Delta design, wanting criticism!

Hey reprappers!

I'm currently working on my own Delta design and I want feedback from others to get new insights. My Delta is all based on 3 main aspects: Rigid, Nice and Quality.

I have a basic frame now, and I'm willing to hear your reactions on it. What do you think is good, can be better, is waaay off or is super smart!

Some information on the design:
- Corners are all aluminium, so no printed parts here
- Effector and carriages are the only printed parts for now, wich will be printed in XT-CF20 on 100% infill. Super rigid!
- 12mm magnetic balls are used here, still hesitate about going with 14mm
- N52 magnets used
- MGN-12 carriages
- Dimensions: 40x40x40 (triangle) 700mm height
- 220mm diameter printing area, 350mm height
- Driven by a DuetWifi board, 0.9° stepper motors, 24V, and an AC powered heated bed.
- Delrin used as rod end to rotate around the balls
- IR sensor for probing.
- E3D Hotend

Please feel free to come up with suggestions if you think some of these can be better!

Questions:
- What extruder to use? Was looking into Bondtech or E3D Titan
- What distance between the parallel rods, now it is 46mm, is this important?

I hope to get some input!

Arm Length and Length of short sides? 40x40x40 what?

Reduce the ball to 10mm its more than enough and its shave some weight... 10mm ball and 3/8 magnet go well together , I have that setup and never once disconnect even when I was doing 3500 acel , 150mm/sec with high jerk.

Also make sure your cup allow the ball to go through the cone into the magnet space a little, so you back ( if tight enough)/shim/glue the magnet to have the proper clearance, I find its way better than build a clearance into the cup, its require more time to setup like checking for connection with a multimeter between ball and magnet and then push them a little till the connection is broken but you make sure by doing this you have the very minimum clearance between them and its important to have a proper magnet pull on the balls and its also less complicated for a machinist to do and reduce the reject and price.

On your design you use full length of mgn rail, you dont need to use that long, you will never slide below the carrier home position so m you may same some money there buying shorter rails

About the distance between the ball its a personal choice I went with 60mm but close to 50 still do the job, the more space you allow the better effector stability but its cost you printing space with a bigger effector. it can help minimize the tilt problem I guess. the more space you allow also get your balls closer to the next set of arms and its also good for stability.

In your design the effector is blocking the flow of the cooling fan for the head.. not idea since its look like the effector is thick. This can be OK for a alu effector made out of 3mm plate but your effector im not sure about that its also gonna create blowback and will make the fan work pretty badly. That the reason must delta put it under or on top of effector never in middle.

Your picture are all in black and are hard to see... do you have a proper tensioning system?

Thank you for your valuable input. The reason I placed the hotend higher is so you would see effector tilt less in your prints. But you have a good point.

Quote
Andreas15
Thank you for your valuable input. The reason I placed the hotend higher is so you would see effector tilt less in your prints. But you have a good point.

It looks like you could raise the hotend up a lot further which will reduce effector tilt inaccuracies, and improve airflow at the same time. I think you could just make the tube standoffs longer and you'd kill two birds with one stone.

After chasing effector tilt issues on one of my deltas all last week, I can definitely say that it's worth trying to get that nozzle as close to the effector as possible.

The only issue is trying to get a good layer cooling fan arrangement when the effector is in the way of the duct.

Updates! Good flow for the hotend, nozzle close to the effector (1cm under it) and 2 fans for cooling the filament. The fans are not directly pointed at the nozzle, but I feel it doesn't have to. I think it's better to cool the filament just after it is extruded instead of cooling it while it is extruding. But we'll see.

Looks great! Are those the Sunon 30x30x10 fans you're using for cooling the filament? I've found them to be excellent. Expensive, hard to find, but excellent when you can get your hands on them. I think that they're close enough to the nozzle to not need a duct.

All you need now is a Z sensor and you're sweet


Are you considering a flying extruder?

Edited 1 time(s). Last edit at 12/19/2016 11:05PM by nebbian.

Thank you! The fans are 35x35x7.5 blower fans. The Z-sensor will probably be dc42's IR sensor, altough I'm looking into a nozzle probing system.

Mmmh, the extruder is still a hard one. The bowden tube will be very long if I go with a fixed extruder. But I personally don't like the look of a flying extruder. Also, it looks like it makes your carriages way heavier because of the load. What advantages has it next to better retraction/extrusion?

next thing I'm going to try is the extruder suspended from a spare motor at the top of the frame and having it reel it in and out as Z goes up and down. it'd still have to have enough bowden to not get swayed side to side by x/y changes but you could theoretically not have to cover much Z travel at least.

For an extruder you could use a RDD extruder (Remote Direct Drive).
This combines the all the good points of a direct drive and a bowden drive, with none (almost) of the disadvantages.
For instance, the Zesty Nimble or the Flex3drive. (PSA I am one of the creators of the Nimble)

They work by splitting the power and the drive function, so the stepper is mounted on the frame and a drive cable transmits the power to the extruder part.
Because there is a gear system inside the extruder unit, there is very little flex or slop, retractions are nice and fast. I use 1 mm retraction for instance.
Your elevated hot end mount can be used perfectly to bolt the extruder to.

How are you planning on keeping the belts tight? Spring mechanism in the top?
I also second the idea of 10mm balls, no need to go bigger at all.

---

Lykle
________________________________________________

Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]

Thanks for your input! I like the idea of an RDD, but they are so expensive.

