DIY 400x400x200 printer, choice question / doubt

Hi,

first post here, I've been lurking a while... I am building and designing-on-the-go a 3D printer with a workspace of 400x400 bed, 200mm Z axis. The bed will move up and down like with a core XY, X and Y axis on top of the machine. I am in doubt about the drive of the X and Y axis. I have found the CoreXY drive system, but I could also drive the X axis with it's own motor traveling with the Y axis. This would make the set-up easier I think. The downside would be the weight to be moved. But the weight is already considerable for a 400mm trave x-axis, adding a little motor and a lot less belt I don't think would matter much. I think I would gain a lot more from using 12mm aluminium rods for my Igus bearings instead of the heavy hardened steel ones I have now. But all this is just my thoughts, not based on any experience. Please advise from experience.

Cheers,

Hugo

Quote
HugoW
Hi,

first post here, I've been lurking a while... I am building and designing-on-the-go a 3D printer with a workspace of 400x400 bed, 200mm Z axis. The bed will move up and down like with a core XY, X and Y axis on top of the machine. I am in doubt about the drive of the X and Y axis. I have found the CoreXY drive system, but I could also drive the X axis with it's own motor traveling with the Y axis. This would make the set-up easier I think. The downside would be the weight to be moved. But the weight is already considerable for a 400mm trave x-axis, adding a little motor and a lot less belt I don't think would matter much. I think I would gain a lot more from using 12mm aluminium rods for my Igus bearings instead of the heavy hardened steel ones I have now. But all this is just my thoughts, not based on any experience. Please advise from experience.

Cheers,

Hugo

CoreXY is pretty nifty in many ways. For one, the tension on the belt means you don't need to use such a rigid Y axis. It's similar to adding wire tensioners to frames, but in a moving system. If the Y axis tries to move one way, the belt tension on the other side will counteract it. This means the Y axis can actually be a lot lighter then you would use for a Cartesian machine. Top it off with the fact that you don't need a 300gram motor and you're looking at big weight savings. Also consider the fact that you will not be distributing the weight evenly on the Y axis if you have a motor. Any backlash in the system will show up even more (but lopsided) if you have one side of the gantry heavier then the other. Place the motor in the centre and your rods will bow from the weight.

You'd be surprised how much difference removing weight makes. Although using Cartesian in the way you are suggesting with the bed in Z makes a lot of sense (removes some banding, less problems caused by weight of print, less likely to be kicked off bed), you don't solve the main issue which is the huge weight of the bed moving in Y. Instead you have a possibly even heavier 400mm X axis which moves in Y. If you are using Nema 17 motors, be prepared to print slow to keep the quality up.

My suggestion would be to make a CoreXY, and if the system really disagrees with you then the modifications to make it Cartesian are minimal.

So is a 400 x 400 heater bed available
or
do you have to cut one on a table saw?

600 mm linear rails out there cheep?

See what hardware is out there -- design around it
unless cost is no problem
then get everything custom made

One down side to the CoreXY, especially for a large printer, is the loss of belt stiffness with the really long belt lengths involved. The mass of a large carriage makes the belt stiffness even more critical. You might want to go to 12mm wide belts if using 2mm pitch, or maybe jump up to 3mm pitch belts, which are a lot stiffer. Here's a link to a Gates datasheet listing comparative belt stiffnesses (tensile modulus):

belt properties

On my 350x350mm printer, I used a modified H-bot configuration because it uses shorter belts, and I've also got a fairly low mass carriage. I used a single 6mm wide MXL belt (similar to GT2) which works OK, but I wouldn't mind it being a little stiffer.

Edited 1 time(s). Last edit at 03/09/2017 01:34AM by LoboCNC.

Theoretically on corexy, the longer belts make a difference but I don't see much of it in the prints. I recently switched from using steel reinforced belts to fibre reinforced on my corexy (300x200 print area) as the steel started to snap, but although the steel belts are much less elastic there is no discernible change to print quality. Going up to 9mm wide belt or 3mm pitch belt is an option but it adds quite a bit of expense and difficulty in obtaining parts (GT2 belt is now ubiquitous in 3D printing). Gt2 belt is spec'd to take a lot more punishment than most 3D printers dole out. I'd be interested to hear how you get on if you do this.

