What would you build if you were starting again now?

[pixelpieper]

@zapta a TMC2209 is about 1.5$ on digikey if you buy large enough quantities and can be SMD assembled while soldering the stepstick headers is fairly expensive - I doubt there are huge savings by keeping the drivers of the PCB.

[o_lampe]

@deckingman said in What would you build if you were starting again now?:

@o_lampe said in What would you build if you were starting again now?:

A big plus for Deltas here, they only need 3 steppers for the whole motion system, where others use 3 steppers only for Z-axis.

I can't buy that argument. People might choose to use multiple Z motors but that's a personal choice and a single motor driving multiple screws is a perfectly viable solution, even with a large heavy build plate. So whether it be Cartesian, or CoreXY then 3 motors are perfectly viable for X, Y and Z. My personal preference is CoreXY because for all moves other than 45 degree infill, then both motors contribute to motion. And for those 45 degree moves which use a single motor, there is an increase in available torque due to the gearing effect of the belt path.

Point taken; we don't have to use more than one Z-driver. But the real life tells a different story.
Some even use two X and Y motors, it must be an unknown virus

[deckingman]

@o_lampe said in What would you build if you were starting again now?:

@deckingman said in What would you build if you were starting again now?:

@o_lampe said in What would you build if you were starting again now?:

A big plus for Deltas here, they only need 3 steppers for the whole motion system, where others use 3 steppers only for Z-axis.

I can't buy that argument. People might choose to use multiple Z motors but that's a personal choice and a single motor driving multiple screws is a perfectly viable solution, even with a large heavy build plate. So whether it be Cartesian, or CoreXY then 3 motors are perfectly viable for X, Y and Z. My personal preference is CoreXY because for all moves other than 45 degree infill, then both motors contribute to motion. And for those 45 degree moves which use a single motor, there is an increase in available torque due to the gearing effect of the belt path.

Point taken; we don't have to use more than one Z-driver. But the real life tells a different story.
Some even use two X and Y motors, it must be an unknown virus

It has been known for people to fit extruders on additional gantries on Delta printers, such that they follow the Z axis. So even Delta style printers are not immune from the "additional motor virus".

[dc42]

It is because there are several types of machines that need more than the usual number of motors that we decided to make the Duet 3 Mini capable of driving 7 motors before needing CAN expansion. We originally designed it for 8 motors using a 3-driver daughter board, but we ran out of microcontroller pins to control the 8th motor.

However, on a Voron or similar machine with direct drive extruder, using a CAN-connected tool board on the print head to connect the extruder motor, heater, thermistor, hot end fan, print cooling fan, Z probe, X endstop and accelerometer saves a lot of wires.

[DIY-O-Sphere]

@dc42
With the restriction that the voron (2.4) uses an inductive sensor, and that is currently not yet supported.
This is exactly the point at which I gave a friend the recommendation for toolboard and mini and he now has to lay 3 additional wires.

[dc42]

@diy-o-sphere said in What would you build if you were starting again now?:

With the restriction that the voron (2.4) uses an inductive sensor, and that is currently not yet supported.

I can think of no reason why you can't connect an inductive sensor to a tool board, in the same way that you would connect one to a Duet 3 main board.

[opentoideas]

so what are the inherent strengths and weaknesses of the current options?

Delta - tall for the build height reliable and proven looks like multi head printing is tricky though not impossible

CoreXY - strong structure but quite large footprint for the build area. long belts and multiple idlers "may" introduce errors as it wears

CroXY - looks the strongest again bigger footprint for the build area penalty for strength is the greater moving mass and complexity of multiple motors

these are the few things I have noticed but they are less important than reliability and print quality. speed is always welcome but subservient to print quality

[DIY-O-Sphere]

@dc42
https://duet3d.dozuki.com/Wiki/Duet_3_firmware_configuration_limitations#Section_Semi_permanent_limitations
"Z probes connected to expansion or tool boards are limited to types 8 and 9."
Or have I misunderstood something?

