Closed-loop control for low-cost 3D printers

The topic above is deliberately open-ended and proposes nothing more than closed-loop control, hopefully to encourage people to think laterally and very widely instead of being shackled by a specific construction.  The range of possibilities is enormously varied, probably infinite.

I will however express my own preferences, which are much narrower and more tightly directed.  Please don't be constrained by the following.

Personally, I think closed-loop 3D printers need to head in the direction of direct drive, avoiding intermediate transmission components as much as possible.  Not only would this eliminate loss of rigidity and the severe problems of slip and play and backlash, but it would also open up the possibility of printing our own motors using pancake designs (effectively linear motors arranged in a circle).

This direction is not in the slightest bit easy, but the elimination of transmission components would make this approach more viable at MEMS scales, which are on the path towards which all engineering is leading:  nanotechnology.  The machinery which builds the machinery which builds the machinery which builds the machinery ... of nanoscale systems is in our grasp right now.  It's going to be an interesting voyage.

Morgaine.

so your suggesting that you use servos with encoders? thats dandy and all but its much more complex to set up being that as far as i know (which is verry little i will admit) there arent any programs as side from mach3 that will read encoder data not only that but the reason most people sue steppers today is because they are much cheaper and simpler to use in simple hobbycraft

if you are talking about for the wave of commercial 3d printers? well Great! i definitely see the benefit to useing encoders to determine position it will provide more accurate movement and possibly a better print quality

Hi Michael,

Good point, thanks!

The feedback at the motor axle could help here because the speed (or better: accleration) of the motor could be controlled much more, thus making the movement smoother, and if you know what the next G-code is, you could take care to slow down the head in time to prevent overshoots. Don't know in how far that's implemented in current 3D-printers. Dead zones might also be somewhat compensated if known... I agree that that's a very hideous EE-domain fix for a Mechanical domain problem....

Michael Kellett wrote:

 

overall feedback can't deal with backlash, dead zones or mechanical hsytereisis very well.

It certainly can deal with backlash, so I'd be interested to read anything suggesting that it can't do it very well.

The reality is actually the opposite:  overall feedback returns an error function that contains (as a component) the actual backlash experienced by the work head, instead of a simplified error function that encompasses only a limited set of transmission components and/or which deals with each axis separately.  The needed information is certainly there.  It's a lot harder to interpret this actual feedback of course compared to simplistic single-axis feedback obtained nearer the motors, but it's the real thing, not an approximation.

In contrast, motor or feedscrew-level feedback isn't really closing the loop at all from the PoV of the work head, and so it can't compensate for unplanned interactions between axes and other structural problems, as it simply gets no data on them.  The consequence of this is that very high quality construction is still mandatory to avoid the unmonitored anomalies from appearing in the first place in such a half-open-loop system, which of course precludes its use for the subject of the topic, namely "low-cost 3D printers".

Morgaine.

First let's get our definitions in order, from Wiki:

in mechanical engineering, backlash, sometimes called lash or play, is clearance or lost motion in a mechanism caused by gaps between the parts. One source defines it as the maximum distance through which one part of something can be moved without moving a connected part. An example, in the context of gears and gear trains, is the amount of clearance between mated gear teeth. It can be seen when the direction of movement is reversed and the slack or lost motion is taken up before the reversal of motion is complete.

Note that bit about "can be moved without moving a connected part" - if the actuator is pushing the load along and needs to reverse the force (for whatever reason) the actuator must move back by the backlash distance. In any real system this takes time and no amount of feedback and no amount of cleverness of control algorithm can do anything about it (the time is set me purely physical constraints). You can try to minimse the effect of backlash by having a very fast slewing actuator but this often results in instability and increases costs. The problem is that in the backlash or dead zone the actuator is not connected to the load so in the time it takes to reconnect the load is uncontrolled.

Clever strategies to try to and reduce the impact are possible (and routinely and instinctively used by manual machine tool operators.). Under some conditions it is possible to ensure that the direction of applied force never changes in a critical place but these restrictions are often inconvenient.

A low cost mechanism may easily be so poor in it's mechnical performance that there is no gain in improving the control. Stepper motors are very poor at fast dynamic control so I can easily believe that there are many cases where control of the stepper motor shaft is no worse than attempting to control the load position, but at least has the advantage of cheapness and robustness (robust in control stability sense).

MK

Michael Kellett wrote:

 

Note that bit about "can be moved without moving a connected part" - if the actuator is pushing the load along and needs to reverse the force (for whatever reason) the actuator must move back by the backlash distance. In any real system this takes time and no amount of feedback and no amount of cleverness of control algorithm can do anything about it (the time is set me purely physical constraints). You can try to minimse the effect of backlash by having a very fast slewing actuator but this often results in instability and increases costs. The problem is that in the backlash or dead zone the actuator is not connected to the load so in the time it takes to reconnect the load is uncontrolled.

