3018 CNC End Stops and Calibration
ฝัง
- เผยแพร่เมื่อ 4 ก.ค. 2024
- I had some requests for a video showing the installation of end stops of the small CNC engraving machines like the 3018.
Please note that all new videos on topics like this will be on my new channel: / @solderandsawdust
Blog Post by Bogdan Berg - www.bogdanberg.com/2018/09/07/...
3D Printer Files - www.thingiverse.com/thing:279...
Required supplies
M3 T-nuts - www.amazon.com/gp/product/B07...
M2 and M3 bolts and nuts - www.amazon.com/gp/product/B01...
Roller Switches - www.amazon.com/gp/product/B07...
Female DuPont Connector Jumper Wires - www.amazon.com/gp/product/B01...
Some Hookup Wire - www.amazon.com/gp/product/B01...
Heat Shrink Tubing - www.amazon.com/gp/product/B07... - แนวปฏิบัติและการใช้ชีวิต
4 years later still helping people like me 🙏 legend
Very explicit and to the point, a nice piece of teaching and impartation of knowledge.
2 weeks ago, i bought this cnc 3018 and i just knew that "unlock" is intended for end stops. Thankyou
Excellent information. Thanks for sharing this with us. I learned a lot of things that will help me as a beginner.
Really nice video. Covered a lot of things, Cheers.
the best explanation regarding the calibration. Longtime looking for such a convincing way to do it right. Thank you for sharing knowledge.
Ouch, this calibration method is a bad advice unfortunately. Sad that this is taken as a rule of thumbs.
If your tool is 1.9mm diameter instead of 2.0mm diameter, you'll measure a square of 100.1mm instead of 100.0mm. And you are going to multiply the 0.1mm error to all the dimensions of your part : the farsest you are from the origin, the biggest will be the error of your part dimension.
I actively recommand to use a "calibration cross" (look at thingiverse, this is good for a 3d printer and a CNC calibration).
The calibration cross comes with an Excel spreadsheet, you'll have to take several measures.
The spreadsheet will give you two numbers : the constant offset (this is the tool diameter error) and a linear offset (this is the step/mm error)
Hope I could help you :)
What an awesome video!
Thanks
THIS IS A LIFESAVER
EXCELLENT! I've been looking as how to calibrate it since it cuts way too small, almost 50% too small. :)
Thanks for the link, very usefull, also your video was great
VERY good. Thank you. Jim
Great tutorial Thank you!
Excellent
Hi, could you tell what size of cable wrap you are using? And don't you have a pullup resistor
on your microswitches and a capacitor to avoid fals triggering? Does your Wodpecker PCB
allready have the pullup resistor and capacitor onboard? I just bought a 3018 with a Cronos
Maker PCB, hell i whish there would be some detailed information on that thing, to make
things even worse there are a lot of different versions of the board i have... Thank you
for sharing. Cheers!
There are a ton of different boards on these. I'm pretty confident that the pullups are just above the header pins see 4:16 I haven't line by line gone through grbl source but the 328 chip has internal pullups as well if enabled properly. So far I haven't needed to cap the switches but it really wouldn't hurt. Usually I use a 30pf paired with a 10-100uf electrolytic with switches that large to avoid bounce. My lead wire to the switches is twisted around itself for most of the run to also help with noise. I ran my spindle wire separate from the others as well see 1:27 The cable wrap came with it, but it looks like 6mm. I'm going to fiddle with this board a bit more and then upgrade to a different board and use trinamic steppers.
Thanks very helpful
with the right bits and lubrication setup will this handle 6061 aluminum? trying to make a few parts for my ar15
Not accurately. With very shallow pass depths they can cut softer metals but they really don’t have enough rigidity to be accurate.
Great video. Thanks! My only question is how do you unlock the machine. You make mention of this when editing gcode. Can you explain how you know if the machine is locked and how you unlock it?
It's a software lock that GRBL machines go into when things go wrong. It stops the machine and most control between the PC and the machine to help keep from wreaking havoc. If you look at time 7:04 in the video you will see the icon in Candle. Other senders have the same sort of thing. Usually the icon will go red. Most definitely the machine will halt and refuse to take commands from the PC. You can also send a $X to unlock. See 'github.com/gnea/grbl/wiki/Grbl-v1.1-Commands' for the official reference.
Are you using the Normally Open or the Normally Closed contacts on the micro switch?
I am using them Normally Open.