Edit:
Got it...Typ8 is similar to Typ5,,,should work....

[sebkritikel]

@opentoideas said in What would you build if you were starting again now?:

so what are the inherent strengths and weaknesses of the current options?
CoreXY - strong structure but quite large footprint for the build area. long belts and multiple idlers "may" introduce errors as it wears

these are the few things I have noticed but they are less important than reliability and print quality. speed is always welcome but subservient to print quality

Instead of a CoreXY, why not an equal volume 'cartesian' (not to be confused with a bed-slinger) printer?

• Arguably simpler belt and drive component routing (X moves in X, Y in Y, Z in Z)
• Shorter belt segments (lengths are only ~2x a single axis length)
• Equally strong/rigid structure
• Marginally heavier gantry moving in the Y directions (since you're moving the X axis stepper)

[opentoideas]

@sebkritikel

No reason, I would put that into the CroXY camp though as the geometry is more of a match just without the second arm but its semantics.

I was just going for the general machine types as there are many different designs in each camp but allong similar principles that i would think the strengths and weaknesses would hopefully match.

Like I said those were just my simple and uneducated views and by no means exhaustive or even correct, just a starting point and I am hoping those with knowledge and experience can explain what the benefits and weaknesses are as they see them.

The more I have thought about it the more I think about DC42's mention that his delta has been running years with the least trouble and while there are many many designs that work in complex and interesting ways the delta at least in terms of mechanical components is simple tried and tested....

This led me to ask the question why all these new super complex machines? Why are they so popular when the delta just works?

[fcwilt]

@opentoideas said in What would you build if you were starting again now?:

This led me to ask the question why all these new super complex machines? Why are they so popular when the delta just works?

Well I don't consider my printers "super complex". Yes, one has 6 steppers (1 X, 2 Y, 3Z) but it does what I want it to. I have no multiple tool printers.

I started with three mini-deltas and then got a classic Cartesian, an FT5. The FT5 print quality was so much better that I gave up trying to get the deltas to generate the prints I wanted and disposed of them.

Frederick

[pixelpieper]

@diy-o-sphere I have an inductive Superpinda on my V2 connected to the 1LC, works beautifully.

[opentoideas]

@fcwilt ok perhaps not super complex but mechanically a delta is almost as simple as it gets and I suspect as a result needs to be a far higher tollerance to ensure everything works. Kits from china are less likely to be as high a tollerance as needed and would make the setup challenging but once done or corrected its not got much to go wrong or out of allignment.

With everything at right angles I can see the box design being easier to get correct but there are a lot of moving parts all wearing differently so tweaking and maintenance may be a long term issue.

Again just thoughts that spring to mind not experience of the machines talking

[dc42]

@opentoideas said in What would you build if you were starting again now?:

This led me to ask the question why all these new super complex machines? Why are they so popular when the delta just works?

The delta is great if you just want one tool with a single Bowden extruder. Some users have also built deltas using light-weight direct drive extruders (which is becoming more practical now that there are more light-weight extruders, such as the Orbiter), and others use flying extruders. But if you want to go muti-head, it's easier using Cartesian and CoreXY designs. That said, another user has posted details of a tool-changing delta on this forum.

[opentoideas]

@dc42 lol I just want to make things.

My current task is a case system for a mining cluster and i have the first of 3 over 30hr prints going now and another 2 or 3 similar prints to follow

I want a printer that "just works" when I want it to rather than spending days getting it to work just right on test pieces each time.

I dont mind spending time making it right and my trusty CR-10 is getting to that point but its not the best starting point and there is only so much to be done before a complete redesign and rebuild is now the only way to go.

For me, I expect by the time i need or want multi head printing it will be established enough that it may warrant another change but for now single is fine... although the # design is interesting from the multiple simultaneous prints as my 30 hour print is 18 parts filling the print bed creating the long print time

[opentoideas]

This is current job. Lost some bed space with the MGN rails but was worth it

[deckingman]

@opentoideas said in What would you build if you were starting again now?:

...............This led me to ask the question why all these new super complex machines?............................

