M-9353-BB302 Injector Voltage Offset work for sharing/ feedback

[Gojira94]

The low pulse offset adder is indeed needed for non linearity of delivered fuel with very low open times. At ls1 pcm data the table spans till 4ms at lt1 pcm it is set to 2ms[Which can be easily expanded to 4ms with some simple patch

That is good news. I can definitely see there being (big) injectors that don't hit linear flow until a point after 1.952ms. I'd be interested to know more about this simple patch, though I wouldn't likely use it for my Firebird. It would be good for other projects, mine or someone else's. Having crawled through this over a couple of weeks, I think I could do it with another Ford injector a lot quicker. I may take a couple more, like a 36# and a 80# just to see how they look in this rough math modeling method...

Also, I noticed you did a lot of work on the EEX xdf project. Thanks are definitely in order!

[Gojira94]

I loaded these settings into my tune. Stone cold start, immediately fired, no issues with idle as it warmed up. The only other change I made in fueling was to resize the injectors from 33.01 to 31.61. I also made a correction to something I found my tuner from 2009 missed. The boundary for Main to Extended spark advance at 4,000RPM wasn't matched. Very minor.

I can tell it's not as rich. It runs warmer but not hot. Before it ran cooler from too much fuel basically everywhere. It now thinks the injectors are smaller than they were so it's adding a smidge of fuel in places where needed to meet target AFRs, I'm sure. I really need to get a wideband setup on this car. My TBI truck runs an LC-2 and has instrumentation out the wazoo but my Firebird has nothing but stock gauges. I have a spare driver's side A-pillar I was intending to poke holes in for a triple pod, may be able to get to that by spring.

I can say seat of the pants tells me there is quite a difference. Throttle response is dramatically better and it revs quicker. Next step is datalogging it, been a while since I looked into that. TTS Datamaster used to have like a 20 free recording limit, I'm pretty sure EEHack also has built-in datalogging. I can at least get some idea of AFRs with the narrowbands...

I'll do a similar analysis of a couple more Ford injectors like I said above. This all is probably the right idea and probably close, but I am not suggesting anyone use what I've done here without some input from people who know a lot more about this than I do.

[JimCT_9C1]

Sub'd - Thanks for sharing your work!

I have been playing with the data for these injectors as well, and found the below links to be very useful for defining the differences between the GM and Ford data:
http://injectordynamics.com/articles...racterization/
http://injectordynamics.com/articles...racterization/

I am still working through the numbers based on the above, with a goal of having these injectors be indistinguishable from the stockers in my 95 9C1.
I still have plenty of work ahead to install, tune, tweak and validate.

Congrats on the real world driveability improvement!

Jim

[Gojira94]

Sub'd - Thanks for sharing your work!

I have been playing with the data for these injectors as well, and found the below links to be very useful for defining the differences between the GM and Ford data:
http://injectordynamics.com/articles...racterization/
http://injectordynamics.com/articles...racterization/

I am still working through the numbers based on the above, with a goal of having these injectors be indistinguishable from the stockers in my 95 9C1.
I still have plenty of work ahead to install, tune, tweak and validate.

Congrats on the real world driveability improvement!

Jim

Thanks for the interest! I'm almost done with the BB-9593-LU47 sheet, then doing the BB-9593-LU80. I'm working toward an Excel workbook that you can just populate with the Ford tables from their datasheets and each worksheet in the workbook does the rest. Except for the voltage offsets and the multipliers for 43.5psi since they don't exist, but those have graphs that let you match/ smooth to the static values for fuel pressure and VBAT increments that are common to all Ford datasheets. I'd also like to do a duplicate worksheet to deal with 4 or more bar pressure for later GM systems, might be able to set that across the board with a filtered cell on sheet 1 of the workbook. Kur4o also mentioned that HP Tuners has larger VBAT offset tables, across 4ms instead of just 2 with the $EE / EEx definitions. I'd like to be able to address other tuning systems' characterizations for Ford Injectors if I can.

