You are not logged in. Please login or register.
SoliForum - 3D Printing Community → The Latest in 3D Printing → And who has the most effective, hassle-free self-leveling technology?
Bed leveling has to be the most critical element of FDM printing. Well, and Z-height. If it ain't level, you ain't printing decent stuff.
So who's self-leveling gee-whiz is the whizziest?
The best auto level feature is one that isn't needed. Using software patches to compensate for crappy mechanical platforms is a last resort type of thing. The actual process is fine and fairly equal as they all do pretty much the same thing. It all comes down to the precision of the probe mechanism. If that is perfect and perfectly repeatable than you have a good system. Couple that with an already solid and level build platform and now you have something.
Second best is one that actually levels the bed rather than compensating by moving the Z screw. this could be done pretty easily with motors driving the leveling screws and a probe running across the bed while doing so to give feedback. it's been done before, but I've never seen it implemented in any major firmware.
My wife suggested separating the bed and xy carriage frames from the printer body then balancing the bed and the xy with separate gyros which I guess would adjust leveling servos. Thus, the print head and bed would each be level at all times with the horizon and each other.
It would probably work great but be a little pricey (two gyros and what, four or six leveling servos? So, about $200 extra for true gyroscopic leveling.
Gyroscopes can't detect level (or "horizon", which I'm guessing you're defining as a line perpendicular to gravity), they measure angular velocity.
Its a bit like saying you'll use a speedometer to make a picture frame level.
Gyros can be used for that (keeping a motorcycle upright at speed, for example) if they are spun vertically. They balance non-moving objects perpendicular to the pull of gravity by spinning horizontally. I guess I don't see how the speedometer analogy applies at all.
you might find this thread useful
http://bbs.homeshopmachinist.net/thread … level-work
Gyros can be used for that (keeping a motorcycle upright at speed, for example) if they are spun vertically. They balance non-moving objects perpendicular to the pull of gravity by spinning horizontally. I guess I don't see how the speedometer analogy applies at all.
You're conflating a a whole bunch of different things. Let me try to explain.
First of all, you said "with separate gyros which I guess would adjust leveling servos". That implies (strongly) that you're using gyroscopic sensors, not gyroscopes (which are just spinning wheels). Neither would help you, but making the distinction is important.
A gyroscopic sensor just measures rate of angular rotation. This is obviously not helpful in terms of measuring a gravity vector (direction and magnitude), so a gyroscopic sensor is out. Rate of angular rotation is what a speedometer measures, though it converts that angular rotation (RPMs of the the car's axle) to linear speed (MPH/KPH) as it knows the diameter of the tire, but it is just measuring the rate of angular rotation - hence my analogy.
A gyroscope is a spinning wheel, which is of course subject to conservation of angular momentum. What this means is that once spinning, you have an angular momentum vector, as a result of the torque that was applied to the wheel to get it spinning. Then you have a torque due to gravity, which also creates angular momentum. The total angular momentum is the sum of these two angular momentum vectors, and as a result it can resist gravity. That was probably complicated, so let me break it down:
If stopped, a gyroscope would fall due to gravity. If spinning, the gyroscope begins rotating around a second axis due to gravity. You can see the effects in this video:
https://www.youtube.com/watch?v=ty9QSiVC2g0
But as you can see in the video, you *cannot* deduce the angle of the gravity vector from the angle of the gyroscope. Again: A gyroscope only yields the angular velocity.
Furthermore, bicycles don't stay upright from gyroscopic effects. This is a common misconception, debunked here:
http://www.popularmechanics.com/adventu … less-bike/
and explained here:
https://www.youtube.com/watch?v=NcZCzr9ExKk
and here:
https://www.youtube.com/watch?v=oZAc5t2lkvo
and elsewhere. Bicycles stay upright from self-steering.
Finally: If you wanted to measure the direction of the gravity vector, you'd use an accelerometer.
Force due to gravity = m * g
F = m * a
therefore, the gravity vector is nothing but an acceleration. An accelerometer can measure the direction, but not accurately enough for bed leveling. An error of even 0.1 degrees in the angle measurement would result in +/-0.2mm of error in the bed leveling! They are useful for rough estimates though, the accelerometer in your smartphone is what tells it to flip the screen when you rotate the phone.
