Custom Printer Build

1

Not sure if this is the correct category, but I couldn’t find a better one. Anyway, a little background first. One of the offshoots of my 3d printing hobby is 1/250 scale RC models of famous ships. I have only had one complete success, a model of the Edmund Fitzgerald, but I have another model in progress, and a few past ones that I call learning experiences. So obviously, the ships that I want to do are pretty long, ranging from approx 60cm on the very bottom end, to over 130cm on the top end. Most are in the 80-120cm range. As I print the entire hull and superstructure, I have had to print in segments up till now. This works, but creates a LOT of extra assembly work and frustration, as well as being weaker than a single segment hull. So enter my idea of a quite specialised FDM printer. A quick note, I am aware of the Creality CR-30, but don’t want to go that route for detail and strength reasons primarily. I am not 100% sure I will go ahead with this build but I am seriously considering it, and figured I couldn’t go wrong asking for feedback from you all.

So my basic idea is this. A 1400mm long by 200-235mm wide bed. The Z axis will probably be 130mm, give or take. The idea would be to do aluminum extrusions along both sides of the bed (I haven’t settled on an exact configuration yet, but it will probably consist of two extrusions bolted together to get the desired profile. I would have v-slot wheel carriages running on the extrusions and a classic ender 3 style gantry would be mounted on these carriages to travel along the length of the bed. I would probably use 4 carriages, 2 on each extrusion, spaced aprox 10 cm apart to provide better stability. I realize this configuration will have a pretty high moving mass, which will limit print speeds, but that is allright with me, I am not aiming for a high performance printer. Doing a long h-bot style configuration might work and would probably be higher performance, but it would limit side access somewhat, which I need, as there won’t be any top access where I would put it if I build this thing.

For the print bed, I am thinking probably 7 rows of screws/bolts 3 bolts across to mount the print bed to the printer, with silicone heat mats in between the rows. This calculator https://jscalc.io/calc/mVSHnXdt0sjpeUFM if it is accurate, tells me 250 watts would take about 7-8 minutes to heat the bed; that is a while, but the prints will be around the week long mark a lot of the time, so that doesn’t seem to be a big deal to me, just as long as it can attain 60°C. So I have a couple questions maybe some here can answer, and I would love feedback, constructive criticism, and suggestions too.

Question #1: could a mosfet be added to any off the shelf mainboard go supply the power to the bed so 250+ watts doesn’t need to go through the board? Likely would be a BTT SKR Mini E3 V3.

Question #2: Regarding auto bed leveling, can Marlin firmware be configured to have a rectangular probe grid of 4 by 23 instead of a square grid? That would be ideal.

Question #3: Any preferred places to buy v-slot aluminum extrusions? Maker parts.ca has extrusions up to 150cm, but I would prefer 160cm. Thanks for reading this long speil, and I am looking forward to seeing your replies!

Edit, added paragraph spacings.

2

Interesting project and hobby.

Questions:

  1. Could a MOSFET be added to the controller board so that 250 Watts is supplied to the heated bed? 250 Watts isn’t an unusual amount of power to drive a heated bed (it’s just over 10A at 24V) so, while you can add a MOSFET, there probably isn’t any need for it. I would say you need to do some more research into heated beds.as the bed geometry is quite unusual and I don’t think there are any single DC ohmic heated beds of this size commercially available.
  2. Can Marlin be configured to have a rectangular probe grid of 4 x 23 instead of a square grid? Sure, it will just take you to go into the code and make the required changes which would probably take a few days if you’re experienced in programming but haven’t made changes to Marlin before so you have to figure out where the code is.
  3. Where to get long V-Slot rails? Openbuilds.com - they have them available up to 2m long and you can have them within a few days.

I’m not sure where you’re getting a 1400mm x 235mm heated bed and I would have thought that you’d be looking at a machined aluminum plate with (four) AC heating pads underneath (like the Voron 2.4) driven by a Solid State Relay.

Along with that, if you’re doing a bed slinger (Ender 3 “RepRap Cartesian” style) you’re going to need an awful lot of space for it and it’s going to be challenge to make it acceptably rigid as well as have steppers (and probably leadscrews) that can move the bed at any kind of speed. I suggest you look at a CoreXY architecture instead and it’s going to be a challenge finding linear rails in the length that you’re looking for.

Very interesting project and one that you should document as you go through it as you’ll have some challenges and if there’s a community following you, they can probably be of help to get you over the rough patches.

