|
|
|
|
Thread Tools | Display Modes |
09-21-2017, 02:53 PM | #1 |
Scooby Newbie
Member#: 250275
Join Date: Jun 2010
Chapter/Region:
W. Canada
|
DW Fuel Injectors Top Feed 850cc (SMALL IPW)
So... to start off... a few things...
1. I am new to EM 2. I am using the Cobb AccessTuner 3. Car is a 2007 STI I am having some trouble figuring out how to calculate the fuel trim (small IPW for my injectors (21S-01-0850-4). The only datasheet that DW has listed for these injectors is the following... http://www.deatschwerks.com/sites/ww...terization.pdf I am clueless on how to retreieve my small IPW from this data, if possible at all. Thank you for your help in advance!
* Registered users of the site do not see these ads.
|
09-24-2017, 11:36 AM | #2 |
Scooby Specialist
Member#: 375462
Join Date: Dec 2013
Chapter/Region:
NWIC
Location: Portland, OR
Vehicle:2006 Impreza WRX STI WR Blue |
You need to determine the equation of the linear line, then approximate the x-offset of the non-linear points to the linear line. The only data you need for this is the pulsed flow table (and Excel).
The problem with DW data is that you don't get a ton of resolution in non-linear region. It's better than nothing, but having the pulse widths of 1.00, 1.02, 1.04, etc. would be much more helpful than just tenths. |
09-27-2017, 09:14 AM | #3 | |
Scooby Newbie
Member#: 250275
Join Date: Jun 2010
Chapter/Region:
W. Canada
|
Quote:
Thank you very much for the response. Would you mind providing an example of sorts? I do not completely follow what you are stating to do. Last edited by johndom4774; 10-10-2017 at 11:21 AM. |
|
10-18-2017, 04:43 PM | #4 |
Scooby Specialist
Member#: 375462
Join Date: Dec 2013
Chapter/Region:
NWIC
Location: Portland, OR
Vehicle:2006 Impreza WRX STI WR Blue |
Based on your datasheet, Fuel Injector Trim (Small IPW)(Max. IPW) should be set to 1.4 0.6.
Filling out the Fuel Injector Trim (Small IPW) table: EDIT - 0. remove latency from DW data!!! 1. Use linear regression to create best fit lines corresponding to the linear and non-linear ranges of x = Pulse Width (ms) to y = Gravimetric (mg/shot) data points. Doing both of these as a straight line will make the following tasks MUCH easier, though slightly less accurate (many OEMs use this method as it is accurate enough for our purposes - if you were using 2000 cc injectors on a factory motor, I'd probably get a little more complex with the math; i.e. logarithmic or polynomial regression in the non-linear trendline). 2. You need to determine the difference in x-offset between the linear flow line and non-linear flow line you've created and the data points chosen as axis values in #4, below. This is one of the big hang-ups some people get stuck on - even though DW only provides data down to 1.0 ms, you can and absolutely should go lower! You need to select a reasonable low-end value and then set the high end to the first linear range value (1.4 0.6, in this case). Note that the OEM table goes down to about 0.75 ms (before latency) - since you're using larger injectors, logically, you'll need to have a lower 'low' value than stock or you'll still be stuck in the 1990's trying to idle a 1.6 liter motor on 1000 cc injectors 3. You need to convert the values obtained in #2 to a percentage increase to the original linear Pulse Width (ms) x-value - why Subaru chose to use a percentage-based comp. instead of a straight adder is strange... but it ends up having the same result: it arbitrarily extends the pulse width at low commanded IPW values. 4. Plug the compensation percentages in to Fuel Injector Trim (Small IPW) - suggested axis values (for this example ONLY) = 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.6 ('0.6' data point should = '0') 5. Profit. Using Excel makes this a quick task (note, this graph still contains latency!): Keep in mind the DW data is given at 3 bar differential pressure and 13.5 V. You'll notice the blue trendline has an x-intercept of about 0.8 ms, which naturally, matches the dead time of the injector at 3 bar and 13.5 V. EDIT: I noticed that small IPW values are take before latency correction by the Subaru ECU, so while the information above is still accurate, the actual numbers I've given will be off! You need to calculate the latency at 13.5 V and 3 bar from the DW data, then shift the entire graph to the left by this amount. Edits made above to account for this - BOLD items are correct! Here's the final chart ("answers" are included under '%'): With the data in the image above, you'll likely be able to inject a lower fuel mass than even the OEM setup allows (reliably). Any idle issues are on you and your tuner now Extra Credit So let's say you wanted to get even more accurate (or have huge injectors and need to)! For reference, even at the lowest corrected small IPW value of 0.15 ms, you will be within 0.01 ms of linear in the above example/solution... which is pretty damn good. The linear equation of our non-linear region above has an R^2 value of 0.98; by using a logarithmic trendline on the non-linear region, you can obtain an R^2 value of 0.9968 (without digressing into a long explanation - the closer to '1' the better) - I've actually included the 'solution' for OPs setup here, as well; it will be marginally more accurate than the first solution, but the differences in final IPW are so small with injectors of this size that you will not be able to notice a difference: And by using a second-order polynomial, we achieve R^2 = 0.9984 (I didn't solve this one - it would take a bit more work): Lastly, for our purposes, a third order polynomial gives an R^2 of approximately 1.0 (again, didn't bother to solve): Last edited by mrsaturn7085; 10-18-2017 at 06:21 PM. |
10-24-2017, 02:41 PM | #5 |
Scooby Specialist
Member#: 375462
Join Date: Dec 2013
Chapter/Region:
NWIC
Location: Portland, OR
Vehicle:2006 Impreza WRX STI WR Blue |
I've received some information that seems to indicate that DW injector small IPW data may not be accurate at all - given the differences between individual drilled out heads (damage seen with an eye loupe), I would have no problems believing this.
The above information is accurate to the datasheet only - trust, but verify. Real world performance may very likely be totally different if the manufacturer-provided data isn't correct. Last edited by mrsaturn7085; 10-26-2017 at 02:27 PM. |
10-26-2017, 10:33 AM | #6 |
Scooby Newbie
Member#: 468311
Join Date: May 2017
Location: I live In NH
Vehicle:2003 WRX wagon World Rally Blue! |
Subscribed!
Details details <3 |
11-07-2017, 02:46 PM | #7 |
Scooby Newbie
Member#: 250275
Join Date: Jun 2010
Chapter/Region:
W. Canada
|
Thank you very much for the details!
I really do appreciate it! |
Thread Tools | |
Display Modes | |
|
|