ElmScan + Customized Scantool Software = data logging!
 

Hello -

Last year I got one of these gizmos that interfaces the OBD-II to a PC :

ElmScan VPW Scan Tool - P/N: 420300
EDIT: URL BELOW IS DEAD LINK BECAUSE PRODUCT NO LONGER OFFERED
http://www.scantool.net/products/pro...products_id=12
http://ecomodder.com/forum/attachmen...1&d=1233886788
Now, it took me a (long) time to get it working right. My Laptop PC (probably) had too many apps running, so the Elmscan was having connection problems. Here is what I did :

1 - Bided my time and found an obsolete 2.5" 4GB hard disk for $10. This way, if I crash the disk proving the system, I don't care because the hard disk is worthless anyway. I have an easy-to-swap hard disk, so I can pop the "base OS" hard disk in, go to work, and swap my "regular OS" back in when I am done.

2 - Installed a base XP OS on the 4GB hard disk in my laptop. That way, there is no contention for the COM1 port.

3 - Downloaded the "scantool v1.13" software that comes with it. Since it comes with source code, I could customize it after ...

4 - ... I downloaded the "allegro" game programming library for Windows/XP :

Allegro - Download - Latest version (download all422.zip)

5 - Used my trusty Visual C++ version 6.0 to compile Allego. I procastinated doing this because I expected it to be a PITA. In fact it was easier than I thought. This is the power of gamer-programmers. They are edge and MAKE IT WORK! Pssst: But, you don't need to compile it. There are also pre-compiled binaries at the URL in Step 4. Stooopeed me did it the hard way.

6 - Once I had the library, I made a Visual C++ 6.0 project for the scantool software and customized it for my SW2 :

Still working on the color pallete, :o :
http://img166.imageshack.us/img166/7...overhs3.th.jpg


These are the only variables that are available for my Saturn :
http://img528.imageshack.us/img528/1...nsw2so8.th.jpg


4 - But the really big thing is that I got it to log data. I am logging in 1 second increments. I post-processed this to be one minute increments in this example result :

Using the original 1-second increment data, I used this to make a *simple* RPM vs MPH chart that shows my engine on coasting :

http://img528.imageshack.us/img528/8...hartcj1.th.jpg


My long term goals are :

1 - Testing only. The Scangauge is still my main man. At this point, I want to see the behavior of mods like my EFIE, the IAT resistor mod, and maybe some aero mods. I wanted this gizmo because I couldn't see the fuel trim or the 02 sensor with my ScanGauge I. Also, no data logging with the ScanGauge.

2 - Get a solid state hard disk. I can get a 2GB hard disk for $60 right now. If I run the 4GB until it crashes, the price will be lower by then.

3 - Define an algorithm that uses the data I am collecting to display real time MPG (Help Me Please :o !!!!!!).

CarloSW2

.

I've been chewing on that one a little too since my elm323 showed up. Here's all I've got so far:

PID 07 -> xx Long term fuel trim Bank 1 % x*(100.0/128)-100
PID 0b -> xx Intake Manifold Pressure kPaA x
PID 0c -> xx xx Engine RPM RPM x*0.25
PID 0d -> xx Vehicle Speed km/h x
PID 0f -> xx Intake Air Temperature C X-40


lastScanTime=now (milliseconds)
onScanEvent{
time=now-lastScanTime
lastScanTime=now
distance=speed(pid0d) * time
"airMoles"=some PV=NRT thing, http://en.wikipedia.org/wiki/Ideal_gas_law
* V=rpm(pid0c)*displacement*volumetric efficiency, P=pid0b, T=Pid0f
oxygen=airMoles* 20% ?
stiochfuel=oxygen/14.7
fuel=stoichfuel*fuelTrim(pid07)

litersPer100KM= fuel converted to liters/distance converted to km

MilesPerGallon= distance converted to miles/fuel converted to gallons

}

ref: http://prj.perquin.com/obdii/
ref: http://en.wikipedia.org/wiki/Ideal_gas_law

dcb -

*Arf*, that's really hard. Time to do more homework.

CarloSW2

Sorry. Try out the opengauge freeduino signal processor then. Let us know if it works :)


(then teach it to read your cruise control switches on an analog port)

dcb -

That's ok. Actually, I see *some* of these calculations already in the scantool source code, so it is definitely speaking the same language. I guess that's another reason why the ScanGauge is so cool. I'll need to read your "opengauge development: gunio technical considerations" thread also.

