Tuning Injector Slopes

Anything technical worth keeping around related to EEC Tuning or just general engine tech info.

Moderators: cgrey8, EDS50, Jon 94GT, 2Shaker, tmoss

Locked
User avatar
cgrey8
Administrator
Posts: 10296
Joined: Fri Jun 24, 2005 5:54 am
Location: Acworth, Ga (Metro Atlanta)
Contact:

Tuning Injector Slopes

Post by cgrey8 » Fri Oct 12, 2007 11:53 am

If you have BinaryEditor as your datalogging software and EEC Analyzer, getting good Injector Slopes is a little easier. Most of this discussion will assume you use at least EEC Analyzer. Calculating slopes with some other tool is not specifically covered here.

So first off, set the injector parameters in your tune to the injectors you are using. Assuming you find a template matching your injectors, the template most likely contains injector high slope, low slope, breakpoint, and Injector Offset vs Battery Voltage. If you cannot find a template that matches your injectors, then start by entering the injector's actual rated flow for the high slope. Enter a low slope value that's ~7% above the high slope. Set the Breakpoint to around 3-4...actually 3e-5 to 4e-5 lb/rev. Assuming you are using a MAF other than stock, enter the MAF curve that describes your MAF. EEC Analyzer contains a number of common MAF curves for both stock and aftermarket MAFs. If you can't get your MAF's transfer curve from the manufacturer or on the Internet anywhere, but you at least know what injectors the MAF was calibrated for, you can try picking a MAF curve of a MAF calibrated for the same size injectors.

At this point, you are ready to try cranking the engine and driving it around if you can. Don't be surprised if the engine doesn't exactly run great though. If it will drive, only drive with light throttle ESPECIALLY if you have a boosted application. Also for starters, keep the FPR set to 39-43.5 PSI depending on what your injector parameters above were rated for.

Once you've been able to get some driving datalogs of your setup, you need to do some analysis of the datalogs. This is where EEC Analyzer comes in. But before you can do analysis, you need to make sure EA has your tune's MAF Curve. If it doesn't, enter it manually and save it for future use. Since the MAF curve is likely to change a lot during the initial tuning phase, it is a good idea to give the MAF curve a name that also includes the date so you can differentiate between this curve and future modifications you make to your curve. Before running a MAF curve calculation, make sure the MAF curve that is selected represents the curve used by the EEC for the datalog you have loaded. This is VERY important.

At this point running the MAF curve analysis will be mostly anecdotal. Even if the MAF curve analysis shows error (and it most likely will), your tuning will not likely begin with the MAF curve, but rather it should begin with Injector Slope tuning assuming you can drive the vehicle to get some datalogs. If you can't drive it, then you'll likely go through a bit if trial-n-error guess work tuning in the driveway until you get something that is stable enough to drive. Once you get past the trial-n-error phase of just getting the engine to run well enough to drive, you'll be doing mostly analysis work using the tools at that point. But don't make a lot of adjustment based on short drives until you have a LOT of confidence in what you are doing and in the behavior of your vehicle. The engine may behave differently in the morning than in the evening. Thus a morning commute into work may have you making one adjustment, but the drive home that afternoon may have you adjusting things right back in the opposite direction as compared to the morning drive. So once you've gotten it in to where it's close, then make baby-steps from here to make sure you don't overstep.

Once your MAF curve is selected as the #1 curve, have EA analyze your datalogs by doing a MAF calculation. The #2 curve can be whatever you want, the same curve, an older curve, or nothing (all zeros). The results of EA's analysis will be represented as the 3rd curve which literally is what EA believes would've run better. But EA is basing this PURELY on the datalog which doesn't always tell the entire story. So DO NOT blindly take what EA suggests and plug that into your tune for the next run. This is where you as the human have to analyze the results of the analyzer. If your whole MAF curve is off by a certain percentage, then the problem is probably not the MAF curve at all, but rather Injector Slopes or a combination of MAF and Injector parameter issues. In EA, a consistent percentage off throughout the MAF range will look like the calculated curve is "peeling" off your tune's curve. The distance between the two will increase as you move to the right of the chart in EA. Most likely you'll need to increase the slope for conditions where the calculated curve is below the actual curve in the tune. Lower the high slope if the calculated curve is above the curve used in the tune. But using the MAF curve analysis, you don't know by how much you need to adjust the slopes/breakpoint/offset stuff. This is where the Injector Slope tuning calculator in EA is more helpful than the MAF curve calculator.

