ColorChange

Guys:

Here is what I am up to. I am talking with the people at AimSport in Italy. I am trying to convince them to write a link to our car’s CAN bus. If they do this, you can have all the instrumentation you reasonably need without having to hardwire all the different sensors (at costs well over $15k), including 4 wheel speeds, rpm, gear, lat g, long g, yaw rate, throttle position, steering angle, and probably all of the engine parameters also, like oil temp, oil pressure, intake charge temp, boost (for turbos), water temp, etc., etc., etc.

The AimSport DAS costs about $2,000 and you get an extremely complete, fully instrumented DAS that has video capability. The system should apply to all 996’s (and Boxters) with PSM, not the GT-3’s or GT-2’s. The recommended system is the MXL Pista.

Here is the link to AimSport http://www.aim-sportline.com/


Please indicate how serious you are interested by responding to this post.

924RACR

Very nice and well worth the effort! But I'm afraid I've got no computers on any of my own cars, so no real use for this... Oh well...

NetManiac

What about without PSM? And as long as the have software for OS X and/or Linux. Sure I would be into it.

ColorChange

I think you will have to have PSM to get all the driver stuff.

RJay

While I think it would be great if they did this, I suspect that the market is limited at the moment as most of the 996/Boxsters (admittedly not all) I see at DE are generally in the hands of novices. When most guys get serious, they move to a dedicated track car which typically is of an earlier vintage. As the cars age, I'm sure more will show up when 996s sink under $30K or so, but as a 996 owner (and one these guys would decided that a dedicated car was the way to go), I'm hoping that won't be for a while yet!

In your conversations with them, did they indicate they could do this? Not to sound too skeptical, but my understanding is that Porsche is pretty tight about their ECU as they (like Mercedes) apparently make a lot of money selling analysis tools to shops. I've only glanced at the CAN protocol, but I didn't see that there were standard definitions for the data crossing the wire, it looked like a pretty dumb transport protocol to me, which likely means a reverse engineering job to actually understand whats on the wire. I presume if CAN had standard definitions it would apply to a wide range of cars, it'd be easy, and likely be already done.

But hell, as a Pista owner I'd love to see them do it even if I can't take advantage of it today.

FixedWing

I've been told that Race-Tech is working on the same thing right now.

Curious which of the parameters you think really useful for the purpose of examining one's driving (vs. analysing the car's performance)?

For me, the list would include:

* RPM
* Throttle position

Lat/long g-loads are provided by the unit and not available on the CAN-bus.

Gear selection can be derived from RPM and speed.

What does yaw rate tell you? Ditto steering angle?

So that leaves only two inputs to be taken from the car (RPM and throttle position).

Of the two, RPM is the more critical. I haven’t looked into this enough to be definitive about it, but I think that RPM is very easy to get from the OBD-ii interface.

I’ve looked into it a little on another car and for that the throttle position is available on one of the inputs to the ECU. Is it not the same for the Porsche?

So although it would be nice to have it all served up in a CAN-bus interface, I wonder if it is really that big of a deal to get the data that is needed?

Stephen

RJay

If the sensor in question is analog, it can be read, digitized and logged by the logger (up to the number of channels it supports). As was discussed here recently in another thread, the throttle position sensor, for example, can just be read directly. The RPM, oil, water etc sensors can just be gotten the same way the instruments are getting them, via existing wires. The problem (and its not one for me) is that you use up your available channels when if you could just understand the ECU data, you could save the channels for more exotic purposes, like IR sensors for each tire or shock sensors, for example. Plus, as CC points out, the install is cheaper not to mention less invasive, which is a pretty important consideration to guys who track expensive street cars like CC.

If you saw the MS video from the BBC referenced in the recent thread opn DAS, I think you'd agree that throttle is certainly as important as anything from a driver improvement standpoint. Sadly the video had no brake trace which would have been extremely interesting, particularly if it showed any left foot braking. The vid also included a steering trace, which made me think that someday I'd like to add that sensor as its highly indicitive of how the driver is dealing with the current state of the car.

Things like speed sensors on all four wheels can tell you a bunch of interesting stuff, particularly brake lockup, drift etc. It would be very nice if you could buy a unit that would largely just plug and play with the car. The time and expensive of installing the logger, depending on how much info you wish to capture can easily exceed the price of the basic hardware. As the AIM will capture 64 digital channels, it would be a lot easier, and cheaper if it could read the digital data rather than the sensors; you could pick and choose what you want to look at from there.

