Underslung Delta Effector

So I've got a FLSUN Kossel Delta. It's had a bunch of problems with extrusion, so I'll be getting a v6 hotend soon. And while I'm upgrading things, I'd like to get a smaller effector, preferably with better part cooling. Here's what I'm looking for:

V6 compatible
Underslung, so I don't lose any build volume
My board can only support two fans, so just one for the hotend and one for part cooling
A cooling nozzle that reaches at least 270 degrees around the nozzle
Screw mounts for non-magnetic ball joints
Some kind of z probe. I'm open to anything that's small and works with an aluminum bed

Has anyone heard of an effector like this? I've been looking on thingiverse and I found this one (https://www.thingiverse.com/thing:2050626). I like the design and cooling nozzles, but I can't use an IR sensor or three fans or magnetic joints.
I also found the anycubic kossel effector with a piezo sensor (https://www.thingiverse.com/thing:2771867). Has anyone had problems with a removable piezo sensor like that?

I've also got solidworks at work, so I could make my own, but I figured I'd get some design input from you all first.

May your lines be straight and your extruders never clog,
Alex

Using two part cooling fans, doesn't mean you need two control ports for them. They are usually wired parallel.
The Mini-IR Z-probe works fine with an Aluminum bed ( with blue tape or other printsurface ). It also works with any controller/firmware I know.

I would actually safe some money and go for magnetic rods and the 'smart-effector' from Duet3D. It's definitely worth it.
If you feel adventurous, you can try something like this instead:



I've changed the design over time and now it works pretty reliable. It's light and cheap.

What is a Kossel Delta. How does it differ from any other Delta?

@o_lampe I hadn't thought of using two fans like that, I've got almost zero electronics experience. I'll definitely look into that!
I was also told an IR probe wouldn't work with an aluminum bed (too reflective) but I have been covering it with masking tape, so that should probably work.
I like that super small design, and I love that it's holding the heatsink at the bottom (I'd really rather not lose printing height) but I'd hesitate to go with that Delrin ball idea. It sounds like more R&D than I would like to be doing.
Also, I'm trying to keep costs low, so probably not on that Smart Effector. Thanks for the suggestions though!

Also, what's the main benefit of switching to a magnetic rod setup? I currently have a small spring between each pair of rods, which pretty much removes any wobble in the effector, so I didn't think they were necessary.

I would not suggest the mag balls unless you really want to make a lot of effector changes, they centre "ok" but they definitely do not print as clean or as accurate as solid rod ends especially at higher speeds (and speed is a big reason for getting a delta). I also would not get too carried away with underslung effector setups, it is pretty rare to high max build height on most deltas, normally the bed width is the restricting factor, admittedly they tend to have far less effector tilt but are far harder to get all the parts you want incorporated.

Plenty of decent aluminium effectors on aliepress that work a treat and even more fan ducts on thingiverse to match them.

Cool, cool, I'll keep my non-magnetic rods. And yeah, I'm willing to lose ~40mm of my 320mm build volume if it means I have a working hotend. I'll look into those effectors and post pics in a while once everything's set up!

Quote
Redemptioner
I would not suggest the mag balls unless you really want to make a lot of effector changes, they centre "ok" but they definitely do not print as clean or as accurate as solid rod ends especially at higher speeds (and speed is a big reason for getting a delta)....

Mag Ball systems are more precise than normal Stlye rods. No slackness in the Rods.
The posibillity of a fast effector change is just a bonus function. Most Delta Users change to magnet system simply because they hande fast accerlations and speeds much better than traxxas style rods

Sorry but that is simply not true, most people go to mag balls because they think they look cool and think of it as an "upgrade" and few change because they need to be able to change the effector out regularly (or think they do). Rod ends are far more stiff and have ZERO play in them compared to mag balls, people who have issues with rod ends do so because they use garbage rod ends (i.e. Traxxis branded). If you spent half the cost of a set of semi decent mag balls on decent rod ends you will end up with a printer significantly more constrained than mag balls. Just grab the effector on mag balls and wiggle it then compare the same to rod ends, it is night and day difference and keep in mind any movement on a delta effects every axis. Mag balls never centre "exactly" and also move ever so slightly on each movement (there is no way to stop the movement), this is easily measured and is obvious in prints when you setup 2 prints identical to each other on the different setups and run the print at a decent speed. Will 80% of users notice the difference, probably not, especially if they have come from cheap Chinese rod ends but it does not change the mechanical facts.

