Accel / gyro project

Can't test your patch, but if you deal with floats, you should give your accum objects a float argument, i.e. [accum 0.]!
Depending on your needs, to capture the stream of values, you could use text, jit.matrix, etc., and even… capture.

accumulate object does not add wenn it receives float in the left inlet.
It just outputs it.

Does that mean that Your giro outputs zero when movement stops ?
About capturing - that dependends on what You want to do with the data collected...
Visual representation ? Or some kind of analysis ?

no it doesn't always return to zero, I need to work in some kind of filter - it should be a high pass for the gyro and a low pass for the accel.

the patch works pretty good at tallying up the angles. my sensor is supposed to read 20ms intervals but the numbers in the floats obviously aren't changing at 20ms intervals does that mean I am missing readings or is it just that the number animation can't keep up but the calculations still happen at right speed?

I did hope to be able to rotate a 3d object for a visual representation of the movement, but I can't seem to get the object to move in the correct angles?

But the most important part of the detail is to record the numerical outputs from pitch/roll, gyroscope Acc(x,y,z) then I was going to work out total linear acceleration vector and record this, then anything that is not 0.98 is movement other than gravity.
I want to translate it to spreadsheet format so in a manner that the results would be transferable would be good.

thanks

I would record raw data coming from the device, using timestamp output
to serve as for example address in coll, or even record it into audio buffer.
Later one can play it back, do calculations or whatever.
And ... the timing problem could be related to the filter You use -
averaging 5 readings

How do I do that?

that would be to get the raw data from my experimenting then process it from the extrapolated file.
so take the output terminals - convert to signal?
then how do I convert the audio back to numerical data?

I use buffer sometimes if massive data collection is needed.
It is a matter of scaling incoming data into floats that fit in audio buffer
using peek to read / write.
But in Your case that is not necessary.
You could use timestamp, which is as I understand in miliseconds
and use it as coll address. You can link that coll to cellblock.
I don't have any phidget to test, but there is good help at :
https://www.phidgets.com/?view=api
-----------
Here is simple patch with coll - cellblock :

Timestamp must be int, otherwise coll will not accept it for storage.
I don't know what gets out of the phidget, but if it is float,
You have to convert it to int.

I tried that patch, I can't get any numbers to input into the cellblock. do I need to do something to connect the coll to the cellblock?

the Phidget patch uses unpack to separate out the outputs there are 9 plus the timestamp
3 accel, 3 gyro and 3 magnetometer.

so I put number boxes between the unpack and the zl.rot
Is this correct?

Have a look in this patch :

The idea is to get the last item in the list, which is timestamp,
make sure that it is integer and not float, then move it from
last to first position of the list, and store the list into coll.
Coll address will become timestamp, and first column in the cell.
There would be alternative ways of doing so, like running
a metro - counter combo as timer, and use the list without timestamp.
If timestamp resets to zero every time one restarts the phidget,
than You get nice timeline.
You can play that back from coll using for example {clocker 1}
which would count in ms.