As for corexy, its a clever mechanism (as is the ultimaker style gantry [i0.wp.com]) and allows for a relatively light x and y axis. Moving the bed in Z makes sense the movements are small and slow.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Thanks for all the replies. I answer chronological:

@Origamib; thanks, I understand what you mean.

@cozmicray; I indeed design around stuff I can get my hands on easily / cheaply. There are heated beds out there 200x200mm, four would make a 400x400 bed. But I am planning a glass bed with nichrome wire and Kapton tape heating. I have made heaters before for other purposes. It might even become a 230V~ heating, if I can get my hands on a cheap 1:1 transformer.

@LoboCNC; I have planned to use the 10mm GT2 belts. I’ll look into 3mm pitch.

@DjDemonD; I’ll let you know how the 10mm GT2 belts hold up if I go that route. Which I will probably do due to availability. I most parts from Banggood and AliEpxress.

The system will run on 24V, RAMPS with TMC2100 drivers and NEMA 17 motors. I have designed an extruder carriage Mk8 style but both carriage and extruder are made as one part. I will go with the CoreXY set-up because of the above advise.

Cheers,

Hugo

I guess its just priority
Heater maker or 3D printer maker???

The good Digital Dentist has convinced many -- aluminum tooling plate

Glass a good thermal insulator 400 x 400 with nichrome wire
it will look like the Alps when heated up??

Flat hot bed = good prints

Perhaps stainless tubing in ceramic bed
with super heated steam from your neighborhood reactor???
Controller would have to control steam inlet (servo valve)
Steam outlet --- cook hot dogs --- keep buns warm

You can get silicone heaters at that size for very cheap. I have a 280mm 240v heater I bought for £20 from aliexpres.

Don't stick a heater to a piece of glass. It doesn't work very well.

Here's a Taz printer with a glass bed and heater glued directly to it:



After the glass broke, I converted it to aluminum and put it on 3 point leveling instead of 4. Here's the same heater glued to a piece of 1/4" cast tooling plate:



Here's the whole conversion process for that printer: [www.instructables.com]

Not sure about flatness? Here's 85 wheels for nerdy derby cars printed in one go, nearly edge to edge on 1/4" cast aluminum bed plate:



Here's a 250+ mm diameter print in ABS printed a the same machine:



Still not sure about flatness?

https://vimeo.com/157222577

https://vimeo.com/157784843

Any questions?

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Thanks for all the info. I sort-of scrapped the tooling plate because of the weight. I was calculating 1/2", not 1/4", and I was still working on a Y-axis moving bed design when I considered the tooling plate. Hauling that about would not be sensible. I never considered it again when choosing the CoreXY-like design. I guess it's back on the priority list.

Cheers,

Hugo

1/4 tooling plate (I use 6mm) works wonderfully and I can thoroughly endorse it. It does work best on a delta where it does not need to move at all, but it can be managed on a moving Z platform. I use a counterweight for my corexy 1.5kgs, essentially my belt driven Z axis has a fair amount of mass but is easy for a nema 17 to move and hold.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

For info, my Core XY has a build area a tad under 400 x 400 and 740 in Z. The main frame is 600 x 600 x 1,000. I can print stuff like this [www.youtube.com] and this [www.youtube.com] and there are some other things on my Youtube channel. The bed weighs in at around 7kg (10mm thick aluminium topped with 6mm thick float glass) and is lifted by 3 screws driven by a single Nema 17 via a continuous belt. As configured in the videos, the Y axis including the X axis weighed in at 1,900 gms and I regularly print at 90mm/sec. All belts are 6mm gt2 - never a problem.