A little bit of commonsense looking at how the 2 connections work should clearly bring you to the obvious issue with mag balls, they are only semi constrained and rely on a magnetic field that can vary to hold them in the same place. If you want to see this yourself, take the effector in and out on a mag ball setup a few times then try and run a print without calibration and you will find that it needs to be re-calibrated, do the same on a decent rod end setup and you will find you won't need to re-calibrate.....

Mag balls also wear at a faster rate but you don't generally notice this wear due to the magnets removing the play and this results in print inaccuracies without an easy way of identifying it. Why do they wear faster, well that is because the magnets have to apply a significantly higher force on the joints in order to over come gravity as well as hold the joint in place well enough not to fall off during movements as well as being open to allow dust/print particles to enter the joint and acts as an abrasive. Rod ends wear much slower (assuming you are using decent ones) but when the do they start to introduce play, once they wear enough, which you can easily feel if you are doing proper maintenance. This then allows you to change them out before they start to cause inaccuracies assuming you do something about it, the rod ends also have a "wiper" to prevent particles getting into the joint. I only have one printer left with mag balls (out of 4) and that one is only on mag balls still because it needs to be able to swap in and out the laser heads, I have done this because these 4 printers run nearly 24/7 and when you do this you soon find the problem with mag balls when you are doing over 600 hours of print time a week.

I point out though, there is nothing wrong with mag balls to hold effectors in place, they are just not as good as rod ends. If you are a hobbyist and like them and won't be printing 24/7 or need to change out the effector regularly then use them, if you plan to pound on your printer all the time, have no plan to swap out the effector or want the most constrained setup you can then go with the rod ends.

Edited 2 time(s). Last edit at 07/18/2018 03:55AM by Redemptioner.

There are good and bad versions of rod ends and also of magnet effectors.
Those cupped magnets are surely not strong enough, but Haydn's mag-arms are even with the best rod ends.

Did you know that there is a radically different solution, which escapes most of the problems that affect magnets or rod ends?
A system very light, intrinsically accurate and insensitive to wear and tear ?

It's Zatsit's living hinge system... (shameless autopromotion)

www.zatsit.fr

Yeah, there's good and bad versions of magnetic and rod arms. I currently have rod arms and a limited budget for upgrades, so I'll stick with them.
Here's my final plan for upgrades:
E3D V6 1.75mm Bowden 12V
This effector (https://www.thingiverse.com/thing:1093345), possibly modified to hold a microswitch instead of an optical endstop
This fan mount (https://www.thingiverse.com/thing:1137927), because I really need good part cooling. Or any cooling, for that matter.

Maybe I'll get the Aus3D Differential IR probe in the future, but again, budget constraints. I'm a poor college student rn.

Also, @M_Xeno, that does look super cool! Maybe I'll try those arms on my next printer, whenever that is!

Quote
Redemptioner
Sorry but that is simply not true, most people go to mag balls because they think they look cool and think of it as an "upgrade" and few change because they need to be able to change the effector out regularly (or think they do). Rod ends are far more stiff and have ZERO play in them compared to mag balls, people who have issues with rod ends do so because they use garbage rod ends (i.e. Traxxis branded). If you spent half the cost of a set of semi decent mag balls on decent rod ends you will end up with a printer significantly more constrained than mag balls. Just grab the effector on mag balls and wiggle it then compare the same to rod ends, it is night and day difference and keep in mind any movement on a delta effects every axis. Mag balls never centre "exactly" and also move ever so slightly on each movement (there is no way to stop the movement), this is easily measured and is obvious in prints when you setup 2 prints identical to each other on the different setups and run the print at a decent speed. Will 80% of users notice the difference, probably not, especially if they have come from cheap Chinese rod ends but it does not change the mechanical facts.

I can only assume that either you have no experience of magnetic joints, or you have only experienced ones that were not properly constructed. Properly made magnetic joints - such as the ones I have, made by Haydn Huntley (the variant with machined Delrin ends) - have zero play.

Ball-and-eye joints such as Traxxas, Igus and all-metal joints tend to be either too stiff or to have some play. If they start off being just right, when they wear they will have play. Maybe you know a brand that wears well, but I don't think you've told us what joints you are using.