Awesome Video thanks, using this as a guide to fitting up my endstops and calibrating my 3016. (have had a few overruns its not good for the machine)
I note in the GRBL "handbook" that $$5 is for limit pin invert. would it be possible to use that to change the normally open contacts to using normally Closed (which on blogs seems to be the preferred set up.
Again, thanks and a great vid
Happy you find the video helpful. According to the grbl docs yes the $5 variable does just that. Off the cuff I would think that an an active high (trigger on low) could help with spurious line noise but would need a smidge more debounce time / filtering (variable $26). It's just my personal experience with arduino projects that trigger low switches are just a shade more noisy than trigger high when using the internal pullup resistors alone. It would depend on the quality of the switches also.
@@everlitesv i am very new to this so will go as your vid outlined. and it works amazingly.
Now to the fun part of actually learning how to use it
You only need one stop on each axis at the zero end. You can use soft limits for the other ends
You are correct. It was one of those 'while I'm doing it, i might as well' moments.
Can i know how the best feed rate and plung rate setting for this machine
The feed rate I used in this video is 100 mm/min with an 80mm/min plunge, spindle at 100%. That was in softwood (pine) with a 3mm 2 flute endmill and a pass depth of 0.5mm. I would say to go slower with harder material. The 'Best' rates are a more complicated answer. In theory feeds and speeds can be calculated mathematically, but that math assumes a near perfectly rigid machine. These little machines just don't have enough structural rigidity to handle proper chip loads that the tools are capable of. In my larger cnc I can use that same 3mm tool at 16k rpm, 800-1200mm/min feed with a 1.5mm pass depth (in softwood) all day long.
They also have websites where you enter the stepper motor information and it does the calculations for you
Can you share this info please?
Hello Friend, excellent video. I hardly find someone who uses the same grbl controller as me. I can tell you that I am in the process of upgrading my 2418 CNC. I bought it. First the laser broke, then I bought an Xcarve and a 40 Watt Co2 Chinese laser. But now with this unemployment I set about turning my 2418 into a superminiCNC. Well I will tell you that the steps of the X and the Y axis are correct in 800,000, what happens is that you did not subtract the diameter of the milling bit with which you cut the sample of the 100 X 100 mm square. Actually the perfect short machine 100 mm. But the GRBL controller is just a Gcodes reader. In the program used to generate the Gcode, Example must be assigned. any bit, but the cut must be outside the line, so that the measurement is exactly 100 mm. If you assign drills on the line, you must consider that the diameter of the bit will be 50% greater than the total value of the diameter of the bit. remember. over the line. minus 50% X & Y positive plus minus 50% X & Y negative. For a total of a full total diameter of the bit in use. And if you are cutting inside the square you have to add double, because the bit is cutting completely inside. This is a common mistake. The vast majority of the time they do not take these factors into account. And there is speculation that the equipment malfunctions. Believe me I learned this hard. Making dies for boxes, where the human eye could not see a thousandth of an inch. And the machine used 1/32 bits to engrave the wood and make the grooves for the cutting blades.
I'm not sure if I read you completely but, yes it needs to be an outside cut in your CAM software.
I have a problem with the endstop. If I run x y or z manually with the software the endstops work, but wenn I try homing position the z goes up and over the endstop. I can't figure out why?
That is a known problem if you have an older board from grbl 0.8 era. In 0.9 the pins (D11 and D12) were changed to allow for variable spindle speed. Here is a video showing one way to work arround. th-cam.com/video/aXcsRKDZsv4/w-d-xo.html On my board there are two pins that work for each axis limits I just didn't install the lower Z. If you look closely at 4:16 in my video there is an empty space between the Yen and Zen, that is the second Z on my board. Boards on these can vary a lot though. Google 'grbl z limit pins' and there is a bunch of info on this. Myself, I would first try to use a multi meter (with the machine off and unplugged completely) and continuity trace where pin 12 on the actual ATMega chip come out on the pin header.
For better working you have to use coaxial cable( antenna cable) protected from electric noise caused by the spindle. In some machines when you turn on the motor it can cause the fake activation of an end switch
I agree. It can get bad enough to cause missed steps as well. I just haven't made it that far in my mods yet. Small (30pf) capacitors across the terminals of both the spindle and separately on the limit switches can also help with this. And maybe a 100uf electrolytic on the spindle terminals as well. I'm also looking at some stepper smoother boards to prevent false / skipped steps from noise.
Where do you connect the shield to, as I can't find an earth connection on my machine?
