Different way to fight inertial force?!

The usual way to fight inertia is to make the printer-frames as stiff as possible and sometimes even screw them down to a heavy baseplate.

What if we try to lead a huge part of the inertia forces away from the frame?
I thought of using four ball caster rollers under the printer, so the printer can move freely. Add a little damping to the movement with 1/10th scale rc-buggy dampers.

Would that help my weak steppers to run faster without shaking the hole printer apart?
-Olaf

Frame stiffness is about preventing unwanted motion of the extruder nozzle relative to the print bed because that affects print quality. If you put a floppy frame on a set of wheels, you still have a floppy frame and the extruder nozzle is still going to move relative to the bed. A floppy frame already moves when other parts of the printer move, so it won't improve anything to let the whole frame plus a wheeled dolly move. If there were some way to make the frame move as a whole you might have something. Wait, there is just such a way- make the frame rigid!

Think about the X axis. If you were to mount the motor on springs, the body of the motor would rotate a little each time the motor accelerates. When that happens, the extruder nozzle doesn't go quite as far as it is supposed to. When the motor stops accelerating, the springs would "unwind" and return the motor to its "home" position, and deliver the nozzle to its target position. The problem is the extruder is extruding plastic the whole time all this spring winding and unwinding is going on. Imagine what that is doing to the surface of the print- since you're modulating the speed of the X axis, you're going to modulate the width of the plastic line being laid down which will probably result in ripples in the surface of the print. A floppy frame is roughly the equivalent of mounting the motor on springs. Now imagine this scenario happening in two dimensions at the same time where you have X and Y axes moving. The Z axis motor probably doesn't contribute much to the error, but when the whole frame is flexing because of torque applied to it by the X and Y motors, there is probably some Z component in the extruder nozzle's motion as well.

If you want to prevent unwanted motion between the nozzle and the bed, you have to make the printer frame rigid. Doing so causes the torque on the motor bodies to to be coupled to the printer frame as a whole, which is much more massive than the motor body and so moves much less. The extruder goes where the controller tells it to go because it doesn't have a choice- the springs it has to play with are just too stiff.

My printer has a very rigid frame and has operated on a wheeled cart (from a school auction- it was used for TV/VCR/DVD) for over 2 years. The cart wobbles and shakes a little when the machine is running. I have occasionally had the machine on a solid work table and have compared print quality and found absolutely no difference.

There is no substitute for a rigid printer frame.

Thx for the explanation DD,
but stil...
If a boxer has to take a punch while his head is fixed, he will go K.O.
But, if he has the chance to move his hole body away from the punch, he will see 12th round.. maybe
-Olaf

A boxer isn't trying to extrude plastic in a perfect pattern while he's taking punches.

Quote
the_digital_dentist
A boxer isn't trying to extrude plastic in a perfect pattern while he's taking punches.

It was a lame example of newtons 2nd law about "action and reaction ".
-Olaf

The other option is to reduce the moving mass as much as possible.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

As long as you are moving any mass at all, and you always do, a rigid frame will always be better than a non rigid frame.

In your analogy with the boxer: in the case of a 3D printer, it's as if the boxer is punching himself in the face, and it won't matter if his body is fixed in place or not.

I think I can start to see where you are coming from but the Boxer analogy is throwing things out.

if you want to reduce for example the inertia on the Y bed when decelerating from a move in one direction then it's more like stopping a train.
if you place an immoveable object in front of the train to stop it then the rest of the train and everything on/in it wants to carry on forward. to reduce the effect of the inertia on the passengers you would have their chairs on sliding spring loaded rails so their deceleration is separate to the train and over a longer distance.

for your idea of using rollers on the bottom of the printer this would reduce the force of the frame against the table (or the train against the immoveable object) by having the frame move in relation to the table (or have the immoveable object move with the train as it collides). but in the case of a 3d printer we want the hotend nozzle and Bed to not move in relation to each other except in precise movements (like with the train it is the passengers we want to not move due to inertia) so unless you can have the carriage/bed move the hotend or print surface to the desired location without inertia then that wont work.

The way to reduce inertia is to decrease the mass of your moving object, or limit the acceleration of your object so there is less energy in the object for deceleration.
so we can print slow with heavy carriages or beds. or we can reduce the weight and print at faster speeds. or we can have a strong frame in order to decelerate our objects faster and remain within the resolution of movement we want the printer to have.

Quote
Zerker
so we can print slow with heavy carriages or beds. or we can reduce the weight and print at faster speeds. or we can have a strong frame in order to decelerate our objects faster and remain within the resolution of movement we want the printer to have.

It isn't a matter of "or". You keep masses minimal, and you print at speeds/accelerations that your motors/controllers can handle, and you use a rigid construction including frame, rails, bearings, etc. All three are important contributors to print quality and speed and even reliability.

I was meaning in regards to current printer designs, attempting to explain why we are heading towards stiffer frames.
i.e if you want to increase speed you need to firm up and or lighten up or continue at current speeds.
AND works too though

The way to fight inertia is to reduce moving mass

Counter weight is easiest way to reduce inertia of the Hotend.
I saw one Deltabot 3D Printer that was using counter weight on the net.

