Datalogging RPM signal

[Jari Porhio]

 Sorry to bring up such an old topic, but I haven't found time to read
the DIY-EFI list recently, just happened to see this topic.

> Date: Fri, 21 Dec 2001 10:20:54 -0300
> From: "Tomas J. Sokorai Sch." <tsokorai at xperts.cl>
> Subject: Re: Datalogging RPM signal
> 
> Hi, I'm writing an opensource program to plot HP curves and monitor/log 
> engine RPMs. I'm using a voltage divider to pick the 12V signal from the coil 
> input. The input is recorded with my soundcard using the Line input.
> I'm having troubles with trying to figure out how to count the spikes, 
> because the signal I get is VERY far from a square wave :). And with my 
> ignition, when the RPM goes up, the signal amplitude goes DOWN ....so no 
> static treshold for counting spikes :/.
> The input for the HP measurement I'm using is the tire diameter, gearbox 
> ratio, differencial ratio, vehicle weight, nr. of cylinders (for calculating 
> each 360 deg. engine turn)
> The formulas I'm working with are based on converting the engine rotations in 
> distance traveled. Then I apply the basic motion formulas to calculate speed, 
> acceleration, force, work and power.
> I'm calculating the drag from the desacceleration... I know this isn't 
> completely right, but I don't know the Cd of my truck, so I wanted to have a 
> rough estimate.
> My program is in it's very beggining stages, anyone wants to participate in 
> such project?.  It's being programmed under Linux with KDE gui.
> 
> 
> Tomas Sokorai.

 Hi Tomas,

 Have you seen my rather rough try at

http://www.iki.fi/eppu/bin/rev.html

 This is exactly what you're doing, only that those programs are
mainly command line tools with unbelievably poor user friendliness.

 The revolution detection program goes through the signal to find the
spikes, and done like it is, it's not very easy. If I had to do this
again, I'd pay more attention to hardware signal conditioning before
converting the signal to digital.

 There are of course some advantages in a set-up like this. You can
directly use a very simple ignition pick-up, like a standard timing
lamp pick-up. And the signal gives more information than just the rev
count, like spark burn time and coil charging time.

 So, do use the rev kit if you need it, the sources are there. I don't
have the time to polish and improve the kit at the moment. It won't
probably compile right out of the box, and is basically a handful to
use.

 But the results are quite nice, see my revtest archives (see the
tests calculated using rev 1.30 and later). The equations used to
calculate the engine torque have quite a long history, so there may be
some inaccuracies in them. But for comparative work, like checking how
your engine modifications work, it's a cruel tool.

 Mike Glover uses some core parts of the rev kit in his Windows-based
StreetDyno too.

 In the future I'd like to do it all again using a slighly different
approach. Calculating the vehicle speed from engine revs through all
those different drivetrain ratios etc. is a very error-prone method.

 I'd like to use a method where the vehicle speed can be extracted
directly using for example a radar or some other signal available on
road (GPS, others...). Commercial systems usually use an acceleration
sensor, but for what I've seen, they have their own problems with
accuracy and signal noise. Note that I'm not saying they don't work.

 This system would probabaly be based on the PDA/Windows CE platform,
or other ready-made platform like that. Which should allow an easy
porting to laptops/tablet PCs.

 I don't have the time to start this now, but if anyone is interested
in participating, let me know. OR, IF ANYONE HAS INFORMATION ON
EASY-TO-USE RADAR MODULES, please tell me. I think Siemens did have a
module like this, but I don't know about its availability today.

_______________________________________________________________________
 Jari Porhio   http://www.iki.fi/eppu/

 NOTE: In the future I can be reached by e-mail at jari.porhio at iki.fi
       (eppu at iki.fi will also do). My accounts in domains *.tut.fi,
       *.samplerate.* and *.flextronics.com may/will expire shortly.
----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org