For tuners, a Holy Grail of Injector voltage Offset

[Hog]

Good info above, I would like to add in reference to resized injectors, there is the belief that when using an injector with known offsets, that if you take the same injector and get it resized, that the previously known offsets are still applicable. This is incorrect.

[XfireZ51]

Here's the PW calc:

BPW = BPC * MAP * T * A/F * VE * BVC * BLM * DFCO * DE * CLT * TBM

Where

BPW - Base Pulse Width

BPC - Base Pulse Constant

MAP - Manifold Absolute Pressure

T - Temperature

A/F - Air Fuel Ratio

VE - Volumetric Efficiency

BVC - Battery Voltage Correction

BLM - Block Learn

DFCO - Decel Fuel Cutoff

DE - Decel Enleanment

Hog,

Also, increasing fuel pressure on an injector will necessitate increasing the injector offset. Takes longer to open the injector. So even if you know the offset for the injector at its rated FP, increasing the FP changes the correct offset.

Scott,

Your 20ms is at 100% DC. Typically you'd like a DC% of about 80 at WOT. So that drops the PW to 16ms at 6000rpm,
12.8ms@7500rpm w an 80%DC.

I can see that the offset for the Accel injectors I am using, the offset is incorrect because my VE table has that "bathtub" curve
at low rpm. right know I have "tuned around that" but I need to bump the offset further to smooth that out. I could make a change to the IB v PW table and modify the smaller PW only leaving the rest alone. In either case, I'd need to re-do the VE Learns tables.
Changing the Voltage Offset would richen the AFR globally although more so at the bottom than the top.

[scottfab]

Thanks for the reply.

Originally Posted by XfireZ51

This may help in understanding why tuning benefits from having the correct Injector Offsets. From a Thirdgen thread

http://www.thirdgen.org/techboard/po...545-post3.html

"There is no set formula for flow versus Injector offset.
Wish there was, but it is a little more difficult than that.
The offset (or "deadtime") is the result of several factors, which include:
o mass of the pintle and its lift off the seat
o fuel pressure
o coil inductance
o magnetic circuit characteristics - function of materials and design (like eddy currents)
o injector driver circuit design

The list of why there is lag/delay/deadtime/latency etc are all valid. But "why" is not an issue (with me). It is a given. The other given is that this delay (OT and CT) vary from vendor to vendor AND even inj to inj if the internal resistance (actually impedance) varies too much.
(BTW not all injectors have pintles)

I do disagree that there is no set formula for computing flow vs offset introduced by OT and CT. Well maybe there isn't one but that does not mean we can't come up with one. Tuning based on empirical evidence is fine but there can/should be a way of reducing that to theory (formula) From that one can determine the max effect of OT and CT on what the ECM is doing. Is it an order of magnitude less or can a delay be large enough to in fact swamp the ECM?

There's lots to consider. Not just supply voltage (regulated or not), and ambient heat and fuel pressure/temperature and plenum heat soak and all the other variables mentioned in the list but I submit other factors we could come up with.

Originally Posted by XfireZ51

So depending on these, there is a correction curve based on battery voltage that takes in account the above, which defines the "deadtime" as the opening delay time minus closing delay time, in milliseconds.

This is developed for each injector type and is then coded into the calibration. Any change in the above parameters can affect this offset correction curve in the calibration.

Again the theory of operation here is not in question. It's a given. I could even add a few variables left out. What I am missing right now is what the max pulse width variation the ECM can dial in (form most lean to as rich as possible). Having those numbers in ms is critical for determining the effect of each injectors OT and CT.

Originally Posted by XfireZ51

If you have a test bench (with an oscilloscope and an accelerometer) you can set the injector up and then feed it increasing PW using a function generator and using the accelerometer can determine how long at various voltages it takes to open and close the injector. This then becomes the battery correction curve, and then the BPW after that is the actual PW that provides the fuel."

Another good explanation of injector offset:

http://www.thirdgen.org/techboard/po...341-post8.html

"The spike you are seeing is the "peak" portion of a "peak and hold" injector pulse. This is a hardware function and has nothing to do with injector bias. The reason for the bias is because the injector requires a set amount of time to open to the point where fuel flows. This time can be as short as a few hundred microseconds for a TBI injector to over one millisecond for a saturated port injector. Say for examplethe ECM calculated that on a port injected setup, it requires a 1.5mS pulse width. Since it takes 1mS for the injector to open, fuel will only flow for .5mS or only 1/3 of what is needed. If a bias of 1mS is added to the final pulse width, a 2.5mS pulse occurs minus the 1mS for the injector to open and fuel flows for 1.5mS. Just what the ECM commanded. Make sense? "[/QUOTE]

Yes but again I take this theory of operation as a given and stated on the last post that we need real numbers. For instance. Given the example you listed. If 1.5ms was needed but only 0.5ms was happening due to OT wouldn't the ECM via the O2 sensors see a lean condition and make the necessary corrective action? (BLM) Or is there something I'm missing?

Originally Posted by XfireZ51

Scott,

Perhaps you could use your oscilliscope to help in developing these correction curves for the various injectors available for the LT5.

Example of VE table w Injector Offset too low. Creates a "bathtub" area of VE.

I'm in discovery on theory of operation so as to understand the possible worst case effect of OT and CT at the moment. From there some actual lab kind or work could be formulated. Here is what I've discovered (or already knew)
1. The ECM through short term and long term tables makes changes to the injector on time/off times.
2. The injectors have not only OT delays but CT delays as well.
3. Many other factors (no complete list yet) have influence on the injector performance. The list is TBD.

What is missing besides filling in the TBD list? Assuming #1 above cannot compensate for the effect of #2 then #3 needs to be completed to get a formula. From there mitigation ideas could be investigated including ECM table corrections and/or injector supply voltage regulation etc. If nothing else I find it entertaining to get complete understanding of such things. Makes for interesting dreams.

[scottfab]

Originally Posted by XfireZ51

Here's the PW calc:

BPW = BPC * MAP * T * A/F * VE * BVC * BLM * DFCO * DE * CLT * TBM

Where

BPW - Base Pulse Width

BPC - Base Pulse Constant

MAP - Manifold Absolute Pressure

T - Temperature

A/F - Air Fuel Ratio

VE - Volumetric Efficiency

BVC - Battery Voltage Correction

BLM - Block Learn

DFCO - Decel Fuel Cutoff

DE - Decel Enleanment

This is good stuff for sure. Thanks! I think I've been exposed to it before in passing. I just was not receptive to thinking that deeply about fuel issues at the time. What was the source for this info? LT5 supplement? Or is it in the FSM? Once I find out the units used for each variable I could run some iterations.

Originally Posted by XfireZ51

Hog,

Also, increasing fuel pressure on an injector will necessitate increasing the injector offset. Takes longer to open the injector. So even if you know the offset for the injector at its rated FP, increasing the FP changes the correct offset.

Seems like with higher FP you'd get larger flow rates too.

Originally Posted by XfireZ51

Scott,

Your 20ms is at 100% DC. Typically you'd like a DC% of about 80 at WOT. So that drops the PW to 16ms at 6000rpm,
12.8ms@7500rpm w an 80%DC.

I see. It would seem then that a delay of 1ms would introduce about 1/12 error of the over all value. At lower rpm it would be even less % wise. Interesting stuff for sure.

Originally Posted by XfireZ51

I can see that the offset for the Accel injectors I am using, the offset is incorrect because my VE table has that "bathtub" curve
at low rpm. right know I have "tuned around that" but I need to bump the offset further to smooth that out. I could make a change to the IB v PW table and modify the smaller PW only leaving the rest alone. In either case, I'd need to re-do the VE Learns tables.
Changing the Voltage Offset would richen the AFR globally although more so at the bottom than the top.

Do you know what the OT is for the stock injectors?

[XfireZ51]

Originally Posted by Hib Halverson

Being that this is a tuning thread, let me hijack it a bit.

What do you guys do for engines forced to run on E10?

Do you work with E10's different stoich or do you just tune for a certain lambda value from your wideband?

Hib,


The topic you bring up should probably be its own thread. It involves using Lambda for tuning if operating in O/L or
manipulating the NB O2 swing points to arrive at the ~14.3 stoich using BLMs for C/L.