@@anthonygregson719 If you are referring to using coaxial cable, for the limit switches I would use the center conductor for the signal side and the shield for the other. If you look closely at the woodpecker board the entire bottom row of header pins are ground. Physically keeping the spindle wires separate also goes a long way to solving the noise problem. If you really wanted to ground the machine you could run a wire from the chassis to a metal water pipe or something like that and the ground to the chassis. My particular machine isn't that terribly noisy but I ran my z axis stepper and limit switches separate from the spindle wires. There are some 3d printable standoffs to keep them separate as well. Personally if my motor was noisy enough to cause consistent real problems I would put a couple of caps across the motor first.
Hi I want to try that CNC with Alphacam. Do you know if that controller is compatible with fanuc (like the mech 3 software ) or other thanks .
I'm not sure what post processors are available on Alphacam. The stock board on theses is GRBL based.
It would be nice if you could tell us the machine coordinates you end up with in the gui when homed. Also the exact pins for min/max connectors. I notice you $130, $131, $132 seem a little off. How important is that setting?
At 4:19, I have a close-up of my board (not all are the same) the pins are labeled Xen, Yen, and Zen. In my setup, I put switches on both X sides and both Y sides with only one on the Z.
The $130s are for when you are using soft limits. When you have the physical limit switches installed, you use hard limits. But, if you are using soft limits they are very important.
At 7:11, there is a shot of the coordinates after homing. All are zero. That is what homing with switches is all about. We are defining that particular position as 0,0,0. Now, when you position your bit over your actual workpiece and zero that out, then the work and machine coordinates will no longer be the same. I hope this helps.
Very good content! However the camera movement combined with the not working autofocus are a really rough edge
hi
everlitesv , is it the same as the arduino uno with shield v3 board ? about calibration settings. (cnc 3018 nano board no detect / broken, replace arduino uno board with shield)
If I understand you...can you replace the Woodpecker board with another arduino based CNC board? Yes you could. Unos with a CNC shield (3 axis) or Nano based CNC control boards should all work. Some CNC shields need special attention to some jumpers though. I'm not 100 percent on the polarity/phase of the motor connections as they are so I would check that against what your replacement requires.
thanks
Stepper motor can rotate not smooth :)
Hopefully you are still monitoring this. I tried your G-code settings for setting limit switches, changing to =1 and switches
worked great. Unfortunately, the spindle motor no longer turned on and I could not "send" the program to the machine.
Changed the settings back to =0 and all was well. Any suggestions??
That could depend on the version of GRBL that is on your board and how old (version wise) that your board is. In grbl .09 there was a pin assignment change that in 0.8 was for a limit switch and in 0.9 and newer that pin is what drives the spindle. That usually just messes with the z axis switch though. Some boards can get the benefit of homing by setting $22 = 1 and setting $21=0 to turn off hard limits if that works for you. Once a proper homing can be achieved the soft limit functions will work for better safety as long as it is enabled and the actual max travel dimensions are saved to the firmware($130-$132). If you are running a newer firmware like 1.1e,f or g also check $32 and make sure it is 0.
@@everlitesv Thanks.
@@everlitesv Checked, I have V08.1.0 . dated 2019. I will try changing the =1 settings one at a time and see what happens
@@everlitesv Thanks for responding. I fixed a number of problems by upgrading to Candle v1.1.7 and grbl 1.1e. Can't go any higher as Easel only works with the earliest v1 of grbl. Have looked at Denvi/candle.wiki on github and nowhere,
starting with v01, does he explain the function of the keys. Some are basic but a Vid explaining how each works and the later upgrades would be welcome. Do you know of such a site?? I'm sure a lot of folks would pay for that.
I try posting questions on the wiki but never get a response.
@@johnhill2594 Thanks I will look into it.
Hello Friend. I am in a project to make a cnc table. I still have to wire the race limit switches that go to the card. I have a diagram of the card and the limits. You can help me ? I can send it to you by email. I would greatly appreciate it
In my opinion
If you measure the inner square, you Miss a "half of diameter" router bit.
Any way 👍 clear sound! No annoy music.
Intuitively it would seem that way. In the CAD / CAM software, you define the actual bit size and make a tool path that cuts on the outside. The CAM software takes the tool diameter into account when calculating the final tool path for the g-code.
when trying to cut a pattern I set up the machine and set x & y to 0 at the center of the pattern. When I start to cut x axes goes to stops on 0 end. Nothing seems to change this
.
