ECU Explorer and Performance Measuring

[SpeedNeeder]

I was thinking about getting a GTech or... can ECU explorer be used to measure 0-60, 1/4 mile, estimate HP, etc?

[lookslikeanevo]

no

[TDC Joe]

You can get relative timing differences off of the ECU Explorer but nothing will tell you HP figures or lateral g's etc

[TSiWRX]

ecuEdit (which is now capable of logging as well as editing) current beta release does offer a dyno function. Not sure how accurate it is, though - and keep in mind, too, that it's still a work-in-progress.

No 0-60/quarter-mile times or lateral-G figures, though.

---

However, if using the OBD-II tap is your preferred method of extracting information:

OBDII Software, Vehicle Diagnostics Software

 

This program has been verfied by sublo, a fellow LGT brother, through direct contact with the manufacturer, to be compatible with our LGTs, and in addition to datalogging, is capable of both dyno and 0-60/quarter-mile figures.

 

I am uncertain, however, if braking performance or lateral-acceleration can be measured by the digimoto interface/program.

[j0ebert]

Is this the ecuEdit (EcuEDIT - Customised and Enhanced Code) you are talking about? I don't see anything for our car on there. I think i'm going on a 400 mile roadtrip as a passenger in a Legacy this weekend, so with my laptop and my cable (I hope it gets here today) I want to try as many programs as possible.

Edit: I think I found the one you are talking about on Main Page - OpenECU but the link to ecuEdit isn't working right now. Anyone try ecuFlash on there yet? It shows it's okay to flash the 04-06 Legacy GT.

[MDTURBO]

With the correct attributes and a level road surface (and a radar detector to keep cops away) a road dyno package will be BETTER than a shop dyno. You are calculating loads for air and ground resistance. That being said, you'll also wind up doing over 100 mph on each run and this of course is not safe, thus why the world has shop dynos. lol.

[rob]

I am not sure if there is any way to get accurate dyno info from any scan device. The problem is having a data sampling time fast enough to give useful acceleration results.

 

Anything that uses the VSS or ignition pulses (like "The Road Dyno") is accurate and repeatable given a controled environment. If you do your runs in 2nd gear, then you can easily find a place to do the runs. I have used The Road Dyno with great success.

 

-Rob

[MDTURBO]

^ true, but if you can get a sample of 500ms that is pretty good. Your average dyno software is going to be doing the same thing.

 

Regarding 2nd, the gearing is shorter and will wind out quicker thus less data points per rpm. 4th (in a 5spd) is usually a 1:1 ratio and thus the desired gear to test in. Unfortunately 3k to redline in 4th is about 60mph to 125ish-mph in our car (or my car)

[Jon in CT]

I am not sure if there is any way to get accurate dyno info from any scan device. The problem is having a data sampling time fast enough to give useful acceleration results.

 

Anything that uses the VSS or ignition pulses (like "The Road Dyno") is accurate and repeatable given a controled environment. If you do your runs in 2nd gear, then you can easily find a place to do the runs. I have used The Road Dyno with great success.

This is all true. For those interested in a freeware option, check out StreetDyno v1.0. You'll have to build/buy a little device that'll let you "listen" to ignition system pulses as MIC input, though.

 

For those who want to use their logger data to generate dyno plots, I've found an interesting option near the bottom of the page at SUBARU IMPREZA WRX TURBO, which the author calls "ECUTEK versie 1.1.0 - juli 2005." It's consists of some MS Excel macros that post-process data from EcuTeK's DeltaDash Road Dyno. Basically, you log a single parameter, engine speed. You need to massage your ECUexplorer/ECUedit output .CSV file a little. The time units need to change from milliseconds to seconds (I think) and some prefix information about tire size, weight, gear ratio, temperature ©, Cd, frontal area (sq meters), etc. needs to be added. I've included a sample of such a file as an attacment below (rename it to *.csv).

 

The only problem is that I don't run MS Excel. I use the OpenOffice 2.0 freeware and none of the buttons work (missing script), so I can't say how well that stuff actually works.

 

Oh yeah, almost forgot to mention it. My Dutch is pretty weak, too, and it's all in Dutch.

 

Just remembered that a guy with the screen name of west_minist on NASIOC and OpenECU has been able to get those MS Excel macros to work. Here's an example:

http://forums.openecu.org/files/v1_dyno_273.png

STI7UK-EMS-Export.pdf

[rob]

^ true, but if you can get a sample of 500ms that is pretty good. Your average dyno software is going to be doing the same thing.

 

Regarding 2nd, the gearing is shorter and will wind out quicker thus less data points per rpm. 4th (in a 5spd) is usually a 1:1 ratio and thus the desired gear to test in. Unfortunately 3k to redline in 4th is about 60mph to 125ish-mph in our car (or my car)

 

1/2 Hertz? At 2000 RRM, your ignition pulses at over 8 Hz....that's 16x the resoltuion at the lowest possible RPM point. I don't think that's enough resolution to be useful.

 

2nd gear testing has been good enough for my analysis. Using the Scanmaster 3 and The Road Dyno I was able to collect enough info to calculate BSFC on my RS. That's a very handy number to have.