3

10-4 on the bed wattage, that makes sense about lots of printer beds being that high draw, I guess I didn’t think of that.
Experienced in programming I am not, but I guess, with help, it sounds like I could overcome that hurdle possibly.
I will check out open builds.
My plan for the bed would be getting an aluminum plate from a local metal working shop. I think that they will have such. The bed itself will be completely stationary. A bed slinger of that size would certainly need a lot of room! The plan would be to have the gantry do all 3 of the axis. I was thinking 6 40w(or the nearest wattage to that available) silicone heating pads of about 200 by 200mm, I can tweak the bed size plan a little as necessary to fit the heat mat sizes that are available. The 21 screws that would be used to mount the bed to the printer frame should hold it fairly rigid, and allow for a little ability to heat the bed up, then adjust to compensate for the warp that could happen.
My primary reasons not corexy with linear rails is the expense and side accessibility, plus, I really don’t need a high performance printer. I won’t be printing a huge number of ships. I am willing to slow down to get good results. Thanks for the suggestions!

4

I like your idea, a long slender printer, purpose-built. I know you can order the AC bed pads in custom size as well. If you went with a BTT-type configuration you may be able to control multiple SSRs from it as well.

We also have all the Vslot and C channel open builds extrusion in up to 2.5 Meter lengths. The only issue is we cannot ship it, would have to be local pickup. We can also cut and tap as per your requirements.

Anything you see on open builds we will have. This includes the mending plates, motor mounts and all associated hardware.

5

Thanks Jason! I saw you had shorter extrusions on your site, but unfortunately, I am too far from you guys to come pick it up. :cry: Open builds has the extrusions I would need. Which BTT board can control multiple SSRs? That would likely allow a more even bed temperature, and would reduce power supply size requirements too if I went with 120V AC heating instead of 24v DC.

6

you could actually use a couple of them, the most universal would be the SKR line, they have the add on cards that would give you additional outputs. Mostly I would be looking at a combo of the following

If you need additional motor controllers you can add the following

Alternatively, you can use the Octopus, it has a ton of PWM connections on it.

Your firmware would be completely custom so that is going to be a bit of a feat all in of itself. but its possible,

Another different route would be for you to use the DUET line with REPRAP firmware, with this you can enable CANBUS and really control as many SSRs as you want to.

7

I have seen a few of these large custom printers on you tube, this link is for one on Thingiverse.
He even shows how he set up the bed ( 4 regular beds into one).

8

Hmm, clearly I need to go down the rabbit hole of research on boards, firmware compiling and the like. I’ve not had a lot of luck with firmware compilation up till now, let alone a completely custom setup. Looks like a fair bit of learning involved, which would not be all bad, as far as that goes. No doubt that controlling the temperature in multiple segments would be the technically best option. I’m not sure how much effort to expend on the “best” or go a simpler route with the entire bed controlled by one thermistor and power supply source. Don’t let perfect be the enemy of good, as “they” say. I will have to find a balance there. I do plan on insulating the underside of the bed, probably with a foil backed insulation if some type. That should reduce energy requirements once brought up to temperature, and to some degree, help keep the bed a more even temperature if I do a good insulation job. As to multiple steppers, that reminds me of a question. Can two steppers on a single axis be controlled off one stepper driver? On my enders with BTT SKR Mini E3 V3 boards, I can adjust stepper driver voltage. If I up the voltage and use a y harness, can I use two steppers off of one driver? I might do that for the axis moving the gantry along the bed.

9

I will have to check that out!

10

You’re going to have to be careful about having multiple heaters and sensors attached to the same heated bed as there will be interactions as I’m not sure how you will do your PID tuning while ensuring nothing goes out of control.

The technically best heater implementation would be a “MISO” (Multiple Input, Single Output) PID which would have a thermistor on the bottom side and one on the top side (probably at a corner) that would ensure that the bed warms up evenly in the least amount of time without exceeding any maximums. I’ve been in a couple of discussions about MISO PID implementations but I’ve never seen code that works properly - nobody with a background in digital control theory seems to have looked at this problem. I understand the theory behind it but I don’t have the practical experience writing advanced PID code.

Personally, for this application, I would use one heater control and one thermistor (probably put in the bottom middle) and then wait an hour or so for the bed to heat evenly. I would also use a heat camera to ensure that the heat is fairly evenly distributed. You’re not looking at that much power and it’s better to start as simply as possible.

It’s probably a good idea to put insulation on the back of the build plate as there’s a lot of surface area there for the heat to escape. It would also help the heat distribute more evenly and faster.

