hsinkoyu / fs100 Goto Github PK

Hello,

I have seen on the FS100 documentation that only network input signals are writable and there was nothing written about general purpose IO. I was wondering if reading and writing general purpose IO is possible on FS100. If not, i would like to know if there is an alternative on controlling general purpose IO using python or something else.

After checking the FS100 code and datasheet. It is a bit unclear on how to move the robot using absolute pulse values. For example: home pose robot independent from adjusted robot tool in the pendant.

I was trying to move the robot in a straight line instead of an arc; more like moveL in its pendant. I browsed the FS100 datasheet to find Straight absolute position operation as a movement type. I could not find any explanation on this move type. Can you illustrate it please? I tried using this move type, but the movement was not as expected; either not moving at all and giving an error or going to another pose far from the given pose.

Dear author,

When I tried to connect to the FS100 Controller, we confront an error "ConnectionError: timed out". I have already set the " #82015 CMD REMOTE SEL" on and set the controller to remote mode.
So I would like to ask what setting have you made on the FS100 Pendant before you connect.

Thank you for your time and may you have a nice day.

Hello,

is it possible to read/write a variable of IO type, addressing only a single bit?
Thanks

Hi I am trying to write an I/O from the fs100 code.

I am activating the output signal port=3002, when I send the command I received the error:
failed writing the variable, err=0xa001

Checking the error it seems an Instance error, however the rest of the variables works with the instance=1

Any idea what is the issue? Do I need to change the configuration of the robot to allow the writting of I/O?

Kind regards,
Alejandro

I'm using read_position to read robot's pose
https://github.com/hsinkoyu/fs100/blob/master/fs100.py#L1258-L1264
I find that its sampling rate is around 50Hz.
Is it possible to increase this sampling rate of robot's pose?

Good morning,

First thank you for your work!
I am currently testing the function read_torque, however I did not find neither here or in the fs100 HIGH-SPEED ETHERNET SERVER documentation what are the unit of the function output. While testing I received positive and negatives values so I assume it cannot be % of power if yes, maybe the decoding is incorrect?

Do you have any idea?

Maybe it would be nice to consider renaming the library and the repository from fs100 to something like pyhse to indicate what protocol the library implements. Because it's not limited to FS100 controller, you of course know that. We are using it with YRC1000 controller and it works with YRC1000micro too. Maybe also with older controllers, I don't know.

Then there is a possibility of making a standard Python package out of it, apply some versioning scheme and publish it to PyPI so that everyone interested can easily install it via pip.

What do you think? Of course such effort means more time required to do the changes and maintain them. I can offer some of my time, I have no experience with publishing to PyPI but I learn rather quick. So maybe...
Regarding the name change, that would of course mean that all the links to the repository will stop working. Still people will find the library with Google and GitHub tags as I found it. And there will be a new way through PyPI.

Is there any particular reason for using string keys in returned dict values for axis identification? I mean 1st_axis and similar keys. It seems a little strange, I would expect just a list of objects or primitive values.

There is also a problem related to the configuration of the robot controller that you communicate with. Not every controller has 7 or 8 axes like you expect in some functions. I think that you addressed it for one of them in #5, but the same problem is present in other functions. I believe that I experienced it with read_torque for example.

Would it be a problem if we change the interface and return a list from those functions? Or should we provide two functions, one returning a list and the other (using current name and signature) returning this dict for backward compatibility (perhaps with keys according to the real number of axes)?

I also believe that it would be nicer if the functions return a tuple with result code and dict/array instead of accepting pre-constructed dict or list that will get modified.

I'm open to contribute a PR with all this, but wanted to discuss that with you first. Maybe there are reasons behind this. You are using the library at work or you did it for another purpose?

Thanks.

Hi again

I was wondering if you know about the possibility of including joint/linear interpolation in the movements sent from the High-Speed Ethernet server. This option is available in the pendant so why not in via High-Speed Ethernet