End Effector Tilt on Anycubic Kossel + interesting developments in the world of 3D printing
ฝัง
- เผยแพร่เมื่อ 27 ก.ย. 2017
- While working on getting auto-bed leveling working, I figured out that the end effector tilt is undeniably a problem. Also, I talk about some really neat and brand new innovations in 3D printing.
My video about inductive sensors being heat sensitive: • All 3D Printer Proximi...
Josef Prusa saying exactly what I was saying only 5 months after me:: • First Look at Original...
Just want to let you know how much I love your Anycubic Delta series. I don't think anyone really spends as much time as you do on the Delta and it's a huge help to those of us with one.
Hey I'm having. This weird tilt issue also and I'm getting from this video that magnetic zero tolerance, effector arms is the solution. Is this correct?
Actually the delta can be a lot more exact and cheaper. Just remove those angle arms, and just add two regular shafts spaced . You just have to re calibrate the arm length , there is no need for the arms to go on same center point as long as it is not offset too far as all joints still will be equal. This was my " ahaa" for a while ago, and i removed all wobbly things and just drilled two shafts angled 90 degrees from eachother and works much better
On my 3rd set of rods from Anycubic, going magnetic as well. They have been nice enough to replace them, just going nowhere though.
Any way we can get an update on the magnetic rods?
How repeatable is your bltouch? What's the standard deviation? Thank you
VERY HELPFULL THANKS
If you use springs between each pair of diagonal rods, it will help take up any play in the eyelets. Use a pair of springs, one near each end, per rod pair.
Kudos on your inductive temperature variance! When I saw the video I agreed totally. I got new 10-36v inductive sn04 sensors after your video. Btw, although my heated bed is aluminum, I line it with a thin steel plate 22G under my glass surface. The fact that sensor reads the steel and not the hot/cold aluminum helps the sensor enormously and it does not affect readings as much. I think this is why Prusa is using steel plate now.
You should be using these they are Capacitive so they sense the glass, wood your finger whatever gets in there way and best of all they are adjustable and damn cheap. www.lazada.co.th/1-10mm-approach-sensor-cylindrical-capacitive-proximity-switch-npnno-dc-6-36v-intl-29598510.html?Orderconfirmation&Email&LineItem
my problem disappeared when I got rid of that trigorila board - which had multiple issues including forgetting the Z height setting and went to a re-arm with lv 8729s with Marlin 2.xx been like it 2 years and not had to recalibrate or had an issue since
Did you make a video of the mag etic rods etc? If so which is it?
I have had the worst luck with getting new magnets in the mail to replace the ones that I broke. Hopefully the latest order will arrive soon so that I can finish the build on the Anycubic Kossel.
Design Prototype Test thank you for your reply. I look forward to your video on the magnetic links, it would be interesting to see how much weight they can cope with if you decide to add extra heads and fans etc
I am very happy about my Anycubic Kossel Pulley version from gearbest after some adjustments: The GT2 belts should be a little tight, like a string , and the spring tensioners don't work. I use the adjustable rodmounts from banggood.
The Rods should be pushed together with a spring or rubberband, eliminating almost any play in them. The Pulley should have a counter part or the wheels tilt : thing:2494178.
I am now trying to get AutoBedLeveling to work, I use the Optical IR Probe works great +-0.02 mm ish, but not the Marlin 1.1.5. adopted by ? to Kossel Delta. it reads/computes rubbish, however very consistant, -0.5 mm - + 1.5 mm over a glass bed . Will test with vanilla 1.1.5 or 1.1.4.
Unified Bed Leveling I try with all points zero.. and there's a naughty bump at one point + 2-3 mm printing . Head goes into air like a bump!!
I used this to calibrate mine, great help. www.thingiverse.com/thing:1837929. BTW Min_bed_temp should be 0 or you get Bed Temp Error if no HotBed. Or disable HotBed sensor.
I would like to see tech stuff as well/ Like designing, building things. Also building and programming a CNC machine would be cool. I'm thinking of making a laser cutter for foam board (for RC planes) and have been looking how to do it.
Also: David Crocker (developer of the Duet3d controller board) fought the effector-tilt problem on one of his builds. He wound up putting a bulls-eye bubble level on his effector to see the problem. His blog is here: miscsolutions.wordpress.com/2016/06/25/more-delta-printer-upgrades-wifi-and-silence/ Check the image at the bottom of the page.