That's a very actuator-focused view (instead of being concerned mainly with the work head), and it's also a time-focused view whereas our main concern is that the work head be in the desired position (time is not the key requirement).  It's entirely normal to trade off time (speed of operation) to lower the cost of end products --- this is why low-end 2D printers are slower than higher end ones.

In order to get the work head to a desired position in axis X, the only requirement on the drive train is that this position is reachable in a monotonic traversal using actuator X.  It is not a requirement that the position be reachable by tiny bidirectional increments of the actuator which would litter the workspace with dead zones --- backlash in a given axis is taken up at the start of a movement in one direction in that axis.  The control loop doesn't even need to know that backlash exists in the drive train, because all it's concerned with is reaching its seek position --- backlash just looks like some extra sluggishness in reaching the desired point.  If the backlash is severe enough that it's noticeable at the work head as bumpiness in the extruded plastic at points where there is change of direction, then the control software just needs to reduce the feed rate at the start of changes of direction.  This is no big deal, and it shouldn't be presented as a terminal stumbling block.

I think we may be looking at two very different goals, one (the topic of this thread) which is creating 3D printers at very low cost by closing the loop at the work head to compensate for less than ideal components, and two, making industrial machinery with good open-loop performance, which is extremely interesting in its own right and contributes hugely to the thought process, but is not the end purpose in this thread.

Morgaine.

shabaz wrote:

 

I was thinking of using a cheap micrometer end and some sort of camera measurement since they usually have their markings etched quite clearly.

As camera resolution gets better and better fueled by consumerism, the same sensors give us new capabilities at low cost for machine control as well.  The 1920 horizontal pixels of 1080p is almost 11 bits of resolution horizontally, and with some creativity one could use the 2200 or so diagonal pixels to get a little bit more, but on the whole it's still low resolution if one is thinking of 2D camera sensors as a cheap means of providing position feedback for 2 axes of a 3D printer by observing it from the top.

By Nyquist, a 1920x1080 array of sensors at most allows you to sample 960x540 data points (adjacent pixels 1 and 0 to detect a single period), which at the commonly sought lowest FFF feature resolution of 0.1mm provides only a 96x54mm manufacturing space, and that's assuming that everything else is perfect.  Needless to say, perfection is rarely achieved.  On the whole this approach of using a camera isn't quite there yet, as it doesn't match the product of workspace size times lowest feature resolution of the open loop approach, for now.

It's a pity that those 2 millions pixel sensors aren't available in a nice long line for us.

Linear camera sensors are available as well of course, but they don't get the volume pricing benefits of 2D arrays.  It's worth keeping an eye on them though, since no-contact position sensing of the work head is a very desirable approach.

Morgaine.

PS.  A 2x2 array of 1080p cameras would put us more or less on a sensing par with the Size*Res of current-day open-loop FFF printers.  The Pi's camera board is just about cheap enough for it, although using 4 x Pi just for sensing would be pretty silly.

It's the difference between kinetics and kinematics.

If the dead zones are known, then you can easily add that in your control scheme. Not perfectly, but still. Otherwise, adding an accelerometer to the end of your tool head might tell you a lot about the accelerations in the head, and can at least tell you when it starts moving. That, combined in a filter with a stepper motor and a control algorithm might already give a great performance boost.

Victor Sluiter wrote:

 

Otherwise, adding an accelerometer to the end of your tool head might tell you a lot about the accelerations in the head, and can at least tell you when it starts moving.

Yes!  That's an excellent additional piece of information that the printer firmware can use to do a better job, and at very little cost.  It's probably true to say that just about every kind of data acquired at the work head can be of benefit as feedback.

Morgaine.

PS.  As an example, low cost physical construction can result in very nasty undamped vibration modes, which can be annoying, harmful to effective resolution, and reduce lifetime.  Accelerometers with adequate bandwidth can help you detect and avoid equipment resonances.

By ignoring the accumulated experience of machine tool design you are doomed to repeat all the mistakes of the past. 3D toothpaste style printers have to control speed and precise position of the print head in order to manage the 'thread' of semi molten plastic that they are depositing. They absolutely must cope with direction changes while maintaining positional accuracy and they must control speed because it isn't possible to change the extrusion temperature rapidly.. They don't have to move fast but extrusion type 3D printers are already very slow.

