RomRaider

this is the log file that i use

i will try that when i get a chance

Airboy, some of the 5EAT guys asked about using the dyno function. I added a MPH parameter to the dyno function. The parameter has a "None" button so it can always revert back to using rpm. IIRC, the 32 bit ecu's use the ABS sensor to log vehicle speed. From my own 5MT datalogs, I know that the mph and rpm match very linearly, so I'm fairly sure that the LGT 5EAT will be able to use this.

I don't have the calibrations for the LGT 5EAT, but they'll still be able to use it as a tuning tool for comparing one run to the next run. It'll be like a dyno chart using an induction pickup to read rpm.

BTW, I also added a total timing chart and a table rescaler. I deleted the load and rpm rescaler from the other charts. I didn't like the way shepard's method would screw around with some of the data. I use the 2D rescaler to change the MAFv table. It's very handy to reduce resolution in areas that don't need it and increase resolution in areas where there are erratic behavior.

[edit - bug]

Are you sure about this? The 32bit ECUs list a "VSS" parameter (which I've had to tap the wire for, hence my knowledge of this) that supposedly pulses 4x per revolution of the center diff. Picking up the speed from the ABS tone wheels would contradict this.

Not saying I'm right, just saying there's different data out there.

I'm not sure. It's what I heard.

However, I know that the LGT 5MT is accurate. I also looked at the LGT 5EAT mph and rpm graphs and they look like what I expect. It's accurate enough to see when the torque converter locks up and what the stall speed is. Before Airboy completely rewrote the road dyno spreadsheet, I tried it with some LGT 5EAT and the road dyno did work as expected.

Here's the reference about the ABS sensors. Jon [in CT] is very knowledgeable about stuff like that.

http://www.legacygt.com/forums/showthre ... post467214

I think that post contained the critical fact: CAN. He mentions the ABS sensors being used to derive car speed in the context of a CAN system. It wouldn't surprise me if the pre-CAN 32bit ECUs still derive speed from the VSS as the Impreza service manual says but the new CAN-based systems have moved on.

No. He is referring to CAN as in car area network. I am sure that post is referring to 2005-2006 LGTs. I do not believe he is referring to the new OBDII-CAN interface.

I had not realized that CAN (which I was, as you say, expanding to Car Area Network) had any valid application outside of the new-model cars as an ODB-II replacement.

However, to further muddy the issue, if you look in the 2005 Impreza Service Manual (I've other years too, but this is the one I had in front of me), page 5MT-38 shows where on the tranny the VSS is and how to replace it and then page 5MT-76 shows the internal gearing that couples the VSS to the geartrain inside the tranny.

Yet, in seeming contradiction to this, page IDI-11 states that:



To exit the realm of the tech spec, and enter the realm of the practical, things get even more confusing. I have my VSS wire tapped. I know from real-world calibration that each pulse of the VSS is worth 1/2420th of a mile, which is about 2.18ft. Multiplying this by the 4 pulses per rev that the manual claims, this would lead to 8.73ft per rev of the wheel or 605 revolutions per mile. However, the wheel actually covers about 6.30ft per rev, or 838 revolutions per mile. This discrepancy seems to imply the VSS signal seen at the ECU is NOT the pulse of the ABS tone wheels.

Sadly, the math doesn't come out any better for the VSS being in the center diff and the math is too involved to be worth posting here.

In other words, NFC.

Thanks for the additions, Michael. The LGT must have different way to get vehicle speed because, as I mentioned before, for my '06 WRX, RPM and MPH don't track each other very well.

Sorry. There was a major bug in the spreadsheet that I modified for 5EAT. It calculated accel and hp but then stopped there. The calculations for torque and the graphs were based on the wrong column.

I heard there is another revision with total timing table.

Is that ready yet...........Mike??????????

I don't know if it's ready, but it's available for checking. It's in the post right above yours.

Scroll to the far right and you will see a spreadsheet called TotalTiming. It gets its data from the base and dynamic advance table. Scroll to Column AR and you will see a plot.

I tried it seems to work nicely, it says 5eat only, but it seems to work OK, what is for the auto only?

The Dyno function is for the 5EAT only although it seems to still work okay with the 5MT.

I changed the MPH and RPM smoothing. MPH is either obtained from the log or it is calculated from the rpm and gear data. Then the MPH is smoothed.

The dyno function that Airboy wrote smoothes out the RPM first.