Lots of advice required for large dual x carriage printer!

Hi Everyone, this is my first post but I am guessing I will be pestering you all for quite a while!

I am currently working out what type of large scale printer to build. A bit of background - I have a small company and I want to 3d print parts I currently make with aluminum extrusions. It would be a huge benefit to me if it works. I have a small DIY printer at the moment and have some experience with building CNC routers (V-Slot OX) and am pretty expert at cutting and drilling aluminum.

I am sorry for asking so many questions but a) I work 60 -70 hours a week which limits my time for research – by the end of the day I am often too tired to even eat. b) I just cant afford to make an expensive mistake and wind up with a useless machine that wont do what I need.

The parts I want to print are simple but very big! One is just like a picture frame approximately 50 cm long and 40 wide made from a square 15mm wide and 15mm high. Its just a empty frame which I plan on printing with PETG. The other part is similar is size as the frame and consists of a 15mm wide border/base and some supports rising 2cm from the base and then across about 10cm toward the center. This is to support fans and power supplies and will be printed from Carbon Fiber reinforced PETG. This part will need supports to print and I am currently waiting on some PVA to test with a dual hotend.

I am not doing anything with small details and plan on using 0.5mm nozzles, maybe even 0.6mm

Ideally I want to use a double X axis carriage to print 2 copies of the same part at the same time as this effectively doubles the throughput. Because of the size of the x axis which will be about 110cm I am considering a CoreXY design (C/D Bot looks good!)to keep the moving mass down. It also has the advantage that I can enclose the printer (it will be in a cold approx 15C room in winter) but leave the motors at least partly outside the enclosure for cooling. I will be using Boden extruders. Ideally I am looking for a build with good instructions which can be scaled and modified to my requirements.


Right, now for the questions/advice requests!

Base Design

Which would be better for my application, Core XY or standard Cartesian? Is the length of the belts in a big core XY design likely to cause problems with stretching or backlash? Any recommendations on conventional or steel core belts?

Construction

I plan on using the V-Slot system as I have experience with it and the quality is excellent. Where possible I plan on using aluminium parts (for example corner brackets and reinforcement) rather than printed parts. Where I do use printed parts I will use carbon fiber reinforced PETG as it is very rigid and prints flat and warp free (love this stuff!). Any comments, suggestions on this?

Electronics

For electronics I am planning on using the MKS Rumba board with touch screen and using their CD opto isolators to connect to TB6600 drivers. This will allow me to use higher current motors and keep the cost down.
While I like the look of the features on the Duet board I dont see much advantage of 256 micro stepping and current is limited to 2.4 amps so I probably coulnt run 3 amp nema 23´s with it. Its also expensive and doesnt have the level of support or forums of the marlin boards.

Motors

Motors – I am planning on using PrimoPal motors – I used them for my OX machines and they worked very well and have proven to be very reliable. Customer service is also excellent. I am thinking of using the same motors as before which are Nema 23, model is PHB57S56-430 which has the following specs:


Current 3.0
Resistance 0.8
Inductance 2.4
Holding Torque 110 N.cm
Detent Torque 3.5 N.cm

Rotor Inertia 280 gm cm2
Bi/Unipolar Bipolar
Weight 680 gms
Length 56mm


The full range of nema 23 motors is here:
[www.primopal.com]

The nema 17 motors are here:
[www.primopal.com]

Any comments or suggestions about the motors? Should I just stick to nema 17 like most people or go with the nema 23?

Heated Bed

I am thinking of using 4mm glass for the heated bed. Would it be best to have a single piece of glass about 60 x 100 cm or 2 separate ones 60x50 cm and level them individually? If using 2 pieces I can control the temp of one of them with a PID controller. I would use aluminium except I am not sure if I can get it machined flat around here and the cost would probably be pretty high. Assuming I could get either glass or aluminum, which would be better and what would be the best finish for the aluminium?

