msec. In the grand scheme of things that is almost nothing. My idle pulse width is ~3.5 msec at hot idle. During deceleration the pulse width drops to 2.2 msec. I would need roughly double the injector size before the short pulse adder really came into play. As said before this would make a difference on an engine with really large injectors like a turbo build, but with reasonably sized injectors its not something that would ever come into play. Even with double the injector size, it would still likely idle above 1.6 msec and any issue in deceleration could be compensated for by activating DFCO for a brief period. Ford did exactly that on the mid 90s Cobras with the 351 to get them past emissions and smooth out surging that might otherwise exist from inconsistent fueling during deceleration.
Last edited by Fast355; 02-07-2023 at 11:01 PM.
This is what I am running for a short pulse adder. With the offsets calculated for 58 psi and using the high slope flow rate calculated for 58 psi, runs perfectly. I have a LS3 MAF on it in a 100mm tube. Runs on the Lingenfelter MAF transfer for a 100mm MAF like GM calibrated it.
seriously i wouldn't worry about it on an LT1. it's just not a precise machine.
even injector voltage offsets, if you don't have them, are no big deal. unless your alternator shits the bed it only ever uses two cells of that table anyway. just rough in the injector constant and throw the offsets in if you have them.
it's like obsessing over tuning the far top right corner of your VE table, which unless you have a standard transmission and floor it from a standstill in 6th gear, you absolutely can't reach, and even if you did, fueling accuracy in that region doesn't matter.
in most of my tuning i never touched that and they were dead on accurate with no hiccups in drivability
focus on things that make a difference. kur4o's EOIT stuff probably makes way more of a difference than all this
Agreed! Some guys can stress out to the nth degree, but it does not really matter much. The better the data you have the better off you are; that being said it is not the end of the world. Tune it so that the air/fuel ratio is right and you are golden.
This was with Ford data out of a stock Ford PCM converted to GM which some people claim will make the GM run like poop.
https://youtube.com/shorts/TRVDXmOpSHA?feature=share
there definitely are temperamental machines that require incredible precision.... with high compression and tons of boost you can often be .1 lambda point or 1 degree of timing in between scalding power and death. you'd figure that you'd want to really nail your injectors down, but guess what, even on those machines people often aren't screwing with injector parameters that much. my 25psi boost buggy has a 32 bit ecm with only 5 cells for injector latency and that's it, and injector timing probably is all thrown through some proprietary equations that match with variable intake cam timing. when i put completely different injectors in there, i just left that table alone too.
Actually, after going through the exercise of sketching out the slopes yesterday, I'm no longer missing that point. You are correct. As I'm looking at the low pulse adder table for my '01 LS1, it's a typical adder table. Though I didn't originate it, I guess I'm guilty of spreading bad information.
I know the feeling there. It threw me off a bit too having just been messing with and converting injector data. Then a light bulb clicked on when I saw those Injector Dynamics graphs.
Posting Permissions
Reply With Quote
Bookmarks