AX22 versus Driftbox accuracy revisited
12-29-2006, 01:40 PM
#16
Burning Brakes
Join Date: Jul 2001
Location: Orange Park Acres, CA
Posts: 982
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Red9
Maybe you can use the data in court! LOL
12-29-2006, 06:51 PM
#17
Addict
Lifetime Rennlist
Member
Lifetime Rennlist
Member
Sorry it took me a while to come back.
I did some analysis today with the data provided by KPG and TB993TT. The findings were quite revealing....From lack of revelation!
The test consists in the following..
Same run datalogged by TB on both units, the Race Technology AX22 and the Drift Box. both units were fitted in the same car.
The objective of the analysis is to verify that the data recorded by the loggers can in fact be checked based on relationship between acceleration recorded, speed and time.
I use the data extrapolated for every hundredth of a second (speed, long Gs and time) and through a mathematical formula, I compare the speed obtained vs. the speed recorded by the datalogger's GPS for every 1/100th. The numbers should be quite close if there was no slope or the datalogger was not fooled in some way through incorrect calibration while driving or movement during the run etc..
So far on all the tests that I had performed on the DBox data, only one was accurate, the others were all showing irrelevant numbers, those other runs were in fact extremely fast.
On this run recorded by TB993TT, the test proved that actual GPS speed, and calculated speed are very close (2mph at the end of the run), meaning the data recorded on the Dbox for that run is correct. While the long Gs do vary vs. the AX22, the relevance is really small.
The same test performed on the AX22 also showed very close results between calculated speed and GPS recorded speed, there is a speed variance at the end of the run, which is due to the slope that was present during the run.
At the left hand side, the Y axis shows the scale for the variance between calculated speed and datalogger(DBOX) speed in mph, and the bottom scale, X axis is the speed in mph at which the car was traveling... As you can see the variance is minimal, never above 0.05 mph. This contrasts with most of the runs that I have seen by other users in private where the variance was in the 10s of mph.
Thsi other chart compares the speed calculated for the AX22 vs. the DBox, NOT the speed shown by the datalogger. The difference can be due to many variables, mainly the way the data is extrapolated in each of the units (the Dbox does not extrapolate exactly by 1/100th. of a second) and the existing slope during the run. The difference is small nevertheless.
Net net what this tells me is that the Dbox runs SHOULD also show the same accuracy level when analyzed using the methodology I follow as the AX22, and if they don't, it means the run was not properly done. Both units have the same accuracy levels as seen by TB993TT's runs.
Hope this helps.
I did some analysis today with the data provided by KPG and TB993TT. The findings were quite revealing....From lack of revelation!
The test consists in the following..
Same run datalogged by TB on both units, the Race Technology AX22 and the Drift Box. both units were fitted in the same car.
The objective of the analysis is to verify that the data recorded by the loggers can in fact be checked based on relationship between acceleration recorded, speed and time.
I use the data extrapolated for every hundredth of a second (speed, long Gs and time) and through a mathematical formula, I compare the speed obtained vs. the speed recorded by the datalogger's GPS for every 1/100th. The numbers should be quite close if there was no slope or the datalogger was not fooled in some way through incorrect calibration while driving or movement during the run etc..
So far on all the tests that I had performed on the DBox data, only one was accurate, the others were all showing irrelevant numbers, those other runs were in fact extremely fast.
On this run recorded by TB993TT, the test proved that actual GPS speed, and calculated speed are very close (2mph at the end of the run), meaning the data recorded on the Dbox for that run is correct. While the long Gs do vary vs. the AX22, the relevance is really small.
The same test performed on the AX22 also showed very close results between calculated speed and GPS recorded speed, there is a speed variance at the end of the run, which is due to the slope that was present during the run.
At the left hand side, the Y axis shows the scale for the variance between calculated speed and datalogger(DBOX) speed in mph, and the bottom scale, X axis is the speed in mph at which the car was traveling... As you can see the variance is minimal, never above 0.05 mph. This contrasts with most of the runs that I have seen by other users in private where the variance was in the 10s of mph.
Thsi other chart compares the speed calculated for the AX22 vs. the DBox, NOT the speed shown by the datalogger. The difference can be due to many variables, mainly the way the data is extrapolated in each of the units (the Dbox does not extrapolate exactly by 1/100th. of a second) and the existing slope during the run. The difference is small nevertheless.