What would happen if you attempted to tune purely out of the MAF curve and ignored Injector Slope/BP/Offset values? This will be problematic because adjustments to the MAF curve are intended to adjust for error in the MAF curve relative to actual MAF behavior. When doing MAF analysis, EA assumes that any error in AFR is because of error in the MAF. But since Injector Slope error can also affect fueling, you need to get 1 right or at least relatively close before you can do analysis on the other. Lets say you have Injector parameters with a fair amount of error that causes small Injector PW conditions to be too rich and larger Injector PWs to be lean. Well if you were at a low RPM and flowing say 100kg/hr, EA would think the MAF was too rich and try to drop 100kg/hr to lean things out. But when it got to a part of the datalog where the RPMs were higher but at a lighter load, and MAF flow was around 100kg/hr, EA would interpret this as a time when the MAF was too lean and try to increase the MAF value. Well now you have conflicting conditions for 100kg/hr. In some cases, the MAF runs the engine too rich. At other conditions that same flow runs the engine too lean. At yet other conditions, the AFR is probably closer to correct. So point is, you cannot change the MAF curve at 100kg/hr without negatively affecting some running conditions. Before you can make changes to the MAF curve, you need to be sure that all conditions for a given flow rate are going to be consistent so that if you make an adjustment for a rich condition, you correct/improve all running conditions that produce that flowrate.

So run an Injector Slope/BP analysis in EEC Analyzer. Keep in mind the Breakpoint value calculated is dependent on how many cylinders your engine has. So if you are calculating for a V6 or L4, you need to configure EA to know this so it calculates the BP value correctly. For details on how the BP value and cylinder count are related, refer to the post below that goes through the math. After running a slope analysis, you'll see the thin line that represents EA's suggestion for slopes on top of a very jittery region that's a visual representation of the datalog. The thin line that represents your slopes should follow through the "center" of your datalog. The high slope, low slope, and BP that does that best is the values you need to use EVEN if those numbers are not the same numbers that EA calculated on its own. You can always type in your own values after the calculation is done. After entering new values for hi/lo slope & BP, hit <ENTER> to get EA to repaint the graph with your slope/bp values. The high slope tuning should be fairly straight forward. It will represent a majority of the chart. The low slope and breakpoint are a little trickier. The low slope will shift the high slope line up and down. So if you found a high slope that visually follows parallel with the trend in your datalog, but it is above or below the datalog, then you use the low slope and breakpoint values to get the high slope to run through the center of the datalog from bottom to top.

Now if you find that you have to enter a low slope that's much higher than your high slope, then this is often an indication that your Injector Offset curve is adding too much. Lower all points on the Injector Offset curve by the same amount to get the low slope to drop in value. Some applications, you can lower the Injector Offset curve enough that you can literally run the same value for both high and low slope. If you can do this, more power to you. This simplifies things a LOT. But don't be upset if you can't get this.

Likewise, if you find the low slope is below your high slope, this may be an indication that your Injector Offset curve isn't adding enough pulse. Increase all points on that curve and see if this makes your calculated low slope increase a bit.

You know you have good Injector Slope/BP values when you can load up a datalog from any drive, run an Injector Slope analysis just to get the datalog to display, enter your tune's high/low slope and BP values and have your values "fit" in your datalog. At this point, you aren't interested in what EA is calculating, only that your values still fit the datalog. But just because you have good injector slope/bp values does not mean the engine is going to run better. This just means you can stop messing with injector parameters and move onto the MAF curve.

These are very generic descriptions but come close to what is happening. Your results may vary. Also keep in mind, close is generally good enough. But if you are a perfectionist, one thing you can do is use EA's Mapping tool and do further analysis with LONG drives in Closed Loop. Map the average MAF & average LAMBSE. You'll notice that different RPM/Loads will produce similar MAF values (i.e. high throttle at 1500 will produce about the same flow as a lower throttle value at 2500). What you are looking for is consistency. If you have 1500 indicating rich LAMBSEs and 2500 indicating lean LAMBSEs for the same exact MAF airflow, then guess what? You aren't going to be solving that problem by adjusting the MAF curve. If you try, you'll solve the problem in one of those RPM/Load ranges, but make the other worse. The Injector Slopes/Breakpoint are usually the reason why you wouldn't be within say .75 LAMBSEs here. Once you get a LAMBSE map in Closed Loop that shows very near 1.00 lambda (14.64 AFR), then I'd say you are done with both MAF and Injector Slope tuning. However if you still have some annoying driveability problems, then you'll need to attack those one at a time and determine if they are problems caused by mechanical problems or just gremlins in the tune related to stabilized fueling, transient fueling, Closed Loop LAMBSE drift approaching a stop causing lean tip-ins, spark issues, or idle.
...Always Somethin'

89 Ranger Supercab, 331 w/GT40p heads, ported Explorer lower, Crane Powermax 2020 cam, 1.6RRs, FMS Explorer (GT40p) headers, Slot Style MAF, aftermarket T5 'Z-Spec', 8.8" rear w/3.27s, Powertrax Locker, Innovate LC-1, GUFB, Moates QuarterHorse tuned using BE&EA

