LinuxCNC EtherCAT for reals this time - New Control Cabinet and Leadshine Servos
ฝัง
- เผยแพร่เมื่อ 7 พ.ค. 2024
- Precision Motion Control with or without EtherCAT www.leadshine.com/
CNC Spindles Made in Germany www.spinogy.de/en/
Control computers www.welectron.com/Boards_1
More from me / marcoreps
My LinuxCNC machine config github.com/marcoreps/linuxcnc...
00:00 Intro
00:59 Control Cabinet
03:46 Leadshine Servos
06:05 Electrical
10:28 Software
12:25 Sensorless homing
13:29 Encoder battery
14:39 Installation
16:48 RS485 VFD comms
18:58 "First" chips
20:48 Motor tuning
22:39 Results
#cnc #leadshine #linuxcnc - วิทยาศาสตร์และเทคโนโลยี
I only understand about 1% of what your talking about, but your sense of humor is outstanding! Keep up the great work! 😀👍
I'm with you on this @thevespap200e been with Marco for so many years now and just love his content.
I'm not alone then.
At least the stuff from 3:45 was self-consciously technobabbly (even if an accurate description of what happens). :)
I'm continuously in awe of the level of technical capability that comes out of the open source community.
Glad to see someone else using Debian with the XFCE desktop. It is also very cool seeing open source software like LinuxCNC competing with proprietary software
I’m sure I speak for many, but thank you for documenting all this!!
Nice to see some EtherCAT in a DIY application. For a hobby project, that's really impressive. The whole configuration and parameterisation with EtherCAT, especially if you want to drive DC synchronously, is a steep learning curve.
I work for a company that manufactures servo amplifiers, motors etc. and I was a bit perplexed about the encoders. If they were "real" absolute encoders, they wouldn't need a battery to store their current position. To me, this smells more like normal incremental encoders that are battery-backed on the servo amplifier side. A proper multiturn absolute encoder quickly costs 5 to 10 times as much as a simple incremental encoder.
It could be that, or it could be that it needs to count the total number of revolutions, an absolute positioning encoder can't determine 1.1 rotations from 2.1 rotations for example.
I can't say how Leadshine does it but I know for sure how Omron does it. They have a Single Turn Absolute encoder built into their absolute servos, that way they can tell the exact position for one turn, additnally they count how many full turns the motor has made so far, which multiplies by the absolute encoder resolution plus the fraction of the current turn gives the absolute postion. With that the just need to update the memory for a full turn and hold that using the battery power. The newer series 1S is even fancier and doesn't need a battery at all. The use the energy generated by the rotation of the motor to update the absolute rotation counter. The only thing I don't know is what happens if a motor goes back an forth over the tipping point of a full turn by a few increments (does this generate enough energy to update the memory every time!?)
@@bouni8506 Good point about the single turn absolute encoders, didn't thought about them.
Single turn absolute encoders could use a battery for multi turn counting, but it's probably an incremental with drive side keeping track. And to the other person, you are describing a weigand sensor for multi turn counting. They dont use energy produced by the motor, they power themselves. Look up weigand effect.
@@bouni8506 as long as you feed electric power to turn or hold the motor, you should not need any turning or motion at all.
A 4th axis is happening, wow.. impressive!
The 4th Dimension 😵💫
@ 0:39 we are gonna fix that, by ignoring the spaghetti monster, and plugging in a usb extension cable.. fastest marco reps video EVER XD
When my hip replacement business can start for real. 🤣
You never fail to make me laugh and learn. 😅
That's just a "cover" to throw everyone off the true goal: precision and accuracy measured in angstroms for neural implant surgery!
@@shazam6274 😂👍
For some reason, i checked your channel for new videos just today, and while watching your old content, you treat me to this. Thanks marco
I do the same!
2:34 We always used some Loctite 273 to fix the bolts to the cabinet. So they don't come off.
8:35 flyback diode ? Will prevent the brake to close fast enough, so use a varistor instead. Brake will work much faster!
9:03 Battery backup! Normally you can turn the motor without loosing position counting. But if you remove the motor from the spindle, all is lost! So don't forget to write down, how and where to do homing! If you need it some years later, all is forgotten and lost in time... [Blade runner]
13:16 Yes homing against a fixed block is usual today anywhere, but I don't like it! I still think there's an overload to the ball bearing nut of the spindle, even with reduced currents, because of momentum! I prefer and recommend switches ! And I know how to use them, to bring all accuracy...