For heating I am planning on using 220V AC silicone heater pads switched with SSR´s. I would use several smaller units arranged for even heating. Back of the bed will be insulated with cork sheet or pure wool felt (its very heat resistant).

Hotend

For the hotend I am planning on using a copy of the E3D Chimera which has highly polished Titanium tube with no teflon liner, which should resist the abrasive effect of the carbon fiber filament. Added bonus is I can print anything with no heat limits! I just received (today) one I bought to test and the quality is excellent with high quality machining and bright internal polish of the titanium tube! Its also easier to mount than the E3D V6 style.

Extruder

I am planning on using boden extruders to keep the moving mass down. I have found these extruders which have 2 synchronised and gear driven feed rollers. I think they should greatly reduce feeding problems which would be great for 12 hour print jobs! What do you think? Anyone have experience with them or similar? Are they worth the extra money?

[www.aliexpress.com]

This looks like an improved Bulldog design.
[www.aliexpress.com]


Not so crazy about this one as it uses a steel gear to drive 2 nylon ones, for me mixing metal and plastic gears isnt a good idea.

[www.aliexpress.com]
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Ok, I think thats enough questions for now ; )

I am open to any and all suggestions as nothing is set in stone. The main thing is good quality prints with high reliability.

Steel core belts for coreXY will not be good unless you use very large pulleys. Steel core belts are not as flexible as glass core belts and a lot of your motor's torque will be spent just bending them around the pulleys. If you are worried about stretch, use 9 or 10 mm wide belts instead of 6 mm belt.

NEMA-23 motors tend to vibrate a lot more than NEMA-17, so if you can get sufficient torque from NEMA-17, it's better to use them. Microstepping >16:1 will reduce noise and vibration.

Heating a glass bed directly is a bad idea- you'll need aluminum under it to spread the heat, and that doesn't work very well, so skip the glass entirely. Just use cast aluminum tooling plate- it comes milled flat, and with large diameter nozzle, you'll probably print in thick layers, so a piece of 8 mm MIC6 will be flat enough to print on. Put a thin layer of PEI on it, and mount it on a good leveling system and it will be flat enough to print edge to edge and prints will stick. It's going to expand a lot when heated, so design the mount to allow that. See: [drmrehorst.blogspot.com]

You're going to need a very high powered heater to get that bed plate up to temperature. 0.4-0.5W/cm^2 should be adequate to heat up without having to wait too long to reach print temperature.

The "improved BullDog design" photos don't show what you need to see if it's really an improvement over the BullDog. The BullDog's only real problem besides its weight was that there was a gap between the drive gear/pinch roller and the tube that fed filament to the hot-end. I can't see from the pictures if there is a gap that will allow filament to buckle. I have run a BullDog XL on a printer in almost daily use for close to 3 years and it has had only 1 jam, has never stripped filament, and only a couple times buckled the filament in that gap. A single drive gear and pinch roller works fine and all the double drive designs are a solution in search of a problem. I've been using Titan extruders recently (for the last year) and they have proven as reliable, but have no gap between the drive gear and the feed tube to the hot-end, so the filament can't buckle. That makes them better suited for flexible filament than the BullDog XL that requires careful tweaking of pinch roller pressure and very low print speed to ensure reliable filament feed. You probably won't be printing flexible filament with a Bowden setup, so maybe the gap isn't a problem for you.

---

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

What is the budget for your printer are you from the uk
I would go with coreXY, I'm using the HyperCube Evolution as a base.

I'm biased of course, but here are some things I think you ought to be aware of:

- I think dual X carriages on a CoreXY printer has only been done using Duet electronics. Ask in the CoreXY forum at [forums.reprap.org] to check.

- Higher microstepping will make your stepper motors and drive system much quieter - and as digital_dentist says, Nema23 motors are noisier than Nema 17 anyway.

- 2.4A is 80% of the 3.0A rating of your proposed motors. It's not common to run stepper motors above 80% of rated current anyway, because they get very hot if you do. You can also get 2.8A Nema 23 motors.