[kur4o]

You can look at the injector patch thread here.

http://www.gearhead-efi.com/Fuel-Inj...njector-tables

There is an xdf that contains the patch. You can use whole patch or only low pulse extend patch and than populate the table with your values.

I do have a better updated version of the patch but it is not set up in a xdf yet. It reads voltage from injectors at a spare pcm pin and some improvements at stratup. If you are willing to try it, I will do some xdf for it. I run it for 2 years now with zero issues and much better overall drivability.

There is also some screenshots with table size and format for ls1 pcm. If you want to add support for that size tables too.

[Gojira94]

You can look at the injector patch thread here.

http://www.gearhead-efi.com/Fuel-Inj...njector-tables

There is an xdf that contains the patch. You can use whole patch or only low pulse extend patch and than populate the table with your values.

I do have a better updated version of the patch but it is not set up in a xdf yet. It reads voltage from injectors at a spare pcm pin and some improvements at stratup. If you are willing to try it, I will do some xdf for it. I run it for 2 years now with zero issues and much better overall drivability.

There is also some screenshots with table size and format for ls1 pcm. If you want to add support for that size tables too.

Sure, I'd very much appreciate a chance to try the patch you already have. I'd like to know more about the version in development and how it uses the spare pin. Sounds very interesting. For 0411 PCMs in HP Tuners or any other platform's definition table I'd just need to see the range of values in the table and understand any rounding or incrementation they do, or if they take any decimal value entered exactly. I've never tuned anything OBD2, but have walked a friend through tuning his H/C/I 2000 Tahoe, so I have a copy of his current tune and VCM editor on a computer. I see these parameters and I'll start studying them in more detail to get my head around it:

Flow rate vs kPA (flow rate)
Flow rate mult vs volts (flow rate)
Min injector pulse (limits)
Default injector pulse (limits)
Offset vs volts vs vac (offset)
Short pulse limit (pulse correction)
Short pulse adder (pulse correction)
Boundary (injection timing)

[kur4o]

There is unsused pin that reads 0-25v. Possibly left for fuel pump voltage but unused in code.

I repurpose it to read injector voltage for better precision. It needs a wire from pcm to one of the injector feeds[close to engine] I was looking at logs of that voltage and you can see how voltage drops at wot or rapid changes. I had to do it since the ign feed for pcm is very inaccurate going all around the car dropping the voltage the pc sees. At some cars can be good but at least on mine there is some voltage drop that cant be fixed, even with better wiring. You also needs to run fixed fuel pressure all the time.[Disconnect the vacuum reference tube].


The patch uses 3 tables from ls1 calibration.
low pulse offset,
injector offset [vacuum vs voltage]
fuel flow vs vacuum.

Since the tables don`t match in size, some interpolation is needed between lt1<->ls1 tables.

[Gojira94]

Here are the 47# and 80# sheets. Had to zip them because .xlsx still isn't supported...

Values you need to input are filled tan. Other colors are there to visually represent important things like minPW, slope, static values from TunerPro, etc.
All voltages are assumed to be 14V for the sake of sanity.
Fuel is modeled to be ethanol blended 93 octane gasoline, non-winter blend, 5-7% ethanol content, with specific gravity of .740, 571.5623862cc/gallon.

The first 4 tabs only require values from the spec sheet from Ford, plus one cell you need to fill for 43.5#. The graph is there to help visually find the value needed, consistent with the overall shape. It will be very, very close. 39.15 and 50.03 are 10.88 apart and 43.5 falls about 45% of the way between the 2 points.