Perhaps you are referring to the ball-and-socket joints used on SeeMeCNC and similar machines, where the sockets are held on the balls by springs? I think they can be made to work quite well, however there is a problem. They need to use quite a strong spring between the rod ends to hold the sockets in place. This typically causes the rods in each pair to bend towards each other, because the springs are some little way in from the rod ends. The extension of the spring, and hence its tension, varies with effector position. That in turn causes the rods to bend towards each other by an amount that depends on XY position. And that causes the effective rod length to vary slightly with XY position - which messes up calibration.

Quote
Redemptioner
A little bit of commonsense looking at how the 2 connections work should clearly bring you to the obvious issue with mag balls, they are only semi constrained...

What do you mean by semi-constrained? They are as constrained as any other type of joint, until you apply sufficient force to separate the ball from the socket.

Quote
Redemptioner
....and rely on a magnetic field that can vary to hold them in the same place.

Vary? How?

Quote
Redemptioner
Mag balls also wear at a faster rate but you don't generally notice this wear due to the magnets removing the play and this results in print inaccuracies without an easy way of identifying it. Why do they wear faster, well that is because the magnets have to apply a significantly higher force on the joints in order to over come gravity as well as hold the joint in place well enough not to fall off during movements as well as being open to allow dust/print particles to enter the joint and acts as an abrasive. Rod ends wear much slower (assuming you are using decent ones) but when the do they start to introduce play, once they wear enough, which you can easily feel if you are doing proper maintenance. This then allows you to change them out before they start to cause inaccuracies assuming you do something about it, the rod ends also have a "wiper" to prevent particles getting into the joint. I only have one printer left with mag balls (out of 4) and that one is only on mag balls still because it needs to be able to swap in and out the laser heads, I have done this because these 4 printers run nearly 24/7 and when you do this you soon find the problem with mag balls when you are doing over 600 hours of print time a week.

A nice feature of magnetic joints is that they have to wear a lot before you get any play. There is a gap between the magnet and the ball stud, and the sockets would need to wear sufficiently for the magnet end to touch the ball before you get any play. I appreciate that you run your printers for more hours than most users, but have you tried a set of Haydn Huntley's magnetic joints with the Delrin ends? It sounds to me that you have only tried poorly-made magnetic joints.

Quote
Redemptioner
I point out though, there is nothing wrong with mag balls to hold effectors in place, they are just not as good as rod ends.

Good magnetic joints are (from my experience) much better than poor and mediocre rod ends. I can believe that good rod ends are better than badly-made magnetic joints. But you haven't convinced me that good rod ends (whatever they are) are better than good magnetic joints.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

@Redemtioner

What a nonsense.
Only magnet joints have ZERO play. They have constant contact with no play.if they have no contact anymore its just because the magnets are not strong engough for the mass. I recently looked how fast my Delta with DIY Magnet rods ccan print and was able to bring it up to 450mm/s with 5000mm/s acceleration without the effector causing problems before the extruder gave up

I don't know what you use for rod-ends... all I've ever had in my hand was nothing for a long time. China stuff is crap..Traxxas just a little better..Igus are nice but have a too small working angle. In the Rod End you have play everywhere... Either the screw fits through the eye and has play or it does not fit... The ball in the Rod End either runs very easily and has play or it has no play and is stiff. And the wear on Rod ends is much bigger than on good constructed magnet joints

Quote
dc42
I can only assume that either you have no experience of magnetic joints, or you have only experienced ones that were not properly constructed. Properly made magnetic joints - such as the ones I have, made by Haydn Huntley (the variant with machined Delrin ends) - have zero play.

That simply is not true David, I have one of your effectors (well I had 2 but sold one) and you can indeed feel the movement if you grab it and give it a wiggle (also has flex in the PCB material which is need for the FSR to work but that is another issue), again I stress this probably is not enough to effect most users as they are never going to print fast enough or need the part accuracy high enough to care but it does not change the fact it is there. I won't get into it further as anyone who has a decent setup in both styles can see the difference, suffice to say you should have a chat with a mechanical engineer who has a background in joints they will confirm my position.

Quote
dc42
What do you mean by semi-constrained? They are as constrained as any other type of joint, until you apply sufficient force to separate the ball from the socket.

