"Candy Apple" printer first pic and video

Thanks to everyone here for the advice, especially dc42.

Previous forum topics for printer:

First Build Hiwin CoreXY
Duet controlled CoreXY - Install Procedure

Design ideas explored:

1. Use all Hiwin style linear slides for all movement to increase precision
2. Reduce part count
3. Reduce weight of linear moving parts in XY plane to minimum (currently 173gm, will be 150gm shortly)
4. Box design for easy future enclosure (to control print environment temperature, air convection, and add reverse pressure fume filtration)
5. Maximize build volume for minimum exterior dimensions (build 20x20x27cm in 27x31x36cm box)
6. 32 bit processor in the print controller

I believe that the weight of moving parts in the XY plane is the lowest of any printer with this build volume (or maybe any build volume). Anyone else know of a lighter movement?

I am hoping to show it in person at MakerFaire NYC September 26 & 27, but I just finished building, so may be too late for them to give me a booth. If anyone has suggestions on how to get a late application approved, please let me know.

In the picture, the dust covers were printed on a bed without tape, so they warped. I will be printing new dust covers shortly.

Sooooo... Finally the picture:



and video of first 2 vases printed:

YouTube video

Maybe you can tell why I am calling it Candy Apple?

Thanks again everyone,
Patrick

Edited 6 time(s). Last edit at 09/01/2015 04:38PM by patrickrio.

It looks very interesting. May you post pictures of some prints?

What do you think about linear guides after trying them?

Hi,

Very nice ! It looks so clean. I love the covers.

++JM

I am trying to edit a short video of it printing a white vase.... looking for a free, easy video editor to download right now for windows 10.

Thoughts on linear guides:

There is no play in the movements. They work great. Print errors will not be coming from the linear movements themselves, which is all you can ask of them.

Durability is still being tested. Since tolerances are closer in the linear guides, I feel it is important to control even room dust as much as possible... hence the dust covers.

I did some rough durability calculations based on expected forces, and the expected durability is supposed to be more than 1000km of movement if kept correctly lubricated. We will see.

Edited 1 time(s). Last edit at 09/01/2015 11:05AM by patrickrio.

Quote
patrickrio
The expected durability is supposed to be more than 100km of movement if kept correctly lubricated.

100km seems very little, I was not aware of such a short life of linear guides.
Even supposing a low average printing speed, let us say 10mm/s, divided by the two available directions X and Y (Z is negligible, and this is anyway an unrealistic optimal case just to obtain an upper bound), we have 5mm/s per axis * 3600s/hour = 180m/hour. 100km is then 556 printing hours, or about 24 days of continuous printing.

What is the formula to calculate their life given the load?

Edit: I thought I read 100km while it was 1000km. With 1000km we have then 240days of continuous printing. This sounds good for me (at least for my average daily printing time).

Edit 2: 10mm/s is however unrealistic for some printers I guess, 50mm/s may be closer to reality, which gives about 50 days of continuous printing. Again, I would have expected more...

Edited 2 time(s). Last edit at 09/01/2015 11:17AM by cristian.

sorry, missed a zero on that... 1000km. And that was using the maximum force the motor can apply continuously, from one side pull only, with an exaggerated pull angle than the printer actually pulls.

In actuality, the motor will not pull even close to 100% of the time at full power, and CoreXY actually balances the forces when it is pulling in one direction.

1000km should be a minimum based on the calculation.

I suspect that environmental dust and lubrication will be the real limiting factors, but as I said, we will see!!! I am testing!!!

If durability becomes an issue, I would increase the size of the slides which would increase durability.

I pushed the limits on this design to get very low moving weight in the XY plane... if durability is an issue, I would back off.

Edited 2 time(s). Last edit at 09/01/2015 11:35AM by patrickrio.

Quote
cristian
What is the formula to calculate their life given the load?

Found an example of formula here.

yep, those are the formulas..... They end up being a rule of thumb anyway, because there are so many more variables than those formulas take into account... I outlined only a few in my case above. Also, I am using Chinese slides for the prototype and the calculations I used were actually for another brand. The calculations will not be exact anyway. (why mess up sizing during prototype design with expensive slides... use the cheap ones first I say...)

Edited 2 time(s). Last edit at 09/01/2015 11:40AM by patrickrio.

Quote
patrickrio
I suspect that environmental dust and lubrication will be the real limiting factors, but as I said, we will see!!! I am testing!!!

Thanks for sharing your experience. After several informed advices that I received in this forum section I am also planning to use linear guides for a coreXY, so I am very interested in your build.

Quote
patrickrio
(why mess up sizing during prototype design with expensive slides... use the cheap ones first I say...)

This is actually a very good thought I missed. It may be worth starting "cheap" with some Chinese guides and then buy the serious ones when the design is stable.

Which rails did you buy and where? I had a look at the threads you linked but I missed the post where you say where you finally decided to buy them.

Hi,

I just posted a video of the first two vases printed. I put it in the first post of this thread via an edit. sorry for the poor video editing quality, but my recent windows 10 upgrade has disabled the normal video editors I used, so I had to download something I have never used before.... very messy.

