My Run With A Gtech In My N/a

[Merlin]

  ed-au-ba-xr6 said:

that's cool - im assure you im not lying, the run was wirtnessed by three others and repeated twice. g-Techs have also proved to be within 0.01 secs of a track timer on multiple experiments - my mate bought the G-Tech as he is a mechanical engineer undertaking a project that requires him to test the 0-100 runs of certain cars - he chose this particular model for its accuracy.

as for the show pony comment - c'mon guys im just teasing - no offence ment. Not trying to say the NA is better or anything (I would buy a T in a heartbeat) just showing off what the NA is capable of.

[faygun]

Merlin that's very impressive

go the team N/A

[Guest jc20b]

im in for team n/a xr6 that's if you guys want a 17 year old in the team

I need to start modding my xr first exhaust then will wee what else is best bang for my buck

[Guest TIC-302]

Talking about G Tech's, I bought mine in 1999, the first type G Tech

I took it to Eastern Creek racway to check it out

At that time I had a modified Liberty RS turbo

The car ran Consistant Low 13's on th 1/4mile (13.2-13.3sec on the street)

and 0-100 in high 4's(4.8,4.9sec)

The day I went to the Creek it was hot and didn't get a good start

1st run 13.80sec and the G Tech showed 13.89

2nd run 13.90 and the G Tech showed 13.85

the mph was alot higher than the Creek timing, I was running 98mph and the G Tech was showing 103-105mph

So MY G Tech is very accurate !(apart from th e speed)

So now My EF XR8 auto runs, 0-100 in 7.7sec and 0-400m in 15.7sec

My 1985 RX-7 with a carby 13B , manual, extractors and 2.5" system

runs 0-100 in 7.2sec and 0-400 in 15.2sec.

[Guest fordmuscle_83]

that's an amazing time to 100 merlin! Sure u only have herrod exhaust? Do you have an aftermarket airfilter?

Your car must be putting out around 200kw to get a time like that.

Pete

[turbotrana]

You might be lucky and get a G-tech that's calibrated well. My experience is that they are favourable by up to half a second and a few mph. Nevertheless they are a usefull tool. But really you shouldn't be saying my car did this and that on the Gtech cause it dosn't mean alot. If you said my car did this stock and did this with an exhaust then that means something. For qtr mile times on a stock car just read your Wheels or Modern Motor Mag

[Guest TIC-302]

  turbotrana said:

You might be lucky and get a G-tech that's calibrated well. My experience is that they are favourable by up to half a second and a few mph. Nevertheless they are a usefull tool. But really you shouldn't be saying my car did this and that on the Gtech cause it dosn't mean alot. If you said my car did this stock and did this with an exhaust then that means something. For qtr mile times on a stock car just read your Wheels or Modern Motor Mag

I am just letting people know my experiance and posting some times

yes, the EF XR 8 is modified(slightly)

and yes the RX-7 is modified, anything else you like to know?

I also got 90% of Motor and wheels magazines from the early 70's to 2003

just for turbotrana

MY G TECH IS VERY ACCURATE

[Merlin]

  fordmuscle_83 said:

that's an amazing time to 100 merlin! Sure u only have herrod exhaust? Do you have an aftermarket airfilter?

Your car must be putting out around 200kw to get a time like that.

Pete

Thanks Pete - As TIC-302 stated, G-techs are quite accuarte despite what some members may believe.

To me the time isnt really that amazing - I mean a stock manual XT does 0-100 in 7.1 secs (according to Motor), so around half a second off to 100 is a good result - I was hoping to just get into the 6 second bracket, but did it quite convincingly.

Despite the exhaust I assure you the car is stock - herrod stated I would get around 8-10rwkw with the exhaust so im guessing somewhere around 190-195kw at the flywheel.

PS: Now that double demerits are over, im going to have a few more runs - apparantly theres an abandoned airstrip somewhere near Windsor....my mate and I are going to try and find it on Friday night!

[TURBO4LT]

Im using the Race Technology AP-22 and the best time I have got from my T is 6.2 0-100km . and 14.3 @ 160.5 . I ran it on my mates G Tech and I got better results of 14.01@ 165 so im not sure which is more accurate?????

[bkofoed]

Well I'm in the process of building my own g-tech like device from scratch. So it may be worth explaining the theory in the hope of understanding the errors involved:

The basic principle is this:

A single chip, dual-axis accelerometer (which is a triumph of seminconductor manufacturing- a device with a silicon weight suspended on silicon springs over a silicon sensor, all embedded in a chip the size of a match head)

Take said device, (in my case, and ADXL202E which is a +/- 2g dual-axis device, same as in the AP-22 and possibly G-tech range), connect it to your favorite microcontroller, add a screen and some buttons, and throw together some great maths and hope for the best.

As the device measures acceleration in it's two sensitive axis, (both static and dynamic), you can determine when the car starts to move by a spike in accelaration.

At this point, kick off your 1/4 mile timer.

Read the value of accelaration like mad (I.e. in as short a code loop as possible), and store the results with the timestamp of when they were read.

Every few tenths of a second, integrate the results, and from physics we know integration of Acceleration = Velocity. Assume a standing start (actually we can measure it by calibrating the device at power-on to assume that is a stopped condition), and whacko- we know how fast you're going.

Integrate velocity, and you'll get --- distance!

So we now know, how long since your launch, the current speed, and current distance.

When velocity is above 100km/h, find the closest time slot (or average two) and report that- ding - 0-100 time.

When distance reaches the 400m, do the same and report Time and ET.

Sounds easy right?

Well, it's all good in theory but the G-tech/AP22 people earn their money in the algorithms to accurately read the real acceleration.

Remember the sensor is extremely sensitive- it can detect your pulse through your hand and measure vibration we can't even see. So it takes a bit of maths (sliding windows, moving averages, trending, etc name your favorite statisical modelling tool) to get the real data out.

Aside: This is the very same sensor used to trigger airbags, just waiting for a very distinct accelaration change profile - a crash infact

Sources of errors:

- measurement of the output of the accelerometer device (usually done with a pulse-width technique)

- number of samples per second (smooths the graph out and reduces ripples/jitter)

- the device itself (though some include a calibration factor in the factory, but they are more expensive and used for earthquake monitoring applications)

- the accuracy of the time source in the micro (crystal based- so usually insignificant)

- the precison of the maths used (integration will need to be done in floating point, so 8bit resolution? 16bit?)

I hope to include some calibration in my device, and I have an overpowered chip for the job so I can do more number crunching- time will tell if it's any better than a guy with lots of coffee and a stopwatch, and a painted line on the road....

Hope this helps

Later

BK