Data acquisition usefulness

These questions go out to all those of you who love their data-aq systems and credit them for gaining seconds.

Background:

I am interested mostly in monitoring engine functions (specifically EGT/RPM/Throttle position) in order to quickly optimize proper jetting instead of chasing it all weekend long, finally getting it pretty damn close on the Sunday warm-up. I also want to optimize tire size and gear selection without having to ensure that I glanced at the tach at the exact same point before braking everytime. I also like the fact that I can get a tach with programable shift lights and lap timer built in.

My questions are more towards the driver evaluation and fastest line:

The GPS based systems claim accuracy of 9'. How can one evaluate a line, braking points, exit speeds, segment times if the data can be skewed by +/- 9 feet. I am looking for .1's of a second, not seconds (although that would be great).

Are the systems that generate track maps based on math and accelerometers more/less/same accurate? Surely, they have tolerances as well.

If I utilize a wheel speed sensor vs. time to determine position and I take a longer line through a corner that results in a higher average speed through the corner but the EXACT same amount of time through that corner...doesn't the system think I am now further from the starting line and therefore have gained time vs. position?

Lastly, how accurate do you believe the systems that have theoretical best laps are? I understand these are based on the addition of your best segment times. Again, if the segment start/finishes can be off by 9' each they could be dead on, or 18' feet off. Where is the repeatability there? Also, I can create a 'faster' segment time if my goal is to only go faster in that segment. With no concern about what that blast may do to the entrance to the next segment. Do this to each segment over a series of laps and the system can predict a lap that will never be attainable.

Am I better off getting a basic system that will give me lap times/rpms and egt and consider all the other stuff "fluff" unless I spring for one of those pro systems that cost more than my car (out of the question).

Thanks,

I'm on the steep part of the learning curve with my DL-1 but here's what I think I know at this point...

The system resolves both the GPS and internal accelerometer inputs and, if the track trace is to be believed, it is accurate within a couple feet. I could identify the lap that I dropped the left front tire off the outside of Oak Tree at VIR, as opposed to the ones where I just rode the edge.

I took my car for a drive on the street to do the coast-down tests necessary to program the system that computes engine power and was putting some big wobs into the wheel to see how the rain tires' grip peaked: I could SEE the lateral acceleration trace get fuzzy where I put in too much steering and the cold front tires scrubbed the pavement rather than just gripping.

I'm not entirely convinced about the "best theoretical lap" thing simply because, depending on where your segment breaks are, you might have to give up more in an adjoining segment, that you gained by optimizing just one.

As I said, I have a lot to learn - both about the system and my driving - but I am absolutly convinced that data helped pretty much all of the drivers when we did the VIR enduro.

K

I'm with Kirk--- My data aq system has the display with tach, all the data displays, which are configurable, and the ability to time laps, segments, etc.
I agree that the whole track mapping thing seems a little spooky, so the jury is out on that.

But...(and I just got mine barely up and running for the ARRCs), just the display with the lap timing function, configurable tach, and data displays are easily worth the price. Price a tach and lap timing system of equal capability and the total is nearly the same as the whole data aq system.

I really think it's a no brainer for the empirical data you will get instantly, which makes your track time SO much more useful. The neat things these devices are capable of is endless. Want to know the actual charge temp of your cold air induction system? No problem! Or if it is restricting airflow at high rpms? Easy!

Like anything there is a learning curve, but I wish I had done it long ago....

(The one downside is that my system needs a laptop for full functionality, but these days, there are lots of options there as well)

Check out the other threads for some good info as well...

------------------
Jake Gulick
CarriageHouse Motorsports
ITA 57 RX-7
New England Region
[email protected]

[This message has been edited by lateapex911 (edited February 07, 2005).]

Kirk is right, the typical performance of the GPS systems is better than the advertised performance - I guess we need better marketing people!

Having said that, I generally tell people that the speeds, times and G forces are the main things to look at (in terms of the whole vehicle variables). The map is mainly to provide context for the rest of the data, and it does that very well.

