26

Re: New bed level probe idea

Another thing that has been done is a simple lever with magnets to hold it at the up or down position.  A bit of gcode causes the extruder to move past something that snags the lever, causing it to extend or retract.

Here is an example, using a screw to pull the lever out and put it away. - http://www.youtube.com/watch?v=uk29TIPouLk


I've done the electrical connection between the nozzle and bed for setting the Z tab before.  The problem with using it for probing is your build plate is no longer easily  swappable.  Also you would want to have a wire brush and a wipe routine.

27

Re: New bed level probe idea

IronMan wrote:
Hazer wrote:

Wow, you guys are taking it too far. Thermal expansion on the nozzle?

Servo problem is easy, the probe would have an hardstop on the down position. Servo is commanded farther than that but can only reach the (known) distance of the hardstop. Probe is at a reliable distance and retractable.

What I am curious about is how well this works once Marlin has the data about the z axis. Think about it, the Z axis would be going up and down constantly during the print....alot. Is this supposed to go on for every layer? For every move? Better build an anti-backlash Z rod setup for the stock machine while we are at it.


I have a better idea: Put 4 servo motors under the bed attached to leveling screws. Use the data to level the bed using the servos (I am thinking they would have to continuous rotation with rotary encoder).

Like that idea...In the video above, the system seems to be just logging the data; just wondering if the Delta style printer can make the necessary adjustments automatically since the z-level is controlled by each of the 3 arms...basically creating a custom level plane every time you print.

http://www.pololu.com/catalog/product/536 $14 servo. Marlin reports they have options for using continuous servo already (using duration instead of angle).

Honestly, I think this is all over-kill. Its also a classic example of trying to use complicated electrical control methods to compensate for a relatively easy mechanical design flaw. You should never have to level your bed more than once. If you do, there is a big problem with your mechanical design.

Now, someone mentioned Z-height adjustment on every print, that I can understand especially if you swap your glass. But instead of adding servos or probes and adding extra control methods, simply put the Z endstop on the carriage instead. Instead of hitting an adjsutable screw on the case, you would be stopping on the actual printing surface itself. Swap glass, tile, garolite, or right to the aluminum kapton surface.

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

28

Re: New bed level probe idea

Hazer wrote:
IronMan wrote:
Hazer wrote:

Wow, you guys are taking it too far. Thermal expansion on the nozzle?

Servo problem is easy, the probe would have an hardstop on the down position. Servo is commanded farther than that but can only reach the (known) distance of the hardstop. Probe is at a reliable distance and retractable.

What I am curious about is how well this works once Marlin has the data about the z axis. Think about it, the Z axis would be going up and down constantly during the print....alot. Is this supposed to go on for every layer? For every move? Better build an anti-backlash Z rod setup for the stock machine while we are at it.


I have a better idea: Put 4 servo motors under the bed attached to leveling screws. Use the data to level the bed using the servos (I am thinking they would have to continuous rotation with rotary encoder).

Like that idea...In the video above, the system seems to be just logging the data; just wondering if the Delta style printer can make the necessary adjustments automatically since the z-level is controlled by each of the 3 arms...basically creating a custom level plane every time you print.

http://www.pololu.com/catalog/product/536 $14 servo. Marlin reports they have options for using continuous servo already (using duration instead of angle).

Honestly, I think this is all over-kill. Its also a classic example of trying to use complicated electrical control methods to compensate for a relatively easy mechanical design flaw. You should never have to level your bed more than once. If you do, there is a big problem with your mechanical design.

Now, someone mentioned Z-height adjustment on every print, that I can understand especially if you swap your glass. But instead of adding servos or probes and adding extra control methods, simply put the Z endstop on the carriage instead. Instead of hitting an adjsutable screw on the case, you would be stopping on the actual printing surface itself. Swap glass, tile, garolite, or right to the aluminum kapton surface.

^ elegance.

Though, I suppose you are never sure until you check.

Former Solidoodle employee, no longer associated with the company.

29

Re: New bed level probe idea

Having the Z limit fixed to the housing would eliminate a flexing wire as well.

30

Re: New bed level probe idea

Because I was trying a different bed and I was curious about it, I tried a "manual" probe by hooking up my dial indicator and running it around the bed taking lots of measurements on a 10mm grid. I learned all sorts of possibly relevant things.

For one, you really don't want the probe dragging around on the bed surface causing rocking. The only way I could get consistent readings was to lower the bed, move the carriage, then raise the bed too much and lower it a little again.

Simply raising the bed also cause rocking, the raise, then lower cycle seemed to be needed to fix that.