I just noticed this problem last night i.was having a hard time leveling and saw my effector tilt when it was probing yes it actually was shifting during the probing. I saw you commented about how laborsome EZ bed is really I used it once and got 20 prints off before needing to calibrate again. Hopefully the magnetic joints will work there is a link on thingiversw for blue eagle labs rod arms check it out it has effector and carriage mounts.
I also think smoothie might be a better option you can get an MKS smoothie board for like $35 bucks the community finds it cancerous but IDC personally. I am hoping once I get this thing stabled through some of your efforts I can up the size in the X and Y direction and maybe the Z
Either way great observation and I look forward to your magnetic joints
Harrison Hawkins Thanks Harrison! Between you and Levi Blaney I can't tell who is my greatest supporter.
Design Prototype Test No problem I notice great content when I see it and this is better than great!
Good luck with your magnets, i try 2 years ago same kind chinese magnet ball, work but only with slow accellation and slow mouvement, if i push the print speed faster then 40mm...the effector fly away
Christian T Ha, ha! I thought that, then I read your comment. I think maybe a way to fix the play is to put a thick higher temp grease on the ball joints. Have to get it worked in behind the housing that holds the ball.
Why when you show products in the video, is there no link in the description? to same.
I had the same issue with two of the ball end joints. I was able to fix it by drilling a hole in two pieces of flat aluminum, sandwiching the part in a vice, and tightening just until no play is felt. I let anycubic know about it, but who knows if they did anything.
Can you please explain this better? I'm quite interested
@@stellari81 The idea is to squeeze the covers of the ball tighter.
I have a question.
Since you got the Linear Plus, why didn't you use the brass spacer on the effector and the top mounting
Without those brass spacer, the rods will not move correctly.
It's been a while since I was thinking about this printer, but as I recall all that matters is that the distance between the pivots is the same at the top and on the effector. Because I upgraded my vertical trolleys to be aluminum for the linear rails, I had a narrower spacing, and therefore I removed the brass spacers to maintain consistent spacing.
Hi, any chance for me to get your cooler modeling?
I've seen people set up auto leveling using an electrified bed/nozzle tip before so that idea isn't exactly new. But yeah the pcb idea is definitely a better way to go about it.
And flexible build plates have been around a long ass time. But PEI coated spring steel is amazing; PRUSA never fail to deliver. Can't wait to order the MK2.5 upgrade kit for my machine!
I kinda want to buy a delta now after watching all these videos just to mess around XD
It's totally different. Those old electrified designs required a metal printing surface to complete the circuit between the bed and the nozzle. This design senses the physical touch of the nozzle to the bed, but instead of activating a switch the strain gauge registers a change in its reading (like a scale with a weight suddenly applied) and the controll board knows the nozzle has just touched the bed. It will work on glass, PEI, BuildTac, etc. The problem I realized after I made this video is that strain gauge printed directly into a PCB might be extremely sensitive to temperature changes. I would love to get my hands on one to experiment.
Design Prototype Test, I had the same thing, I just decided to go as someone told me here to use the EZbed leveling system just for this printer, it works perfectly. Does cost another 30-40 USD, but cheaper than the BLtouch and works.
This system works with these sloppy ball joints because it probes directly under the nozzle tip which largely eliminates effector tilt issues. If EZbed wants to send me a unit to test, I could possibly say good things about them, and apparently their system does work. However, I am not all that impressed. Mostly because you must be connected to your computer, and you have to place the sensor by hand. It is very effortful. Once I get the BLTouch working it will just do the auto calibrate automatically before every single print with no extra effort on my part.
Can this be solved with a preload on each pair of rods? At minimum help relieve the problem. Like a rubber band between the pair of rods.
Edit: the magnets would accomplish the same thing, IMO.
I've seen this patch in use before. I'm skeptical of it's effectiveness. How do you ensure absolutely the same rod lengths when you are making the rods? It's not easy to pre-load with wet epoxy.
Design Prototype Test if you want precision, I believe linkages (ie. Deltas) are difficult to manufacture.
But pretty much all systems would require some feedback or other form of control to appropriate the error. It's up to us to decide the degree of such control.
I apologize. I sort of skipped through a minute or two of the videos. I might have missed a crucial bit of what the exact problem is and what you want to do with it.