Quote
deckingman
For info, my Core XY has a build area a tad under 400 x 400 and 740 in Z. The main frame is 600 x 600 x 1,000. I can print stuff like this [www.youtube.com] and this [www.youtube.com] and there are some other things on my Youtube channel. The bed weighs in at around 7kg (10mm thick aluminium topped with 6mm thick float glass) and is lifted by 3 screws driven by a single Nema 17 via a continuous belt. As configured in the videos, the Y axis including the X axis weighed in at 1,900 gms and I regularly print at 90mm/sec. All belts are 6mm gt2 - never a problem.

Sir!
That is impressive work!!

Quote
SatorCodex

Sir!
That is impressive work!!

Thanks but I didn't make the post for any sort of plaudits. Just to demonstrate that even with a 600mm x 600mm frame, with belts around 2.4metres in length and shifting 1.9kg of mass, standard (ebay's finest) 6mm GT2 belts and Nema 17 motors are perfectly capable of achieving good results. The printer has been running for about 6 months and there haven't been many days when it has been idle, so many hundreds of hours of use. Never had cause to change the belt tension (or anything else for that matter). I've recently changed the design of the XY axis so I replaced the belts at the same time but to be honest, there was absolutely nothing wrong with the old ones that I've taken off.

Thanks for all the input. I've just bought aluminium rods instead of the steel ones for the moving X axle, reducing weight considerably. To keep things simple, I plan to use these:



Any comments? I like it to be spring loaded as it keeps tenstion, but I wonder if it is stiff enough.

Cheers,

Hugo

I'd say avoid those springs, they look on the face of it to be a good idea, and they're cheap as dirt as you can get them out of clothes pegs, but they will reduce your print quality. When the motor turns the belt, especially at high jerk/acceleration the spring has to compress slightly before the belt moves, this is especially pronounced during rapid oscillating moves.

Try something like this:
re-orientated_tensioner_big_part.stl
re-orientated_tensioner_small_part.stl
[www.thingiverse.com]

There's also this type which I haven't tried but which works like the spring tensioner, but its adjustable and importantly fixed in place at a set tension.
[www.thingiverse.com]

Edited 1 time(s). Last edit at 03/20/2017 08:43AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
HugoW
Thanks for all the input. I've just bought aluminium rods instead of the steel ones for the moving X axle, reducing weight considerably. To keep things simple, I plan to use these:

Timing-Belt-Tensioner-Spring.jpg

Any comments? I like it to be spring loaded as it keeps tenstion, but I wonder if it is stiff enough.

Cheers,

Hugo

As long as you don't exceed the bias force of this type of spring, it won't add any additional flex to the belt. However, ad DjDemonD points out, high jerk/acceleration moves may exceed the bias force and then cause defects in your prints.

What?
X Axle? X guide rod?
Aluminum rods? How long?
How will they be straight?

How are they machined, polished?

What type bearing will NOT shred them?

Really? saving weight on load bearing piece that just has momentum in Y?

Quote
HugoW
Thanks for all the input. I've just bought aluminium rods instead of the steel ones for the moving X axle, reducing weight considerably. To keep things simple, I plan to use these:

Timing-Belt-Tensioner-Spring.jpg

Any comments? I like it to be spring loaded as it keeps tenstion, but I wonder if it is stiff enough.

Cheers,

Hugo

Subtlety is not your strong point, is it? I use Igus sliding bearings, combined with Igus hard annodized rods. Proven combo. 2x 500mm steel rod is 886 grams, same in alloy is 307 grams. That is 579 grams less to move about when moving the X axle in Y direction. That is a saving in energy so an enhancement in accuracy for the machine. The Powered hotend and sled weigh too little to make the alloy rods bend or sagg. The Y axle rods will have more load to cope with, but since they are stationary they will be steel.

[www.igus.com]

Cheers,

Hugo

@DjDemonD; that's a nice option but I cannot 3D print, yet. I found these triggered by your post:

[www.banggood.com]

I think I can cut off the side things and use it.

