I have a question about using this sensor as a main reference to nail some glitchs under the bonnet.

I need to get rpm from the crankshaft sensor via math channel. As far as I know we have to use the math fuction freq(channel) to acomplish this but all I get is a flat line in my scope ( 4423 and 6.9.14.16).

The fuction I am using is: (freq(A)*60)/60 of course this is equivalent to freq(A) but anyway all I get is a flat line in the upper side of the screen.

I was been told that it is useful the derivative of the signal and try to get the peaks on the derivative to get the time of each falling edge on the signal and knowing the crank angle between the falling edges, do the calculations to convert time to RPM and then filter the signal missing points. Any suggestion?
I think this is the strategy TiePie guys use in their program

It would be nice being able to calculate RPM many times for only one crankshaft revolution so that we will be able to see individual changes as long as every single cylinder contributes to engine power. I think the current software version is able to do it but I don't know how to do it yet!

Another issue is about the trigger for this signal. I used to use the "interval" type and condition "greater than", so I noticed than sometimes works. So I would like to know what is the best trigger for the crankshaft position sensor inductive type as well.

Thanks in advance

Dario

Last edited by sigoaprendiendo on Thu Sep 18, 2014 6:55 pm, edited 1 time in total.

First try without the brackets in your math function - freq(A)*60/60. Secondly - have you specified a RPM table in your function, lets say from 0 RPM to 6000 RPM.

Look at this waveform, which are taken from a VW 1.9 TDI.

You are right, it is 60-2 teeth case. I will take your advice and try again. One mistake a made is to limit the the RPM in my table function to 1000 RPM, now I know why a flat line appears in the upper side of the screen.

I took a look to your wave form it is quite right, but I mean something more accurate like that, I know that with Pico we can do it. Check this shot so lets see how we (I mean all of you guys in this forum) can figure out how to "improve it" as Autonerdz have. Lets see this captures from Autonerdz

I downloaded this pictures from Autonerdz, by the way

Attachments

As Mr. Tom Roberts stated, such a kind a detail was the key to solve the problem with this car as the dropping down in the math channel doesn't follow the crank signal voltage pattern

As Mr. Tom Roberts stated, such a kind a detail was the key to solve the problem with this car as the dropping down in the math channel doesn't follow the crank signal voltage pattern

For more detail from your maths channel increase the number of samples across the screen as with the Autonerdz guys.

1MS is a good "all round" one size fits all for most automotive sensors. This now allows for greater detail in your maths channel hopefully revealing the change in crankshaft speed during each revolution.

When setting the "range and units" for your maths channel" for Dario 0-1000 rpm was too low if you need to analyse engine speeds above 1000 rpm (Hence the flat line at 1000 rpm and above )

For Kim 0-6000 RPM covers the whole engine range but with reduced detail in specific engine speed ranges.

With Kim's settings you could increase the scaling factor at the bottom of the scale to see more detail within your maths channel

Alternatively, should your running issue be relevant at idle speed only set your "range and units" for your maths channel at 0-1000 rpm.

In both cases increase the number of samples and utilise the zoom for further detail

I downloaded this pictures from Autonerdz, by the way

Dario,

You have copied materials from the Autonerdz forums without permission and posted them here. Everything on our forums is copyrighted and you are in violation of your member's agreement by posting them in another venue.

You are welcome to share the materials by simply posting a link to the discussion on our forums. Like this: