Lyman V5 : Filament fluctuation analysis [math help needed]
Posted by Frans@France
|
Lyman V5 : Filament fluctuation analysis [math help needed] March 04, 2015 05:28PM |
I'm trying to building and test a Lyman V5 filament extruder for some time now (in dutch: [forums.reprap.org]). In fact I'm rebuilding it to bring down the costs and trying to improve performance and stability. I'm using the sensor that [Filip Mulier] designed (http://www.thingiverse.com/thing:454584)
A M700 commando has been added to the mackerel software allowing me to record all parameters of the extruder with some software connected via the serial port. This version of mackerel will be available as soon as all bugs are solved. A version with some changes is already available here [github.com]
The client that records the data will also become available. Some screenshots (windows version)
Long introduction
almost done 
I have recorded 9 sets of data for the extruder running at different speeds (2.0 2.5 and 3.0rpm) and for each extruder speed I have set the puller at a fixed speed so that the filament would have a relative fixed width (1.8 2.0 2.2mm) Each run was 5000mm long
I did a FFT on these 9 data sets and was surprised to find the exact frequency of the extruder in the data. It was rather easy to find as that frequency had the highest magnitude
I then removed that frequency from the data and inversed the FFT only to find to my surprise (almost) no change in the data. So this means (if I understand all correctly) that this extruder "heartbeat" has no significant influence on the total fluctuation.
What I would like to know however is how can I correctly calculate the total contribution of this heartbeat to the fluctuation?
What I did was to remove all but this "heartbeat" frequency and then inverse the data. This showed a wave with a upper and lower limit somewhere between +/- 0,002mm - 0.007mm depending on the data set. Is this the correct way to calculate this??
A M700 commando has been added to the mackerel software allowing me to record all parameters of the extruder with some software connected via the serial port. This version of mackerel will be available as soon as all bugs are solved. A version with some changes is already available here [github.com]
The client that records the data will also become available. Some screenshots (windows version)
Long introduction
almost done I have recorded 9 sets of data for the extruder running at different speeds (2.0 2.5 and 3.0rpm) and for each extruder speed I have set the puller at a fixed speed so that the filament would have a relative fixed width (1.8 2.0 2.2mm) Each run was 5000mm long
I did a FFT on these 9 data sets and was surprised to find the exact frequency of the extruder in the data. It was rather easy to find as that frequency had the highest magnitude
I then removed that frequency from the data and inversed the FFT only to find to my surprise (almost) no change in the data. So this means (if I understand all correctly) that this extruder "heartbeat" has no significant influence on the total fluctuation.
What I would like to know however is how can I correctly calculate the total contribution of this heartbeat to the fluctuation?
What I did was to remove all but this "heartbeat" frequency and then inverse the data. This showed a wave with a upper and lower limit somewhere between +/- 0,002mm - 0.007mm depending on the data set. Is this the correct way to calculate this??
Sorry, only registered users may post in this forum.