After adjusting rail pressure, the injection time will be reduced. Because of this, the SOI could be tweaked.

I hope somebody can verify these maths and if so, hopefully come in useful to people elsewhere:

Notes:
- This is using LINEAR interpolation which is not 100% accurate for this purpose! Hopefully admin will comment on weather or not it is accurate enough. I tried everywhere for a non-linear interpolation calculator but they are impossible to find!

For this example, I will work with the following parameters:

Code: Select all

RPM is at 2250
IQ is at 60
Rail Pressure before adjustment: 1100 BAR
Rail Pressure after adjustment: 1155 BAR


First, I need to find the injection duration for both before and after rail pressure adjustment.

The injection duration maps IQ axis do not match perfectly with the rail pressure and SOI axis, so I must use interpolation.

Code: Select all

Before (1100 BAR):
I have a 1100BAR value on the BAR axis for injection duration. Just need to interpolate IQ:

x = Value for IQ 57 RP 1100BAR: 1.045
y = Value for IQ 61 RP 1100BAR: 1.135
z = Value for IQ 60 RP 1100BAR: .....?.....

First find the delta (d): d = (60 - 57) / (61 - 57)
So d = 0.75
(where 60 is the IQ I want to find the duration for, 57 is the lower bound that I have and 61 the upper bound that I have)

Now using delta (d) I can find the value I want (z): z = ((y - x) * d) + x
So: z = Value for IQ 60 RP 1100BAR: 1.1125


Code: Select all

After (1155 BAR):
Now I have to interpolate both BAR and IQ as there is not a value for 1155 on the BAR axis or 60 on the IQ axis! So this is even harder!

v = 1155 (the rail pressure we are using)
w = value for IQ 57 (a) RP 1100 (e) BAR: w = 1.045
x = value for IQ 61 (b) RP 1100 (e) BAR: x = 1.135
y = value for IQ 57 (a) RP 1200 (f) BAR: y = 0.959
z = value for IQ 61 (b) RP 1200 (f) BAR: z = 0.993

(are you keeping up? haha my head hurts)

First find the delta for Rail Pressure 1155 (d): d = (v - e) / (f - e)
So d = (1155 - 1100) / (1200 - 1100)
So d = 0.55

Now I can find the following injection duration times
h = IQ 57 (a) RP 1155 (v)
i = IQ 61 (b) RP 1155 (v)

h = ((y - w) * d) + w
h = ((0.959 - 1.045) * 0.55) + 1.045
h = 0.9977

i = ((z - x) * d) + x
i = 1.0569

So finally I need my result. r will be the timing duration of IQ 60 at RP 1155.
The delta for IQ is 0.75, worked out in the previous code block. (yes I am getting lazy now)
r = ((i - h) * 0.75) + h
r = 1.0421


(Reminder this is for RPM 2250 and IQ 60)
Now we have our Injection Duration BEFORE the rail pressure increase: 1.1125ms
and the Injection Duration AFTER the rail pressure increase: 1.0421ms

The difference in time is 0.0704 ms or 70.4 us.

Now using the SOI calculator found here: simple-soi-calc-excel-spreadsheet-t27
Using value -70.4 us (minus because the time is shorter)

I conclude that my SOI at 2250 RPM and IQ 60 should be -0.9504 or -1

Mods, what do you think? or have I just wasted 3 hours :roll:

Just to check:
Have you applied the correct factor to the injector duration map? (0.976 iirc)
Also double check the time unit that injector duration is in.... I thought it was micro seconds not milli seconds ;)

Also for your calculated ingition angle, is that BTDC or ATDC?

just had a quick read at the maths and seems to make sense. You say you may be interested in non-linear interpolations, that is easy to do with just about any data manipulation program. I'm not an Excel expert but I am ready to bet that you can also do that in Excel / Calc.
Which kind of interpolation would you like to implement? A non linear third order fit would be more than enough for your aims.
You can do it very easily with programs like Mathematica too, unfortunately you have to learn the Mathematica language tho :D
If you are familiar with the C language then the ROOT language (from Cern) is based on that and is very good for doing this kind of data analysis.
But you may also end up using wolframalpha.com... it's very limited, slow and the sintax may at times be funny, but I think you can make it fit the data with the desired (ideally infinite) precision.

