Print bed flatness

How about adding a sensor to the tool head which measures the nozzle height from the bed.
Anyone who has tried to print on a warped bed will know that a small curvature of say, 0.5mm in 50mm is enough to disrupt the raft, and therefore prevent printing.

If we had a height sensor (laser based, eddy current, or some kind of feeler probe. micro-epsilon)
Then the curvature of the print bed could be mapped out and compensated for via software, by extruding a thicker raft in certain areas.
This would increase the ease of reproduction, as there is less of a need for very flat print beds. Im thinking here of using old chopping boards or plastic pannels. A hdpe board thinner than 1cm (some say 8mm) thick is practically unusable due to its inherent curvature, and warping from use.

Also it would allow the reprap to self calibrate its nozzle height, leading to an easier to use machine.

Maybe the nozzle itself could be used as a probe, by using the steppers as force gauges, they have this tactile feedback for some commercial robot arms. This would also double as a safety mechanism to prevent head crashes or running into finished prints.

Just throwing the idea out there, maybe the open source sensors guys will have something.

I've been thinking about putting a force gauge between the thermal barrier and the extruder drive block. In normal operation, this would measure the pressure in the extruder chamber allowing the firmware to estimate the flowrate and adjust the feedrate to match. I was thinking it could be used to home the z-axis as well and it should be able to map the bed as you suggest.

---

---

Darwin clone, Gen 2 electronics, Arduino Duemilanove w/ AtMega328, 5D Firmware, Pinchwheel extruder
[www.codeerrors.com]

An IR range sensor can be had for about $10. This would be easy to calibrate and easy to mount near the print head. Mapping could be done accounting for the offset, and it would provide a decent print head measurement. UltraSonic range finders would work as well, but are more expensive. Neither of these have the accuracy of a laser, but they're a heck of a lot cheaper to implement...

If you knew precisely the curvature of your print surface, you could compensate in software by "warping" each layer of the print to match the underlying bed. This would allow raftless prints like nophead is doing, but without the tedious calibration and expensive super-flat aluminum.

In other words, forget making the raft thicker, just make the layers of the part you're printing be of constant thickness.

Unless your bed is REALLY warped, you'd never notice the difference. Except you'd have higher quality prints with less work.

---

--
I'm building it with Baling Wire

I think a probe for measuring the inaccuracies of the bed, both in terms of height/bending, but also in terms of mis-allignment of the X-Y axis is a great idea.

I'm a bit worried about the precision of a cheap IR range finder, but Nophead dug up a link for something which could be used as a homegrown probe:

Nopheads original blog:
[hydraraptor.blogspot.com]

The probe, which inspired Nophead

[www.brusselsprout.org]

Maybe that could be adapted for our use

So for bend Z-axes we could mount these horizontaly, home to a specific edge and meassure the distance to the edge?
As threaded rods are never perfectly straight this could allow making
higher quality printers with cheaper materials.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

MarcusWolschon Wrote:
-------------------------------------------------------
> So for bend Z-axes we could mount these
> horizontaly, home to a specific edge and meassure
> the distance to the edge?
> As threaded rods are never perfectly straight this
> could allow making
> higher quality printers with cheaper materials.


That was my initial thought, the firmware need to be modified, so the results of the scan can be transferred back to the pc, for further processing.

anton Wrote:
-------------------------------------------------------
> MarcusWolschon Wrote:
> --------------------------------------------------
> -----
> > So for bend Z-axes we could mount these
> > horizontaly, home to a specific edge and
> meassure
> > the distance to the edge?
> > As threaded rods are never perfectly straight
> this
> > could allow making
> > higher quality printers with cheaper materials.
>
>
> That was my initial thought, the firmware need to
> be modified, so the results of the scan can be
> transferred back to the pc, for further
> processing.

For Z you either only need a way to store it or adjust the
feed-rate on the fly until you are above the highest point in the bed.
For X and Y the firmware can do the adjustments itself.

So, no back-channel required.
Of cause with a longer range this could also double as an
alternative toolhead for 1-face 3d-scanning.
Like mounting an existing object to the bed, scanning it
and then printing things onto it.
(overhangs >60° without support? repairing aborted prints?
...)

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

@MarcusWolschon
I like the idea of dynamically adjusting the feedrate, but won't there be memory problems storing the height map of the bed?

Also, where should the height map be stored when the machine is powered down? I could well imagine that scanning the height of the bed at any decent resolution would be time consuming.

anton Wrote:
> Also, where should the height map be stored when
> the machine is powered down? I could well imagine
> that scanning the height of the bed at any decent
> resolution would be time consuming.

If it is permanently attached,
what is the problem with scanning and adjusting
while you print?

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

Silly me