For the belts I want a system that I can easy tension them with a screw. Something like what the Atom 2.0 uses. It has to be clean and rock solid. I have added a picture, it are two pieces that slide over each other and can be regulated by a screw. One screw (that also locks the carriage) locks them down. It is hard to create a good, easy to use mechanism without adding to much compexity. But I personally don't think my system is solid enough right now. I'm always happy to hear your toughts!

EDIT: For the magnetic balls, I think I'm going to stick with 12mm. It could be too much, but I don't like the look of 10mm on such a large kossel. Remember that beauty is a key on this machine too

Edited 1 time(s). Last edit at 12/22/2016 06:23AM by Andreas15.

Consider putting the motors and other electronics on top. It makes it easier to wire and also makes servicing it easier.

Quote
Andreas15


EDIT: For the magnetic balls, I think I'm going to stick with 12mm. It could be too much, but I don't like the look of 10mm on such a large kossel. Remember that beauty is a key on this machine too

A 220 mm printing diameter 700 height is not a big printer, ( why not go 300 and 1M this wont really add more cost but if you need more printing area you will have it+ 300 heatbed are easy enough to find around I saw more of those than 220 one).
Lot of us use 10mm with 300+ diameter and 1M+ tower and no they do not look tiny or out of specs.at all, I tell you again if you worry about pull force and that why you need bigger balls, 10 mm and 3/8 magnet have plenty of pull force, no need to go bigger and limit your acceleration with a heavy effector. I bet no one can tell the look between a 12 or 10 balls but your printer can tell the heavier effector, I hope your not going with bigger magnet too.

Why not go with a bigger magnet? It's not because "everyone" uses 10mm I should, I guess...

I chose 220mm because I do not want a giant printer next to my desk. Were will it end? 500mm and 3m height? If I choose for 220mm printbed I probably have a reason for that, no?

And where do I say that I worry about pull force?

Andreas15, if you want good joints then use angle ball joints. They will be much lighter then magnets and they will withstand much bigger forces. Magnetic joints are crap ... especially if you plan to use high speeds and accelerations. If they are small enough to be light then they are weak. When they are strong then they are heavy too.

Also read this: [reprap.org] ... especially the part about TES coefficient.

What is a good target TES? My delta's TES is 253.5.

Please give me any evidence that magnetic connections aren't suitable for the job...

etfrench: I do not know about any particular target to aim for. When designing a new machine, just design it so that it as high as you can achieve.

Andreas15: Their ratio holdingForce/weight is small. If this is not good enough argument for you then just use magnetic joints and be happy with the choice you like most

Simply adding a spring loaded tensioner for each tower between the effector and carriages will give you any amount of holding force to weight ratio desired. @Andreas15: Consider adding a mounting point for these on the effector and carriages.

etfrench: Spring tensioner does not decrease the weight. Quite the contrary.

to everybody: But if you intend to print slowly then both holding force and weight are not a big deal.

That is just a weak argument. You can only lay down a certain amount of filament in a certain amount of time, it's not that a Delta printer suddenly is able to lay down plastic faster. For accelerations and travel speeds, yes indeed, you need strong connections. Saying "they don't have a strong hold" is just talking bullshit. Yes, offcourse, they will have way less hold than fixed connections. But the question is not if has a lot of hold, the question is wether it is enough hold for the application of a delta printer or not. From what I see on Youtube, this forum. A lot of people print with very high speeds and have not yet popped a magnetic link off.

Edited 1 time(s). Last edit at 12/24/2016 09:00AM by Andreas15.

Neodymium magnets replacing the ball joints should reduce friction. Probably not an original thought but it popped in my mind.

Edited 1 time(s). Last edit at 12/25/2016 11:57AM by John Anderson.

Quote
John Anderson
Neodymium magnets replacing the ball joints should reduce friction. Probably not an original thought but it popped in my mind.

That is a good thing, right?

Quote
hercek
Andreas15, if you want good joints then use angle ball joints. They will be much lighter then magnets and they will withstand much bigger forces. Magnetic joints are crap ... especially if you plan to use high speeds and accelerations. If they are small enough to be light then they are weak. When they are strong then they are heavy too.

Also read this: [reprap.org] ... especially the part about TES coefficient.

I have been using mag joints for two years now with no problems. I also now have a titan extruder mounted on the effector for direct drive. No problems.

1/2 in spherical magnets. 12lb holding force each. More than enough. I could probably get dual extrusion going if i wanted to.

Edited 1 time(s). Last edit at 01/25/2017 05:19PM by MrBaz.

How heavy are they compared to my ball joint which weights about 7 grams (all included for one joint) and can hold about 12lb of force too?

Edited 1 time(s). Last edit at 01/26/2017 07:24AM by hercek.

Never bothered weighing them. Irrelevant in the long run. I print for quality, not quick ugly prints. The mag joints have zero lash and will never develop any play that has to be accounted for. Teflon pads make for very low friction and smooth running. I run a 350mm diameter bed with a 400mm print height. Mag joints make for easy maintenance and swaps.

I normally print around 65mm/s with 1500mm/s^2 accel with no problems. Travels are 350mm/s. I could print faster if i wanted to.

Edited 4 time(s). Last edit at 01/26/2017 02:46PM by MrBaz.

Agreed. If you do not intend to print quickly and with high accelerations then magnetic joints will do.

Hi Andreas,
Have a look at this discussion, as it describes what you wanted to do for a calibration sensor.
It uses a peizoelectronic sensor to measure the nozzle tapping the print surface.

[forums.reprap.org]

I added a small modification so the Nimble would fit on top. Or you can use the bowden variant.

---

Lykle
________________________________________________

Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]