Make sure you have the origin in your CAM software matching where you zero the origin on the physical machine for each project. I usually work from the center and top of the workpiece myself. Also, make sure your cut area in your CAD/CAM is not physically larger than the machine. If you have never had anything cut correctly make sure your steppers are driving in the right direction.
@@everlitesv Everything you suggested is right. The pattern is only 4" square. everything is going in the correct direction. Is there a way to reset the machine to factory defaults?
Do you happen to know where I can find end parts for the other side? Like right X axis?
The end stop/limit switch mount parts were 3d printed. Many of the other parts are available on Bangood and Ali Express depending on what model 3018 you have. On this one they are just 25x25 extrusions available in many places, just not cut to length. The 90-degree braces and pillow blocks that shape the gantry and hold the 10mm rods are available in many places including Amazon. I have seen the plastic corner brace on both Thingiverse and Bangood.
@@everlitesv thanks for the reply! I saw your link to the thingiverse page but there are only end stops for one side (like X-axis left and Z-Axis top). Or am I missing something obvious? For homing one corner is enough but I want to protect my machine in both sides
@@petermuller608 I used a program called Autodesk\Meshmixer to create mirror images and relocate the flat surfaces to print better on my 3d printer. This way you have a right - left and front - back mounts.
@@geraldwardell4873 thank you I will try that!
what happened to the print
I have the same one come how hard to build
Not very hard. Just go slow and make everything even and square. Using a metric tape measure with a mm scale helps.
How do you connect 2 limit switches on one axis,I mean the diagram.
At 4:17 in the video it shows the backside of the control board. It has pins for Xen, Xen, Yen, Yen and Zen (there is actually a secret second Z axis pin for the bottom). The top set of pins are your hots and all of the bottom pins are grounds. Just hook up your X axis switches to the Xen pins, your Y axis switches to the Yen pins and the Z axis to the Zen. (On this control board) the order of each pair does not matter and the polarity does not matter. The $ settings set up which direction things are.
@@everlitesv thank you.
You said to solder wires to the COM and NO......shouldn't that be NC?
Thanks for sharing!
I just verified that mine are connected to Normal Open (NO.) Seems like firmware has the pins trigger on a high not a low. NC should lock the machine in a continual alarm state. Could be tested with a single jumper wire.
Edit that. I suspect I have the pin trigger state backwards but not the physical connections on my switches. They are definitely Normal Open. I'd have to verify with a meter but the one row of pins is most likely always hot and when the switch closes it shunts to ground. Could look for it in the GRBL source code as well.
@@everlitesv You are correct... I just tested it and it would put the machine in continual alarm state as you stated if wired NC.... I stand corrected! =:)
Hi and thank's for your video.
Unfortunatly, like each many tutorials i see, I do not see any explanation of the wiring connectors on the board.
It is here the everything starts. As an electronician, this is the minimum required before even we talk.
I see many people burning their boards because of the many confusions made and the non-understanding of things.
Most of the time, each tuturial is made just to explain how to turn 2 or 3 screws!!!
Plus there a confusion : i see on my board 4 connectors : 3 for x y z limits and one called STOP.
Who did explain clearly what are all thoses?
Thanks for the comment. My appologies if this didn't cover your needs. It was my first 'real' video for youtube so I made some mistakes. One reason that I did not go too deeply into the board connections is that here are several different control boards out there and they are not all the same physically, but electrically will be the same or GRBL will not run on them without a custom compile. Also, at roughly 4:12 in the video I make a very clear shot of the connectors on my board.
On my board the entire bottom row of pins are all ground. That may be different on yours, but by default the pins on the MCU are active low and they can be tested with a multi meter. The connections should be made with one leg of the switch tied to ground and the other leg to it's corresponding labeled pin(Xen, Yen, Zen). In your case the pins are labeled slightly different with just X,Y,Z. The one labeled STOP on your is most likely for an emergency stop button and would be wired the same way. There are other pins to use as well. There should be a pin available on most boards for the Z-probe.
Even if the pins were unlabled, you can test the pin config using a volt meter and verifying that the all +V to ground while watching the console output. The voltmeter itself may allow enough signal through to trigger the various tested configurations in the software so also keep an eye to the console output while testing. Another way would be to use a 10k resistor (to limit the current below any damage levels) with a female dupont cable and probe each pin to ground while watching the console for output.
I hope this helps.
Oh my God.....not another mobile camera shot....I'm getting quite motion sick......bah.
It is like a rollercoaster! Make the video with photos instead!