They are only semi constrained because you are relying on a magnetic field to hold a ball in a semi-cup socket, so only about 30% of the joint is constrained "ish", and only from a small range of angles of load that aligns with the cup. A rod end is 100% constrained as the joint is captured on all axis, no matter the angle of the load it cannot move outside of the constraint.

Quote
dc42

Quote
Redemptioner
....and rely on a magnetic field that can vary to hold them in the same place.

Vary? How?

I will not go into all the fundamentals about magnets and especial spherical magnets or using magnets on a sphere, but I will point out a couple of issues. Firstly magnets are effected by heat, the stronger the magnet the more it is effected, the angle the magnet is at compare to its pole alignment and the alignment of the electrons in the ferric coupler means the strength of the bond changes as well. Add a little friction and momentum to the joint and you always have some movement (gets worse as abrasives get in between the mating surfaces). Sure it is small (it is exagerated the bigger your printer dimensions are) and again I stress for most people "who cares", but if we are talking about the mechanical fundamentals of the joint you simply can't get around this fact. The upside is the magnet will pull further into the socket as it wears so the play stays pretty much the same (actually it technically gets better the more wear as the magnet gets closer to its ferric coupler) except your dimensions start to get out of tolerance and you won't easily be able tell this has happened (some would argue that is a good thing but as an engineer I would argue it is not).

Quote
dc42
Good magnetic joints are (from my experience) much better than poor and mediocre rod ends. I can believe that good rod ends are better than badly-made magnetic joints. But you haven't convinced me that good rod ends (whatever they are) are better than good magnetic joints.

Each to there own but physics don't lie. Your argument here does not really stack up though as a good electric car is better than a "mediocre" petrol car, a good apple is better than a "mediocre" apple and same as a good rod end is better than a "mediocre" mag ball, but a good rod end always trumps a good mag ball. I am not going to get into an argument about what and where you have bought the rod ends you have used, suffice to say you are unlikely to get decent ones from eBay or a Chinese website. Rod ends have been around and in common use since the 20's, the simple fact is they are the only 100% constrained multi-axis joint and as a result they produce zero play at all angles and are the preferred joint type in applications where they need 100% constraint of the a joint. As a result there are a number of amazing rod end manufactures out there, don't think about 3D printers when looking for rod ends, think about where else they need to use 100% constrained joints (hint: aircraft, boats, robots...)

Quote
Barracuda72
Either the screw fits through the eye and has play or it does not fit

I don't think you understand how a rod end works, how it should be mounted or how the rod end itself actually stays in place. Not knowing how to correctly setup a rod end connection is not the fault of the rod end, I suspect the next thing you will be doing is complaining to us about the same thing on a bearing which is to say you don't use a standard bolt to mount them....

Quote
Barracuda72
was able to bring it up to 450mm/s with 5000mm/s acceleration without the effector causing problems before the extruder gave up

In your own words "What nonsense", you have done yourself a disservice making ludicrous statements like that.....lol

Quote
M_Xeno
Did you know that there is a radically different solution, which escapes most of the problems that affect magnets or rod ends?
A system very light, intrinsically accurate and insensitive to wear and tear ?

It's Zatsit's living hinge system... (shameless autopromotion)

www.zatsit.fr

Interesting designs, especially the flying extruder with the multi-axis rotation, it is an ingenious way of taking the load off the PTFE tube while maintaining a constant resistance to the filament path especially for larger delta's. The dimensional constrain over the towers to ensure the best possible alignment with the Z tower is also a great idea, although doesn't the bar extending out past the Z tower that is only held on with 2 bolts put you at risk again in adding a "skew" back into the towers?

What rate does it build heat in the joint, I see it is rated to 120*C and most materials will exceed that temp when continually flexed pretty quickly?

Would also be great if you have a link to the joint running at a decent speed.

Quote
Barracuda72
I recently looked how fast my Delta with DIY Magnet rods ccan print and was able to bring it up to 450mm/s with 5000mm/s acceleration without the effector causing problems before the extruder gave up

Which is a rather pathetic speed and poor acceleration for a delta. It does not give much trust in magnetic joints.

Anyway magnetic joints are zero play if the rod end never "rattles" on the ball. You can achieve that by using small accelerations.