Adding an accelerometer won't help unless you have an actuator which can operate in the backlash zone. If you consider biological systems they use multiple sensors and multiple actuators, very complex control systems (brains) and require a vast (multi million cycles) amounts of training. So far we have had very little success in copying such systems but done quite well in finding alternative methods. For example replacing panel beating with press tools.

Back to the 3D printer - you can't always add a better control system at an existing mechanism and get the performance you want - it is almost always necessary to design the mechanics, sensors and controls as a complete system with due reference to the spec.

With that in mind, what is your performance target for your improved 3D printer ?

Michael Kellett wrote:

 

With that in mind, what is your performance target for your improved 3D printer ?

Nanometer resolution at terahertz deposition rates, what else?   The only limits are those we impose on ourselves, or try to impose on others.

Admittedly we're not in that operating area yet, but of one thing I am certain:  saying that we can't do it because 300 years of experience tells us that we can't is totally doomed to failure.  Future 3D printers won't look anything like current ones nor like CNC tools of the past, so quoting past experience is not particularly useful.  In contrast, control theory is extremely robust and applies at all scales, so closed-loop systems is a very good place at which to start.

Michael Kellett writes:

 

Back to the 3D printer - you can't always add a better control system at an existing mechanism and get the performance you want - it is almost always necessary to design the mechanics, sensors and controls as a complete system with due reference to the spec.

That goes without saying, which is why the starting point here is closed loop control, and everything else is up for grabs.  I expect that there will be some desire to retain some of the components of past designs, but my inclination towards direct drive certainly doesn't fall into that category.  I'm fascinated to see what emerges, but it's going to need some lateral thinking.

Morgaine.

Your wilfull misunderstanding of my earlier comments does you no credit.

As you are well aware, I have not, at any time, suggested that improvements are not possible.

"The only limits are those we impose on ourselves, or try to impose on others" - this isn't engineering but nonsense - and as an excuse for ignoring the work of countless other engineers it's arrogant as well.

MK

Your wilfull misunderstanding of my earlier comments does you no credit.

As you are well aware, I have not, at any time, suggested that improvements are not possible.

"The only limits are those we impose on ourselves, or try to impose on others" - this isn't engineering but nonsense - and as an excuse for ignoring the work of countless other engineers it's arrogant as well.

MK

Michael Kellett wrote:

 

"The only limits are those we impose on ourselves, or try to impose on others" - this isn't engineering but nonsense - and as an excuse for ignoring the work of countless other engineers it's arrogant as well.

That's false as a fact and ridiculous as an accusation.  And your personalizing of the discussion is uncalled for as well.

It's by ignoring past assumed limits that we've put people on the moon instead of still lying huddled and shivering in caves.  The "valued experience" of experts assured us that the earth was flat, and later that it was the centre of the universe --- see how far that got us.  Minds as bright as Einstein's weren't sufficient to bring us quantum mechanics, so other people had to add that to our state of knowledge.  And in the fields of engineering, materials and methods and understanding improve continually.  Everything changes, and limits derived from partial understanding fall away as our understanding improves.  They are entirely self-imposed limits, and to consider them absolute is to not understand how science works.

Arrogance was a pretty negative and irrelevant thing for you to bring up here, but if you're looking for arrogance, consider your own belief that nothing will supersede the experience of the last 300 years or your own knowledge of lessons from CNC --- that's arrogance to the point of comedy.  It's also arrogance to view engineers as high priests preaching unquestionable gospel and wisdom.  There is no such thing --- everything that we know evolves in relevance, and we have so many ad hoc rules of thumb in engineering that treating them as conditional is always advised.

A better approach if one is interested in the future is to embrace the key principle held by scientists and a core M.O. of engineers who don't have have a closed mind --- the scientific method, which in an engineering context equates to "nothing is sacrosanct".  Best engineering practices and the most cherished theories are only as good as the next development that improves upon them.  This is as true in machine tool operation as it is in everything else.  It is expected and unavoidable in domestic 3D printing because we're barely on the first rung of that ladder.

To answer your last point specifically, it's important to take into account existing experience and best practices where relevant, but only where relevant.  Indeed, the whole point of past experience is to identify the conditions that make known difficulties relevant, so that we can bypass them.  If a known limitation is likely to bite us if we head down a certain road, then the answer is not to head down that road.  And that's exactly what this thread is about, since as Ben Heck said (and I agree), there is not much mileage available in the current open-loop designs for making 3D printers significantly cheaper.

I'm not sure why you're trying to naysay future development starting from closed loop design.  You've certainly not presented any argument for why it's not a good way forward, and you haven't bothered to answer previous posts that addressed yours either, so it looks like you're simply begging for a fight as always.

Future development is done in the context of past experience.  Past experience should never be employed as a shackle.

Morgaine.