The other thing I found was that the 3d graph I made of the data points in gnuplot looks horrifyingly like the hood of a car :-). Gotta try thicker glass. There were definite "dents" around the clips used to hold it down. (The ceramic tile I was trying was definitely not flat - it is much thicker in the center than the edges).

31

Re: New bed level probe idea

Hazer wrote:

You should never have to level your bed more than once. If you do, there is a big problem with your mechanical design.

That depends: if you are swapping out different pieces of glass or tile, are the surfaces flat and parallel? Or are some of them perhaps flat, but not really parallel? Of course you could lump non-parallel surface bed material into the category of "problem with mechanical design" :-).

32 (edited by Hazer 2013-10-13 03:45:35)

Re: New bed level probe idea

Glass should be parallel. Ceramic tile, not so much.

I think my personal opinion makes me biased. For starters, I have kept my mirror clipped on since I first installed it. As soon as a print finishes, I use a razor blade to pop it off easily, and then run another print. Ian mentioned once before, more print jobs leave the bed in good stick condition. Every once in awhile, I scrape the mirror clean (I save on acetone) and give it a spray. So my personal preference leads me to want a stable bed base.

But other people like swapping their glass. So for them, level could end up going out if the glass is different in any way.

This auto-leveling keeps growing on me. Especially with the Solidoodle.

So imagine this: a small metal ball maybe 1/3 inch diameter housed in a tube that lets the bottom hemisphere open. The ball would be just below a plunger type snap switch. This is then mounted onto the underneath of the SD carriage. The idea here is that the ball would be just below the tip of the nozzle. It would trigger the auto-level program, but then during a print, the ball would just 'roll' over any collisions with the print (remember, it is just sitting there like a mouse ball being held in the air). The only force down would be gravity and the switches light actuation spring force. Now there would be no need for a solenoid or servo.

Think about a specifically printed version of this:

http://www.amazon.com/Morris-Products-R … ler+switch

Or you could get this http://www.amazon.com/Diameter-Inductiv … ity+switch and glue some thin metal sheets just under the glass at the probe points. Contactless probe.

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

33

Re: New bed level probe idea

It may be overkill but I still like it. How could you not, it's a sweet idea!

SD2 w/ mods: Rumba controller, ATX PSU, SD pro case, glass bed, with more to come...

34

Re: New bed level probe idea

http://www.soliforum.com/post/38893/#p38893

Firmware is done. Latest marlin, all the solidoodle patches and tweaks.

I reckon 'babystepping' kills the auto-levelling in terms of immediate reward smile

Anyway, let me know if anyone has any dramas on the software side of things.

Will work for Printrboard variants too - just use motherboard 81 and dance the printrboard firmware dance...

35

Re: New bed level probe idea

So the ball bearing running all over just laid down extruded
hot polymer wouldn't be a problem?

I can't find spec on prox sensor   trigger at =/- 1cm ?
at the same distance all the time?

Russian roulette bed leveling?????


Hazer wrote:

Think about a specifically printed version of this:

http://www.amazon.com/Morris-Products-R … ler+switch

Or you could get this http://www.amazon.com/Diameter-Inductiv … ity+switch and glue some thin metal sheets just under the glass at the probe points. Contactless probe.

36

Re: New bed level probe idea

Do you have a switch idea for a non-delta, non-bowden extruder

I really don't understand your switch
does it close on XY move or Z move?

Tnx




IanJohnson wrote:

Another thing that has been done is a simple lever with magnets to hold it at the up or down position.  A bit of gcode causes the extruder to move past something that snags the lever, causing it to extend or retract.

Here is an example, using a screw to pull the lever out and put it away. - http://www.youtube.com/watch?v=uk29TIPouLk


I've done the electrical connection between the nozzle and bed for setting the Z tab before.  The problem with using it for probing is your build plate is no longer easily  swappable.  Also you would want to have a wire brush and a wipe routine.

37

Re: New bed level probe idea

For the retractable probe, there would be a probing switch at the end of a lever.  The extruder would move past something like a bolt that would snag the lever and cause it to be pushed down as the extruder moved by, lowing the probing switch into position.  When the probing is done, the extruder would move past the bolt from a different direction that would cause the lever to be pushed back up.  That is the principle used in the video I posted.

Using the nozzle itself to do probing shouldn't be hard on the Solidoodle.  Because the X carriage has a loose hold on the back X rod, the nozzle already can move up and down a little.  If you set the bed to home a little too close, you can see it push up on the nozzle and rock the extruder back.  There should be a way to position a switch at the back of the carriage that will  trigger when the bed pushes up on the nozzle, and a spring that will keep the extruder tilted forward when the bed isn't pushing up on it.