The absolute position error of the GPS data is mostly a slow drift, and really doesn't seem to affect the time and speed data measurably. It does mean you can't use your data system for accurate surveying though!

Comparing DL1 lap times to AMB transponder times, I've generally seen agreement within one to several hundredths - certainly better than the tenth you mention. And the segment times are much better than beacon/wheelspeed systems (assuming 1 beacon per lap, the pros often use many beacons) because of the reason you mention, plus tire slip (big factor) anf tire growth.

And I actually think that the "best theoretical lap" function is pretty realistic - usually. Yes, it's certainly possible to screw up one segment to benefit another - but with thoughtful placement of segment markers and not using too fine a segment division, this can really be kept to a minimum.

I'm biased in favor of the GPS approach, but really you should find any of the full-featured data systems a valuable tool.

Al Seim
Action Digital Race Data Systems
www.actdigital.com

Al and others. Thanks for the responses.

For now, I guess I'll go without the GPS or wheel speed track mapping features.

Per the Action Digital website FAQ's:

"How accurately is position measured?
With good GPS reception, positional accuracy is about 3m (CEP)."

My "research" would indicate that both methods (GPS/Wheel Speed) have severe limitations. GPS accuracy and wheel speed variations due to wheel slip or lock up/spin, change in diameter with speed/load. Change in line resulting in 20 extra tire revolutions over a 3 mile course...etc. They are great at providing you with a pretty reference map in order to help communicate approximately where a certain habit/problem is occuring. Much easier to work with than "2786 ft from start/finish" or "23.7 seconds from start/finish". But, if the goal is to accurately see the effects of minor changes in line or inconsistancies in driving then they are only valuable if your lines vary enough that you have other issues to worry about.

I was also looking for a steering wheel position sensor to evaluate my smoothness or lack thereof. Here the limiting factor would be the collective amount of slop in the steering (between the steering wheel and tie rod end). I get about 5* of steering input from straight ahead before there is any movement in the tires. So, I can be slick smooth and have a staircase for a steering wheel trace. Or I can be sawing away at the wheel and it might look much smoother. How much of it is kickback in the wheel? I guess I could put a rotary pot on the wheel and a linear pot on the tie rod and look at BOTH traces and see what is really happening? Or look at the lat g's for a nice smooth steep ramp up/down with very little correction in the transistional areas and nice, high, long peak areas.

Are the sampling rates of different channels on the DL1 programable? For example if I want RPMS and EGT at 10HZ, but oil temp at 1HZ is that doable?

I don't know the answer to the sampling question but, for what it's worth...

http://it2.evaluand.com/downloads/oaktreeoff.jpg

K

Quote:

---

Originally posted by Knestis:
I don't know the answer to the sampling question but, for what it's worth...

---

(Kirk, you sure that's not just bad driving...<grin>?)

Conversely, from that same car, same weekend, same driver, same session. Data aq was turned on, car left the hot pits, drove around like a madman a bit, then returned to the same slot along the wall:

http://www.kakashiracing.com/temp/vir1.jpg

Another one, even more telling. Car turned on in paddock spot just behind the pitlane spectator chainlink fence area, drove to a break in the wall mid-pit lane to the interim cold pit lane, drove out and around the track like a madman, and returned to hot pit lane slot:

http://www.kakashiracing.com/temp/vir2.jpg

That's a LOT better than 3m resolution, folks...



[This message has been edited by GregAmy (edited February 07, 2005).]

Quote:

---

Originally posted by GregAmy:
(Kirk, you sure that's not just bad driving...

---

Hey, now - that's no way to talk to the boss, Massimiliano. I'll put you in the car after "l'incidente" and make you ride in "la sede in pieno dei feces."

That's "poopy seat" to you and me, Rusty.

K

Daryl:

Off the top of my head, I'm pretty sure that the sample rate for all the data channels is set at once. But having said that, using CompactFlash for data storage means that storage capacity isn't an issue, so the "extra" data really doesn't hurt you.