14:44 keyed shaft: You use them for conveyor belts mostly! If the motor is turning one direction only. But in a CNC with turning left and right, you will always see some tolerances in that key piece. Which will bring you faults if clamping isn't sufficient. So use sufficient clamping instead and get rid of that key piece!
Great work though! Have fun
10:35 I love the "ethercat"
Great work! Thanks for the mention, I've forked your repo to have a look at your homing module. Its quite amazing how far Linuxcnc's ethercat support has come over the last few months. My ethercat mill is going great tool
This is peak content. Deep technical dives, machining, robotics, bang-on humour, and none of that perpetually overhyped inyourface yelling with constant sponsor influence. I may have just found a new favorite channel! Thanks mate.
Cool to see some more vids on ethercat :)
Also I don’t think 30% passive z axis load is that bad. The machine I run at work sits at like 45% which is also completely normal for it
Great stuff Marco
Great machine you made there, can't wait till 4th axis!
Great video, thanks. The hip replacement business line made me laugh :)
Sometimes freewheeling diodes are not used on brakes because it slows the engagement time, though probably not important in this application. I'd be surprised if the drives didn't have some sort of protection built in.
Yeah I agree but i must say that I dee servos in industrial application s that have better EMI shielding so the brake has less of a disruption then for Macro
@@bashaaksema94 indeed, it is kind of odd they combined it all into one tiny connector.
I *love* the soundbite at 0:33 😂
good catch!
Nice rittal cabinet mate 🔥 as an industrial cabinet designer i think you did great
been missing this series. Manna from heaven!
Finally, another Marco Reps video!
A real industrial style enclosure setup... Great job!
Phew.. Thought for a moment you quit TH-cam-ing.. Always looking forward for your videos. I'm currently building my own CNC so every bit of info from you is priceless. Vielen dank!!!!!!!
You're getting closer and closer to the point where I'm just going to want to order an entire kit that replicates what you've made. It's wonderful that you've turned your considerable attention to detail towards DIY CNC. Watching you put together some excellent control electronics is great, but seeing you methodically chase down what will no-doubt be nanometre-precision machining is sublime!
It is really starting to look like one of the best DIY setups out there, but I'm a bit afraid I don't have even 10% of Marco's expertise to follow that.
Respect Marco. Serving information in your manor isn't an easy task nowadays. By far my favorite creator. Looking forward to your new video ;)
Wow, what an amazing video. Thanks for all the information!
Thankyou for sharing on github. I will definetly have a peek when i'm messing with ethercat drives
Love the CNC content
Your level of knowledge on things like this blows my tiny mind....
Great stuff - both the hardware and the video!
Satisfying af. Stay safe, Sir.
Great video as always! Love the humor! ❤😂
Next to always great content I m enjoying the SFX greatly hehe
Hi been enjoying this series. In the stall home setting mode that you used , Meldas ,in their home setting sequence would then back off of the hardstop by about 20mm
Fanuc had 2 marks 1 fixed and 1 moving . position the axis the required distance away from the marks (if I recall about 3 rotations to find the zero/index mark)
then jog past the marks.
Builders would include urethane bumps at the ends to prevent metal to metal crashes
Marco, layman's comment here; use lubrication when cutting aluminium. Otherwise your endmill chokes as an ultimate result. In the meantime your surface finish deteriorates. You can unclog your endmill using a strong base solution. But that doesn't correct your lost workpiece and heart damage when the endmill suddenly decides to hold position and kill all movement... Wonderful content as always magician. Best! Job
EtherCAT is a remoto I/O system which run on a 100 MBit network. What starts (and ends) with a normal network card by sending a ecat package what goes to all ecat modules ,in series, and then back to the same network card. The time delay is one bit (10 nS) per module. The outputs are set when the stream is coming by and the status of inputs is filled in (in the stream), so the returning message has all data of outputs and inputs.
If I were your axis control unit I prefer the constant load by gravity over the changing load by gas spring compensation. The later will only be perfect in exactly one position. Maybe add some lightness to your Z-Axis.
You might be right! I've done a bit more manual tweaking in the meantime and got better results with the same static load
@@reps Former industrial CNC field service engineer here.
The practice with industrial cnc is either a constant force counterbalance or a servo motor large enough to handle the weight.
counterbalance can either be a weight and chain system or a hydraulic cylinder with a fast acting counterbalance valve (actual product name).
both of these counterbalances systems are used in relatively slow moving systems.
For high speed and small rapid feeds, the motor only is the most common approach.