38

Re: New bed level probe idea

Seems to me like this guy's got it done!  Truly leveling the bed using only 2 servos...on the SD it would no doubt be the 2 front leveling screws...

http://www.youtube.com/watch?v=Ppx_kGyQorc

SD2 - Stock - Enclosure - Heated Bed - Glass Plate - Auto Fire Extinguisher
Ord Bot Hadron - RAMPS 1.4 - Bulldog XL - E3D v6 - 10" x 10" PCB Heated Build w/SSR - Glass Plate
Thanks for All of Your Help!

39

Re: New bed level probe idea

Now that I like

SD2 - Glass Bed, Fans on PCB and Y motor, Custom enclosure
Slicer - Simplify3D

40

Re: New bed level probe idea

Heres an idea for a probe switch.
http://www.adafruit.com/products/166?gc … MgodFjgArg

Mount this between the extruder servo/fan and the x carriage.

I just ordered one to try it out. We will see what happens. Now I just have to figure out how to control the servos. Im thinking and I2c servo board? Or do you think I could just use some pins on the Printrboard?

41

Re: New bed level probe idea

Another idea (don't know how good it is, mind you). I stumbled across this gadget on thingiverse:

http://www.thingiverse.com/thing:44572

What if, instead of merely a spring, you had the pressure sensor above sandwiched between 2 plates with the spring pulling them together.

Not only would you maybe get a less wobbly print head, but you should also be able to notice a sharp change in the resistance as soon as the nozzle makes contact with the bed. With micro stepping on the Z axis, you might even be able to read the level exceedingly accurately.

That's the theory, anyway.

42 (edited by dubbsd 2013-10-29 09:44:53)

Re: New bed level probe idea

what about incorporating strain gauges into the bed
http://morf.lv/modules.php?name=tutorials&lasit=19

http://smdsensors.com/Products/default. … ce+Sensors

Ultimaker S3.

43

Re: New bed level probe idea

IronMan wrote:

Seems to me like this guy's got it done!  Truly leveling the bed using only 2 servos...on the SD it would no doubt be the 2 front leveling screws...

http://www.youtube.com/watch?v=Ppx_kGyQorc


Any chance you downloaded that video? Shows up as private for me.

44

Re: New bed level probe idea

Tomek wrote:
IronMan wrote:

Seems to me like this guy's got it done!  Truly leveling the bed using only 2 servos...on the SD it would no doubt be the 2 front leveling screws...

http://www.youtube.com/watch?v=Ppx_kGyQorc


Any chance you downloaded that video? Shows up as private for me.

No, I did not...it shows up as private for me now too.  No doubt he took it down for legal reasons...he's probably pursuing a patent!

SD2 - Stock - Enclosure - Heated Bed - Glass Plate - Auto Fire Extinguisher
Ord Bot Hadron - RAMPS 1.4 - Bulldog XL - E3D v6 - 10" x 10" PCB Heated Build w/SSR - Glass Plate
Thanks for All of Your Help!

45

Re: New bed level probe idea

Do you mind explaining what he did? I would assume he needs 3 servos to level the bed.

46

Re: New bed level probe idea

Tomek wrote:

Do you mind explaining what he did? I would assume he needs 3 servos to level the bed.

He only used 2 servos.  He had a 3 point manual screw-leveling system (like our SD's); what he did was first home near the single screw location (on SD, the one in the back, no servo attached) and he had a proximity probe that set that reading as "System Zero". 

Then he went to the first of the two other servo-driven screw locations (front) and with the z-table homed to the newly defined System Zero point, whatever the probe detected as the Delta was then used to engage the servo to bring that point to match the System Zero +/-.  He then repeated this for the 3rd position.

I hope the explanation was clear enough...

SD2 - Stock - Enclosure - Heated Bed - Glass Plate - Auto Fire Extinguisher
Ord Bot Hadron - RAMPS 1.4 - Bulldog XL - E3D v6 - 10" x 10" PCB Heated Build w/SSR - Glass Plate
Thanks for All of Your Help!

47

Re: New bed level probe idea

Ok! It sounds to me like you're saying that the back was determined as the reference, and the two servos adjusted with respect to it.

Made me think, if was feasible, how cool it would be to have a detachable continuous rotation servo motor that just turned the screws. Then you'd really be rocking, and at half the price. but it would be much harder to center it.

48

Re: New bed level probe idea

As Morpheous would say:

Yessss...

SD2 - Stock - Enclosure - Heated Bed - Glass Plate - Auto Fire Extinguisher
Ord Bot Hadron - RAMPS 1.4 - Bulldog XL - E3D v6 - 10" x 10" PCB Heated Build w/SSR - Glass Plate
Thanks for All of Your Help!