I'm pretty pleased with my delta calibration.

Not perfect, as nothing ever will be, but it's a great start: [www.youtube.com]

Printer frame is made from Misumi extrusion and linear rails, parts from Werner Berry (shopspt.com) and some custom pieces. I tried to keep the custom parts count to a minimum. Electronics are Duet 0.8.5 with dc42 fork of RRF (obviously). After I put this through the paces I'm going to refine the plans, make it as cheap as possible and release the CAD assembly, BOM, etc with an open source licence for anyone who wants to build one. Currently as it sits, it's a little bit more expensive than an Ultimaker 2. I'm thinking of organizing a group-buy (once the design is refined some) if people are interested in order to get some of these in people's hands at a lower cost.

Any thoughts and questions are welcomed.

I see magnetic joints are really popular. Which is something I do not understand.
So what is the maximum acceleration you can achieve with it?

I asked the same question in the "New delta" thread but I did not get any response :-(

I think any magnetic joint system with sufficient magnetic strength is totally adequate for all delta setups. The limiting factor for acceleration seems to be the length of my arms, which are very long at 440mm. Also, the "maximum acceleration" is dependent on subjective qualities. Some people don't mind ringing artifacts and the like caused by high-acceleration and max-jerk settings. I prefer to minimize all artifacts. I'm going for precision and print quality over speed. I'm still able to run average speeds which I am happy with. I use 2400mm/s/s acceleration, but with a low jerk setting. In reprapfirmware-dc42 I'm using 480mm/min for my max instantaneous speed change. Print time estimate from Simplify3d of 1 hr is 4 minutes off, the print being longer. Deltas aren't necessarily able to go faster when printing than cartesians, but they can certain do travel moves lightning quick.

OK; if you find some time to try higher accelerations then I would like to know when magnetic joints start to rattle or disconnect. I do not care about artifact from springy belts or not stiff enough tower/carriages. Just what accelerations good magnetic joints can handle.

E.g. I use simple ball joints and acceleration of 7200 mm/s², 20 mm/s jerk.

I've tried 9000mm/s/s and the magnets did not pop off. I do not think any acceleration will make it pop off. I could likely turn off acceleration, command a 150mm/s move and either it would make the move very jerkily or the motors would stall. I've stalled the motors by hitting the endstops and no magnets popped off. The only time I had a magnet pop off was when I forgot to home the towers before running a calibration test gcode file that didn't itself home the towers. The effector crashed into the tower, and one magnet joint did pop off... I want to add some leashes onto the arms so I can strap them to the effector and carriages for a fail-safe.

Dear n8bot,

I noted in your different posts tha you use Duet0.85 with Simplify3D. It's my configuration.

However, I got to some difficults with S3D and temperature control. If I define Bed temperature and hot end temperature in S3D setup, when I launch the printing, hot end and bed don't warm up. I have to define and start the hot end's temperature in the webinbterface on the Duet controller and then when temperatures are reached, I launch printing with S3D. Is there a specific configuration on S3D to control temperature ?

Thank you very much.

Laurent

Ok, thanks. 9000 mm/s² should be good enough for most people. So proper magnetic joints have only slight weight disadvantage to proper ball joints.

hercek, what type of ball joints are you using? I'm planning for a larger scale prototype and will not be using magnets in all likelihood. I'm not sure why the magnets would necessarily weigh more than a typical spherical rod end, though. There is the ball and the steel rod end. Are you comparing the weight to plastic rod ends like the Igus ones?

Laurent, I always have heated up my bed and nozzle via the paneldue and web interface before a print. I also have never used simplify3d as a host controller, and have no plans to. If you mean initiating a gcode file generated and saved from s3d to the sdcard, then I believe that behaviour works as intended in regards to setting temperature. I can't try right now as my printer will be working for another 56 hours or so.

Edited 1 time(s). Last edit at 11/01/2015 09:02PM by n8bot.

Quote
lolocaledo
Dear n8bot,

I noted in your different posts tha you use Duet0.85 with Simplify3D. It's my configuration.

However, I got to some difficults with S3D and temperature control. If I define Bed temperature and hot end temperature in S3D setup, when I launch the printing, hot end and bed don't warm up. I have to define and start the hot end's temperature in the webinbterface on the Duet controller and then when temperatures are reached, I launch printing with S3D. Is there a specific configuration on S3D to control temperature ?