Net net what this tells me is that the Dbox runs SHOULD also show the same accuracy level when analyzed using the methodology I follow as the AX22, and if they don't, it means the run was not properly done. Both units have the same accuracy levels as seen by TB993TT's runs.
Hope this helps.
12-30-2006, 07:02 AM
#19
Addict
Rennlist Member
Rennlist Member
Thread Starter
Jean
You have just saved many people loads of $$$ - the new cheapo Driftbox (the one that doesn't measure drift just acceleration) should be more than adequate for accurate acceleration time measurement.
I am truly amazed that anyone can get incorrect readings -WTF are they doing ? ( KPG suggested I do a run with my sig other moving the unit around etc -unfortunately that wasn't going to happen, normal people get upset at these kinds of Gs particularly if your test track is quite narrow
)
To me it is quite simple - read the data on the 60-130 thread, ask on the board where others think you should be based on the known factors 4WD/weight/mods/gear changes etc then do 3 or 4 runs - check the slope and the runs should be within ~0.5s of each other.
You have just saved many people loads of $$$ - the new cheapo Driftbox (the one that doesn't measure drift just acceleration) should be more than adequate for accurate acceleration time measurement.
I am truly amazed that anyone can get incorrect readings -WTF are they doing ? ( KPG suggested I do a run with my sig other moving the unit around etc -unfortunately that wasn't going to happen, normal people get upset at these kinds of Gs particularly if your test track is quite narrow
) To me it is quite simple - read the data on the 60-130 thread, ask on the board where others think you should be based on the known factors 4WD/weight/mods/gear changes etc then do 3 or 4 runs - check the slope and the runs should be within ~0.5s of each other.
12-30-2006, 10:18 AM
#20
Pro
Join Date: Aug 2005
Location: Michigan
Posts: 726
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by TB993tt
Jean
You have just saved many people loads of $$$ - the new cheapo Driftbox (the one that doesn't measure drift just acceleration) should be more than adequate for accurate acceleration time measurement.
I am truly amazed that anyone can get incorrect readings -WTF are they doing ? ( KPG suggested I do a run with my sig other moving the unit around etc -unfortunately that wasn't going to happen, normal people get upset at these kinds of Gs particularly if your test track is quite narrow )
To me it is quite simple - read the data on the 60-130 thread, ask on the board where others think you should be based on the known factors 4WD/weight/mods/gear changes etc then do 3 or 4 runs - check the slope and the runs should be within ~0.5s of each other.
You have just saved many people loads of $$$ - the new cheapo Driftbox (the one that doesn't measure drift just acceleration) should be more than adequate for accurate acceleration time measurement.
I am truly amazed that anyone can get incorrect readings -WTF are they doing ? ( KPG suggested I do a run with my sig other moving the unit around etc -unfortunately that wasn't going to happen, normal people get upset at these kinds of Gs particularly if your test track is quite narrow
) To me it is quite simple - read the data on the 60-130 thread, ask on the board where others think you should be based on the known factors 4WD/weight/mods/gear changes etc then do 3 or 4 runs - check the slope and the runs should be within ~0.5s of each other.
Jean , your efforts are much appreciated as well...
12-30-2006, 12:17 PM
#21
Rennlist Member
Join Date: Mar 2004
Location: behind the Corn Curtain
Posts: 2,314
Likes: 0
Received 0 Likes
on
0 Posts
Upon TB's initial post concerning the Driftbox, I purchased the PerformanceBox, hoping its' accuracy would hold up in the Rennlist Court. 
So this is all good news for me.
One note: If you are a Mac user and are not running the latest w/Intel chip, you will not be able to do graphics - you will still be able to get text. Software, such as Virtual PC, will allow you to do graphics. This was all fully disclosed up front by the US distributor - nice guys to deal with.
Noah
So this is all good news for me.
One note: If you are a Mac user and are not running the latest w/Intel chip, you will not be able to do graphics - you will still be able to get text. Software, such as Virtual PC, will allow you to do graphics. This was all fully disclosed up front by the US distributor - nice guys to deal with.
Noah