Newton's three laws of motion fails in rotational body.
Fast rotating body creates weight reduction effect around it. Eric Laithwaite's famous Chrimas lecture had some experiments.
That was career killer to Eric Laithwaite as a promising scientist.
Two opposite rotating magnetic fields reduces weight around it too.

Free fall speed of a strong magnet inside non-magnetic metallic tube gets slow down.
In the context of conventional physics, they explain it using Lenz's law or Eddy current simply dismissing relationship between
gravity and magnetic force. Interesting?

Electrogravitics is well-known technology to manipulate gravity.
Out of curiosity I bought a scientist's book that explains how electrogravitics works. That was one big rabbit hole to me.

Scifi fantasy stuff is OK, but since it doesn't work, you may as well study witchcraft. That doesn't work either.

Isn't real physics interesting enough?

Interesting side note... I personally believe, that light speed is not the end, like common physics teach us. It´s just the beginning of something else...

Back to topic:
I think the ball caster friction is to big to make a real difference.

What if I let my printer swing freely? Hook up some cords at each corner and route them to a single hook in the ceiling? Like a 3D-swing.
Or let it float on water? That would eliminate the tendency of a swing to center itself.

Ì just don´t know, how to prove the concept. Printing a cube with increasing speed and compare that with the same gcode printed on the ground?
-Olaf

Zerker and the digital dentist gave you pretty much perfect advise.
Forget about SciFi and the whole idea of somehow suspending the printer to somehow reduce the inertia of printer internal movement is just stupid, Newton would yell at you for that
1. Reduce the moving Mass
2. Build a stiff frame
3. Make sure the drive system hasn't to much give

If silence is the goal then also look at 32 bit electronics and high microstepping drivers. This will give a smoother movement.
Also check for using bushings instead of bearings, they are not only more silent but also weigh less.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

Here's an interesting idea... 3d shock absorbers

A spring isn't a "shock absorber" because it doesn't have any damping.

Most printers have a frame that is floppy/flexy (aka a spring) and they derive some rigidity (at least at the base of the machine) from being set down on a hard, massive surface. If you put a spring under each foot of the printer you'll allow it to flex any where it wants to. That's a bad idea, though it might reduce the vibration coupled to the table and make the printer run quieter (which was the point of the original design).

If you really wanted to improve the print, but didn't want to go to the effort and expense of actually fixing any of the problems (which begs the question how much do you "really" want to improve the prints?), you could borrow a trick from the audiophools and put spikes under the printer. If you wanted to both improve the prints and reduce the vibrations coupled to the furniture, put a massive slab of something hard under the printer with spikes on the printer feet and then put springs under the massive slab.

None of this screwing around with springs/spikes/massive slabs of I don't know what is going to fix the real problem of a nonrigid frame and flexible guide rails, but it might make you feel better about your printer, at least for a while. That's pretty much how it worked for the audiophools...

What the digital dentist said.
To get rid of the rest of the noisy vibrations i had, i glued a couple of 50cmx50cm tiles upon another and put soft silicon feet under the stack. The printer has simple hard rubber feet and stands on the pack of tiles.
It is overkill, but at the same time super silent. Since it gets rid of the low frequency rumble my wife complained about it was worth every minute of effort.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

o_lampe has a point. Endlessly saying that stiff frame equals obeying Newtons laws of motion does not make it so. Take an idealized Mendel Prusa I - the one with the threaded rods. If the extruder and X carriage is accelerated along the rails it tries to deform a rectangle consisting of the X rails and triangle joining rod (horizontals) and the Z rails (vertical) which couples at the top to a second deformable rectangle consisting of the triangles (vertical) and the top and bottom joining rods (horizontal). If the bottom is secured then the result is that all forces act to deform these rectangles into rhombuses (or whatever the flexing of the rods and plastic parts will allow.) The result will be that all of the deformation will result in the extruder not being where it should be. On the other hand, if the bottom of the frame is able to move freely, at least some of the force will result in the bed being able to move away from the extruder in the desired direction. I believe that this is what o-lampe was saying.
There are supremely accurate dental milling machines made by a company called Renishaw. These were designed by using a good understanding of physics - including Newton's laws. A rule based method of simply making everything massively rigid would also have worked - but the resulting machine would not have fitted on a desktop.
Mike

I wasn't able to find that Renishaw system, but it is certainly an option if you happen to be a manufacturer of high end measuring systems.
If not, you have to work with what technology you can pay for or construct yourself.
So please lets discuss this in the proper context.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

A professionally engineered system that sells for $50-100K can be simulated in great detail using FEA/physcial simulation software, and the design optimized and refined. The average hobbyist doesn't have access to such software, and doesn't usually have the skills to use it even if they did have access. Most also don't have knowledge of and access to machine tools to fabricate the device. When you consider the man-years that go into design and fabrication of that machine, does a single person have a chance of doing likewise in less time (they probably will in 20 years because of technological improvements)? The principle of "overbuild" becomes the next best option.