Going back to the bed leveling issue.
I have tried the induction type leveling gadgets on both my Printrbot simple and the Solidoodle, and I think they're more trouble than they're worth. The real fix is leveling your bed to start with.
Additionally, I think that in practice, by adjusting the height of the nozzle to accommodate a skewed bed, it promotes a model is is printed "wrong", with layers that are all skewed. The vertical wall is straight, but the layers in that wall are "skewed.
On the other hand, if your bed is not level, but it's close, the only thing that will happen is that your first couple of layers will be overly squished on one side, but as the models keeps building upward, it will be fine. Take a look at the picture below. The first 2 layers are partially squished and possibly incomplete, but after that the print is normal, and straight.
^That!
A good platform that has been leveled to the nozzle properly is the way to go. I have not had to re-level my bed for about 6 months...
The best auto level feature is one that isn't needed. Using software patches to compensate for crappy mechanical platforms is a last resort type of thing.
Nice to see we've come full circle with this conversation 
Thanks, Tim. I wa basing my (wife's) hypothesis on observations of the Segway and a now-defunct gyroscopically stabilized Indian motorcycle prototyped back in the 80s. You're right in that I was mischaracterizing the gyroscopic sensor as a gyroscope. I think you see my point though, if some electronic sensor can detect that the bed and/or carriage is out of level, then it could cause servos on the bed leveling screws to adjust both back into level. The object not being to necessarily have the bed level with the horizon as much is to have the bed and carriage parallel to each other. Maybe a laser to measure the distance between bed and print head, talking to adjusting servos?
I agree that, of course, starting with a level bed is best, at least with the technology of today.
Pirvan, you're right about the elf-adjusting Z axis to accommodate an unlevel or uneven bed, it makes goofy parts. And perhaps you've had better luck than I but when my bed was not level even a little bit, the first layer would stick on the close side and not at all on the high side. But maybe my bed was more unlevel than what you mean. My experience is, you need a correct first layer before you can have a good part. This seems particularly true if you're printing support since poorly stuck support tends to come loose and end up dragged around in balls all over the machine path.
However we look at it, gyroscopes be damned, bed leveling is very important and we should alway keep our eye out for a better, more fool-proof way to achieve and maintain it.
A Segway uses gyroscopes and tilt sensors to stay upright. The tilt sensor measures angular rotation with respect to gravity, and the gyroscopic sensor measures rate of angular rotation. With these sensors combined, the microcontroller can tell how far away the Segway is from vertical (tilt sensor), and almost more importantly how quickly it is falling in a particular direction (gyroscopic sensor).
Anyway, even a tilt sensor is not much good in this application, for the aforementioned accuracy reasons. You'd need two sensors, doubling the maximum possible error. Much better to sense the distance directly, with something like an IR sensor:
https://www.filastruder.com/collections … ght-sensor
Next: you don't need leveling servos, your whole printer is a "leveling servo"! You can probe the bed at multiple locations and build a height grid, much like a topographical map. Then the Z-axis runs up and down while printing to put the bed in the right place.
To IronMan/pirvan/wardjr's point though, a bed should not need frequent adjustment. I generally adjust once every 100 prints or so.
Thanks
never had any issues with inductive sensors for leveling. they always worked well for me.
I think they are probably fine. If a bed is rigid and flat the concerns that auto-leveling (inductive sensors) causes parts to conform to a warpy bed seem misplaced.
Here are the issues I came across with induction type sensors:
The induction sensor is triggered by the proximity to the aluminum bed. Whenever you have something else on top the print bed, your print head is further away from the actual metal, and so is the sensor, which obviously can't be set any lower than the print nozzle.
The sensors used for bed leveling are usually low cost tube type inductors with a short sensing distance (4-6mm). So if you have something on top of your printbed, like thick glass, mirrors, PEI, etc, the distance between the sensor and the metal increases and as it gets near the maximum of the range, the sensor becomes more erratic which sometimes leads to a head crash into the print surface.
At this time your solution would be to get a sensor with a longer sensing range (10-12mm), and preferably a better quality one, and you quickly find that the $5-6 sensors don't cut it, and you have to spend around $25-30 for a decent sensor.