Depending on the rod angle to the ball the magnetic force will change a very little (for good designs) or a lot (for bad designs). That is because the magnetic resistance of the magnetic circuit is changing. I think this force change is negligible with a good magnetic rod design. Ball rod ends do not have this problem.

I'm with Redemptioner and vote for angle ball rod ends. Especially because they have extremely good holdingForce/weight ratio compared to magnetic joints. The only cases when you should consider magnetic joints is when you intend to change the head too often or when you need very high free angular movement to the sides. Typical ball joints have freedom only around 25°-30° - but you do not need more than that for a typical delta.

Just avoid the crapy traxas. Get e.g. MP-Jet angle ball joints (Ø 3 mm bronze + plastic). Or industry level POS ball joints (bronze + steel) although I'm not sure how well they are available below Ø 5 mm sizes. Definitely use some ball joints if you want to achieve high accelerations.

@Redemptioner

Indeed, the wooden braces and the spacer T20x20 are not enough to constitute a complete non-deformable geometry, but they are enough to remove the ringging component that could come from the oscillation of the whole structure. There is virtually no effort involved, just the suppression of a potential vibration, so this simple solution has shown that it is adequate.

Hylite withstands 120°C paint cycles in the automotive industry, but not when used to make hinges, which would lose their strength at this temperature. However, I tested them in boiling water for a few minutes, and they survived...

Generally speaking, my prototypes have been running for over two years, and hinges pose no problem of any kind, even at 600mm/s with 5G acceleration!

now you have me interested....it has the potential to really shorten up the arms while maintaining incredibly constrained joints with no maintenance.

I agree that we are talking about basically nothing that is going to make any difference. I like the wood as the support arms, the dampening the wood will add to vibrations is a brilliant idea and some great engineering.

Quote
Redemptioner
That simply is not true David, I have one of your effectors (well I had 2 but sold one) and you can indeed feel the movement if you grab it and give it a wiggle (also has flex in the PCB material which is need for the FSR to work but that is another issue), again I stress this probably is not enough to effect most users as they are never going to print fast enough or need the part accuracy high enough to care but it does not change the fact it is there. I won't get into it further as anyone who has a decent setup in both styles can see the difference, suffice to say you should have a chat with a mechanical engineer who has a background in joints they will confirm my position.

I think there must have been something wrong either with your set of joints, or with your carriages. On my delta, the effector has zero play, and I have to apply a lot of force to the effector to make it move at all. When I do that, I can see the the carriages moving slightly, due to elasticity of the belts and motors. There is no visible movement in the joints.

BTW one thing we left out of the original fitting instructions for the Smart Effector is that the rods should be fitted with alternating N-S poles around the effector.

Quote
Redemptioner
They are only semi constrained because you are relying on a magnetic field to hold a ball in a semi-cup socket, so only about 30% of the joint is constrained "ish", and only from a small range of angles of load that aligns with the cup. A rod end is 100% constrained as the joint is captured on all axis, no matter the angle of the load it cannot move outside of the constraint.

Surely the force applied to the rods of a delta when the carriages move is almost entirely along the rod axis? In which case, the joint doesn't need to be constrained in any other direction, except perhaps for safety reasons if there is a risk of the joint detaching. Just to be clear, I am assuming that the magnets are in the rods and the spherical balls are on the carriages and the effector. I know that a few people have tried the reverse, which is IMO a big mistake, because then the magnetic force is not aligned with the rod.

Quote
Redemptioner
I will not go into all the fundamentals about magnets and especial spherical magnets or using magnets on a sphere, but I will point out a couple of issues. Firstly magnets are effected by heat, the stronger the magnet the more it is effected, the angle the magnet is at compare to its pole alignment and the alignment of the electrons in the ferric coupler means the strength of the bond changes as well.

Yes I am well aware of this (I have a degree in physics from Cambridge). According to the data I found, the strength reduces by about 0.6% per degC. So in a heated chamber at 50C there would be about an 18% loss of strength compared to room temperature. If the strength is adequate in the first place and this reduction is factored in, then this won't cause any problems. But if you wanted to use a hotter heated chamber, then I agree that magnetic joints should not be used.

Quote
Redemptioner
Add a little friction and momentum to the joint and you always have some movement (gets worse as abrasives get in between the mating surfaces).

That's true for all types of joint. But magnetic joints can tolerate this wear much better than some other types.