Member V8-Ranger.com

User avatar
cgrey8
Administrator
Posts: 10296
Joined: Fri Jun 24, 2005 5:54 am
Location: Acworth, Ga (Metro Atlanta)
Contact:

Re: Tuning Injector Slopes

Post by cgrey8 » Sat Jan 09, 2010 9:16 pm

A question about exactly how Injector Slope/BP/Offset is used by the EEC came up. Since this Tech Doc entry only talks about how to tune Injector Slopes/BP info, I thought it might be nice to have info about how the EEC uses this info on the same Tech Doc. Knowing how the EEC uses this info might be useful. Others feel free to disagree with me all you want because I could be wrong. But the following is consistent with what I've studied in the GUFB doc and with my experience tuning Injector Slopes/BP.

It all starts at the MAF. The EEC measures the amount of air coming into the engine via the MAF. It then has a conversion that it applies to convert the amount of air measured to the correct amount of fuel to accompany that air. For gasoline, the conversion is 14.64 parts air to 1 part fuel. In the GUFx strats, this conversion is represented as the Z_AFR scalar. This scalar is not exposed in any other strategy that I'm aware of, but it exists in all MAF-based strategies. With that, the EEC has a beginning point for an air-to-fuel conversion. But what if the EEC isn't targeting 14.64? What if the engine is at WOT and the fuel tables/functions are instead calling for a richer-than-stoic mix? In this case, the EEC adjusts the base stoic conversion to account for added enrichment/adjustment. If there's additional KAMRF learned adjustment on top of the Open Loop LAMBSE enrichment, then that is also taken into account either adding or pulling fuel based on what the KAMRFs have learned. And finally, the EEC adds additional fuel to the mix called accel and transient enrichment (strategy specific). For this discussion, I am abstracting out the accel/transient enrichment since they do not heavily relate to injector Slope/bp stuff and understanding exactly how they work is not critical to understanding Slope/BP/Offset calculations.

Once all the various adjustments have been taken into consideration, the EEC has a calculated amount of fuel that it intends to deliver. Now it just needs a set of rules to convert X amount of fuel into an Injector Pulse. Those rules are defined by the Injector Slope/BP/Offset parameters but the algorithm is what's in question. Injector Offset is a simple adder based on VBAT. So regardless what is calculated via slopes/bp, the offset is the absolute PW adder.

High Slope. The high slope describes the behavior of the injector at a steady stream of fuel flow. Once the injector has fully opened and the fuel pressure in the rail is flowing constant, the fuel follow is constant and relatively predictable. As long as the FPR is making sure the differential pressure across the injector is constant, the injector should continue to behave the same regardless of whether the intake manifold is at high vacuum or at heavy boost. It is possible that an injector pulse is so short that the injector never enters this "steady state" and thus the high slope may not be used at low load/small PW conditions.

Low slope. Unlike High Slope, low slope is always used in the Injector PW calculations. The low slope describes fuel injector behavior up to a certain amount of fuel to be delivered. The theory of Injector offset and low slope is that between these two, you cover for the inaccurate "transients" associated with opening and closing an injector as well as the act of stabilizing the fuel pressure in the rails. During these transients, the injector doesn't flow the same as it does when it is at a steady flow of fuel. In addition, the offset curve often over-compensates and the amount of fuel sprayed blindly thanks to the offset function is counteracted by the low slope. But because these transients last for such a short period of time, there needs to be some indication as to when to "stop" using the low slope and begin using the high slope to calculate total injector pulse width. This is where the breakpoint comes in.

Breakpoint is a measure of fuel. When the needed amount of fuel is less than the breakpoint value, then only the low slope is used to calculate a PW. However when the amount of fuel needed is greater than the breakpoint amount of fuel, then both low and high slopes are utilized for the PW calculation. The low slope is used to calculate an Injector PW up to the amount of fuel defined by the breakpoint function/scalar. In GUFx strats, the breakpoint is a function. Since Ford never used it as a function, Ford converted it to a single scalar as is seen in CBAZA. The remainder of the fuel needed above the breakpoint quantity is converted to a PW using the high slope.

So lets see what that looks like in an example starting from the MAF:

MAF reading is 650 kg-air/hr (roughly the WOT airflow intake on a stock 5.0L@5000RPMs)

The EEC needs this converted from kg-air/hr into lb-fuel/hr. The kg-to-lb conversion is actually done by the tuning software, not by the EEC. But I'll include that conversion here just so you can see it play out from MAF input to Injector PW output.