Possibly because of laziness or incompetence, but let's be generous and say it's simply a lack of understanding. Multiple motors on one or more axes has been mentioned often in this thread. Why would one want to add that complexity? In the case of a Z axis, it's because people think it is somehow difficult to build a bed which is flat, level and stays that way. So rather than building it properly with a single motor driving multiple screws, they add a motor to each screw so that they can use the firmware to make it level. That just adds a level of complexity for no gain. In fact it makes matters worse because every time the power is cycled, the motors jump by up to 4 steps relative to each other so that firmware levelling process has to be done every time the printer is turned on and off. Similarly using multiple motors to drive an X or Y gantry is because people can't figure out how to stop the gantry twisting. So they add a motor to each end rather than fix the problem mechanically.

Or maybe it's just bad design. People build machines or start with basic kit, see what problems there are, then add more parts to cure the problems instead of going back to basics and fixing the fundamental design flaws.

Whatever route you choose, take your time, build it properly, build it rigid, use good quality parts and materials. Those things are far more important than what type of kinematics to use.

In the final analysis, Cartesian, CoreXY, Delta and derivatives thereof are all capable of producing good quality parts. Arguably, it is the hot end and extrusion system which will determine the print quality and speed, so how that hot end is moved around in space is somewhat irrelevant.

[opentoideas]

@deckingman said in What would you build if you were starting again now?:

Possibly because of laziness or incompetence, but let's be generous and say it's simply a lack of understanding. Multiple motors on one or more axes has been mentioned often in this thread. Why would one want to add that complexity? In the case of a Z axis, it's because people think it is somehow difficult to build a bed which is flat, level and stays that way. So rather than building it properly with a single motor driving multiple screws, they add a motor to each screw so that they can use the firmware to make it level. That just adds a level of complexity for no gain. In fact it makes matters worse because every time the power is cycled, the motors jump by up to 4 steps relative to each other so that firmware levelling process has to be done every time the printer is turned on and off. Similarly using multiple motors to drive an X or Y gantry is because people can't figure out how to stop the gantry twisting. So they add a motor to each end rather than fix the problem mechanically.

yup I resemble that remark

the first "modification" I made and the biggest continuing pita. now I know slightly more the error is obvious but it seemed like a good idea at the time

[fcwilt]

@opentoideas said in What would you build if you were starting again now?:

@fcwilt ok perhaps not super complex but mechanically a delta is almost as simple as it gets and I suspect as a result needs to be a far higher tollerance to ensure everything works. Kits from china are less likely to be as high a tollerance as needed and would make the setup challenging but once done or corrected its not got much to go wrong or out of allignment.

With everything at right angles I can see the box design being easier to get correct but there are a lot of moving parts all wearing differently so tweaking and maintenance may be a long term issue.

Delta's are indeed rather elegant which is why I started with them.

And clearly lots of folks can get them setup and working just fine.

I just found them rather frustrating. The FT5 was so simple to get working. I had spent months on the delta's trying to get them to work the way I wanted. I got what I wanted from the FT5 in a week.

Maybe someday I will try a large delta but for now I am having fun exploring different kinds of "Cartesian" printers.

[opentoideas]

I think 3D printing and frustration go hand in hand.

I started with my cartesian bed slinger and also found the first months a frustrating and unsatisfying experience however eventually something must have sunk in as it all started coming together.

For this current job I finally ran out of the PETG I was printing with from the start so figured I might as well use up some of the PLA I had kicking about..... que another frustrating couple of days dialling in to a different although alledgedly "easy" material

Fortunately the frustration tends to be short lived and the satisfaction of a good print keeps on rewarding

While speed isnt the ultimate aim I guess I would want to do better. At the moment while I can push 150-200mm/s I only use that for very rough tests and for any actual production part for use I need to be at 50-100mm/s depending on the part.

While it always feels slow what are realistic speeds for a reasonable finish with a well designed modern printer? I am guessing there isnt a huge improvement to be gained possibly double?