And not to belabor the point, but please don't think that you need 6 inch positional accuracy to make track mapping, along with speed/time/g force graphing, a valuable driver development tool. Ask Greg and Kirk to explain!

Best Regards,

Al Seim
Action Digital Race Data Systems
www.actdigital.com

[This message has been edited by Al Seim (edited February 07, 2005).]

Kirk,

Looking at the top traces it looks like your lines (throwing out the one that is way off) varied by 5m (about 15'). When not racing/dealing with traffic issues I would find it hard to believe that you would find a variance of more than 3'.

Again, from a "how consistant am I?", "and how much can I gain from this line/that line" the track mapping seems more gizmo factor than valuable tool.

There seems to be a reason that the pro-level systems rely on several beacons for accurate info.

I am not debating the merits of GPS vs. wheel speed sensors, as both appear to be major compromises.

Of course, my wallet is the biggest compromise.

For clarification: I wasn't looking for lap time accuracy to the tenth. I am looking for lap time improvements in tenths.

Maybe I'll take my GPS out of my Jeep and stick it in my driving suit and then import the data http://Forum.ImprovedTouring.com/it/wink.gif.

IMHO you are way off if you are looking at the track maps to find the line or more speed. (and you last email indicates this but you get to read my opinion anyway http://Forum.ImprovedTouring.com/it/wink.gif The only reason to monitor the track map is to sort out what corner you are looking at when you are examining the speed vs. distance, (or speed vs. time) and the lateral g graphs. When you study the X Y format of the speed or G graphs it is easy to get lost were you are on the track. This is especially true on a long track with lots of corners and when learning how to read the data. So you can place a marker on the X/Y data and then view the track map to determine where this data is. The track map is a cherry but it's not the reason you order the desert..... http://Forum.ImprovedTouring.com/it/wink.gif

I have a sting pot hooked up on the steering and a simple digital input on the brake. To be honest you don't "need" it. It can be useful but most of the time the extra data is overwhelming. This is especially true between session on a race weekend if you are driving, wrenching, and trying to work with the data acq. Steering input, brake and throttle application are all visible on the lateral and long. G graphs.

So for what it is worth I would pick with GPS but either system draws a track map with limited accuracy. The data in X/Y graphs are more than accurate enough to evaluate line and smoothness.

Dave

Quote:

---

<font face=\"Verdana, Arial\" size=\"2\">...don't think that you need 6 inch positional accuracy to make track mapping, along with speed/time/g force graphing, a valuable driver development tool.</font>

---

No, you don't. The value of the GPS data (and this is what I think Dave just said) is that it is used as a basis for comparison. I use the GPS data to create the track map, so that I can set my timing points along consistent areas (mid-straights), so that I can use those waypoints for comparison within the data itself lap-to-lap. I *do not* use the GPS/track map for determining racing lines, I *do* use the GPS data to determine to where the other data correlates.

The data we used at VIR in October was only that provided by the unit itself (GPS position, two-axis accelerometer, calculations). For 2005 I will be adding a lot of engine/chassis-related inputs, such as brake application, throttle position, RPM, and maybe some other engine parameters.

In addition, I plan to write a series of articles - with pretty pictures - on my experience and use with data acquisition. Al has agreed to proofread what I'm writing for technical accuracy, and I'll make that info available either to a periodical for publication or on a web site. My goal will be to take some of the voodoo out of data acquisition and make the information useful to everyone.

I'm of the position that while data acquisition today is a new game and a competitive advantage, "tomorrow" (probably sooner than you think) it will be yet another required tool for any successful club racer (like the tire pressure gauge, tire pyrometer, and camber gauge...) - GA

Greg

Quote:

---

<font face=\"Verdana, Arial\" size=\"2\">I am interested mostly in monitoring engine functions...in order to quickly optimize proper jetting...</font>

---

Daryl, we've decided that using a chassis data acquisition system is not the optimal way to tune the engine. First, as others have pointed out, there's way too much data involved; it would be overwhelming without someone devoted to it full time. Second, that are too many other variables involved in a racing/practice/qualifying situation such that you would be chasing your tail trying to dial in your engine setup.

