Gripper Communication Guide

The communication interface uses standard 11-bit CAN IDs and MIT control frames.

This communication guide applies to all gripper firmware versions in the 0.3.x and 1.4.x series.

1. Firmware and Parameter Overview

The gripper MIT frame uses physical units rather than the ordinary joint-motor units:

The recommended initial gains are:

kp = 2.00
kd = 0.05

2. CAN Bus Parameters

The gripper uses the same CAN/CAN FD communication method as the joint module. See the Joint-Module CAN Communication Guide for the common communication overview.

The current gripper communication configuration is a 1 Mbps FDCAN nominal rate, a 5 Mbps data-phase rate, and CAN FD BRS. MIT command and reply frames are 8 bytes long.

The firmware uses the received frame format for its reply. Keep CAN/CAN FD and BRS settings consistent between the host and the gripper.

3. CAN IDs

The general rules for CAN IDs, MIT unicast, broadcast, and reply IDs are the same as for the joint module. See:

• MIT Mode Master Control Frame
• CANFD Broadcast Multi-Motor Control
• Relationship Between master_id and can_id

The following only lists the gripper defaults and usage.

3.1 MIT Unicast Control

MIT control frames are sent to the gripper can_id. The default configuration is:

can_id    = 1
master_id = can_id | 0x010 = 0x011

The gripper sends replies to master_id, and data[0] in the reply contains the gripper's own can_id.

In a broadcast frame, node n uses the 8-byte block at offset (n - 1) × 8. The DLC must cover the block for the target node.

3.2 Configuration Register CAN IDs

Configuration register commands use:

CAN ID = (cmd << 4) | can_id

For the default node can_id=1:

Configuration-register commands respond only to the target unicast ID. They do not support broadcast register reads or writes.

4. MIT Control Frame

The MIT byte layout, bit widths, and packing method are the same as for the joint module. See the MIT Mode Master Control Frame and Bit Width Description.

The gripper physical ranges and sign convention differ from ordinary joint motors; use the definitions in Gripper Communication Ranges below.

5. Gripper Communication Ranges

5.1 Position

The gripper MIT bus position meaning is:

p_des = 0 mm   -> closed
p_des ≈ 70 mm  -> open

The recommended physical positions for the current gripper are 0 mm when closed and approximately 70 mm when open. The protocol and firmware protection ranges are:

Protocol encoding range: 0–90 mm
Firmware protection range: 0–80.55 mm

The firmware clamps the target and feedback position to 0–80.55 mm, but the host should use the actual approximately 70 mm open position as the application target. Do not treat 90 mm as the actual opening position.

Position encoding:

p_raw = uint16((p_des / 90.0) × 65535)
p_des = p_raw / 65535 × 90.0 mm

5.2 Velocity

The gripper velocity mapping range is calculated from the max_vel configuration register:

v_limit = round(89.5 × max_vel / 2π) mm/s
v_des   = -v_limit to +v_limit mm/s

The default max_vel=45 rad/s corresponds to approximately:

v_des = -642 to +642 mm/s

Velocity encoding:

v_raw = uint12((v_des + v_limit) / (2 × v_limit) × 4095)
v_des = v_raw / 4095 × (2 × v_limit) - v_limit

5.3 kp and kd

kp_raw = uint12(kp / 5.0 × 4095)
kd_raw = uint12(kd / 1.0 × 4095)

5.4 Feed-forward Force t_ff

The MIT t_ff field represents gripper force in N. Its range is calculated from the current limit:

F_limit = current_limit × 0.07 × 1 / (2 × 0.007125)
t_ff    = -F_limit to +F_limit N

The theoretical default ranges are:

Sign convention:

• Positive force/positive torque closes the gripper.
• Negative force/negative torque opens the gripper.
• Maximum force magnitude: approximately 16.37 N for the 50 gripper and 24.03 N for the 50L gripper.

The actual usable gripping force is affected by the mechanism, friction, supply voltage, temperature, and current derating.

t_raw = uint12((t_ff + F_limit) / (2 × F_limit) × 4095)
t_ff  = t_raw / 4095 × (2 × F_limit) - F_limit

When t_raw is 0x7FF or 0x800, the firmware decodes it as exactly zero feed-forward force.

6. MIT Reply Frame

For the common reply layout and master_id rules, see the Joint-Module Motor Response Frame. The gripper position, velocity, and force fields use the physical ranges in Section 5.

For default NTC decoding, AUX polling format, and polling order of data[6] and data[7], see the MIT Polling Reply Guide. Writing 1 to register 0x6D switches the last two bytes to AUX polling data.

7. Configuration Register Communication

The gripper register configuration function is consistent with the joint-module implementation. For register commands, request and reply formats, save operations, information queries, data types, and write levels, see the Motor Register Communication Protocol. The sections below only list commonly used registers and gripper-specific parameter ranges.

7.1 Request and Reply Format

Register requests use little-endian byte order. float values use IEEE754 single-precision format.

REG_READ request:
  data[0..3] = address

REG_WRITE request:
  data[0..3] = address | (value_type << 8)
  data[4..7] = value_raw

Reply:
  data[0..3] = status | (value_type << 8)
  data[4..7] = value_raw

7.2 Common Gripper Registers

7.3 Register Write Levels

Register writes modify RAM only. To persist a setting across power cycles, confirm status=0 and then send REG_SAVE. mit_aux_enable is runtime-only and returns to 0 after power-up, configuration read, or configuration save.

8. Communication Examples

For node 1, write kd_max=1.0. The little-endian IEEE754 representation of float 1.0 is 00 00 80 3F:

CAN ID: 0x181
DLC:    8
DATA:   6B 02 00 00 00 00 80 3F

Read kd_max:

CAN ID: 0x171
DLC:    4
DATA:   6B 00 00 00

Example MIT command: target position 40 mm, velocity 0 mm/s, kp=2, kd=0.05, and feed-forward force 0 N:

CAN ID: 0x001
DLC:    8
DATA:   71 C6 7F F6 66 0C C7 FF

The example uses p_raw=0x71C6, v_raw=0x7FF, kp_raw=0x666, and kd_raw=0x0CC. Zero feed-forward force uses t_raw=0x7FF.

9. Usage and Safety Notes

1. Read can_id, master_id, config_version, and kd_max before first use.
2. Confirm the gripper zero position, direction, and mechanical limits; 0 mm is closed and approximately 70 mm is open, while the protocol range 0–90 mm is not the actual opening position.
3. Start with kp=2 and kd=0.05, using a low target velocity and small t_ff.
4. MIT control frames should be sent continuously. If can_timeout_ms is non-zero, keep the command period below the timeout.
5. When using broadcast control, reserve the correct 8-byte block for each node and avoid CAN ID conflicts.
6. Change can_id, master_id, and MIT mode only in the allowed state, confirm the write response, and save the configuration when needed.