CarloSW2

The guino isn't going to help too much with an obd mpg effort. The guino/opengauge efforts currently look directly at the injector and the speed sensor, and bypass all the obd business. It is much easier to record how long the injector has been open than it is to get mpg out of an obd port, excepting that you have to tap some wires.

Hello -

I noticed this commercial program from Australia. It is designed to work with the ElmScan hardware. It also claims to be able to deduce instant-MPG from OBD II data, which is what I have been looking for :

OBD 2007 - Installation Guide
http://www.glmsoftware.com/Documenta...de20080822.pdf
Picture of "Fuel Analysis" Gauge :
Attachment 1659

Question: What do you think? Does the equation they are using look valid to you?

CarloSW2

Using only the MAF data is an approximation of fuel economy assuming the car's always running at stoich. That's probably what the SGII does.

However, for vehicles with broad band sensors, their program supports adjustments from the equivalence ratio which should make reported fuel consumption more precise.

Thank you for the thread here... I think I have an idea now why my Scanguage isn't calibrating well: My MAF is failing, and thus, it's not reporting accurately, throwing the SG's tracking off.

On the A4 (or MkIV) TDI engines, the MAF sensor is notorious for failing out. I know from the slight lack of power, that mine is failing out again. So, time to sleep, so I can book more hours at work, and buy a new MAF for my car :P

Hello -

(apologies in advance for some of my "custom" acronyms. I am "programming" the variable names in the post)

I am trying to determine an approximate Instant MPG calculation for my scantool software. All of the following is based on this OBD 2007 documentation blurb :

OBD 2007 - Installation Guide
http://www.glmsoftware.com/Documenta...de20080822.pdf - Page 15 of PDF
Here are my inputs :

- Intake Manifold Pressure (MAP) units: inHg
- Vehicle Speed units: MPH

Here are my constants/conversions :

- AFR = Air/Fuel Ratio = 14.64
Source: http://www.glmsoftware.com/Documenta...de20080822.pdf

- "DOG" = Density of Gas (g/cm3) = (0.71 +0.77) / 2 = 0.74
Source: Gasoline - Wikipedia, the free encyclopedia

- "M2K" = Convert Miles to Kilometers = 1.609344
Source: Convert miles to km - Conversion of Measurement Units

- "KML2MPG" = Convert KM/L to MPG = 2.35214583
Source: WikiAnswers - How many km to 1 litre of gas

- "SPH" = Seconds Per Hour = 3600

- "MPL" = Milliliters Per Liter = 1000

Steps to convert with example inputs from real OBD II data logging :

- Intake Manifold Pressure (inHg) = 17.542
- Vehicle Speed (MPH) = 52.533

1 - Convert MPH to KMH
=> MPH * M2K = 52.533 * 1.609344 = 84.54366835 KMH

2 - "FFG" = Calculate FuelFlowGrams/Sec
=> MAP / AFR = 17.542 / 14.64 = 1.198224044 FFG

3 - "FFL" = Calculate FuelFlowLitres/Second
=> (FFG / DOG) / MPL = (1.198224044 / 0.74) / 1000 = 0.001619222 FFL

4 - LPH = Calculate Liters Per Hour
=> FFL * SPH = 0.001619222 * 3600 = 5.829198051 LPH

5 - KML = Calculate KiloMeters / Liter
=> KMH / LPH = 84.54366835 / 5.829198051 = 14.50348189 KML

6 - Calculate MPG
=> KML * KML2MPG = 14.50348189 * 2.35214583 = 34.11430445 MPG

The MPG number looks "ballpark" correct to me, subject to programming and testing. The most iffy part to me is step 3, where I divide by 1000 to get the "FFL". I am doing this because in step 2, it is not clear that the FFG is in Liters.

Also, the OBD 2007 algorithm calls for a MAF sensor. In my S-Series Saturn, I *think* I have a MAP sensor. My question here is, what's the difference? Do I need to convert from MAP to MAF?

Based on the steps above, here is the output I get. This is based on the data logging from thread #1 :

Pending programming and testing, do you think I am on the right track?

CarloSW2