Instead, we are working on a data acquisition system for our car specifically designed for testing and tuning on the dyno. It would be illegal for a race wekeend, as it does not conform to the existing ECU rules, but it allow us to optimize the engine itself absent these other factors. Once done, fine-tuning can be done at the track.

We'll monitor some basic engine parameters with the chassis data acq system to verify our assumptions, but I would caution you against any major engine changes based on data obtained during a race weekend. - GA

I sure didn't mean to suggest that position data were going to help improve one's line. That was just the substance - position accuracy - of the original question.

Oak Tree is not particularly wide - you could block it pretty effectively by putting a car sideways at the apex - so it seems unlikely that there's 15' of distance between the left- and right-most traces. Because it's so slow, I was willing to let faster cars by at the apex, figuring that on a percentage basis, I was giving away little time by doing so, so I suppose there might have been...

To put a scale on that image, there was "in the real world" probably 4 feet of total variance in the car's left-to-right position where the corner ends and the traces leave the image. I never exited the corner on the far driver's right, I don't think.

The most interesting data for me are the acceleration traces. While peak lateral G's are not particularly important - again, back in the real world - the shape of the curve for each turn is VERY interesting.

In comparing my data with Gregs, for example,
his slope on turn-in was both steeper and smoother than mine. Looking at in-car video, his apex was also consistently earlier so I'm losing time on both counts. He proved that I could be more assertive with the wheel AND turn in earlier without running out of road at the exit.

(That's why he's Max and I'm Gianpiero.)

That said, it isn't necessary to even have comparison data, once one knows what to look for and what the accelerative capacities of the car look like.

K

Buddy Fey's book "Using Race Car Data Acquisition" is a must own for anyone who is contemplating using data acquisition to improve their driving. Some of the info is a bit dated (mostly the hardware and computer stuff) but the analysis of the graphs is outstanding.

As Kirk and Greg pointed out earlier, it's the SHAPE of the charts that give key information, and once you understand the basics of tea-leaf, er, chart reading you'll be able to spend countless hours poring over your own data, looking for those elusive seconds! http://Forum.ImprovedTouring.com/it/biggrin.gif

Greg, I'm looking forward to your book!

MC

------------------
Mark Coffin
#14 GP BSI Racing/Action Digital/Airborn Coatings/Krispy Kreme VW Scirocco
http://pages.prodigy.net/Scirocco14gp

[This message has been edited by racer14itc (edited February 08, 2005).]

I had been looking for Buddy Fey's book a couple of years back and understood it was out of print. Anyone have a copy they wouldn't mind parting with?

Kirk, the scale on your plots (IF I am reading it correctly) is at the bottom corner indicating 1" = 10m on my screen resolution.

The purpose of the track maps appear to be as I was thinking (alluded to in my second post).

It will be interesting to see the shapes of the curves on a Vee. It doesn't accelerate worth a stink, couldn't outbrake my motorhome, so the curves could be extremely flat on that axis. Grip is pretty good, transitional response isn't so bad either.

I understand, and appreciate the advice regarding too much data. It would be neat to 'see' some of what is happening but one has to know what to do, and the rules must allow the means to correct these things. I am a numbers guy and one of the most analytical people I know. But sometimes you just need to drive and not concern yourself with things you can't control.

As far as the EGT goes. Once I get the jetting right, it will probably change very little from track to track. The six tracks range from almost sea level to 4000'+. January at one place might be 40* (it is CA) and August at another might be 110*. So there will be a need to change main/air bleeds a couple of sizes throughout the season. The sooner I can optimize that, the sooner I can concentrate on other things.

The simplest thing would be to add a lap timer and an EGT to my already existing plethora of gauges. But, I too believe DA will be a neccessary tool before too long.

There are much more capable systems for engine funtion monitoring, like the Racepack system on my Brother-in-law's Top Alcohol Flat bottom drag boat. But I don't need my EGT/Fuel Pressure/Boost/rpm/prop rpm speeds, etc sampled @ 1000HZ.

