RE: A Small Simple Robot Arm : Part 2 Initial Software Implementation

beacon_dave — Tue, 30 Mar 2021 23:53:11 GMT

This robot arm on YouTube may be of interest:

fischertechnik 6-Achs-Roboterarm RS 816 - YouTube

I spotted in the latest ft:pedia magazine:

https://www.ftcommunity.de/ftpedia/2021/2021-1/ftpedia-2021-1.pdf

where there is more information.

(There is also an article there on mecanum and omniwheels amongst other things.)

RE: A Small Simple Robot Arm : Part 2 Initial Software Implementation

genebren — Fri, 29 Jan 2021 21:29:30 GMT

Dubbie,

Dealing with servos seems pretty easy, but devil is in the details. As [mention:b0bc65b9ecdc4307bd967592f00e340a:e9ed411860ed4f2ba0265705b8793d05] points out, stalled servos draw a lot of current and can cause a lot of smoke. I build a 12 channel servo controller board for one of my clients. This particular client can use up to 12 servos in some of his products and majority of the servos are high torque servos (potentially high current too). In order to keep the controller board safe, I have designed in current monitoring and a MOSFET switch to remove servo power if the current should get too high. The current monitor is located in the ground return in the servo power path (low side). The microcontroller monitors the current and will disable the servo power if the current is too high or near the limit for too long (software). The limits are setup in the board configuration (store in EEPROM), based on the number of servos, the safe max current draw of the servos, and the amount of time the max threshold is allowed.

Here is the current monitoring and servo power disconnect circuits:

[View:/resized-image/__size/620x318/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-c1b39ff9-9c36-4b30-96e2-12b83d83e425/2046.contentimage_5F00_205944.jpg:620:318][View:/resized-image/__size/277x535/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-c1b39ff9-9c36-4b30-96e2-12b83d83e425/5270.contentimage_5F00_205945.jpg:277:535]

In addition to current monitoring, one of the features of the board is a variable slew rates for the servo signals. My servo pulse generation is repeated every 25msecs. Within this window of time, there are eight frames (3.125msec each). The first six frames are used to generate 2 channels worth of servo pulse. During the 7th frame, any servo that is slew limited is updated. This allows me to slowly update the servo commands between the past value and the newly commanded position. I have options to slew the pulses either in fixed steps of by slowly filtering between the past value and commanded value with a software low pass filter. This slewing prevents from over stressing the servos by preventing jerky motions. With well over 100 units installed neither a controller or a servo has reached the smoking stage.

Good luck,

Gene

RE: A Small Simple Robot Arm : Part 2 Initial Software Implementation

luislabmo — Fri, 31 Jan 2020 20:18:02 GMT

Hello Dubbie, nice update!.

I believe a couple of loops and some ascii conversions could have saved you a lot of coding when printing the menu and also when translating the input to degrees. [View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-c1b39ff9-9c36-4b30-96e2-12b83d83e425/contentimage_5F00_3.png:16:16]

Luis