Sun Tach - How accurate - How to check?

[XH29N0G]

All - 

I have been working on an Arduino based (economical) way to monitor things like A/F (I have an A/F meter), RPM, Manifold vacuum (I bought a little sensor), acceleration (also bought a sensor), and maybe some other things that I can tap in as 0-5 volts, and time.  I'm figuring it will provide a lot of information for pretty cheap.  The issue is whether I can verify that the information is accurate enough to be useful.

My Dilemma:  I'm getting mismatches of up to 15% between my sun tach, the Arduino-based measurement, and a measurement made on an old Craftsman engine analyzer. I have not yet gone through the math of speed vs RPM to check any of these and suspect this is the way to proceed, but I am wondering if some of the collective knowledge can help because I do not know whether there are biases in my Arduino based measurements or if my sun tach is correct.  It is encouraging that the Arduino gives a reading in between the Sun and the craftsman measurements. 

First question: Any thoughts on the accuracy and biases of the sun tach measurements?
Second question:  Any comments on how I am making my Arduino based measurements?  (I am measuring the time in milliseconds between peaks using the PulseInLong function and then accounting for the fact that is it a square wave and the duty cycle is 22.5% for the MSD 6AL tach output (so I use .775))  I am looking for biases or problems I might have since I really am a novice at this.

Thanks.

I'll keep people posted if this works. 

[John_Kunkel]

"A man with a watch knows what time it is, a man with two watches is never sure."

[XH29N0G]

SO I have 4 watches.  two I bought and two I made. 

I think the sun tach is reading high.  To be fair, I pulled the sun tach from its case and installed it in the dash.

I think the next test will be a drive where I mark speed and compare with RPM.

I'll keep people updated.

I just thought maybe there would be someone who would say they had experience.

[XH29N0G]

I am making progress and don't have it foolproof, but I was able to log the RPM and confirm they were correct.  I also managed to capture A/F and vacuum.

Here is what I have so far.  My tach numbers had to be cleaned because I am measuring peak to peak time and sometimes this drops out to give a very short (high RPM) flier. 

But overall, I am pleased with the progress and seeing everything line up.  This is just pushing the pedal down in first and then letting up.  The vacuum rises and A/F drop as RPM rise and then do the opposite when I let off the gas. 

Now I need to replace tires and try it with the full run of gears.

[XH29N0G]

Some more tests and an update.

I decided to add the pulse times on adjacent high and low pulses to get an average pulse length with out assuming the duty cycle.  This seems to work a little better, but it still gives some pulses that are too short by a lot.  I am now filtering out the short pulses.  I may try an alternative function.