Data Acquisition Webinars

[Magpie]

Thanks for the video.

If anybody has a VBO file for the Gardner circuit at Eastern Creek it would be appreciated.

[Dan]

I've only got AIM Files, attached is the DRK that you can open in the Race Studio software and the map file for the circuit that I did my analysis with (6 sectors), a video for the two 1:51.6 laps in the last session is in my signature.

It was only my second time out there in the MX5 but the best time is only 1.5 seconds off the NC record so it should be close enough to get an idea.

Not sure if that helps you at all.

[Magpie]

Thanks. I'll see if the aim software is a free download.

[oztrackdays]

004_1.41.7_BEST.vbo.zip

[Magpie]

Dan and Oztrackdays much appreciated. Aim software is a free download, will need to spend some time learning. I'm already using Circuit Tools so the VBO file was easier to use. Will spend the next week or so looking at them.

Dan this will also assist with deciding on AIM, Racepak or Haltech logging.

[Dan]

Following our chat on PM here’s the track map I made from other peoples videos and data before I first went out there for the first time in my MX5 (the speeds were from Charlie Brown’s thread on his NC and the gear shift points are from his video). I find having a few reference points to start from creates an anchor to learn the track quickly.

Obviously gear shift points and some no brake corners are also going to be dependant on HP and gear ratio’s but it should give you an idea at least and with the speeds you can tell pretty easily you have something to fix with your driving if you can’t keep similar speeds through the corners or you don’t get up to the same speed on the straight following the corners.

The only real mistake I think on that map is that on T5 where it says LHS (left hand side) that should actually say RHS (right hand side)

Here’s a couple of videos that I think are good to watch too:
SMSP ’Turn by Turn guide' with John Boston
https://www.youtube.com/watch?v=lnU36YVVoOU
Phil’s (Charlie Brown) laps (1:50.198)
https://www.youtube.com/watch?v=0dqpM3BTZVs

[mazmad]

Dan are you getting brake or steering data?

[Magpie]

mazmad I have had a quick look at Dan's data and have yet to see how to display some of the fields in the file. I have a day off on Sunday so may spend some time with the AIM Software.

I'm yet to install a steering sensor, I need to find the best place to install. Since I have a shell with no engine but steering intact I'm going to look at where best it would be to place it. Since it is a 5 volt sensor it can be plugged into any data logging system. Will just need to be calibrated. Brake data for me is only pressure, not yet logging front/rear.

[Dan]

Dan are you getting brake or steering data?

Nope I don't get either and I don't have the option to connect either on my AIM Solo do you have those on yours?

I would love to have steering data as it would be good to look at and use it to filter out maths channels as well as identify things like understeer but I don't really miss braking too much as my AIM Solo has an internal accelerometer so I can just use the Longitudinal Acceleration data as a substitute as it's very quick to update and accurate.

Here's a couple of examples of how I can use it to quickly identify things using Longitudinal acceleration as a substitute for braking data in maths channels:

Identify where I'm completely letting off the accelerator when I’m not slowing down which is usually where I didn't slow down enough for a corner so I have uncertainty which makes me let off a bit rather than keeping some maintenance throttle or I am creating oversteer by letting off the throttle (There are some false positives where I let off the accelerator a fraction before the clutch but they are easy to see):
IF(LT(MX5_CLUTCH_SWI,0.20),bit_and((GT(Longitudinal_a,-0.20)),(LT(MX5_PPS,10))),0)
Clutch isn’t engaged, longitudinal acceleration is greater than -.2 and the Accelerator peddle is less than 10%

Identify area's I'm not braking hard enough, where I'm slow transitioning between braking/accelerator or where I'm shifting after braking:
bit_and((GT(Longitudinal_a,-0.75)),(EQ(MX5_PPS,0)))
Longitudinal acceleration is greater than .75g and Accelerator peddle is 0%

[mazmad]

No Dan, don't have those either. And p'd off to be honest, completely false misrepresentation of what data is available in the sales materials. Local distributor was not helpful, other than trying to sell the add ons to collect this data through other channels.

Interesting, so you can create your own readings from formulas? Sounds like a good approximation I will have to look into that.

Btw good time on the south circuit well done.

[Dan]

Thanks mate, my first time out there, lots of room for improvement.

I saw the same materials form AIM. I was misled by it too but I think the sensors might be only in the NC2 or something, not sure. Still annoying!

Yeh you can create formula’s using maths channels, either to display calculations as calculations (e.g. data1 * Data2 / Y) or like those formulas I put before which act as a binary switch e.g. coasting is 1 and not coasting is 0, which means you can just look for long lines of 1’s in that calculation and there is probably time there in not coasting.