After taking into account LAMBSEs, KAMRFs, accel/transient enrichment, assume the EEC is applying a final air to fuel conversion of 12:1 AFR:

650 kg-air/hr / 12 air/fuel = 54.12 kg-fuel/hr = 119.3 lb-fuel/hr

Lets assume 650 kg-air/hr was measured while the engine was at 5000 RPMs.

119.3lb/hr / 60min/hr = 1.9883lb/min

1.9883lb/min / 5000rev/min = 3.977e-4lb/rev

Assuming a V8, there are 4 cylinder fires/rev:

3.977e-4lb/rev / 4 cyl/rev = 9.9417e-5lb/cyl

At this point, the EEC knows how much fuel is supposed to go to the next cylinder to get injected and needs to convert this to an injector PW value. We'll assume the result of of the Injector Offset vs VBAT function is 1ms (stock A9L has 13.0v = 1ms). The parameters to do that are:

High Slope is 19lb/hr
Low Slope is 26lb/hr
Breakpoint is 4e-5lb/rev which is 1e-5lb/cyl assuming an 8-cylinder.
Injector Offset is 1ms as mentioned above.

Since the total amount of fuel needed is greater than the breakpoint amount of fuel, the EEC will calculate a PW based on the low slope using only 1e-5 lbs of fuel...the breakpoint amount of fuel.

1e-5 lb
26lb/hr

or 3.846e-7 hrs = 1.3846 ms pulse contribution from low slope

At this point, there's still 8.9417e517e-5lbs of fuel still left that needs to be converted to a PW. So the EEC takes the remainder of fuel and does the same calculation, but uses the high slope this time:

8.9417e-5lb
19lb/hr

or 4.706e-6 hrs = 16.942 ms pulse contribution from high slope

To calculate a total PW, the EEC adds up the pulse time of the low slope, high slope, and Offset function to come up with a total PW, this case:
1.3846ms + 16.942ms + 1ms = 19.326 ms total PW

And as most people are familiar with, that's about right for a stock injector PW at high RPM WOT. As you can see from above, if you use higher injector slopes, the PW will be lower.

The important things to note here are:
  • The larger the injector, the more significant the Injector Offset function is to the results of your fueling. Since it is a blind PW adder, you really need to make sure the stock Injector Offset function is updated to match your actual injectors. The stock function is quite generous with it's adder values. If you leave it being generous with say 42lb injectors, that means things will run terribly rich at low PW conditions (low loads).
  • The closer the high and low slope values are to each other, the less significant the breakpoint value is. If the high and low slope are equal, then the high and low slope calculations are identical thus meaning it makes no difference what portion of the total fuel is run through the low or high slope calc. As is mentioned above, the X3Z has the slope values set equal to each other. There's lots of speculation as to how they managed to decide that was a good idea. But my theory on that is they used much more conservative Injector Offset values in order to make the low slope come closer to the high slope. And even then, the engine probably ran a little too rich on decel, so they solved that by activating Decel Fuel Shutoff (DFSO) so the decel conditions above 1600 RPMs would simply turn off the injectors and that solves the over-rich (aka emissions) problems at decel.
  • Again, I could easily be wrong with all of this, but the numbers tend to make sense and I have witnessed changes in low slope, breakpoint, and high slope have affects that would be expected assuming the above is true. Your experience may differ.
...Always Somethin'

89 Ranger Supercab, 331 w/GT40p heads, ported Explorer lower, Crane Powermax 2020 cam, 1.6RRs, FMS Explorer (GT40p) headers, Slot Style MAF, aftermarket T5 'Z-Spec', 8.8" rear w/3.27s, Powertrax Locker, Innovate LC-1, GUFB, Moates QuarterHorse tuned using BE&EA

Member V8-Ranger.com

User avatar
cgrey8
Administrator
Posts: 10296
Joined: Fri Jun 24, 2005 5:54 am
Location: Acworth, Ga (Metro Atlanta)
Contact:

Re: Tuning Injector Slopes

Post by cgrey8 » Fri Apr 12, 2013 12:01 pm

Here's an article sent to me by another member of the forum that is also good reading and even touches on the differences between the way Ford and GM handle Injector Parameter details:
Fuel Injector Article.pdf
(486.58 KiB) Downloaded 2008 times


Last bumped by cgrey8 on Fri Apr 12, 2013 12:01 pm.
...Always Somethin'

89 Ranger Supercab, 331 w/GT40p heads, ported Explorer lower, Crane Powermax 2020 cam, 1.6RRs, FMS Explorer (GT40p) headers, Slot Style MAF, aftermarket T5 'Z-Spec', 8.8" rear w/3.27s, Powertrax Locker, Innovate LC-1, GUFB, Moates QuarterHorse tuned using BE&EA

Member V8-Ranger.com

Locked

Who is online

Users browsing this forum: No registered users and 2 guests