If you’re looking to drive a gantry with two steppers (which you can by wiring them in parallel with a “Y” cable as you ask) I’d again suggest CoreXY. The gantry will be lighter (no X axis stepper on it) and you’re effectively using two steppers to drive it.

One final suggestion; your posts would be a lot easier to read if you broke them up by paragraphs. The monolithic blocks of copy are hard to work through when there are multiple points and questions.

11

Ok, I’ll try to do better with breaking it up into paragraphs, I didn’t think of that.

I agree that the single thermistor method would be the simplest. I don’t have a heat camera, but I do have a infrared heat gun, so that should be enough to get the bed temperature sorted out.

As to the dual stepper issue, I get where you are coming from, but side access would be more of an issue then. Another downside that I didn’t mention is I will have limited vertical space, and I think a corexy setup will take up more room on that axis. I’m not really worried about one stepper being powerful enough to move the gantry, as I won’t be pushing acceleration and speed hard at all. I figured I’d ask, because it’s an option I am considering. The other option is just a dual output stepper with shafts to each side to drive the belts moving the gantry.

12

I was initially thinking about “cheating” It a little, You could use Bed temp to control one bed and the Chamber temp setup as another. It just so happens both would be operating at the same temp. That way you could keep the thermistors and controllers separate.

I know with Duet you could rename the label and call it LEFT bed and RIGHT bed and treat them both the same. The slicer would always see it as Bed and Chamber but you would have to do that translation yourself.

About the multiple stepper motors, Yes they will work on the same stepper driver. I have one printer I work on its original design used 4 60mm motors all off the same driver. Yes I know it was a bad idea but now in the newer versions they are all independent.

Quick answer, Can you… YES. Should you… Depends, usually not unless you are out of stepper drivers to use.

13

The chamber temperature input is something I hadn’t thought of. Definitely something to consider, although not all boards have that capability.

As to stepper drivers, that would depend on which board I used as to whether I’d have enough. Considering that I’ll not be pushing it hard, I will likely just go with a dual output shaft stepper to move the gantry.

1 Like
14

I don’t agree with the recommendation to have it built as an coreXY.

Belt lenghts would be enourmous for this buid-size, causing a lot of issues, it would be better, in all regards, to simply build it as a cartesian machine.

15

which filaments do you plan on using with this machine? you’ll need a lot of power to heat up the bed and the build area to succesfully print ABS parts in that build volume.

Regarding control boards, i can recommend the duet 3 and the fysetc big-dipper, RRF will be easy to setup for a custom machine regarding all the documentation there is available.

I would also recommend that you use MGN12 linear rails instead of v-slot wheels with v-slots, you can butt-up linear rails to build up to that size of an axis quite easily and without alignment issues.

given the size and mass of the bed, it would also be best to build the machine as a flying gantry like the voron 2.4.

16

What are the issues that you’re expecting from a CoreXY vs a Cartesian this size?

I’m assuming that the X/Y frame will be pretty heavily made for rigidity and 9mm or 12mm belts should be used or go with a long leadscrew.

17

Pretty much just PLA. Might be the odd bit of PETG or TPU, but the items I would build this printer to print would all be PLA.

I agree linear rails would be ideal, but I am trying to balance “the best” vs “good enough”. There are a few aspects of the printer design that I would likely change, or could be simplified if I went with linear rails instead of v-slot wheels. I need to price out more of the components to get a closer estimate of how much more a decent quality linear rail setup would actually cost. Don’t get me wrong, I realize that building a printer of this size is not going to be free, but me insisting on the best of everything for every aspect of the project will result in costs ballooning. So I am trying to balance that all out. This printer will not be needing to be super high output, so I am quite willing to dial back the print speed if it means the printer costs significantly less(to a point of course). So I will need to get a more precise estimate of costs of both options to decide which path I want to go.

18

I was thinking of something like this as a starting point.

image
image553×553 122 KB

A fixed bed with the print head as the z axis. Rigidity would have to be taken into account with a reduced speed.

19

That is an interesting configuration. What is the advantage of that arrangement vs having the entire x-y gantry do the z axis like the voron 2.4? Either way, the place I would have to put my printer would never fit a printer that used so much space in the z axis.

20

All moving parts are on the same plane parallel to the build plate also less parts because your xy axis arrangement does not need to move in the z direction only the hot end. I like the core xy arrangement but I would probably go with this

image
image800×526 23.5 KB

For each side of the y axis and the x axis for the size you want to build. A 9 mm belt is basically fixed in one place then. How much height would you require ?