Paul Foster <pfoster@gdi.net>
> From your comments it is obvious you are more comfortable with hardware
> than software.
*chuckle*
I write software for a living. I'm not an engineer, I just play one on TV. :)
> Any product that requires you to go inside the "black
> box" to enable the other accelerometer is way too difficult for most end
> users.
Ever set the IRQ and data port on an ISA modem? It's not THAT uncommon.
Edelbrock is marketing their system primarily to drag racers, and I guess they
feel that the other accelerometer isn't needed for drag racers. The second
accelerometer is disabled internally to free up another analog channel.
If you're technically capable of wrangling the wiring harness that the product
needs, then you're technically capable of undoing 2 screws and moving a jumper.
> I have the opposite perspective since I am a software developer
> but I do have a background in realtime computer systems. So I have a
> very great appreciation for what Byron has done with GEEZ. The analysis
> software is the key.
But with all due respect to Byron and his efforts - garbage in, garbage out.
He's only got 2 inputs to work with, and they're both subject to error. He makes
all these wonderful calculations for you, but they're still based on
error-filled channels.
Not that the Edelbrock system is immune to these errors either, but with the
extra data channels, you at least get the opportunity to rectify them.
For example, there's a section on one of my traces that shows a 1.3G braking
zone over .68 seconds. However, in that same section, the speed line drops
15MPH. Doing a little math yields 15MPH=22ft/sec, and a change of 22ft/sec in
.68 secs is an acceleration of 32.4 ft/sec^2 - or about 1G.
That's roughly a 33% error from what I assume would be pitch/roll, and any
incline of the racing surface.
> I personally believe that Edelbrock will NEVER
> match the software you get with GEEZ. Not next year, not next decade,
> never!
Based on what?
GEEZ! does a lot of computed channels, but the Edelbrock guys tell me that the
ability to do computed channels in their software is coming. All you need to do
is know the math involved, and then you can define the same channels in the
Edelbrock software.
What's more, for some channels at least, the Edelbrock beats GEEZ! as it stands
- the RPM and speed traces in GEEZ! are calculated from the accelerometers, not
measured values like the Edelbrock. Any errors induced in the accelerometers
will thus be propagated into anything calculated from them.
I suspect that the GEEZ! data is "good enough" for many purposes, and Byron puts
a lot of sanity checking into his calculated values to try and keep the errors
in line, but I like the ability to look at a speed or RPM trace and trust the
value.
> Fanstastic is not the word for these figure. Bogus would be much more
> accurate. As an engineer who works for a car vendor you should know
> that.
Not being an engineer, I hadn't realized that pitch/roll errors could so wind up
so large. I now know better. Thankfully, the I wasn't using the accelerometer
values for anything other than conversation pieces, so the error in values
didn't comprimise my analysis.
But I DO know software - and I like Edelbrock's more than I like GEEZ!. YMMV.
DG
|