Dynamixel xm430 w350

Hi
I want to control the torque in a system and choose between these two dynamixels.
mx 64 and xm430 w350.
My question is how to control the torque in these dynamixel motors? the data sheet states that the torque control is based on the current value, is there a mathematical relationship between the torque and the current value?
Is this relationship applied through the dynamixel motor by calibration for different torques and currents, or should it be achieved?

Hi,

There is no mathematical model that explains current and torque conversion.

However, you can estimate the torque by reading the Present Current value and compare it with the Performance Graph

In case of XM430-W350, the Present Current value to current conversion factor is 2.69[mA], therefore, if you read 500 from the Present Current, it is equal to 1.345[A].

From the performance graph of XM430-W350, 1.345[A] is about 2.3[Nm].

The current reading from X series is a bit more precise than MX series thanks to the new hardware and firmware design.

However, if you need very precise Torque measurements, it is strongly recommended to use additional Torque sensor.

Thank you.

HI
If I use this diagram, how much does the torque obtained from the diagram differ from the actual torque and what is the error value?
what causes this error?

thank you

The performance graph is a reference graph that you can estimate the Speed, Torque, and Current based on the actual test.
The NOTE under the graph describes how Torque is measured by gradually increasing the load on the output horn.
Like I said in the earlier reply, if you need a precise torque measurement, I recommend to use the torque sensor.

Thank you.

I’m sorry to ask the question again
I know how the graph is obtained.
My question is how much is the error?
thank you

@msvbi,
The performance graph is a generalized trend graph based on the actual test result.
And it is difficult to provide the margin of error as it is provided as a reference to “estimate” the torque - current relation.
Thank you.