What your running in the beginning of the video is the bed leveling routine. What you need to do is the delta calibration routine.
John, thank you for the firmware that allowed me to double check my work! It was very helpful. I ran the G33, with yours, and it gave me the same values as I was getting from my build also with G33 enabled.
is that the G33 command equivalent
Dan Mateescu yes
Is this firmware posted online anywhere?
I can post my firmware. It seems to be functioning perfectly, but I have not been able to verify - my printer does not work as you can see in the video. If you want a copy of that, please go and donate to my Patreon page. This helps me to keep this channel going. Google Adwords has paid me about $0.01 per hour. Without patreon support I won't be able to keep making videos.
Just a couple notes because I love the idea of getting the bugs worked out of this super-affordable machine!
The critical things about the uprights is that they be perpendicular to the bed, equidistant from the center, and each evenly spaced 120 degrees apart. Hard to do! But check out the calibration print by dolphin on thingiverse -- www.thingiverse.com/thing:745523. You print this indicator on your machine, compare to a paper gauge that you also print out, and it points to exactly what adjustments need to be made. Brilliant! (My Anycubic Kossel 'Pulley' is sitting in a box in the front room -- I hope to be doing this calibration over the weekend!)
Ever do it?
Where is the follow up video im interesting in your thoughts on the magnetic joints
The printer is still sitting on my shelf in pieces. Granted I have all the pieces, but I haven't had it running since I made this video. The magnetic balls aren't the problem. They function nicely. The problem is the side to side play in the Chinese Linear rails which I have no way to fix. It's bad. As bad or worse than the ball joint in this video, and make me unenthusiastic about the Kossel. On top of that, I now have the Tevo Little Monster Printer which is so much better, and more interesting. Consequently the Kossel got pushed way down the priorities list.
Hate to hear this. Your the main reason I purchased the kossle. I had good hope's for the end results of all your hard work! Anyways love all the videos. #1 3d printing TH-cam channel IMO. And I watch all of them lol
Are those teeny gaps really going to cause that much of an issue?
Interesting..you are putting a tourniquet on you delta....WOW (Think racheting tied down strap ;) )
Definitely want to see the zero lash magnetic arms in action!!!
You are/were a Marine, so you can't be a medic that's a job reserved for us Navy boys...that makes you a sharp cookie to recognize those were tourniquets tied together. :)
92-98 2841, we had combat first aid training unlike some I paid attention ;)
Semper Fi doc
Actually, I was an Engineering Aid in the Seabees.
@@DesignPrototypeTest My dad was a seabee for 35 years. He didn't want to retire but they said he was in too long.
WHATS THE POINT OF ANYCUBIC TO SELL THESE PRINTERS IF EVERYONE HAS THE SAME ISSUES. I WOULD LOVE TO BUY THIS BUT IM NOT SURE IF ITS WORTH HEATBURN AND HEADACHES
China doesn't care about Branding. If it sells they keep making it and selling it. Even if they know it is flawed, and have a better version already for sale. As long as some suckers are still buying them, they won't discontinue their old products. An even more egregious example of this is the Anet A8. Nobody should be purchasing this printer today. When we have similarly priced printers such as the Ender 3, it is a crime to sell someone an A8. But they are still selling, so China is still making them and sending them to unsuspecting consumers.
The new version has brass bushing to fix the slop.
See that's the problem. You never know what you are getting with these Chinese 3D printers. If it is significantly upgraded it should be called something else. At least call it the ANYCUBIC Kossel Mk5. Good to know that some people don't have this same problem. The rest of us who got an inferior product because China doesn't beta test are S.O.L. I guess.
Update? did you fix the problems?
No. I blame Marlin and the 8 Bit control board. I believe I would have succeeded with a Duet Board.
When we will see magnetic ROD upgrade for that kossel:)???
The physical rods are the easy part of that build. They are all but made. It's getting the firmware to work, or getting the frame right that is the hard part. I never figured out which was causing my issue, and just moved on to other projects. I hear you, and other people have asked for the same. I also just learned about the Klipper firmware which would use a raspberry pi in tandem with the stock control board for more processing power to run the Delta at reasonable speeds. I might finally have a solution that will work for that printer.
@@DesignPrototypeTest I have the FLSun Kossel. It's pretty much identical. When I squared my frame I used a square on the uprights. I squared it to the base and then to my printer bed.