Cheers,

Hugo

That would work and might be easier I'm forgetting you're talking about a corexy. Those inline tensioners might not fit as there isn't always a place to put them on the belts which won't interfere with something. The other solution is to have sliding motor mounts. As long as you adjust the tension on the A and B belts to square up the x axis it will work.

Edited 1 time(s). Last edit at 03/28/2017 08:37AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Yeah, I'll buy two so I can adjust all four ends. Bit overkill, but easy to fit.

Another question; I was just given two NEMA 23 motors. They identical in physical size and steps per revolution, but have a different maximum amps rating. I think I could use those to drive the X and Y system on my CoreXY, if I stay below the motor's rating. Which I will do anyway because it is way above the driver's rating. I don't think it is a problem the motors being slightly different. Or am I missing something?

Cheers,

Hugo

Well in theory its not a problem as long as they both turn x amount when asked to move y distance. As a general rule most corexy's are built with 2 identical motors for A and B (x/y) as it lowers the chance of a bad mismatch, or of one motor getting hotter than the other on a long print or something like that. But build it and see, you'll soon know if X and Y are not moving exactly right by printing a line parallel to X which will either be straight or not. You can always switch to matching motors.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

I just found the motors are pretty much identical, but one is a 4 wire and the other 6. So if I wire the 6 up as a 4, they are identical. Great!

Just bought a lot of mechanical bits and bobs at Banggood to make the X-Y set-up, and ordered 5 TMC2100 drivers from Alixpress. Now I just have to wait another three weeks for the lot to arrive and I can take the next step in building the thing. I'll post some pics when I have something to show, the print-head-and-carriage all-in-one should be done today.

I also decided to start without a heated bed, but leave room for it. I will print only PLA and using tape or glue gets decent enough results to try, I think. Then I can first print a product cooling duct to hold a fan and start printing upgrade parts for my printer. It seems having a printer handy is nice when builing one... But until then, I need to buy or make bits.

Cheers,

Hugo

Edited 2 time(s). Last edit at 03/29/2017 06:40AM by HugoW.

Still need to press in the sliding bearings and cut off the lever a bit more, but basically it's done:













The belt mounting / tensioning devices will go on top of the bearing housings, at the front of the bearing housings threaded holes are made to fix the product cooling fan and shroud to.

Hugo

I'd be wary with the tmc2100's a fair few old pros here have tried to optimise them and got not very far.[forums.reprap.org] Might be worth ordering some A4988s which should be pennies just in case you don't get very far with the tmcs.

Edited 1 time(s). Last edit at 03/29/2017 08:10AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

They came highly recommended on another forum. I'll give them a shot at 24V with the NEMA23s on the XY, if it doesn't work the A4988s are easily sources. BTW, my Arduino Mega and RAMPS 1,4 arrived, with broken fuses. Bummer. Should fix that...

Hugo

There is no way that TMC2100 drivers will drive Nema 23 motors adequately.

A printer as large as 400x400 isn't a budget printer, and IMO it deserves better electronics than Arduino/RAMPS.

---

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].

Yes, I still play with Lego Duplo:





The Y-slides will have the belt wheels on top, I need the belt tensioners to come in and be fitted before I can position the wheels. With the set-up on the table like this, I also thought up a way to fit the end position switches on the X axle, they will be on the X slide. Putting one on each Y-slide will give a lot more moving wires, now I can keep it all together. The end position switches on the Y will of course be stationary on the housing of the printer. All electronics will be on the back of the printer housing, except for the control / LCD pannel, although I will need to make the wiring to that a lot longer.

I've changed my drivers to three A4988 drivers on-board and two bigger off board drivers for the NEMA 23 motors which will drive the CoreXY set-up.

Hugo

Looks good! When you start looking at positioning the pulleys be sure that the belt segments that run between the extruder carriage and the end pulleys run parallel to the X axis and that the segments that run between the end pulleys and the motor and corner pulleys all run parallel to the Y axis. A lot of people miss that and end up with all sorts of problems. The belt segments that run from the motors to the corner pulleys and from corner pulley to corner pulley don't need to be parallel to the axes. See: [corexy.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]