Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    JanStry
    @JanStry
    thank andy, i got step/dir working but its pretty noisy since i'm not using differential signals atm.(i will use a am26ls31cn in the future to send rs422 signals for step/dir)
    about the serial protocols. is there a working protocol on the feedback1 connector?
    issyvarsano
    @issyvarsano
    @JanStry use common mode filters at the motor connectors near the stmbl. It will reduce significant the noice
    JanStry
    @JanStry
    286007123_434861071434959_7374313573173750396_n.jpg
    hi i did some more testing. i build a small circuit with am26ls31 to send step/dir. this works pretty good. with my 24V powersupply everything works great. but with my 320V powersupply step/dir is still not working as it should. i tried sending step/dir pulses from a arduino and from time to time the arduino freezes reboots...and my laptop trackpad is acting up.. i have added a ferrite core on the motor leads close to the stmbl as @issyvarsano suggested. this does not make a noticeable difference in my case.. what is the best way to ground the drive? maybe the problem is my small 320V powersupply.. its pretty rudimental : 4diodes in bridge and a small 100uF(400V) cap.
    issyvarsano
    @issyvarsano
    @JanStry wrong ferite ...
    use Laird-Signal Integrity Products 28B1122-100 , make 2-3 turns
    JanStry
    @JanStry
    thanks you @issyvarsano , the ferrite i'm using at the moment is about 2mm shorter and 3mm thinner than the Laird-28B1122-100 you suggested. (not on the photo obvious) . it might be something else..maybe capacitive coupling of the motor voltages like @Sync- also experienced . any solutions for this? thanks
    long0900
    @long0900
    @tlorincz72 Hi, were you able to read all 22 bits of HG-KR servo motor's encoder
    although the manual say the encoder is 22 bit, my experiment and the hint from stmbl's wiki suggest that it only has 18 bit resolution
    tlorincz72
    @tlorincz72
    @long0900 Hi, I read as 17 bit encoder (with 2,5Mbit/s). Good enough for my projects..
    long0900
    @long0900
    I agree that 17-18 bit is plentiful but the encoder has up to 22 bit resolution, at least that's what the manual says. The more the better, right? ;)
    can you confirm your HG-KR motor is equiped with OBA24R encoder? Thanks!
    tlorincz72
    @tlorincz72
    @long0900 just sent 0x32, and encoder replies the 17bit pos data. I have the same OBA24R encoder on the motor (nearby the matrix code)
    Nico Stute
    @crinq
    afaik you have to send a different byte to get more data from the encoder
    tlorincz72
    @tlorincz72
    @crinq I don't modified the original stmbl firmware. Just works with HG-KR encoder.
    tlorincz72
    @tlorincz72
    @crinq Here is the HG-KR73 config:
    link pid
    link pmsm
    link encm_fb0
    link misc
    load enc_cmd
    enc_cmd0.rt_prio = 2
    enc_cmd0.res = conf0.cmd_res
    rev0.in = enc_cmd0.pos
    enc_cmd0.mode = 0
    conf0.r = 0.8
    conf0.l = 0.008
    conf0.j = 0.00013
    conf0.max_force = 4
    conf0.max_ac_cur = 5
    conf0.mot_fb_res = 131072
    conf0.cmd_res = 2048
    conf0.mot_fb_offset=0
    conf0.com_fb_offset=0
    conf0.mot_fb_rev = 1
    conf0.polecount = 5
    conf0.psi = 0.08
    fault0.cmd_error = 0
    io0.cmdg = 1
    long0900
    @long0900
    @tlorincz72 thank you
    long0900
    @long0900
    @crinq that's strange since the encoder reports 3-byte position, there's enough room to fit the LSBs in these bytes
    maybe I'll try to test all possible command bytes and observe the outcomes
    Nico Stute
    @crinq
    it sends single + multi turn data for the 0x32 request
    you can overwrite the request with encm0.cmd
    long0900
    @long0900
    thanks :) I'll take note of this
    currently I haven't built any stmbl drive yet, just tinkering with FOC concept and implementation
    Marian Omnibus
    @rabrabar
    @JanStry Hey try half-wave rectifier configuration, where neutral (verified and not just one from any two wires in any socket) is directly connected to capacitor minus. The neutral should already be grounded somewhere. I have just had problems with full wave rectifier on square wave inverter power, with one output grounded. It seems just a diode drop away there, but who knows what goes on there with the spikes...