I also saw you mention in one reply about the linear rails having wobble. Mine are pretty stiff. They are not as good as a real IGUS rail but mine had no wobble what so ever. Maybe I was just lucky. I tried a set of the rollers for the 2020 uprights as mine came with a a set of them as well but they had a lot of play and I didn't like them.
I did a lot of upgrades to my printer. One was that I ended up printing new endstop blocks & used a rod to make sure they were all the same distance from the base & bed. I also redesigned the effector. It's one of the first things I did. The one on the FLSun was crap. I also redesiged my Bed Mount. Now my print bed is level with no play what so ever. It's uses a 3 point contact and the 3 sets of adjustment screws are at each tower. There are screws underneath to adjust the height and screws on top to lock them in place. With this system I no longer use auto bed leveling. I found that there is no point since my bed is so stiff and everything is squared up. As long as the bed itself is flat it should be fine if everything is squared and levelled. If not then auto level is a must. But these chinese print beds are not the best and many are a bit warped or are easily warped. It's why I used a aluminum tooling plate when building my clone prusa i3 printer
As for the firmware, I haven't had luck in upgrading it for a Delta style printer with a newer version of firmware. A cartesian is much easier for me to keep up to date. I have also tried many times to set one up from scratch but haven't had any luck for some reason when it comes to my Kossel. Since I haven't upgraded my firmware for my Kossel I have to uses two copies of Marlin. The firmware for my Kossel is so old that it won't pass the Verification if it's not using the right version. I should give it another go though. At some point I need to. In any case, it is a must to get into the firmware for your printer. Especially if you start upgrading things. From the effector dimensions to the rod lengths and especially electrical components, even your hotend.
I did upgrade my stepper drivers to TMC2100. Real ones from trinamic which I bought from digikey. They do make a lot of difference.
Anyways...These little Chinese Kossel printers can be quite good with the right tweaks. Right now I am redesigning a new effector for mine as I am going to test out some magnetic rods. The glue on my original rods failed the other day. So it's time to upgrade that part. The original rods are not the best quality rods around. Definitely something that also should be upgraded at some point for anyone that owns these printers. I really would like to see your thoughts on the ones you have at some point, if you ever get around to it. I personally think they will be a lot better that the regular style. I'll find out soon enough though when I complete mine. 😁
Have you received your magnetic rods?
Yes, I have, and finishing the Kossel build is #3 on the list of videos to make.
Cool
One year passed and I cannot see a smart effector clone
5:55 How do I fix It?
1) wind the foil or metal thread 2) tighten with springs, which will remove the backlash, but increase the output on one side. Аlthough if the hinges play.
I would really appreciate if you could get the printer working without the duet electronics and post a video on that before you upgrade to the duet
That is the plan. :) The Duet is very expensive. I may already have the board and firmware working correctly. We'll see once I fix the rods.
it would be great if you could share the information required to get it working.
Of course man. That's what I do here. :) If you are referring to the firmware, and .stl files, I will still ask for a Patreon donation before I send those to to you. I put in a ton of work making these videos. That Sensor video that I talked about took me 3 solid days of work. That's like 30 hours of time, and I only made like $0.50 on adwords revenue from TH-cam. I literally have a family to feed. If this doesn't start paying my bills, I'm going to have to quit and get a "real" job.
no problem man 5$ is nothing compared to the great work you've done
Thank you.
That clamping solution is a bad idea. You’re placing the structure under physical stress and forcing the the triangle into a shape it isn’t designed to be.
I can't help but ask, what is stair porn?
Poor man didnt know about Klipper 5 years ago💀
The WifiDuet costs more than the printer itself...
You get what you pay for, if you want to turn a bad machine into something good then you have to pay!
does somebody else see the stair porn saved site ?
yooOOOO LMAOO
go to the website its like uhhh
Just in time for payday
tighten the stupid screws. Play means they are not tight.
11:20 that is pounds and not dollars. That would have been more expensive in dollars.
15 mins for nothing...
You didn't learn anything huh? You didn't learn about the eyelet play? You didn't learn about why Magnetic joints being zero-lash are superior? You didn't learn about the Duet Smart Effector? You didn't learn about why Marlin on an 8 Bit board is inferior to Duet (Because it doesn't work)? Tell me, please, because I would like to know. When you clicked on the thumbnail that said "FAIL" what exactly were you looking for that you didn't find?