Hello Nitish,

First answer to your last question - Decreasing sensitivity will NOT BE HELPFUL as you shall loose or reduce the signal amount. What important is to have uniform sensitivity on all of your channels (X, Y and Z).

By looking at the problem you are reporting, it is not clear as how you are attempting to draw the line, either manually by hand or through a machined stage?

Before you start taking any measurements as your observed data, it is important to have all channels calibrated and then you can compare the results.

Also, check if there are ANY REAL vibrations along the undesired axix (say X as in your case) while you are expecting your ACTUAL data from the Y axis line drawing activity.

Alternately repeat the same stroke/ line drawing activity several times and see whether you get consistent data or not?

Also check for any unwanted vibrations the accelerometer might be exposed to.

Thanks with best wishes!

-Prasanna Waichal

Hi Nishith, 

The problem is that you most probably is not aligning the axies of the accelerometer with the motion axes. Meaning that your sensor's x-axis is not aligned with your motion's x-axis. For example, any small deviation from your sensor's x-axis with respect to your motion x-axis will result in your sensor reading a motion that is a combination of x and y axes generating accelerations in both directions. The larger acceleration meaning that you are closer to y-Axis but still have acceleration components in the x-direction. I guess you either  need to check your sensor's  direction and align it, which is hard, or find a transformation matrix between your motion and measurement, in other words calibrate it. 

Good luck 

Thanks @Prasanna Waichal for your useful suggestions.

I am taking the readings with pen by hand. Thereby, i think the probable vibrations produced by the hands could give fluctuations in other axis. But, i cannot make it out that does these vibrations are of such high amplitude in nature that it is as shown in the figure.

Also, can you suggest any method, for checking/determining the real vibrations along the undesired axis? Since repeating the same stroke multiple times prone to give different values as the hand movement is dynamic all the times.

Thanks in advance.

Thanks @Abdullatif Alwasel for your valuable answer. How you suggest the parameters to be looked into while designing the transformation matrix for the alignment of motion axis with sensor's axis. (I am performing the motion along with the axis, hence i think both the sensor axis and motion axis are aligned).

Also, the human strokes doesn't always prove to be straight line always hence it might result in the component in other axis.

Hi Nisith,

 First of all what type of actuators you used in all 3 axes.

next what was the class of accuracy you used and type of precision of LMG.

you can meet me personally at Nano center in MU. In case required.

How stacking is done on each axes is important, rectification if done at initial stage at mechanical level you will get accuracy better or best, later stage.

You will be able to see actual model here in our lab.

 Apart from this it can be used in Graphology i.e hand writing analysis.

I too am looking for similar project.

 But can not take it up at present.

Anyway you did some thing before I can start. Thanks

PFA

You can see red colour part which is penholder. in above pictures.

@Ajay Sanghrajka: Thank you soo much for your kind invitation. I will approach you. And also thanks for your model demonstrations.