- We're probably going to increase the maximum motor current on the Duet soon because a key piece of firmware development to facilitate that has just been completed

- Don't expect to see many Duet support questions and answers on these forums, it all happens at [www.duet3d.com].

I won't comment on the mechanics because it isn't my area of expertise. I'm normally a fan of removable glass print beds, but for a printer of this size they would be too unwieldy to be useful.

HTH David

---

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].

A bit of a step away from what others are saying in this thread, but more geared to your comment of wanting to avoid costly mistakes. Why 3D print these parts? The parts you're proposing to print are big, and will take a lot of time to print. If you're planning to use dual extrusion on them they'll take even longer. Does it actually make sense to 3D print them rather than use aluminium extrusions? What are you trying to save on or avoid by 3d printing them?

If a printer is really what you are looking for, by all means go ahead. I'll chip in my two cents. For a printer that big, the weight of the motor moving on the gantry may be negligible compared to the rest of the mass you're moving. The problems caused by more moving mass might be less than the problems caused by longer belts (stretching, sagging etc.) so it may be better to go with the old replicator style gantries. People also seem to forget that belts have mass too, for smaller machines the mass of the belt is considerably smaller than that of the motor, but as you scale things up, and especially if you switch to (heavier) steel core belts, the tables may just turn.

As for extruders, I think the bondtech/dyzextruder style dual drive is better. I can imagine if the drive gears on the ones you posted have slightly different diameters, they'd have different surface speeds and the filament between them would stretch or develop slack as one drive gear pulls more or less than the other. I don't know how noticeable this effect would be but it might cause issues. I have to disagree with DD that dual drive extruders are just a looking for a problem. For starters your idler is no longer just applying pressure and flattening your filament, it's driving it along. It also doubles your contact area which means you can push harder without stripping or slipping on filament. Or, if you prefer, you can use lighter idler pressure while still getting the same grip. This means less filament deformation which leads to more consistent extrusion and is especially important in bowden setups where having the filament get squashed into an oval can cause excess friction in the bowden tube.

Hi David

I think you may have convinced me about the Duet. Adding up all the bits for the Rumba (board, display, opto units, TB6600 boards) I get about €130 so its not that much more for one of your boards and display. As a registered business in Spain can I get it VAT free?
By dual carriage I meant 2 heads connected together NOT independently controlled so its no problem with a regular marlin board as far as I know (admittedly I have only a few weeks of experience!) especially if you dont mind using a PID controller for the hotend.

Richard

Hi Trakyan

Quote
Trakyan
Why 3D print these parts? Does it actually make sense to 3D print them rather than use aluminium extrusions? What are you trying to save on or avoid by 3d printing them?

The difference is paying a skilled metal worker €100 per unit or pushing a button and coming back in 10 or 12 hours! Plus as I need more throughput I can just add more printers. Finally it allows me to use standardised parts elsewhere rather than custom making them. Total hours per unit will drop from 7 hours to 1 hour or less. So yes it make a LOT of sense!

Yeah, jury is still out on Core VS traditional, so far I have one vote for each! I am worried about the length of the belts, they do have some stretch, even if it is a tiny amount, over 3m or so it will add up. I think steel belts are out though as I dont need the strength and reliability and drive-ability will be better with fiber.

The extruders you mentioned are a bit eye wateringly expensive for what they are! I think with modern machining tech the gears will be identical on the models I posted. I may get just one and a normal priced (cheap) alternative and compare performance. They are very simple peices of kit really so €100 for one is a bit much. There is a company in Isreal who make industrial units and they are the same design internally as the ones I posted so I am hopeful!

You're printing big parts with a big nozzle. How accurate/precise do you need the prints to be? That will tell you how much stretch can be tolerated- probably a lot more than you think.

I used 10mm wide no-name Chinese steel core belt in the Z axis of my printer. The ~4 kg bed assembly hangs from two of these belts. I added an additional 4 kg to the bed and measured the stretch to be 42 um/kg. That was with about 1.2 m of belt (each belt) between the bed assembly and the drive pulleys.