Still Excel / Calc seems the simplest way. I'll look into it later if I have some time, but googling "Excel fit" will for sure put out something interesting!

ross2482 wrote:Just to check:
Have you applied the correct factor to the injector duration map? (0.976 iirc)
Also double check the time unit that injector duration is in.... I thought it was micro seconds not milli seconds ;)

Also for your calculated ingition angle, is that BTDC or ATDC?

nope. i have not applied that factor to injector duration. i am not at home now, but I think it is 0.001 (give or take a 0), hence why I seem to be working in ms rather than us.

can you explain why the factor is 0.976? this seems really random for a timing factor?

Also my SOI is before TDC, so using the result calculated above, -0.9504 would start the injection 0.9504 degrees nearer to TDC.

i will update my factors and above calculations later.

OdinEidolon wrote:Still Excel / Calc seems the simplest way. I'll look into it later if I have some time, but googling "Excel fit" will for sure put out something interesting!

I will have a look. More precision is never a bad thing but I am curious if linear would be enough for this purpose?

rlees85 wrote:

OdinEidolon wrote:Still Excel / Calc seems the simplest way. I'll look into it later if I have some time, but googling "Excel fit" will for sure put out something interesting!

I will have a look. More precision is never a bad thing but I am curious if linear would be enough for this purpose?

Well, I think so, but more precisionis always better ;)
But I honestly think a linear fit would suffice.

Will look at this later:

http://www.ithaca.edu/tpfaff/pdf/CurveF ... hExcel.pdf

Will be interested in what the mods say about my first post. Modifying SOI = brown pants I think

Hi guys,
nice thread, very good to think so detail, that is always good thing.

If there is any volunteer i would suggest to do this (you will get our full support):
Locate this maps:

Desired rail pressure
Torque limiter
Injection duration
SOI
Rail pressure precontrol value

than look at the dependencies.

than one spread sheet in excel with original values
another spread sheet with tuned rail pressure and TL
and the results for ID, SOI and RPPV

that would be very useful calculator for all users who does matter about precision and good work.

I have heard optimum end of injection can be anything from 2-18 degrees after TDC.
Would you not need to find best end of injection first ?

ecuedit wrote:[...]
If there is any volunteer i would suggest to do this (you will get our full support):
[...]
than one spread sheet in excel with original values
another spread sheet with tuned rail pressure and TL
and the results for ID, SOI and RPPV
that would be very useful calculator for all users who does matter about precision and good work.

I'd lke to do that, it would be highly useful for everyone.
but first I have to understand the dependencies, then I need to have time.
right now I'm in the middle of university exams. I'll have free time starting september 13 ;)
be prepared to answer a lot of questions :D
it's going to be a nice project

Relic wrote:I have heard optimum end of injection can be anything from 2-18 degrees after TDC.
Would you not need to find best end of injection first ?

If you raise rail pressure and you inject the same quantity in less time, you'll end injection sooner.
Adjusting SOI like rlees85 is calculating, means bringing end of injection back to stock.

About finding the best end of injections I agree, because stock end of injection is a compromise between performance, emissions, noise.. etc.
We have in mind mostly performance 8-)
But you would need something like cylinder pressure against crank position,
or (a simplier way) a personal steady state dyno to test every IQ and every RPM and get the best SOI that produces highest power for every pair of IQ,RPM. This wold be very time consuming because we can't change SOI in real time on the dyno to see what happens :( instead you have to flash the ecu many many times and compare results.

Ok. I see.

So you'll have to use TLs to lock the IQ across all rpm to do it on the dyno (15 times) ?
That would only be for operating temp SOI I take it.
Then retard from there if needed.