    Marian Omnibus
    @rabrabar
    This may be off topic but I know most of You use LinuxCNC... Does anyone remember the keystick ncurses interface? I never used it but I would like to avoid X on one machine... It looks like it was abandoned for a few years before it went out of EMC. I tried linuxcncrsh and the python interface really is a snail compared to that. I might try to get keystick working again, or create something similar. My question is if I should use emc.hh NML stuff to communicate or is non-realtime hal component talking to halui an acceptable solution? I'm not sure it would cover all needed features, that is one thing but I wonder if any of these be noticably faster than the other? Any other things to consider? Thanks.
    John Dammeyer
    @jcdammeyer
    Is there an easy way to make the STMBL become a constant torque mode drive? The application is a power draw bar control (through a planetary gearbox) that tightens to 24NM but uses 30NM to release. In addition to the constant torque mode a displacement is also required. I've done that sort of thing with Commercial drives and CAN bus messages.
    Nico Stute
    @crinq
    you can set pid0.pos_en, pid0.vel_en and pid0.acc_en to 0 to disable the outer loops
    than set pid0.torque_ext_cmd to your desired torque
    e.g. you can use a mux comp to set it so fixed values depending on a gpio, or scale an analog input or connect it to smart serial
    for a guaranteed displacement i would us it in position mode with stp, disable saturation and pos error and set your desired torque limits
    John Dammeyer
    @jcdammeyer
    @crinq Thanks. Doesn't even really need an extra GPIO pin. The direction can be used to select which torque is required. The gearbox is a planetary 25:1 and the motor about 1.25 NM which is max 31.25 NM and only 28 to 30 NM would be what is required for CCW direction.
    Drop it down to 24NM for CW direction. Then just step pulses up to 20 turns CW but I'd have to monitor if not position error then something to let me know it was stopped. I could use position error but I wouldn't want it to fault and require a reset. Although for the tool changer draw bar that wouldn't be the end of the world.
    Bryce Johnson
    @brycedjohnson
    Did anyone have replacement for the Si8621BB-B-IS isolator? I was going to put an order in for 10 more stmbls on jlcpcb and I think I got most everything else sourced... but the current prices on isolators those are crazy. Like $30 a piece. Was the rated 150Mps data rate a requirement for that data?
    Bryce Johnson
    @brycedjohnson
    image.png
    This CA-IS3721 looks like it might be drop in replacement
    image.png
    Vs the si8621. Appears to both be 150Mbps as well
    andypugh
    @andypugh
    @rabrabar Keystick might still work, it's here: https://github.com/LinuxCNC/linuxcnc/blob/2.7/src/emc/usr_intf/keystick.cc
    But it isn't aware of Joints-Axes changes, so still assumes X =0, Y=1 etc.
    wrenchonline
    @wrenchonline
    Hi everyone, I'm a newbie here. I would like to know if this stmbl needs a DC power supply or an AC power supply?
    Nico Stute
    @crinq
    DC
    andypugh
    @andypugh
    But it can be simple rectified AC
    wrenchonline
    @wrenchonline
    thanks for the reply
    wrenchonline
    @wrenchonline
    After a few days of study I decided to make this crazy machine and I found that I am missing LQH32PN4R7NNC here Is it possible to replace it with LQH32PN4R7NN0L
    Nico Stute
    @crinq
    yes
    andypugh
    @andypugh
    I think that will be one of the less fussy components. It's been a while but email-mining suggests that I have used https://octopart.com/search?q=LPS3015-472MR&currency=USD&specs=0 too. The pads on the PCB are pretty big so a lot of inductors will fit.
    wrenchonline
    @wrenchonline
    @crinq @andypugh thanks. It means that LQH32PN4R7NN0L can meet most of the requirements? I looked at LPS3015-472MR and found that the rated current parameters of the three of them are somewhat different
    hello110300509
    @hello110300509
    Hi everyone, when i read the source code of stmbl, i am confused by the PHASE_90_3PH, PHASE_120_3PH, PHASE_180_2PH, and PHASE_180_3PH. what is the meaning of the phrase?
    Nico Stute
    @crinq
    number of phase wires and electrical angle of the coils