ColorChange

Rjay: Nice job. One real neat use of steeering angle is to plot steering angle versus lat g's. This gives a good indication of how you are handling the car with regard to oversteer/understeer and set-up changes.

924RACR

The only parameters listed above that are PSM-specific are: lateral g's, yaw rate, and steering angle. Perhaps brake pressure too, though I wouldn't expect that it's necessarily sent out on the bus. The rest should be available even without PSM, and I would expect be available (if on the bus) from the ABS ECU and the engine/trans ECU(s). Then it's just the question of CAN id's.

CAN id's tend to be fairly standard within manufacturers - after all, why reinvent the wheels. I expect that different Porsches of the same vintage would have very similar if not identical CAN message matrices. Of course looking at a GM, or Ford, or whatever, in contrast would be completely confusing, I'd expect it to be totally different.

If AIM's really wise about this, they'll develop the technology to tap in to ANY CAN bus - and just leave the message list open to editing/adjustment by the end user. Or maybe they already have that - I don't know - and all this is really looking for is a new AIM unit that will rely virtually entirely on that?

CC's absolutely right about the usage of steering angle, yaw rate, and lateral g's - this is what the entire PSM algorithm is based around, in addition to wheel speeds and such. You'd simply mimic some of those calculations in your measurement/data display, and review.

RJay

Yeah, that is a very interesting use. More fuel for the steering wheel sensor fire. Another thing I've been musing about (perhaps you are too as I see you have the Motons) is that as I'm thinking about installing a set of Ledas or JRZs this offseason, I'm wondering if dedicating a channel at opposite corners to shocks might not be worth while. Not sure I'd know what I was looking at initially, but it could prove useful. When I was at VIR my car was starting to wander a bit on the back straight, it would have be great to see if the shocks were contributing to the obvious aero deficiencies (man, my old clunker is exciting at a buck forty-five) and have the data and adjustability to try to tune it out.

Perhaps this is the thing I'm most excited about with the installation of my unit is the ability to go out to a test and tune day, make a few set up changes and then see the precise effect on the car. I'm sure the guessing about setup will never entirely go away, but at least the results will always be precisely measured assuming I can develop some consistency. Even if I'm a lumox and don't improve at least the car will

RJay

Cool! So if you know the codes you in.

Fully agree, it would be very wise in any case as who knows how often IDs might change from model to model or manufacturer update to update. Are they rommed or are they assigned dynamically?

The other cool possibility I would look into if I were running AIM is the possibility of OEMing the logger to BMW or Porsche who would then intergrate into the car and make it available as a factory option at least for RS/CS and CSL kind of models. Who they hell wants a GPS when you could use that center console screen for lap times eh? I refuse to pay $2k to know where I am, but to figure out how fast I managed to get there...priceless.

BBailey

I've already told them I want it as soon as its available.

fatbillybob

I am extremely interested in this. I have heard of this possibility. What exactly is the CAN bus? Perhaps it would not be too difficult to get an engineer buddy to moonlight make us a connector? What is involved?

924RACR

The CAN bus is the car's communications network for ECU's - think Car LAN, you'll have the right concept. All the messages between the various computers in the car pass on this network. But more specifically, CAN is a particular architecture for that network (standard originally developed by Bosch, as I understand); there are other architectures as well (which don't typically work quite as well), GM's ClassII being but one example. Some vehicles also can have not one but two CAN busses, completely separate networks.

The really fun stuff happens when you've got two separate and incompatible networks in one car. Kinda like having a Mac and a PC network in one building! LOL

FixedWing

I wonder how useful this data really is?

From my perspective, there are two seperate uses for the data. The first is in analysing the driver and the decisions he makes. The second is to analyse the car and how well the car is tuned for conditions.

I believe that much of the data people are talking about seems to relate to tuning the car and not to the driver. I think that is true for the steering angle.

I think that there is a real risk in being burried in data. From the driver's perspective, only a very little information related to the decisions he is making are needed.

Whether the driver is setting the car up correctly for a corner is probably something best judged by the seat of the pants. This should be obvious. And I think it can be judged much better this way than through an analysis of data after the fact.

Stephen