Quote
Redemptioner
Each to there own but physics don't lie.

What physics would that be? IMO you have not demonstrated any physics that makes magnetic joints a poor choice, except at high temperatures.

btw you still haven't told us what brand of ball-and-eye joints (if that's what they are) you think are so wonderful.

Edited 2 time(s). Last edit at 07/20/2018 06:02AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
What physics would that be? IMO you have not demonstrated any physics that makes magnetic joints a poor choice, except at high temperatures.

Magnetic joints have very poor holdingForce / weight ratio compared to ball rod ends. That is pretty solid physics.

How heavy is the magnet and the metal ball in Haydn's mag-arms?

E.g. MP-Jet ball joint weights about 4 grams including the M3 nut for the ball connection (which has bolt built in) but without connection to the rod. I use about 2 cm long M3 steel threaded rod which connects the carbon rod and the ball joint. It adds about 1 gram. I did not try how big is the holding force but definitely at least 30 N. Whether it will hold in a 3dPrinter is no concern at all regardless of all the forces a common Nema17 can provide.

Quote
hercek

Quote
dc42
What physics would that be? IMO you have not demonstrated any physics that makes magnetic joints a poor choice, except at high temperatures.

Magnetic joints have very poor holdingForce / weight ratio compared to ball rod ends. That is pretty solid physics.

Solid: yes. Relevant: not necessarily. I could just as well argue that it's solid physics that 50mm diameter carbon rods are stiffer than 6mm or 10mm ones, therefore nobody should be making delta printers with rods smaller than 50mm diameter.

What matters is whether you can get magnetic joints that have sufficient holding force but don't add enough extra mass to cause ringing that wouldn't otherwise be present. AFAIR, Haydn claims 1kgf per joint (so around 10N). The total mass of my effector and everything on it plus 6 rods is just under 400g. I've just tested homing my printer (successfully, multiple times) with a 1.5kg water bottle slung from the effector. So I appear to have a more than adequate safety margin. The joints on my delta have never detached during a print.

Quote
hercek
How heavy is the magnet and the metal ball in Haydn's mag-arms?

I don't have any accurate weighing scales for small weights, however measurements taken on my postal scales suggest that the ball studs weight about 5g each including nuts and washers, and the 360mm x 6mm rods including magnets and Delrin rod ends weigh about 33g each.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

And still no Link to his Wonder Ball Joints.
till now you have only 2 arguments that all need to use Ball end over Magnetic System

Weight: but the more mass is so margin and no really pro Argument because as DC statet..the magnets have more than enough magnetic force to hold the effectot perfect. And if we take a look at the weight..how much brings your whole effector and rod array on the scale? and have it the same functionality as the smart effector? spare some grams on the joint dont means much if the effector is heavier. Its the whole system that gives good speed and quality

Heat: sure..in high temps they lose magnetic force..but if anywhere in your magic ballend plastic is used..like in the Igus..what will they do in a high heat chamber? getting bigger..getting softer..so much more wear.


Pro for the mag system...less wear..zero play over the whole lifetime ( I use 12mm N52 ring Magnets and not Haydn arms and they are nylon coated) My first attemp was with chromed magnets but metal on metal wasnt good. My maintance on the magnet joints is some fine grease once a month.

Edited 1 time(s). Last edit at 07/20/2018 05:49PM by Barracuda72.

@DC42:

Ok, hopefuly I understood you correctly:

The weight or your rod is 33g +2*5g = 43g per rod and can hodl 10 N.
33g is for carbon with magnets and delrin cups. 5g for metal ball stud (with nut) at each end.

My rod including rod ends and the nuts weights 7g + 2*5g = 17g and can hold 30 N definitely probably much more (I would guess it at about 50 - 100 N but I never tried).
Carbon part has 6.25g and one rod end with everything needed has 5g.

Ball joints save 6*(43-17) = 6*26 = 156 grams on the weight of moving parts and can withstand much more force.

A common Nema17 will give you up to 40 Ncm. Or with 16 teeth T2.5 pulley about 63 N at the belt at most. Which is more than 10 - 20 N your magnetic arms will be able to hold on one tower. You actually can disconnect your arms during printing if you would ever stress your printer hard enough. There is no chance it would happen to me.

