Odd circle test results (Page 21 of 24)

Just out of curiosity, AOYOU3D, what printer do you actually own?  Most of the time I see your posts bashing the Solidoodle, telling us how much better designed and robust the Makerbot is, so I'm curious if you even own a Solidoodle?

Because if you did, you would know that the Solidoodle frame robustness is not an issue, there are 1001 other problems that creep up, but frame rigidity is not one of them usually.  I have yet to see any post from someone who tweaked their frame.

so... no answer then? and makerbot is also in the 'low price market/budget price' which just means under $10k in 3D printer talk. with "high priced market' being more like the normal commercial/industrial use range of $30k to $500k area.

Maker bot used to be as cheap as Solidoodle (Sam C used to work for them back then) and it started about $1,000 as an assemble yourself Plywood cabinet kit.  then they used a very similar metal frame as Solidoodle and tripled the price for a pre-assembled unit when Sam decided he could do better for the public to get into 3D printing for under $500 pre-assembled.  and the mods you say addressed the issue for some, had very little to do with the frame thickness (which by the way you never posted Makerbots frame thickness) so...

which costlier manufacturers 3D printer are you trying to promote here? instead of helping Solidoodle owners bring back the perfect circles that most previous Solidoodle models with the same "flimsy frame" have produced for 3 years now?

because we could buy 6 Solidoodle 2 or 3 printers b4 spending near what a Makerbot costs. or do every mod on the Soliform for a Solidoodle printer plus upgrading it to duel E3Dv6 extruders even. and produce way better, finer detailed prints than any Makerbot can. thats why we didnt take out a loan to buy a Makerbot in the 1st place.

and if you want a low cost 3D printer with the thickest rigid frame for under $10,000 there are at least 6 that beat Makerbot. but if you want a Makerbot clone for Solidoodle price there are 3 that I know of. so your constant telling Solidoodle owners the way to get desired performance from their printer is to buy a Makerbot is NOT the definition of "helpful advice" on these topics. we have seen all your posts since joining this forum and none have been useful advice for topics posted in.

We talked the desktop 3D printers, not industrial 3D printer, such cost $10,000,
or $100,000 in 20 years ago.
1/8" (3mm) thk. PVC is enough rigid than 1mm thin sheet metal wall.

If you follow my suggestions to modify SD printers,
You will spend less money to get much better results.
So forget the Makerbot, just discuss my suggestion
1,2,3, right or not?

I would save this for last even if it will cost less, I didn't buy the SD4 with stock enclosure to throw the enclosure away or cut a hole in it. I would still try this, but not now, I need to buy and wait for a coupler to get delivered.

Links?
I've even made a topic asking people what kind of GT2 belt I can use to replace the looped XML Y motor belt.

I'll ask again, can you show proof for this?
Because I've disproven this theory right now by bending the rod left and right while my printer is printing a circle, gives no visible difference.

Contacted Neil, the guy who worked on the hysteresis commands. He stopped working on it as there were issues with some prints and he doesn't have the time. He said backlash is something which can be fixed in software but surprisingly nor Slic3r or Repetier Host have anything like that yet.

Well aren't we in luck -_-

(
I don't know where pirvan got his firmware source but the one here doesn't seem to have M98 and M99 set for hysteresis. https://github.com/mlaws/solidoodle2-marlin )

Cool! when you try all this on your SD3 or SD4 post pics of progress and circle test prints so we can see how it works for you. ;^)

Coming back with info with the possible software fix of the circle problem. Thanks to grob for helping me with this.

So you'll need adrian's fork (version) of the Marlin firmware. The M98 and M99 were never added to the official firmware.
https://github.com/ozadr1an/Solidoodle-Marlin_v1

use M98 to get current values and "M99 Xmm Ymm Zmm Emm" to set the values.

I can't confirm this is a solution yet, as I can't measure how much backlash I have at the moment. But putting some values and testing the circle test, it does deform the parts where the dents are located.
I can only verify this after my dial indicator is delivered in few weeks.

I guess I can also measure the mm though not as accurately, by scanning my circles again with a 2d scanner and using  tealvince's calculation which he posted few pages back:

When I type stuff like 2 or 3 mm, I get a lot of deformation.
And when i calculated for myself, I got 4.57mm which is even worse. I'm probably caluclating something wrong.

E = backlash (mm)
R = circle radius
L = length of flattened section.

E = R - sqrt( R^2 - L^2)
E = 82 – sqrt( 6724 – 729 ) = 4.57

"M99 X0 Y0 Z0 E0" (default):

"M99 X0 Y4.57 Z0 E0"

Too much

And here's it decreased 10 times,
"M99 X0 Y0.45 Z0 E0"

Still a bit much. But seems with correct numbers it could fix backlash.

EDIT: Damn used a wrong scan (old). My backlash is actually around 1.96 mm with this formula.
But even that is too much as I even tried 0.45 in the last photo.

Yeah, I messed up with the scan photo I had to use...
But notice my edited post.
How did you measure 0.2? Im calculating something wrong obviously.

Thanks.
I tried experimental results, 0.12mm seemed good for me. Waiting for my dial indicator to arrive to I can measure the exact backlash left.

Here are my results:
(inner circle is about 1.5 cm)

Opinions?

All I did so far was reducing the friction so i could tighten the belts some more, by replacing side and pillow bushings with bearings. There are probably other ways like replacing carriage like jagowilson did. Probably Solidoodle's fault for using poor quality bushings or weak Y motor. It was impossible for me to tighten the belts enough otherwise, it couldn't handle the friction.
Then I "removed" the rest of the backlash in software (firmware) by making the Y motor move 0.12mm more when changing direction to compensate for the backlash.

Even if I'll get a dial gauge I think finding the "sweet spot" by readjusting the Y motor time and time again for me at least is impossible and I don't see a reason for the extra trouble if you can fix the remaining trouble in software.

Slic3r developer doesn't seem to agree with me that allowing people to fix "hardware issues in software" quick and easy is a good thing and refused my feature request for adding a backlash compensation option somewhere in Slic3r. Guess he doesn't think 3d printers should be owned by people who don't know enough to fix the backlash on their hardware and/or don't want to spend the money and time on upgrades and testing. Fair enough. https://github.com/alexrj/Slic3r/issues/2431

You'll need an unofficial firmware forked by Adrian which has this command and reflash your motherboard if you want to try this out.
https://github.com/ozadr1an/Solidoodle-Marlin_v1

Hi, glad to see this discussion.  I'm revisiting backlash myself since I decided to finally get away from Skeinforge/Pronterface for the 'correct' RH/Slic3r install (SD v2). 

Am I reading the tests you did above correctly where you left a zero value for the E, which I thought was the backlash setting, and only had an offset in X and Y?

Incidentally if I send a manual M98 command it does echo X Y Z E results back to me...makes me wonder if at some point back in time I did flash to the indicated firmware.  No problems re-doing it if I have to  - just make sure I write down my calibrated extruder stepper setting first...