On the topic of Injector Offsets, I realized I should have better explained the value of the using accurate offsets. The Injector Bias is one of the variables used in calc'ing the PW at any given RPM/MAP. Particularly for the purpose of emissions, but also driveability etc. the ECM needs to account for the electrical characteristics of critical components like injectors. Voltage to the injectors isn't regulated per se, and so this affects the performance of the injector. Even in a well maintained vehicle, we will have conditions where voltage will vary, ie fans kick on, lights, accessories etc. In order to compensate for the voltage variation, and so the PW realized, the ECM needs to know the "lead time" required by the injector in order to deliver a consistent fuel flow. Inaccurate offsets can be one reason why the motor seems "happy" at one point, but then rough or ragged at other times. VE tables provide a base fueling and INT and BLM will modify that fueling given feedback from the O2 sensor. But there's a lag to this. The ECM does not immediately modify the fuel. It takes a certain amount of time for the INT to affect the BLM. And that's if yopu are running C/L. In O/L, there is no modifying of the PW based on O2 input. So, IMO, correct offsets are even more important.

Attached are the Inj. Comp tables for voltage and by PW.

[Hog]

Originally Posted by XfireZ51

Hog,

Also, increasing fuel pressure on an injector will necessitate increasing the injector offset. Takes longer to open the injector. So even if you know the offset for the injector at its rated FP, increasing the FP changes the correct offset.

.

Yes, agree 100%, increasing fuel rail pressure affects both the opening delay and the nonlinear flow region of injection.

Non Linear flow regions

• A frequent problem involves the difference between rates calculated for tables to perform calibrations and the actual flow rates. Errors of 12 percent are possible. Additionally, flow rates at idle and cruising can differ regardless of the presence of fuel injectors designed to perform at the same rate.
  
  

Simply using Bernoulli's equation (the flow rate increases with the square root of the pressure ratio-


-isnt enough, a calibrator must also factor in injector dynamics such as offsets as well. Basically the more accurate the info about your injection hardware, the more accurate your fueling will be.
Garbage In=Garbage Out.

[XfireZ51]

"
-isnt enough, a calibrator must also factor in injector dynamics such as offsets as well. Basically the more accurate the info about your injection hardware, the more accurate your fueling will be.
Garbage In=Garbage Out."

Hog,

If there is one point I would make re: Tuning, you hit the nail on the head. The fundamental operation of the motor and the baseline values for the parameters need to be as close to optimal as possible. The objective of the tune is to minimize the amount of intervention the ECM must exert in order to achieve proper performance.
As you point out, there's enough systemic variation due to tolerances found in the components, electrical supply, fuel supply etc. The ECM needs to deal w that. Now if you compound that w improper or inoperative components, incorrect settings etc, there is too much variance for the ECM to address. Accurate baseline information is critical(IMO) to a smooth running motor.

[KJL]

Originally Posted by XfireZ51

"
-isnt enough, a calibrator must also factor in injector dynamics such as offsets as well. Basically the more accurate the info about your injection hardware, the more accurate your fueling will be.
Garbage In=Garbage Out."

Hog,

If there is one point I would make re: Tuning, you hit the nail on the head. The fundamental operation of the motor and the baseline values for the parameters need to be as close to optimal as possible. The objective of the tune is to minimize the amount of intervention the ECM must exert in order to achieve proper performance.
As you point out, there's enough systemic variation due to tolerances found in the components, electrical supply, fuel supply etc. The ECM needs to deal w that. Now if you compound that w improper or inoperative components, incorrect settings etc, there is too much variance for the ECM to address. Accurate baseline information is critical(IMO) to a smooth running motor.

Forgive me, I am way out of school here but what is the objective of achieving a perfect offset? Fuel economy? Power? How do you access/program this information/data?

[XfireZ51]

Originally Posted by KJL

Forgive me, I am way out of school here but what is the objective of achieving a perfect offset? Fuel economy? Power? How do you access/program this information/data?

Voltage utilized for operating the injectors is not regulated. It will vary based on load put on the charging system. In the meantime, variable voltage affects the performance of the injector which in turn makes fuel flow variable. For consistent performance and emissions, fuel flow needs to be consistent. The Injector Voltage Offset provides the ECM with the time interval to use in order to compensate for the lag created by less than optimal voltages.
If you have stock injectors, no need to do anything. However, many of us use other than stock injectors and so the electrical characteristics at various voltages needs to be accomodated to maintain a consistent fuel flow. In a previous post, I have attached a .jpg of the tables used in the LT-5 calibration that are used for this purpose.

[efnfast]

So Dom, all of this gets addressed with a custom tune?