Looking at the data I've found that coasting is a bad habit of mine for slowing down in fast corners rather than braking and getting back on the accelerator quicker. Looking at the data last time I went out at SMSP GP it looks like there is about 0.8 seconds in turn 1 and turn 5 just braking instead of coasting which I found comparing laps I tried braking vs ones I coasted and that formula shows it up clear as day. Pretty sure there is time in corporate hill too but I haven't tried braking up there to be able to see for sure, it's not lifting exactly but a while of low % throttle and I think I can find time in braking instead.

Here’s the data from my run from turn 4/5 that shows how much faster braking is than coasting in that corner (these are 4 laps done in succession with the red line being my PB lap) - You can click the image to make it bigger/clearer:


If you have data on circuits I’m happy to have a look for you and send you advice based on the data comparing your laps if you want and also show you if I’m faster than you anywhere if I’ve got data on that track too (will keep it private obviously).

[Magpie]

I'm yet to see if maths channels can be used in Circuit tools, however in Log works (Innovate) they are very powerful and it is possible to 'nest' formulas. The concern with using G forces for brake/accelerate is that they are (at least in my understanding) not as accurate as separate sensors rather indicative.

[mazmad]

Thanks Dan, I will have a look into that. It is always hard at the track to find the time to do it there and then - need to have a data guy there just for that! - but while I don't have the car running it's perfect homework for me! I seriously think a few of us should pitch in occasionally and have someone there to do some analysis and bash our heads and point out what we are doing wrong.

Turns out AIM just got it wrong and it isn't configured for the NC, it may be coming but like the wifi downloads - another misleading sell job - it could be a while, Italian time.

I will check out your analysis thanks. I had looked at throttle as I didn't have brake, for the same reason as you, and saw I was heistant in some spots so worked on either being full go or hard brake.

Fair point Mark. At this point better than nothing at least.

[Dan]

I'm yet to see if maths channels can be used in Circuit tools, however in Log works (Innovate) they are very powerful and it is possible to 'nest' formulas. The concern with using G forces for brake/accelerate is that they are (at least in my understanding) not as accurate as separate sensors rather indicative.

G Forces calculated from an internal accelerometer that is calibrated and is fixed (mine is on a mount which is drilled into the dash) should be accurate.

You can tell that AIM data I posted before is at least accurate in indicating braking by looking at the correlation between speed and the Longitudinal G's (Longitudinal_a in the data), whether the amount of G's is perfect doesn't worry me as it still indicates softer and harder braking.

A brake sensor would be better but I wouldn't say that It's not accurate.

Thanks Dan, I will have a look into that. It is always hard at the track to find the time to do it there and then - need to have a data guy there just for that! - but while I don't have the car running it's perfect homework for me! I seriously think a few of us should pitch in occasionally and have someone there to do some analysis and bash our heads and point out what we are doing wrong.

Turns out AIM just got it wrong and it isn't configured for the NC, it may be coming but like the wifi downloads - another misleading sell job - it could be a while, Italian time.

I will check out your analysis thanks. I had looked at throttle as I didn't have brake, for the same reason as you, and saw I was heistant in some spots so worked on either being full go or hard brake.

Fair point Mark. At this point better than nothing at least.

That's bad news about the brake and steering data!

Yeh, it's hard to find time at the track to do it and it would be awesome to have a data guy! I have pulled out my laptop a couple of times but I feel a bit self conscious about being perceived as a try-hard doing it at the track

[Magpie]

Dan I'm not saying it is not accurate, rather it is indicative. There are other forces that are acting on g force sensors.

This is a plot going into T6 at SMP South

at 0.49 Gear change almost a second between on off throttle (purple)
at 3.77 off throttle on to brake (green)
at about 5.00 bad rev match as I eased off on the brake (bad technique)
6.39 start turn in and trail braking
8.19 Apex (166kg weight transferred to right side - red line)
9.09 on gas, but only 30%
10.5 80% throttle and weight transferring to the left.

This shows a lot of areas for improvement and areas of coasting. The bad rev match caused the car not to slow and I know from experience that the brakes will take about another 100psi before lockup. Also the initial shape of the brake curve is not as steep as I normally have. So soft application, bad rev match all cost time. Then the period of nothing at the apex to getting on the throttle.

Hopefully I will get the steering sensor installed and start to get a better indication of turn in etc. However the rear suspension sensors (light blue dark blue) are very handy at indicating changes in weight transfer. AT 5.39 the sensors are starting to separate whereas before hand they were almost constant. Where they switch at 10.5 means the weight has gone from one side to the other.

The red line is lateral weight transfer Logworks formula is MC(WT;kg;)=(Sideforce*1147*.29)/1.65
Sideforce = lateral G channel
1147 = weight in KG
0.29 = centre of gravity
1.65 = width

Session 3 by Eipeip, on Flickr