Thank you very much.

Laurent

I use S3D and the duet and I have never encountered this problem.
I often use S3D to change the bed/hotend temp during the printing, as to a temperature test for each new filament/color a print 50mm high 10 by 10 rectangle and set the hotend to change temperature each 50th layer automatically (using 0.2 layers)

Quote
n8bot
hercek, what type of ball joints are you using? I'm planning for a larger scale prototype and will not be using magnets in all likelihood. I'm not sure why the magnets would necessarily weigh more than a typical spherical rod end, though. There is the ball and the steel rod end. Are you comparing the weight to plastic rod ends like the Igus ones?

I use MP Jet ball link (Ø 7 mm, M3, M3 short): [www.bigplanes.nl]
One weights about 4 grams. Yes, they are from plastics and brass. They are sold completed and do not have any play. I bet they can hold many times more than magnetic joints of the same weight

I think ball links have a very bad name in 3d printer community because of most people using traxxas which are sold separated and almost nobody uses the right procedure to joint them. And even if joined correctly, maybe traxxas is not a good quality.

If I would need something stronger then I would probably go with POS 5, it has 14 grams and can handle forces up to 3300 N. But with Nema17 steppers your forces will peak at about 50 N and my guess is that the MP Jet ball link can handle this with ease.

I agree that traxxas joints have given people a bad impression about spherical rod ends. The kind you showed look very interesting, indeed! I might need larger scale than the ones listed there, though, for my future prototypes. Have you run your machine with those specific rod ends? And it performed well?

Yes, I use the MP-Jet links on my current machine. About 8 kg of filament was printed with them and I did not notice any degradation of their quality. Still tight and play free. I doubt I will wear them out anytime soon if ever. Notice that even RC helicopters (e.g. T-Rex 500) use plastic ball links to control main rotor blade angle. The main rotor turns at 3000 rpm. Our 3d printers are nothing compared to that.* My point is there are high quality plastics ball links out there.

* Edit: Actually it can be comparable with that. If you do small infill (about 1 mm long or shorter) at speed 100 mm/s and very high acceleration or jerk then you can get near to that.

I have never seen traxxas joints but my impression from this board is that they must be the crappiest ball links ever and our community stumbled on them. What a pity.

The only thing I do not like about my MP-Jet ball links is that they do not have an Hex Socket opposite of their M3 screw. That would allow me to tighten them more easily with an allen key. I need to do it with thin tip pliers now. It is important to tighten them well so that they do not loosen when printing. Or other option is to glue the thread. MP-Jet has a version which uses separate M3 screw. Those would be easy to tighten but then you need some conical washer with its smaller side being rather sharp. That means the washer cannot be from ABS plastics.

Edited 1 time(s). Last edit at 11/07/2015 07:04AM by hercek.

Quote
hercek

Quote
n8bot
hercek, what type of ball joints are you using? I'm planning for a larger scale prototype and will not be using magnets in all likelihood. I'm not sure why the magnets would necessarily weigh more than a typical spherical rod end, though. There is the ball and the steel rod end. Are you comparing the weight to plastic rod ends like the Igus ones?

I use MP Jet ball link (Ø 7 mm, M3, M3 short): [www.bigplanes.nl]
One weights about 4 grams. Yes, they are from plastics and brass. They are sold completed and do not have any play. I bet they can hold many times more than magnetic joints of the same weight

I think ball links have a very bad name in 3d printer community because of most people using traxxas which are sold separated and almost nobody uses the right procedure to joint them. And even if joined correctly, maybe traxxas is not a good quality.

If I would need something stronger then I would probably go with POS 5, it has 14 grams and can handle forces up to 3300 N. But with Nema17 steppers your forces will peak at about 50 N and my guess is that the MP Jet ball link can handle this with ease.

Thanks for the tip, ordered a set for a delta (2 packs), I went for the ones with a hole and a bolt.

Today I have a set of Haydn's magnetic arms and balls, but I'm going to build myself a second delta (slightly larger) so they'll come to use.
Earlier I had traxxas and I was not to impressed by the quality, some were loose and some didn't move properly (sort of twitchy), but I'm very satisfied with the magnetic ones although I haven't tried with higher accelerations than 3000mm/s².