Now when you do not run your printer hard enough then magnetic rods are just fine ... of course.
That is what I'm telling often here. Many people build delta printers but run them as slowly as Cartesian. Partially maybe because of (sometimes) poor magnetic joints.

Is the difference relevant? It depends how hard you run your printer.

Edited 1 time(s). Last edit at 07/20/2018 06:11PM by hercek.

Quote
Barracuda72
till now you have only 2 arguments that all need to use Ball end over Magnetic System

Actually, I have only one argument for ball joints: holdingForce / weight ratio.

Not sure whether you want the link for the ball joints from me. I use MP JET Ball Link (Ø7 mm, M3, M3 short). The first ones I tried were good so I do not have experience with other kinds on a 3dPrinter. I already posted the link many times on this forum: [www.bigplanes.nl]

@hercek
My post was actually meant more for Redemptioner.

Anyway..you are right in holdingForce/weight ratio the ball joints win clearly
But how much holding force we need? Each of the Haydn arms can hold 1kg (10NM) thats 6kg on the system and we talk about 156g more weight?

What speeds and accelarations do you print and what we need to get the system above its limits? I have already lost the Effector several times but only if it was a very bad print and the nozzle got stuck on the print. And then I prefer the Effector hangs freely in the air afterwards (I use a flying extruder) instead of bending my airbrush nozzles.

I know thatball joints..another hobby of me are RC Cars where we use high quality rods too..sure its in the dirt and they wear lots more but they have wear and need to replace quite often.
Your systems looks really light, how long are your rods, which material? guess they are shorter and have a smaller diameter thant DC´s rods..

@Barracuda72:

I would say that that your holding force should be about the same as the force steppers can produce. Maybe half of that. That is if you want to stress your printers with high accelerations. If you run small accelerations then it does not matter.

You cannot just add 10 N from one arm and claim the system has holding force of 60 N. In reality it is more like 5 N to 20 N depending how the forces are applied.

156 g is a significant part of the 400 g of the whole movable system. The difference means that you probably can run the system at about 13% to 38% higher accelerations with the same ringing. Look at M_Xeno\s printer. He has extremely light movable parts and he can run at 6G. That is impressive! It would be good to see how ringing looks like at 6G and some usable speed at which extruder can still manage ... at least 150 mm/s.

I typically run at 9000 m/s² and 120 mm/s printing moves. 250 mm/s non-printing moves. Check this post:
[reprap.org]

The arms are 299 mm long, external Ø 5 mm, carbon fibre.

Longevity of both ball joint and magnetic will not be a problem. Notice that good quality ball joints are used for controlling pitch of RC helicopter rotor blades. These joints can withstand about 30 - 100 hours before the first signs of wearing out. And they are loaded heavily compared to a 3d pritner. The rotors run at about 3500 RPM. I expect at least 200 hours on a 3d printer for these simple plastic joints. One can get full metal ones. Those will last for ever and will probably be lighter than the magnets. I did not search for full metal ones.

Just do not buy the traxas crap.

It doesnt matter how fast the Rotors rotate..if you take a look at the Heli mechanic you can see that there not much movment in the ball joints..they just adjust the pitch and yaw of the rotor blades. They have much more movement on a 3d printer.. more travel over the bed. In a heli the movement is just few mm depending on the heli class and believe me..a model heli pilot who use it for aerobatic competitions ( where the stress is really higher) will change it after one competition weekend..because if they fail..it costs him a Heli 1-3 grand worth. Had enough of these crazy enthusiast in my model club.
Anyway you are right with the holding force...that was my mistake.
Xenos printer..he brought the concept of lightweight to the top. but it may have other drawbacks. if some of the hinges fail..how costly will it be to replace? i dont know...he just looks too skinny to me. would be afraid that i break something if i touch it too hard

btw..6g acceleration? are you sure? that would be 58.800 mm/s² (1g is ~9800mm/s²) not sure if i should believe that

Edit: just read that thread..he tried 3g but didnt work because te extruder..he prints normally at 6000mm/s² and for a speedtest he tried 10000mm/s² so far away from 6G


btw..your rods are much thinner and shorter than DC´s or mine..mine uses 6mm carbon rods and lenght is 380mm so sure your is much lighter...

Next we could talk about why you only use 5mm if the 6mm are much stiffer

You see..i dont think that there is a final truth in that question.. Light is good..but functionality also..at the cost of weight...and so on.