Also,the vertical axis will sometimes use a lower pitch ballscrew. I see that your motors have a 3000rpm rated and 6000rpm peak,
will a lower pitch work for you?
An other consideration of Z-axis cut quality is the stiffness of the Ballscrew and the support bearings . Any amount of float
will harm cut quality , and an incorrectly set up counterbalance will make it even worse . the counterbalance lift should
always be less the the weight of the spindle assembly.
Finally one of the most cute and clever counterbalance systems I have seen was installed on a sherline mill.
A small bore air cylinder connected to a 2L plastic soft drink bottle . the system was charged with enough
air pressure to acheive the desired balance , and because the 2L bottle is is both expandable and a much
larger volume than the cylinder it is almost perfectly constant force and Zero inertia.
@@reps Some drives allow a tuning parameter called motor bias that is meant to provide a constant torque to compensate for things like gravity
That quiet "Dumb ways to die" killed me xD
Beautiful cabinet! 🎉 Even passes my OCD😅.
Rittal enclosures are so nice 🙂 Looking good there!
Definetly
Interesting insight on how to setup a CNC mill
Just did a post on LinuxCNC forums about this.
Thank you.
mr. dog shoe guy I take it? thank you :)
@ YES ! 😃
Wait, what's your nic there?
First of all, thank you very much for the entertaining video!
Just a few little things that I noticed:
1. I would recommend that you fuse the secondary side of the power supply. In the best case, I would use a selectivity module or at least a semiconductor-based fuse. For example, the LOCC box from Lütze would be an option for you, as you don't have a lot of space in your control cabinet. Depending on your power supply's peak current, you can also use a normal circuit breaker to achieve a decent level of protection.
2. I would also recommend grounding the 0V directly at the power supply to detect and switch off short circuits of 24V wires against the housing.
3. Since you are from Germany, I would also recommend adapting the wire colors next time. according to EN 60204-1, red is for AC control circuits (24V DC in your case) and black for AC load circuits (0V DC in your case). It would be more correct to design the 24V wires in dark blue and 0V in dark blue / white or also in dark blue.
Are you switching the STO via etherCat because your connector is not plugged into the servo drive?
Some fire content right here 👍🏽
Ive also made the journey of trying to neatly fit everything in compact electrical cabinets. Nowadays Im very glad that I can comfortably sit with a collegue inside every machine's electrical cabinet, it makes many things much easier. But I also now have the immense floor space that this requires. Its very crazy to see to how much performance you get nowadays for very reasonable money
That's cool to see Beckhoff show up here, used to use them a bunch as a cheap alternative to national instruments that also didn't involve that awful labview software.
Looks great!!!
I love EtherCAT once you get it working!
another dope video!
I am indeed envious.
Excellent
Just awesome!!!!
finally some cnc content again ;)
Great video Marco.
Do you do any consulting in regard to cnc control systems?
Such wonderful toys!
I like the RC cross wrench in a drill that's pretty funny
finaly a usefull vidjeo
"That's right! It goes in a laser-cut square hole!"
😂😂😂😂😂😂😂
"Quirks and Features". 😂 Love it.
Waiting for 5 Axis, but great work!
Maybe the anti parallel diode for the brake is directly integrated in the motor.
But a tasty video again.
Amazing video yet again. Two questions - you show the use of an MPG pendant - could you show which model and how to hook it up to the EtherCat bus? (or do you run it off some non-EtherCat leftover in your LinuxCNC?). And secondly, where do you feed the RS485VFD comms from? Is it from some EtherCat to 485 interface or also from a leftover pre-EtherCat previous interface card?
Hello,
you made a really nice video, very good work.
I would like to make an identical assembly, but with the EL7 series because I need a 3ph 400v power supply
Is your github config for the EL8 series also usable for the EL7 series?
I have difficulty understanding the differences in functionality between these 2 series...
regards
I'm just working also with an EtherCAT PLC and a Spinogy Spindle. I noticed that omron does make EtherCAT Modules for their MX2 devices. Maybe you could add the spindle also to the EtherCAT control? I didn't tried to control it that way for now, but evtuelly this can be nice too get even more feedback from the Spindles Inverter like Current/Torqu or load on the mills.
Anyway very nice Video! The multiplierer Hack for the Omron Inverter maybe makes it into our CNC-like Machine. Thanks for sharing this!
As others have said, and I think I've said before on your videos, I barely understand what's going on, but ayy lmao, this time I think I understand a few things from my prior experience as a CNC machine "operator" (glorified robot babysitter, if I'm being honest) - I liked learning about the maintenance and behind-the-scenes stuff. Certainly in part thanks to my roommate whom is a controls engineer.
