stmbl servo drive setup with smart serial.

Probe!

another update

HI
I would like to tinker with the STMBL setup a bit before deciding on getting some boards made.

For a start, i would like to flash the STMBL firmware to two development boards (like these www.robotics.org.za/6970622937432?search=stm32) and establish the connection to a Mesa Sserial port. Is that even possible?

Henk

I am no expert but this link is where the Deb's hang out

gitter.im/rene-dev/stmbl#utm_source=noti...unread-notifications

More stmbl action

Hi.

(Hope it's OK to post this here rather than a new thread, please move if needed)

I'm wrestling with STMBL connections to linuxcnc again, and having a little more success. I'm trying to figure out the STMBL config - the file stored in flash memory and accessed/configured via Servoterm.

I've seen the lcnc config example andypugh has posted a couple of times, I'm using the bare-bones lcnc config from rene-dev supplied as an example (the same one, I think) and am stepping through building a stmbl config by adding one component at a time and seeing what happens.

I gather that spi, yourself, andypugh, obviously rene-dev and maybe cakeslob have stmbl drives and have managed to get them working with lcnc. What I can't find is any guide to the STMBL HAL components or how to string them together - there are a couple of examples on the gitter channel but rarely explained or commented.

My test rig is Linuxcnc 2.9.2 -> Mesa 7i94 -> STMBL -> old Fanuc Model 20F (A06B-0352-B731 - 23Nm, 20A, ~150V, 2000rpm) with quadrature encoder, as well as C1/C2/C4/C8 commutation output (unused atm).

I'd be grateful if anyone could post their config files and delighted if anyone can shed some light on the workings.

Thanks Andy,
I have been searching the config examples to try to make sense of their hal system, but it's very stab-in-the-dark. I'm hoping for a look at a config which specifically is known to work with linuxcnc.

Right now I have a minimal config, which shows up the stmbl's hal pins in lcnc halshow, releases the dc brake on pressing Enable/F2 in Axis, and will respond to an Axis jog at least a little (by oscillating a little and then giving a saturation fault/amplifier fault in servoterm/lcnc respectively). So it feels like it's a matter of tuning, maybe.

So far if I add my measured/calculated motor parameters to my config, I get saturation faults or the servo makes noises like an old dialup modem. I'm proceeding by adding in hal lines one at a time (from the a06b config you linked and from others that seem like they might be relevant) and testing, but it's very slow going. E.g. if I add 'link enc_cmd', which is in many of the examples, it faults immediately...and really all I have to go on is that it's often linked...it's a bit of a cargo cult way of doing a config.

Would you be able to post your stmbl config?

Right, that's good to see. I guess you have 2 types of feedback, and all entries containing uvw related to configuring the 2nd (3x hall sensors?) so not relevant for me.

The line: conf0.mot_fb_rev = 1 appears to mean 'reverse feedback direction' having just tried it out in my config, now removed.

In case this is useful to anyone else, here's my current config, which responds to lcnc jog keys but as mentioned, unless the jog speed is around 3mm/min it responds with a quickly increasing oscillation followed by a fault:

link pid
link pmsm
link enc_fb0
conf0.mot_fb_res = 8000
conf0.mot_fb_offset = 1.16
link sserial
conf0.polecount = 4
link misc
io0.out0 = fault0.mot_brake
rev0.in = sserial0.pos_cmd
sserial0.pos_fb = enc_fb0.pos
term0.wave5 = sserial0.pos_fb
term0.wave6 = sserial0.pos_cmd

The same config, with my current values for motor parameters included:
link pid
link pmsm
link enc_fb0
conf0.mot_fb_res = 8000
conf0.mot_fb_offset = 1.16
link sserial

conf0.r = 0.1
conf0.l = 1.082
conf0.j = 0.012319
conf0.max_ac_cur = 20
conf0.max_vel = 209.4395
conf0.max_acc = 10
conf0.max_force = 23
conf0.psi = 0.191666666

conf0.polecount = 4
link misc
io0.out0 = fault0.mot_brake
rev0.in = sserial0.pos_cmd
sserial0.pos_fb = enc_fb0.pos
term0.wave5 = sserial0.pos_fb
term0.wave6 = sserial0.pos_cmd

But, as mentioned, adding those in gets me the modem noises. Going to go through the parameter measurement/calculation process again and see if I get different numbers this time.