Hey Chad,
Here's a link to Autometer's new Data loggers and their best system "Pro System" is only 500hz which is half of what the RacePak is and they are also using the CAN bus system.
http://autometer.com/datasystems.aspx
I'm replacing this http://autometer.com/cat_data_tachdetail.aspx?pid=10 with the RacePak system and I've used it for about 10 years with a two magnetic collar and its only 100hz so I think the RacePak with a 1000Hz and an 8 mag collar should be a whole lot better.
Like Jay mentioned. 90% of what I'm looking at is in the first second of the run and the rest is at the gear changes which all take place in the first half of the track. After that the clutch is locked up and we're just on cruise control from that point on.
If the Racepak or whatever satisfies your requirements and is at an acceptable price, then by all means use it. I was just trying to explain how you could be limited. Certainly my own bias was revealed, my apologies if you were insulted. I think you might have missed where I said 'It does offer a nice one stop shopping package, though, and may be quite appropriate for your application.'
Racepak may work fine... or Autometer, or any other brand. My main gripe with Racepak is how they are advertised as being the latestgreatestomfgdatawhutsit when they are not, and that the price tag is quite shocking for what you get. Personally if the price tag is going to be that high, I'd start with Pi
http://www.veracitydata.com/NewFiles/PiProductLine.html, but the packaging and setup may not be in line with what you expect. Perhaps piece your own together from parts at
http://www.mccdaq.com/, but I'm guessing you want a packaged solution. Innovate also offers excellent products. Many product similarities exist between brands, simply because they're often rebranded reference designs at extremely different price points, just like many other electronic products.
Personally, I find the resolution of Racepak to be wanting, but I'm pretty damned picky, and accustomed to doing data analysis in the >=12 bit range, and 10KHz/channel and higher sampling rate. Seeing the rise and fall (and rate of change) of sensor values from nominal to maximum is useful to me. Knowing the event period and scale is an important part of deciding how to measure. High resolution does make data storage and transmission path considerations a part of the package, some el cheapo off the shelf stuff simply doesn't play along. The principle I live by is that you can always mathematically smooth or compress a data set, but you can't just 'imagine' where dips, spikes, or frequency changes may exist; you must measure them.
For example, my cousin and I are trying to sort the clutch on his 9000 RPM bracket car. The eight magnet driveshaft collar wasn't enough, so we're changing the sensor to a 60 tooth trigger wheel and moving it to the input shaft, to give us a better picture of the relationship between the flywheel and the input shaft when the clutch is released at the line. His 'box' (a used Pi Indycar piece from Menard) is quite capable of handling the data stream, where a Racepak would not. It's overkill, but the parts were on hand and offered the necessary resolution.