An alternative to a gas spring would be a pneumatic actuator with a fixed air pressure supplied via a canister mounted elsewhere on the machine. I own an old MHO Series 18 CNC and this is the solution it uses. Works well. I could provide photos if you'd like.
For fans on those cabinets, is positive case pressure (fans on intakes instead of exhausts) with intake filters better? PC builders do this to keep dust from creeping in the cracks.
Marco !
Ethernet truly is the best standard. I’m seeing computers being released that are ditching specialised connectors for high-speed and high-power USB-C, which just ticks me off. We already had a high-power high-bitrate standard. I want to charge my laptop via PoE, I want to connect to my monitor by its IP address.
Cascading B16 circuit breakers without selectivity brings no benefits. It's likely that both the B16 in your cabinet and your breaker box will trip at the same time, or randomly.
cool encoders
Marco, this might be one of the sexiest vids from you. Thank you. Looking forward to order a new hip from you!
You can still switch easily to RPI5.
To get extra ethernet ports you could just use din rail 4-5 port switch.
You might even get din rail mounted RPI5 enclosure.
Niceeeeee 😀
Have you tried the new generic CiA402 driver written by Scott Laird? Scott(and others) have poured a ton of energy in to this lately and it looks very promising! I'm going to try it on my CTB drives later on when i'm done with the wiring of the controlpanel for my machine...
It would be great to see a bill of material or parts list. Great video
3:37 All who has OCD will love you for that!
Could the z axis surface markings be due to z axis tramming? Awesome content by the way👍
not sure i can wait another 5 years for that hip replacement marco.
Nice video👍👍👍.
What is the difference between these ELM2 series servo motors?
ELM2S
ELM2H
ELM2M
ELM2L
Hi Marco, look into Triamec servo drives. You'll like them.
Beautiful work as always. Those encoder batteries are really slick and I'm curious how they accomplish such low power draw. ASIC?
Ye it's interesting, unfortunately the ICs are all anonymized. In retrospect I think I might be clipping on the current range while turning the motor by hand, so it could be some deep sleep µcontroller that only wakes up with encoder activity. But still the encoder must be powered all the time and that alone is remarkable at this low power
@@reps Usually its a Multiturn Absolute encoder which knows the absolute (angle) position of the motor and just keeps track of the number of full turns, so it just needs to update once the encoder rolls over. Thats at least how Omron does it.
@@bouni8506 multiturn absolute encoders don't need battery backup. They typically use a weigand sensor to keep track of multiturn counts.
@@shumanchu1Interesting, I'm not an expert in encoders and that's what our Omron sales rep told us how these encoders work. Nice to learn something new , thank you!
very nice set-up , fast easy to plug you will good for 10 years, :) or more because this will be the new standard no analog no pulse/dir with another command juste a single wire to hook up all :)
How does the leadshine "one-click" software handle tuning on a gantry style machine with two motors for one axis? Is it pretty straightforward? Love the videos!
That's what I was also wondering, as I'm looking for some drives and motors for a gantry style machine
30% static load seems pretty normal tbh, no? our cnc lathe at work is also sitting around that just holding up the turret; no gas spring to be found there.. i think anyway lol
My grandfather always used to say that the most provisional part of any system always last the longer, guess those 3d printed supports are proof of that!
Jesus that 3:38 reference im dying laughing here :D
I think on those Rittal enclosures you can flip the door hinge to the other side.
Got to know the CiA-402 statemachine stuff through the controllers made by Nanotec.
Now that looks like Made In Germany
What was that about adding a glass scale? Very interesting motors!
15:35 thats exactly what dr. dflo did to his large format 3d printer the other day to support a massive pellet exturder z axis, might be worth looking into
0:33 That reference is especially relevant considering one of the dumb ways to die was "doing your own electrical work".
Couldn't stop myself from giggling when I Heard "dumb ways to die!" in the background of the into, and so so true :D We've all made those dumb decisions on a project in the name of time/testing.
8:59 the diode doesn't have to be directly at the coil. It is desirable to have the freewheeling diode at the node where the dI/dt is the highest. This is usually at the switching transistor.
Metallschrauben im Kabelkanal🤯
Nice doug demuro joke you sneaked in there
You can't escape the algorithm
Could you say specifically what model did you use for the phoenix distribution blocks, please? :)
Isn't top surface finish mostly dependant on the angle of the Z axis relative to the XY plane? I don't see how new servos could improve that