The VoltageOffset tab will automatically graph the Ford spec sheet points for 6, 8, 10, 11, 12, 14 & 15V, You need to flesh out the graph with the remaining values in the tan cells. Once satisfied, paste those 32 GM values into TunerPro. Hit the compare button to see what TunerPro does with rounding the values to its nearest increments (of about 15.267 - 15.3). Paste those values into column C. The graph in the LU47 and LU80 I've provided show the variation "delta" between calculated/ hand drawn data curve and how TunerPro rounds to the nearest increment of 15.3usec. TunerPro values are overlaid in red on top of the calculated and hand entered values in blue. Below 6V probably isn't important, even during cranking. None of my cars have ever turned over with less than 10.5 - 11Vbat on an extremely lucky day...

HiSlope and LoSlope tabs are much the same as the first 4, where you enter the Ford sheet values and craft a value for 43.5# with the assistance of the graph.

The PWAdder tab requires the most effort but it pulls almost everything needed from the other tabs. Primarily you deal with columns C and H. Column C represents the TunerPro 'Low Pulsewidth Adder' table. Columns A & G are the TunerPro xdf points with some added below 488usec to reach zero for graphing down to the offset. Orange below, green above, yellow on either side and red at the breakpoint. You need to add rows to create 1) the MinPW in ms and it shows the delivered fuel per pulse and how many cc/min that is, for the graph at the bottom; 2) the breakpoint fuel mass in both high and low slopes. This is the intersect between low and high slopes, which we will cross on the graph by offsetting the high slope until the breakpoint fuel mass in each column are overlaid on the graph. The offset, or time it takes for the injector to actually begin to open is built in, pulling from the InjOffset tab, using 14V and putting those values in column B.

Watch the formulas when you insert rows in and fill down the formulas on this sheet. Things like:

=(A11*LoSlope!D5)+(C11*LoSlope!D5) will increment "D5" to "D6" so make sure those 2 D5 stay D5.

=G7*HiSlope!D5 will do the same, so keep it "D5"

Columns E & J will be fine and preserve their formulas when you fill down or up the formulas (cc/min at PW "A" Lo and cc/min at PW "A" Hi)


"Bending" the low slope up to the high slope to get the offset values:

First, insert rows for offset and breakpoint (breakpoint mass on both Low and High slopes as noted above). Tweak the column A and G values for this to get it to match for 5 places to the right of the decimal point. The actual breakpoint mass at 43.5psi is noted below the Low table, in red. (Note - on either side of that, there are for reference, the low slope and high slope flow rates in cc/ms at 43.5psi).
Next, look at H3. It's something like =G3+0.xxxx, where 0.xxxx is the artificial offset we're adding to the Hi slope so it will intersect the low slope at the breakpoint. Just dial it in and fill down the formula =G3+0.xxxx from H3 to H36 until the red and blue data points for the breakpoint overlap. The zoom function in the View menu in Excel is very helpful between 200% - 300% for this.
Now that the Hi slope and Lo slope breakpoints overlap, we can start "bending" the Lo slope up until it overlaps the Hi slope line. Again, 200-300% zoom is great for helping you get the lines to overlap almost seamlessly. The idea is that when we convert back to usec for TunerPro's LT1 xdf we're no more than 10-15 millionths of a second off in calculations and the delta the xdf value increments produce.
Start entering values into C3, C4, C5 and so on, zooming and tweaking until the lines are almost seamlessly stacked, up to the breakpoint value, which will be 0.

I'm interested in looking now at the smaller injectors, in the 19# range for the sake of curiosity.

Disclaimer:
Again, I want to be clear that this is theoretical work and assumes my math modeling is correct and I haven't screwed anything up in Excel itself. Anyone and everyone with more math and injector science skills please chime in and help correct any missed assumptions or math, etc.

[Gojira94]

***EDIT: Corrected sheets per issue below uploaded to post 16.***

I found an 'oops.'

Some calibration summary sheets have different max pressure values in BKCOMP, LSCOMP, HSCOMP and OFFCOMP. Some have a highest value of 60.03 and some use 64.96, like the LU47 and probably some others. It's not a used value other than for graphing but I'm going to tweak and upload a new version of the sheet to indicate it's a value that needs to be entered from the calibration summary sheet.