Until FEA/physical simulation software is readily available and extremely easy to use, just ask one question of each part you plan to put into the machine: are you certain it will be adequate to do its job? If the answer is "no", find or make a "more" part- more bigger, more rigid, more torque, etc., until you can answer "yes". Then, once you think you have a good set of parts identified, you figure out if they'll work together or what you'll have to do to make them do so. That is how you overbuild a printer (or anything else) and get a good result.

I spent a couple years designing and building my printer, then swapping out and redesigning parts of it, always aiming to improve print quality and reliability. In hindsight I probably could have avoided some of the dead-ends and outright mistakes had I stuck more closely to the principle, but mistakes teach more than successes, so I would not consider the time and expense wasted. The expense and loss of security are the trade offs you always make when you decide to design/build something yourself that pushes the limits of your knowledge, instead of buying a kit and having someone else tell you how to put it together but not why it is designed the way it is.

Quote
Srek
I wasn't able to find that Renishaw system....

[www.dentalaegis.com]

Quote
Srek
...So please lets discuss this in the proper context.

In precisely what way did I not discuss this in a proper context? o-lampe asked if mounting the printer on a compliant platform (ball caster rollers) would help. This is a technology which is not very expensive or unmakable - precisely the proper context. My pointing out that he was right is also the proper context

Mike

Quote
leadinglights

Quote
Srek
I wasn't able to find that Renishaw system....

[www.dentalaegis.com]

Quote
Srek
...So please lets discuss this in the proper context.

In precisely what way did I not discuss this in a proper context? o-lampe asked if mounting the printer on a compliant platform (ball caster rollers) would help. This is a technology which is not very expensive or unmakable - precisely the proper context. My pointing out that he was right is also the proper context

Mike

Sorry, the link does not include any information on the technology behind it and i couldn't find any more information on how the renishaw mill works different than a Roland, which is a simple cartesic system. I am none the wiser except that i now know the example you show is for a machine that mills away a couple of mm³ of an item that is already only the volume of very few cm³. I still wouldn't call that the proper context since all of this does not help o_lampe to build a better printer.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

Ok, now it get´s interesting.

From : No, it doesn´t help to: No, you can´t do it ( good enough ) is a big step forward...

I agree with Srek and DD, it is worth making the frame rigid and reduce moving mass, but at reasonable cost.
Spending 100$ extra for a steel frame where the hole printer was 320$ is out of question for me. Because I still have these weak 0.4A steppers, which would need an upgrade too, after making everything rigid and heavy ( bed ).
When I can try to get a similar result with less money, I try that road first

My goal is to build a cheap printer.( with a good print quality at medium speed. )
-Olaf

Edited 1 time(s). Last edit at 07/12/2015 04:15AM by o_lampe.

In most cases more stiffness can be achieved with just some small design changes when it comes to frame connections and basic frame design. For example a cubical printer frame made from thin extrusions can be made a lot stiffer by adding cross bracing, which can be some cheap steel wire. If you create the printer from printed parts it helps a lot to have access to an FEM software or similar, or just have a friend who is a mechanical engineer. Mechanical design isn't magic, a good carpenter can give a shitload of good advice.
If you have the opportunity go and check a local hacker- or makerspace. Usually you will find people with some experience on mechanical design that are willing to help and it is a lot more fun to do this stuff together.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

Srek, if I would redesign the prusa i3 frame, I would use a sandwich method for the mainframe.
Put two thinner sheets on front and back of the z-steppers and redesign the z-stepper- and top z-rod mounts to connect to both sheets.
This would reduce flexing a lot, right?
-Olaf

Sheets tend to bend unless folded or stiffened by attached rods. Personally i go with cubical extrusion designs mostly.
The i3 was never a design i gave much thought, since i did not like it on first sight.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

If I were trying to improve an existing printer, I would list the problems I know about and assign some sort of weight to them, then attack them one by one starting with the ones that have highest weight. Start with outright print failures- what causes them in your printer? Then move to quality issues (which can be another form of print failure). It's a bit arbitrary, but having a system that helps you rank the problems can make the process easier and will help you get the most improvement for the least cost. Don't worry if some of the weights are wrong. As you think about each part, think about what its ideal behavior/performance should be in the context of a 3D printer and about how that can go wrong and what the part is really doing in your machine. That includes the electronics as well as the mechanical stuff. Don't be afraid to push and pull on things and watch for unwanted movement- print quality ultimately comes down to controlling the position of the nozzle relative to the bed. Anything that allows uncontrolled motion of either one relative to the other is a problem.

One test you can conduct is to have the machine print some long straight lines on the bed and while it is printing those lines, start pushing on the frame at different points and observe what happens to the printed line. Put a finger on the extruder carriage and apply a little force and do likewise with the bed, and observe the result.

Edited 1 time(s). Last edit at 07/12/2015 07:37AM by the_digital_dentist.