Cheap Performance Meter

[Rusty]

Following on from this thread I thiought I'd start a proper thread about a cheap performance meter using a GPS enable mobile phone, Excel and some maths.

I have a Nokia N95 and have downloaded the Race Chrono package from www.racechrono.com for free. I am going to purchase a Qstarz BT-Q818x (extreme) 5Hz GPS bluetooth jobbie that will link to my phone and have been looking into the maths to calculate wheel hp and fly hp (guestimate anyway) and think this could be a winner. The total cost is £50 for the bluetooth GPS receiver and this can enable sat-nav on your phone too.

The calculations centre around instantaneous measurements of speed and acceleration (I assume calculated from previous data points).

P = Ft.v
Ft = Feng + Fair
Feng = m.a
Fair = 1/2.rho.Cd.Af.v^2

From this you get that

Ft = m.a.v + 1/2.rho.Cd.Af.v^3

Where:-

P = Power (W)
Ft = Total force (N)
Feng = Engine force to accelerate car (N)
Fair = Aair resistance (N)
m = Car mass (kg, my S2 is 1525kg + 25kg fuel +100kg me = 1650kg)
a = Measured acceleration (m/s^2) - I'm going to look at how this is calculated and try some averaging techniques if appropriate.
v = Velocity (m/s)
rho = Density of the air (1.25 kg/m^3)
Cd = Drag coefficient (0.32 Coupe, 0.35 Avant) - ???
Af = Frontal area (Coupe 1.89m^2, Avant 2.04m^2)


So that's the power at the wheels that the car is creating.

From RR data I have calculated losses with engine speed for 4th gear between 50 and 120mph. These giove conversion factors for Patw to Pfly of between 1.2 (4k) and 1.5 (7k). For lower gears I would assume these losses are less due to mechanical advantage and more in the higher gears due to the reverse effect. SOme calibration runs in 4th will show what I can expect the car to produce at the wheels and thus allow for conversion factors to be applied.

All in all I think there this should give useful results and is 1/5th the cost of a GTech (not the same accuracy I grant you) and can do data logging and also lap timing too. What a great piece of kit. The back calculation of performance data is a bonus and use of the data logging mode will yield any analysis you like.

When I'm up to speed with it all I'll see what sort of error can be expected and how the factors affecting the error can be handled.

HTH

[S-Express]

sounds interesting, what if your phone already has a sat nav device in it? does that save £50 or will it not be exact enough?

[MattS2]

Hi Andy.

I downloaded this yesterday and had a little play on the way home last night.
I just had a go at logging the journey but that was all.
I'll be following this thread with interest.


Matt.

[Rusty]

My N95 already has an internal GPS but there are a few problems:-

1. positioning in the car can mean that signal isn't as strong as you'd like
2. it's not that accurate
3. it's only 1Hz and from playing yesterday it really is nowhere near accurate enough for a car like the S2 where a lot happens in a second in 1st/2nd gear.

Here's a graph of a leisurely run to 50mph, you can see that the data points are way too far apart to give any meaningful results.

[Nj]

Christ almighty this is a bit more intense than my coupe trailer thread. I'm out of my depth (but very much interested )

[Rusty]

I'm looking forward to having a play, I don't mind number crunching if it's 1/5 the price and it means that I've got over half a recaro left in the bank

[Nj]

From the site:
FAQ: Which Phones does RaceChrono work on?
A: RaceChrono works on phones with Nokia S60 2nd edition feature pack 1 and above, or Nokia S60 3rd edition.

Did you just install it onto your N95 with no problems?

[Rusty]

No problems at all. The only setting I had to search for was to set the GPS to internal, that took... 10s ish Great piece of kit, if it turns out to be **** on then I'll donate some pesetas to their cause.

[Nj]

hot diggidy dawg. I'll give it a go on my N82

[A80Avant]

finally got my bags back from france and i can perform some measurements...

i may end up getting the g-tech as well though...

[Rusty]

Yeah, this isn't going to be as accurate as the more expensive kit but that doesn't worry me as if I want accurate power measurements then I'll go to the rolling road and get the whole car checked over. This offers a quick and esy test of components and also offers data logging and lap timing for track days or just good old thrash cross country. It's got otehr nice features like altitude (from teh GPS) and teh option of linking to the OBD for RPM, T Posn and other such things. Great program, wel done to the makers for doing it for free too.

[A80Avant]

i wouldnt say excactly that going to a rolling road will give you accuracy.... none at all actually
Even if the dyno is set up properly you can never get an excact figure as to what the car is doing on the road simply because on 4th gear the velocity and amount of air entering the engine when you are on the road is more than what the fans on a dyno can produce.So you basically are getting less power than what the car is actually making on the road.

Dynos are simply a way to reproduce a result before and after a map or an upgrade.BUT it must be done on the same day and under the same specs given to the dyno program.Otherwise you are doing nothing.

Basically if you do lots of tests on the same road on pretty much the same conditions you are probably getting much better figures than what you would on the dyno,AND you are not paying for it!

[Rusty]

I've used the same dyno for the last few times and so it's about comparing results.

Have ordered the BT-Q818x for just under £50 so see what it's like and how I get on.

[S2central.net]

Good stuff Andy - I love a nice spreadsheet !

Let us know how you get on with this - in terms of accuracy and repeatability...

[Rusty]

Yeah, accuracy will be down to how you get the data off it. There are pre-written routines for all performance benchmarks and the longer the time period, the more accurate it'll be as the error is down to the refresh rate of the GPS combined with the positioning accuracy. With a 1Hz system you'd have to guess at the start and stop for each test using the nearest data points or by extending the best guess gradient to axis intercept. My 5Hz GPS receiver will make theis a lot more accurate. For lap timing they reckon on +/- 0.25s on 1Hz and +/- 0.15s on 5Hz, with performance testing this will be a greater improvement. The GTech Pro I had gave results with a span of about 0.5s so anything around that will be as valuable but the difference here is that I can download to a PC and have so much more information to check out and can just go for a drive and data log and then edit it all down as much as I want.