Getting Started on STMBLv5 Documentation

Pinouts

Mon, 01 Jan 0001 00:00:00 +0000

Pinouts #

Interfaces supported by the current firmware (with the config template or component that reads them):

Mesa SmartSerial = full-duplex UART (link sserial)
Encoder = QUAD (link enc_fb0, component enc_fb), optionally with analog SIN/COS tracks (1Vpp)
Resolver = SIN/COS (link res_fb0, component res)
Hall / commutation signals UVW (link uvw_fb0, component uvw)
Mitsubishi (mit 02) = half-duplex UART (link encm_fb0, component encm)
Sanyo Denki = half-duplex UART (link encs_fb0, component encs)
Tamagawa SmartABS = half-duplex UART (link smartabs_fb0, component smartabs)
Fanuc serial = half-duplex UART (link fanuc_fb0, component encf)
Yaskawa Sigma serial (component yaskawa, used by the conf/sgmph-*.txt configs)
DMM = RS485 (link encdmm_fb0, component dmm)
EnDat = half-duplex SPI (component endat, experimental, no template loads it)

The hardware can also carry SSI, BiSS (half-duplex SPI) and Hyperface (SIN/COS + half-duplex UART), but there is no firmware for them.

Servoterm

Mon, 01 Jan 0001 00:00:00 +0000

Servoterm #

Servoterm (servo terminal) provides an interface which allows editing of the drive HAL configuration. It also provides a rolling graphical representation of any chosen parameter in the HAL, which can be a great aid to tuning and motor setup.

Servoterm is now available as a C++ application. You can download it from this GitHub repository.

Installation #

To install Servoterm, follow these steps:

Clone the repository:

git clone https://github.com/STMBL/QtServoterm.git

Navigate to the project directory:

Feedback

Mon, 01 Jan 0001 00:00:00 +0000

Feedback #

Incremental Encoder #

An incremental encoder operates on the principle of quadrature encoding, meaning it uses two channels of data to provide accurate measurement information. When the encoder is rotated or displaced, both output signals A and B change in a specific sequence that indicates direction and magnitude of movement. However, since it doesn’t maintain an absolute position reference, it only provides relative changes from its initial state.

Motor

Mon, 01 Jan 0001 00:00:00 +0000

Motor #

It is possible to use autotuning to estimate the motor parameters.
You can still use this guide if autotuning fails.

Motor parameters #

Unfortunately, servo motor datasheets often lack a good description of the parameters. There is a difference between line-to-line and phase values. Here are listed the most important motor parameters and how to determine them.

Resistance (conf0.r) #

Measure line-to-line resistance R
Convert to phase resistance conf0.r = R/2

Inductance (conf0.l) #

Measure line-to-line inductance L
Convert to phase inductance conf0.l = L/2

If you can’t measure this, a rough guess is conf0.l = 0.001 × conf0.r (a 1 ms electrical time constant). Better, measure it with link id_pmsm (see Tuning).

Tuning

Mon, 01 Jan 0001 00:00:00 +0000

Parameter Tuning #

STMBL can measure most motor and axis parameters itself. Each measurement is an identification component (idpmsm, idacim, iddc, idtune, idm, ids, iddac) loaded by an id_* config template. It takes over the power stage, runs its test and prints the results as conf0.* = ... (or hv0.*, acim_*, vf0.*) lines in green for you to append to the config.

Requirements #

Before You Start #

Which stages for which motor #

Motor	Stages, in this order
DC motor	id_dc, then id_mot, id_sys, id_pid
PMSM with encoder	id_pmsm, id_tune, id_mot, id_sys (optional), id_brk (optional), id_lpf or id_frf (optional), id_pid
Induction motor (ACIM)	id_acim (standstill, optional rotating test), then run it with V/f, V/f with an encoder or FOC; with an encoder also id_mot, id_sys, id_pid ( speed mode for a spindle)
Sensorless PMSM / ACIM	identify as above, then obs_check, sl_pmsm, sl_acim, id_sys_sl

On an STMBL v5 HV board also set the overcurrent trip threshold.

Linuxcnc

Mon, 01 Jan 0001 00:00:00 +0000

LinuxCNC #

Here is an example of a LinuxCNC configuration for one axis, using Mesa smart serial.

STMBL side #

The drive config needs link sserial after the motor and feedback templates, for example:

link pid
link pmsm
link enc_fb0
link sserial
link misc

The sserial template makes the drive a smart serial remote with the card name stbl. It:

links fault0.en = sserial0.enable, so LinuxCNC enables and disables the drive, and sends the drive fault and its fault number (sserial0.fault_code = fault0.last_fault) back to LinuxCNC
feeds the position and velocity command from LinuxCNC through linrev0 and vel_int0 into the controller, and sends the position from linrev0 back to LinuxCNC
derives an acceleration feedforward from the host frames: vel_int0.acc_out (the change of vel_cmd per frame, held until the next frame, 0 while the watchdog holds, clamped to conf0.max_acc) goes through rev0.out_dd to pid0.acc_ext_cmd. It is on by default; vel_int0.acc_k scales it (1 = as commanded, 0 = off)
places the host frames on a smoothed frame clock, vel_int0.sync_k = 0.05: a frame that arrives an rt tick early or late (host jitter) no longer puts a sawtooth of vel_cmd x the timing error on the position command (10-20 um at 200 mm/s on Y). The mean offset is 0, so no latency is added; sync_k = 1 snaps to each frame as before (see vel_int)
runs sserial0 with rt_prio 2.3 and frt_prio 2.0, and sets sserial0.pos_advance = 0.0003 (tune it per axis, see the last section)
raises fault 1 (CMD_ERROR) if no valid packet arrives for 5 ms (sserial0.timeout, 100 frt cycles)

For a position axis (not a spindle) also set these two, both 0 by default:

Servoterm protocol

Mon, 01 Jan 0001 00:00:00 +0000

Servoterm protocol v1 (float scope, capture, header, ack) #

Drive -> host on the USB CDC port. Text and binary share the stream. A binary frame starts with a marker byte in 0xFB..0xFF, and every payload byte of a 0xFD/0xFC/0xFB frame is in 1..250, so markers are unique there and text (ASCII/UTF-8) never contains them. (Base 250, not 253: digits 250..252 would have been bytes 0xFB..0xFD.)

Encodings #

d250: an unsigned integer as base-250 digits, most significant first; each digit is stored as digit + 1 (1..250).
f32: an IEEE754 float’s 32 bits as a 5-digit d250 (250^5 > 2^32).
u16: a 2-digit d250 (0..62499).
u8: a 1-digit d250 (0..249).

Frames #

marker	name	payload	length
0xFF	legacy 8-bit sample (term0.fmt 0)	8 bytes, v + 128 clamped 1..253 (was 1..254; 254 = 0xFE collided with reset)	9
0xFE	reset	none: clear the scope	1
0xFD	float sample (term0.fmt 1)	u8 seq, 8 x f32	42
0xFC	header	6 x f32: proto (1), kind (0 live, 1 capture), period [s], channels (8), count, pre	31
0xFB	capture sample	u16 index (0..count-1), 8 x f32	43

Float values are (wave + offset) * gain, the same number the 8-bit path rounds to a byte before adding 128. The host divides by 128 to get today’s normalised -1..1 (it can go beyond ±1, never clipped).