Well, that's all cut and dry... until you add a PCB heater like I have. I have a PCB heater on top of print bed, and it has 3mm glass on top of it. The minimum distance from my print nozzle to the metal print bed is actually 6.5-7mm. That forced me to get a sensor that supported at least 10mm.
Anyway, when I first tried it, it worked great, until I turned on the bed heater, and found that the sensor was very susceptible to electromagnetic fields. Whenever the PCB bed power was on, it would scramble the sensor. I don't remember if it was triggering to high, or too low, but suffice it to say, after a lot of trial and error, and way too much trouble that it was worth, I just took the stuff off, and went back to the original switch.
I still have the sensor on my Pritrbot Simple, but it's not in use, it too uses a standard switch, and manual levelling.
Inductive sensor didn't work at all with the mirror on my FT kossel - the only time it would "sense" anything was when it was outside the perimeter of the mirror and that only happened on one side of the bed - thus making it useless and rather pointless.
I removed the inductive sensor and calibrated the effector to the fixed bed via the firmware - no more "leveling" required. It just works.
The best auto level feature is one that isn't needed
^ this, without a doubt.
on the SD's, I hardly ever have to tinker with bed level anymore (unless I have to get rough with it removing larger parts, lol). Getting them set up well to start with - a nice, flat, glass surface, and good Z height adjustments - go a long ways towards eliminating the need to fuss with it.
I see. Well, right now I am trying the Print Tite material which adheres to the print bed. So, I have to either wait till the bed cools while to easily remove the part or "get little rough" with it.
I'm convinced that rough handling does indeed take the bed out of level. It's to be expected too, since the leveling screws have no locks.
I have ordered 4MM x 50MM flat head socket screws and wing nuts from McMaster. I will replace the stock leveling screws with these. Once I have the bed nicely tweaked, I will tighten the wingnuts on the lower end of the screws and see if that doesn't keep the bed level even during enthusiastic part removal.
If I'd thought it through, I'd have stuck the Print Tite onto a mirror for removal/rough handling.
Anyway, I'm hopeful the wing nuts (jam nuts) will keep the bed level more or less permanently.
Rough handling will take the bed out of level. I use glass sheets, and swap them in and out between parts so the bed doesn't get abused.
Here is an interesting design, that moves the bed and locks it into place-
https://groups.google.com/forum/#!msg/m … 1unBHWeyYJ
The bed is supported on springs which are around smooth tubes rather than screws. It's a little hard to see what is going on with the levers, but there is a gap between the lever section and the mounting section below it. When the end of the lever pushes outward, it bends and grips the tube which is supporting the bed. The bottom of the gear is a cam which is shaped to press each lever in turn.
When beginning leveling, all the levers are released. The head moves to the first of 3 leveling points and moves to a specific z height, low enough to push the bed down against the spring. While the nozzle holds the bed down, the motor turns the gear, moving the cam into position to press the first lever and grip the tube, holding that point of the bed in place. The nozzle rises and repeats the process at each of the other points. After printing, the cam rotates to a point where all of the levers are released again.
The inventor of this process did it on a Rep2, using the gcode controlled fan pin to drive the motor. Plenty of controllers have enough pins to run this without using the fan. I'm not sure what the limit switch is for, most likely there are bumps on the cam to trigger it at each position the motor needs to stop.
interesting concept with the motor driven gear & cams...
Interesting indeed. It goes through the leveling dance before every print?
Is this the first time we've seen using the head itself to set the bed height? It does look like I could still knock it out of line trying to free a stubborn part. I guess that's why it apparently does it every print, which would be time consuming.
Hopefully the wing/jam nuts on the leveling screws I mentioned earlier will offer a solution to Operator Induced Unlevel Print Bed Syndrome. Be sure to ask your doctor is wingnuts are right for you.
Check out the Robox approach to auto bed leveling
https://www.youtube.com/watch?v=51VsyV4i7OM
SoliForum - 3D Printing Community → The Latest in 3D Printing → And who has the most effective, hassle-free self-leveling technology?
Powered by PunBB, supported by Informer Technologies, Inc.