Beside that i still wait for the link to the balljoints from Redemptioner ..and how much weight his effector with rods in total have

at least you're a friendly person to discuss with. it´s a pleasure to discuss about delta effector mechanics..even if its not in my native language

Edited 1 time(s). Last edit at 07/21/2018 05:38PM by Barracuda72.

Quote
Barracuda72
It doesnt matter how fast the Rotors rotate..if you take a look at the Heli mechanic you can see that there not much movment in the ball joints..they just adjust the pitch and yaw of the rotor blades.

Ooooo, but it does matter how fast the rotor rotates. Helicopters do not tilt the whole main rotor when you move the collective stick. They only change the pitch of the blades and the pitch of any given blade will be different when the blade rotates rotates by 90° or 180°. The angle joints on helicopters do one full cycle every rotor revolution anytime the collective stick is not centred. And especially the competition pilots almost never have their collective stick c entered. Even sustained direct helicopter flight requires the stick to be not centred. Otherwise the helicopter would be slowing down.

The range of movement on a helicopter is smaller but they are moving more quickly and they are more loaded than our 3d printers.

I talked about this with a heli completion pilot myself. He told me that the ball joint wear is not an issue. One will crash his RC helicopter much sooner than the ball joints wear out. Maybe pilots you know crush much less than the pilot I know

Quote
Barracuda72
btw..6g acceleration? are you sure? that would be 58.800 mm/s² (1g is ~9800mm/s²) not sure if i should believe that

You are probably right. I cannot find the 6G reference now.

Quote
Barracuda72
btw..your rods are much thinner and shorter than DC´s or mine..mine uses 6mm carbon rods and length is 380mm so sure your is much lighter...

Ø5 mm rod diameter is plenty enough for 30 cm rods. Any rods thicker than that is a waste. Well it could be good to have maybe Ø 6-7 mm rods with thinner walls to keep the same overall weight. I even posted some computation about that on this forum: [reprap.org]

I guess magnetic rods use such a thick carbon tubes so that they can insert the magnets at the end.

My whole point about magnetic rods is a bigger picture. They are a contradiction. One builds delta for a speed. Because otherwise one would build a cartesian printer which is much easier to calibrate and troubleshoot. But if the printer is built for speed then why use magnetic rods? They are more expensive and specially they slow down your printer if you want to keep the same printed part quality.

but about what speeds are we talking?

150mm/s printspeed.. 5000-8000mm/s² accelrations? Magnetic system can easily handle that..

400mm/s print speed? 2G accelerations? Yes..maybe its too much for magnetic rods...but who print so fast on a daily base and not only to test the system limits?

We started this discussion becuase "someone else" statet that mag arms and the duet smart effecotor ist crap..
I dont want miss it anymore...so much comfort...perfect 1st layer..no matter if i triy a new nozzel..chage from normal hotend to volcano....airbrush nozzle..just hit print and the printer do the rest and give me a perfect layer. for that comfort im willing to have more weight because even now im far away from the limit, this system have.
And i guess i could print easily at our all day print speed without hitting the limit

No one has said magnetic joints are not a legitimate connection for delta's, it has only been pointed out that rod ends are a better mechanical joint for multiple reasons.


@DC42
Your effector flexes when used with a flying extruder, there is enough flex to allow the extruder to pull enough load/flex into the effector to cause nozzle to crash into the print on the release of every retraction, it makes little difference to the print quality (makes the retraction scare a little larger), but it does cause a small build up of filament on the nozzle that can start to cause stringing and adds to the maintenance.

Quote
Barracuda72
And still no Link to his Wonder Ball Joints.

I use FK Bearing rod ends myself, but like I said, there are numerous suppliers of decent rod ends for high performance needs, especially in high performance vehicles like planes and boats, google is your friend........

Hercek has provided you with even more examples why rod ends are superior, and having played with heli's as well I can also confirm on decent setups they run rod ends that last the life of the heli and then some and these rod ends are put through millions of cycles under immense load spinning around with the blades without ever generating any play.

Putting that aside they are also generally easier to install and setup, most arms you can buy can have an inner diameter that you can match to the thread of a male rod end to make setting up the arms a breeze, they are also much easier to measure mount placement and make an adjustment if needed. That is not to say you can 't buy complete mag ball kits that are easy to install as well.