I sourced the linear guides from AliExpress. Search for Hiwin or Linear Guides and all the sources show up.

Quote
patrickrio
I sourced the linear guides from AliExpress. Search for Hiwin or Linear Guides and all the sources show up.

Thanks, I was in particular curious to know the price range you chose for your linear guides. On aliexpress there are cheap (from few dollars) but also very expensive linear guides, inlcluding Hiwin.

The video is well done and the prints seem fine. I think I saw the bed flex a little a couple of times, probably because of some ridges of plastic, but besides this everything looks stiff and there are no visible vibrations, although it is impossible to see if there are imperfections in the print at that distance. In any case linear guides seem to work very well.

I bought the cheapest slides, and a few of the supposedly real HIWIN guides. The quality seemed the same to be honest. I have one of the supposedly real slides in the build to see if it has better durability than the others. Seems the same. Maybe I got a fake too, and just overpaid....

I did look at and inspect in person some of the very high quality slides from Nippon Bearing, THK, CPC and IKO. These bearings were definitely higher quality than the Chinese guides.

All of the slides from the different manufacturers have the same general shape and size (for the same sizes of course).. within fractions of millimeters of each other. They are essentially an interchangeable part in my design.

Edited 1 time(s). Last edit at 09/01/2015 06:19PM by patrickrio.

Thanks. I think I will start with the cheapest ones too, so I won't cry too much if I ruin them.

Yes, there were a few "Bed Flexes" there. I was in a hurry to make a video so I can apply to Maker Faire NYC today (done).... so I filmed before doing a full bed leveling. I also have not set up the auto bed leveling compensation feature of the Duet board yet. The first layers were laid down too close to the bed, and the filament squeezed beyond the edges of the nozzle on the first layers... took to the third layer to quit scraping.

The vase layers are completely perfect on the big rings, and were a little messed up on the small rings because I was moving too fast for the plastic to cool before the next time around. I would get perfect results if I slowed down a bit for the small rings.

There is no banding variation or vibration showing up AT ALL in the prints... unless there is a first layer scrape of course.....

One observation of my printer.

Since I made the print head and X linear guide very small, and covered every other moving part with dust covers: You don't have many moving parts to look at, and they are all small. It is definitely not as exiting to watch as a delta.... or even an Ultimaker.

However, I am going to make the next version a little taller next time, and it will be able to print objects 215x215x320mm with only a 2cm increase in height. The same as an ultimaker 2 extended - approximately.

Nice printer!
What's the frame made out of?

it is made out of MDF

Hi,

Since I am waiting for delivery of a new controller, I thought I would work on some other parts... Here is a picture of all the XY moving parts being weighed... 173gm.

How are you connecting the x and y axis? It looks like the X rail is mounted on a plate, then the plate is mounted on the Y rails?

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

actually, the idlers are mounted on a plate, and the plate and x axis rail are mounted directly to the carriage blocks of the y axis.

Quote
gmh39
How are you connecting the x and y axis? It looks like the X rail is mounted on a plate, then the plate is mounted on the Y rails?

This is the thing noone wants to show

I am searching for a proper method and till now found only ONE 3d-printing dude who shared his method

---

Der 3D-Druck ist tot, lang lebe der 3D-Druck!

Schreibt mich nicht mehr an, ich hab das drucken an den Nagel gehängt.

Hey,

Inspired from your post in Dentist thread - Can you post some more details about your belt paths? I am not sure I entirely understand how you are running them around.

Here is a drawing of the belt paths... sorry for it being hand drawn, but it was much faster than using the computer to make the drawing.....

you will be suprised at how long those shuttles and rails will last you. given the load applied should last you years unless you are subjecting them to high heat or weight. granted the side load they are on isnt the best and why id like to use igus rail system since you can replace the bushings relatively cheap

Hi Patrick,

I've read in your first post that you planned on using 2020 extrusions, did you stick to MDF in the end ? If so, did you encounter any issues with using MDF as a structure ? I find it really interesting how you managed to optimize the build volume / printer's footprint ratio !!! Also is there anything that you'd do differently if you were to start Candy Apple today ?

I stuck with MDF, yes. I think in the long run, that will make it easier to create a heated/abs filtered enclosure. It is heavy though, and parts sizes would not make it good for a mass produced printer I think. My design has a VERY reduced part count as a test of different ideas, and I would back of on this a little as outlined below. HOWEVER, my printer works very very well. the movements show zero vibration or position accuracy artifacts even at high speeds, accelration and jerks. my motors start skipping steps before any vibration or positional errors start showing up in prints. Currently, I need to do some adjustments on retraction and I get some overextrusion at corners at high print speeds due to slicer inability to match deceleration of axis and deceleration of extrusion, but other than these issues, print quality is already fantastic and should improve with more work in these areas.