[ahains]

1/2 Hertz? At 2000 RRM, your ignition pulses at over 8 Hz....that's 16x the resoltuion at the lowest possible RPM point. I don't think that's enough resolution to be useful.

 

2nd gear testing has been good enough for my analysis. Using the Scanmaster 3 and The Road Dyno I was able to collect enough info to calculate BSFC on my RS. That's a very handy number to have.

 

Hey Rob,

I haven't tried it myself, but I think the datalogging option would get you at least reasonably accurate acceleration measurements.

 

The ECU obviously knows each revolution that transpires very accurately (it must to handle ignition), so I presume that the RPM it reports is 100% accurate. 500ms sounds like a pretty lame sample rate, I think I was seeing more like 250ms when I played with it awhile back?

So if you measure the RPM accurately around 4 times per second, you can easily calculate change rpm / second.

 

Armed with knowledge of your gearing ratio, you can go to acceleration / second and calculate G's / hp / whatever.

 

True, you'll require some fill in the dots in your chart since you only get about 4 readings per second. You may potentially miss brief drops in power (say 1/8th second duration) so it might not be good for watching for detonation/issues, but you should get an accurate calculation of torque across that 1/4 second island of time. You can use the other datalog parameters for watching for detonation/issues

 

Can you spot any holes in my thinking?

[Deer Killer]

It's not accurate enough. In fact there is a huge amount of noise in the signal.

 

accuracy != precision. It is precise.

[enthusiast]

']

Oh yeah, almost forgot to mention it. My Dutch is pretty weak, too, and it's all in Dutch.

 

There is a drop down "Taal" -->Engels

 

Everything will be in english. I started to translate it and then saw this at the top.

[ahains]

It's not accurate enough. In fact there is a huge amount of noise in the signal.

 

accuracy != precision. It is precise.

 

To clarify - are you saying that the RPM reported when datalogging is not accurate? That's unfortunate

What did you test it against, an external tachometer?

[Deer Killer]

To clarify - are you saying that the RPM reported when datalogging is not accurate? That's unfortunate

What did you test it against, an external tachometer?

It is not accurate because of noise. Don't need a reference. I've wondered if this is because of of our dual-mass flywheel, but so far I haven't been able to convince anyone to send me a datalog of RPM as fast as possible on an STI, or simlar motor without the DMF.

[ahains]

Maybe I'm just dense but I don't understand what you mean by noise. You first mentioned "noise in the signal", but I'm not sure what signal the ECU bases its RPM datalog values on. I had assumed it was calculated directly by knowing how many ignition events it directed (I don't think it is uncommon for chips like this to support simple incrementing counters in hardware), but you're assertions suggest I was totally wrong.

 

Are you saying it uses something like a microphone pickup to "listen" to the RPM of the engine? That seems like kind of a wonky way to do it. If not, what noisy sensor does it use?

[Deer Killer]

Maybe I'm just dense but I don't understand what you mean by noise. You first mentioned "noise in the signal", but I'm not sure what signal the ECU bases its RPM datalog values on. I had assumed it was calculated directly by knowing how many ignition events it directed (I don't think it is uncommon for chips like this to support simple incrementing counters in hardware), but you're assertions suggest I was totally wrong.

 

Are you saying it uses something like a microphone pickup to "listen" to the RPM of the engine? That seems like kind of a wonky way to do it. If not, what noisy sensor does it use?

I don't know what it's doing. My assumption is that it is counting ignition events, but it turns this into ignition events per minute, and the data becomes garbage.. I've worked on some ways to filter it, but it's not really trivial, and unfortunately that task takes low priority for me. I have an algorithm which I'm pretty sure I can get to work in some fasion, maybe. the resulting accuracy and reproducibility are in question, this would require me to do many actual runs, which I'm not even attempting to do until I get it working 100%.

[highwaydrifter]

With the correct attributes and a level road surface (and a radar detector to keep cops away) a road dyno package will be BETTER than a shop dyno. You are calculating loads for air and ground resistance. That being said, you'll also wind up doing over 100 mph on each run and this of course is not safe, thus why the world has shop dynos. lol.

 

 

More like

 

A shop dyno if setup with the proper dyno, loading settings, and airflow will be the your best bet.

 

I only know of a few of those on the west coast, and havent seen a single one on the east coast (gruppes being one, and then a fwe high end bmw/benz tuners). It is fun playing infront of Moikes fans when they are on..

[Deer Killer]

More like

 

A shop dyno if setup with the proper dyno, loading settings, and airflow will be the your best bet.

 

I only know of a few of those on the west coast, and havent seen a single one on the east coast (gruppes being one, and then a fwe high end bmw/benz tuners). It is fun playing infront of Moikes fans when they are on..

A load based dyno is necessary for proper empirical measurement and tuning of MBT timing, not just for WOT but on the _entire map_, though I doubt few tuners do this anyway. However a road-dyno is much more accurate as to what the car is _actually doing_, especially for things like boost response and map-scaling. Also it is more accurate for power and torque ratings. As long as you have a flat road an no wind you can get a very accurate measurement of total loss (How many HP does that stupid wing lose?) and crank horsepower.