I don't need yaw/pitch/roll axis/damper rates/laser ride height/IR tire temps or even gear position for that matter. Just looking to keep it simple, but have usefull data to learn from.

Quote:

---

In addition, I plan to write a series of articles - with pretty pictures - on my experience and use with data acquisition. Al has agreed to proofread what I&#39;m writing for technical accuracy, and I&#39;ll make that info available either to a periodical for publication or on a web site. My goal will be to take some of the voodoo out of data acquisition and make the information useful to everyone.

Greg
[/b]

---

A little over 2 years later, have you written a book? :)

Damn! I thought I deleted that post...

;)

No, unfortunately. I tried to shop the idea around to a few places and magazines to get some reimbursement (and motivation) and it never materialized. Other priorities prevailed...as an alternative, I&#39;ve also tried to shop out on-track services for driver coaching and data aq services, and that, too, didn&#39;t really materialize.

So, your best bet is to show up at a race where I am and sit down with me and the other folks around us and learn it by doing...!

Greg

These may be helpful...
http://www.waterfordhills.com/dapres.ppt

And the notes:
http://www.waterfordhills.com/dapresnotes.doc

Quote:

---

No, unfortunately. I tried to shop the idea around to a few places and magazines to get some reimbursement (and motivation) and it never materialized.[/b]

---

Reimbursement? :( It&#39;s called a big bank loan. I have yet to learn how to effectively use the DL1 data acq. info., but I&#39;ve already looked into this. I would love to see a book out there that describes how to actually use the data gained. Greg, save me the expense of doing the "Go Ahead - Use Data Acquisition" book. :026:

Quote:

---

doing the "Go Ahead - Use Data Acquisition" book.
[/b]

---

I&#39;ll help! :biggrinsanta:

I&#39;ve been reading the McBeath book and it really needs updating, plus we could a DVD with it with in-car to help people relate what they are seeing in the data to the on-track.

I found Simon McBeath&#39;s book to be "OK", but really nothing more than the obvious. What really needs to be written is something more advanced, along the lines of the stuff you and I discussed in detail last weekend, the same kinda stuff I do with others at the track.

Problem is, that takes time. And, unless I can justify writing a book as a way to attract sponsorship....hmmm....(sounds of gears meshing...and not stripping...)

;)

Quote:

---

I found Simon McBeath&#39;s book to be "OK", but really nothing more than the obvious. What really needs to be written is something more advanced, along the lines of the stuff you and I discussed in detail last weekend, the same kinda stuff I do with others at the track.

Problem is, that takes time. And, unless I can justify writing a book as a way to attract sponsorship....hmmm....(sounds of gears meshing...and not stripping...)

;)
[/b]

---

Sorry to add another level of complexity to your life ;)

I was bored and reading up on DAQs, I like what I saw with the DL1, read a bunch of old posts, came accross yours, and decided to call you out :)

I figure I&#39;ll end up with a DL1, just not this year or next. $$$ is going elsewhere. Maybe late next year, I&#39;ll need the help, but probably &#39;09.

Quote:

---

What really needs to be written is something more advanced[/b]

---

Actually, my thought was that something needs to be written starts with the basics for people like myself can understand. I&#39;ve looked at a few DA books, and just couldn&#39;t grasp it. Once those topics are covered, then go on to discuss the more advanced stuff.

Quote:

---

Problem is, that takes time.[/b]

---

Very, very true. Writing it is the easy part.

Dave and Jeremy, let&#39;s discuss this. I suggest we could leverage the "Go Ahead..." branding, our writing skills, and knowledge into something useful...

Damn you, Rob... ;)

Quote:

---

Dave and Jeremy, let&#39;s discuss this. I suggest we could leverage the "Go Ahead..." branding, our writing skills, and knowledge into something useful...

Damn you, Rob... ;)
[/b]

---

Glad I could help. ;) Just remember who helped drive this when your done, or don&#39;t depending on how well it does :) LOL