Things I would do differently:

1. Plate construction for XY axis. using MDF as my mounting surface for the XY made initial adjusting too difficult. It is just about impossible to get the box perfectly level, and even then, the mdf compresses a little when screwing rails in, making it hard to keep level again. I ended up mounting the rails with very thin washers between the rails and MDF and used additional washers to change rail height at the various mounting points. It was a total pain in the ass. I would also recommend creating mounting lips for the linear guides to make installing completely parallel easier. It may be too difficult to put mounting lips on the plate though, yet to be determined. It also means that I must not print large objects with less than a 0.1mm layer for the first layer. I think you could get level so perfect with a plate that your only limit on first layer thickness would be the plastic qualities of the filament itself. I may already be close to this with .05mm layer thickness on the first layer for small objects, I just don't know because I do not have much experience with other accurate printers.

2. Plates joining X axis to Y axis. My plates are just square steel plates with 2 screws each. The screws go through the single rail and mount into the two holes in each carriage on the parallel linear guides (my parallel guides have only two mounting holes in each carriage). It is a very simple design, but required drilling an 2 additional holes in the single rail so I had holes 8mm apart that matched the 2 carriage blocks...which was a pain in the ass. It also does not give a large moment arm to keep square between the XY axis'... Only 8mm distance between 2 screws is a low amount. It works but I think using 4 hole carriages would be better. Also, it is easier for my design to have the belt put excess sidways pressure on the parallel linear guides which is a HUGE no no. Adjusting out the lateral forces on the parallel guides in a big pain in the ass with only 2 screws at each end. I recommend having AT LEAST 2 or 3 screws at each end of the single guide and using 4 mounting screws in each parallel guide carriage. I think the minimum design should join the X and Y axis with 12 screws, not 4 like I did. Much easier to set up correctly I think. My square and parallel is DEAD ON, but setup could have been much easier with a 12 screw mounting solution.....

3. I used flanged bearings for all of my pulleys. This makes the lips at the edge of my pulleys pretty small which meant that belts could jump off the pulleys more easily if they were not perfectly adjusted for height. I am torn on this. It forces you to adjust pulleys really well so that you have a long belt life, but it was a TOTAL pain in the ass. I wrecked one belt during pulley setup. Of course, my pulleys were dirt cheap!!! Another issue with the flanged pulleys is that they do not have teeth. The last pulleys before the extruder carriage had teeth riding on them so there is a VERY small vibration that gets transmitted to the extruder. Almost imperceptible print variations show up perfectly spaced 2mm apart on my printed objects as a result of this. It is only visible on larger, smooth front and back sides of objects. It gets very slightly worse at higher speeds. I want to emphasize that this wave is VERY slight, and would not even be visible if every other print quality of my printer wasn't so ridiculously good. I think you would not see this error on many other printers because they have other variation errors that are much larger.

4. I would make it taller. The cost to make it taller is only the cost of longer Z rails. The Z rails create the most stable Catilevered Z axis I have ever seen on a printer. The 400 step motors work great with the 16 tooth gears... great resolution and dead simple construction. I might consider a mounting plate for the two parallel guides here too, but longevity is less critical here because the Z axis moves so slowly and only moves up and down one time for each print. If you could get less teeth on the drive pulley it would maybe be better, as you are close to the point of accuracy with this construction, but I am able to do layers that are accurate to .05mm in printing with no banding due to Z axis height resolution problems so.... good enough! It may help that GT2 belt has 2mm distance which makes it easy to select layers with the same number of steps or micro steps. I think MXL belt would make this harder.

5. I am going to TEST a geared extruder with teeth on both drive gear and idler. When I use 1.2mm diameter nozzles, my direct extruder just doesn't have the strength to keep up with higher print speeds... it only can print at about 50mm sec with the 1.2mm nozzle right now before it starts stripping away filament( I backed the amps off so it doesn't strip any more, but now I get skipped steps when I exceed 50mm/sec with this nozzle). However, the existing direct extruder with 78 oz 400 step motor can print 200mm/sec with a .6 nozzle, no problem.

6. Currently, my nozzle is set up with a miniature PEEK heat break and teflon liner. I did not find an all metal hot end solution that met my size constraints. Currently, my hot end carriage has a total width and depth of only 18mm wide by 39mm deep, the area it takes out of build area to operate. Fitting an all metal hot end with a fin cooling fan on it within this width and depth will currently require a custom designed all metal hot end.

Edited 8 time(s). Last edit at 12/16/2015 11:55AM by patrickrio.

Thanks for the elaborated reply Patrick. Working with screws on the "thickness" side of an MDF board can be a PITA (a plate construction seems to be essential), but again the project demonstrated the feasibility of your concepts and the results were impressive.

I wonder how you attached the bed plate to the Z axis (Z carriages), the bed seems to vibrate a bit in the video, some guy pointed it out in the comments and you said it's because of a screw in a 90deg bracket, is the problem still persisting today?

Candy Apple got me interested in coreXY in the first place, and frankly as a noob seeing someone innovating on their first try is quite inspiring. I hope that you'll keep updating us on the evolution of the project.

Edited 1 time(s). Last edit at 12/17/2015 06:20AM by mezora.