---

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

Quote
SplatHammer
Hi David

I think you may have convinced me about the Duet. Adding up all the bits for the Rumba (board, display, opto units, TB6600 boards) I get about €130 so its not that much more for one of your boards and display. As a registered business in Spain can I get it VAT free?

Yes, you can enter your company name and VAT number when ordering.

Quote

By dual carriage I meant 2 heads connected together NOT independently controlled so its no problem with a regular marlin board as far as I know (admittedly I have only a few weeks of experience!) especially if you dont mind using a PID controller for the hotend.

I see, I thought you wanted dual independent X-carriages (IDEX), which allow you to do one large print with dual materials, or two small prints side by side. Anyway, RepRapFirmware on the Duet can handle either configuration as standard.

---

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].

I dont really have a set budget, but I want to keep the cost as low as possible by avoiding bells and whistles and sourcing good quality parts at a low cost.
I am in Spain which can make getting some parts locally a bit difficult - I am probably going to have to get the tooling plate I plan on using for the heated bed from the UK
as I havent found a supplier here yet. Then again it is about 26C here and the sun is shinning as usual and a beer is €1 in a bar ; )

Quote
Digital Dentist
How accurate/precise do you need the prints to be? That will tell you how much stretch can be tolerated- probably a lot more than you think.

The parts I want to print are not detailed (small features, lots of complex shapes etc) I just need decent surface finish and reasonable dimensional
accuracy, say +- 1mm over 50cm. I currently have the FL Sun Cube printer which is a very cheap DIY hobby printer and it actually does an acceptable
job (actually amazing given the price) and produces about the minimum quality I want. I do hope quality will be better with my big machines.

I have a feeling that on the Core/Cartesian issue it is just splitting hairs - if one was far better than the other everyone would know it and would not be making the inferior design!

Hi Dentist

I wanted to thank you for the advice on the hotbed, tooling plate looks exactly what I need! By PEI do you mean Polyethylenimine ?
I looked it up and it appears to be pretty good but very expensive for a large area. I am currently using a really cheap hair spray, its €1.50 for a
huge tin and works perfectly, prints pop off when the bed is cold.

PEI is ployetherimide, similar to kapton. see: [www.amazon.com]
I put a piece of 60 mil PEI on SoM at the makerspace and in over 2 years no one has managed to damage it, and after a recent 10 minute clean-up, prints still stick like they did when it was new. I used 30 mil PEI on UMMD because I know how to remove prints without destroying it. I haven't tried to cut it myself, but I believe you can just score it with a razor knife and snap it off.

---

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

Quote

I have a feeling that on the Core/Cartesian issue it is just splitting hairs -

I find it much better to lower the bed occasionally (CoreXY), than jogging it back and forth all the time. Especially a large bed can be a PITA in a Cartesian printer.

Hi Dentist
Thanks for the advice. I might give it a try, just wish it wasnt so damn expensive. Strangely that material is used in the manufacture of some kinds of paper so I cant understand why it is so expensive. I am wondering if
it is just heavily marked up because it is being sold for 3d printing, kinda like putting the word "gun" in front of anything (like oil) instantly puts the price up by 300%!

I do have one last question for you, as the bed is going to be very heavy at about 15 Kg I think 2 x nema 23 would be best or do you think nema 17s will suffice as its running a screw rod which amplifies the force?
Lastly (ok thats 2 questions) are there any issues with mixing motor sizes - I am thinking on using a nema 23 for the Y axis (moving the gantry) as the gantry will be 1 meter and nema 17 for the x axis (moving the print heads).

By the way I just printed my first prototype light (I make led grow lights) in form futuras "blinded pearl green" PETG and it came out looking stunning with a iridescent sheen like a butterflies wing. Their instuctions are crap though, they give the print temp as 190 - 225 and it took 245 - 250C to come out right finish wise and also to get rid of the globbing and stringing everyone complains about with PETG. Going to try my new full metal titanium hot end with it.