Odd circle test results (Page 20 of 24)

I just remindered your guys few times in some places,
1) SD frame is made by 1 mm thin sheet metal, not
Rigid enough for a timing belt drive, newest Makerbot
Printer has a stainless steel linear guid rail with the
Linear bearings, for X and Y drives, likes the industry linear motion use,
So don't expect your SD printer making a perfect circle,
Remember your SD printer is only for the low cost market,
It is ok to make some toys, some prommotion parts or
Prototypes, etc., that is enough, $350-$500, why you
Expect too much?
2) If you really like to reduce the backlash, make your printer better, you should
Move Y motor to outside of the frame, make a direct drive,
Like this:
Motor>Flexible coupling (one end 5mm--motor shaft, other end 6mm--rear rod)
>>bushing or bearing>>left pulley/belt>>right pulley/belt>>bushing or bearing.
You should use GT2 pulleys and timing belts instead of SD stock pulleys and belts, they will
help to reduce a backlash, but change stock bushings, it doesn't make sense.
If it is possible you can reinforce the frame, make it more rigid.
then you can get a much better circle for your products needed.

dude I never claimed to be a journalist. you get the gist of it if ya dont cut parts out to quote for critiquing

I don't think that's it. I just printed the circle with just 6 cm filament sticking out from the extruder, no filament roll. I think you'd get step skipping if the extruder motor had stress (and that has happened to me when filament tangled).

A proof for this claim?
I'll remind you something myself: some Solidoodles with the same frame print just fine.

What do you think we're all doing right now?

What pirvan said.

I think this is a pretty good test.
In fact, I think a 1 layer print should be good enough for measuring backlash too if you'll attach a pen to the extruder and a paper to the print bed.

I disagree for the following reasons:

The first layer is supposed to be thicker and somewhat squished in order to promote adhesion, for this reason, the first layer usually looks more "filled in" than the rest, and the filament looks like it was a little over extruded.  Subsequent layers do not have problem, and are more accurate because the thickness of the extrusion is closest to the actual number you have in the settings.

Here is a test you can do to prove this.  Print a single wall square, or circle, anything, but print about 15 layers.  Pop it off the print bed, then measure the thickness of the wall at the top, then at the bottom. you should see a difference.  Alternately print a solid surface (100% infill) with a few layers, but stop it before it prints the top layer.  3-4 layers should be enough. Get a magnifying glass and look at the bottom layer, and at the top layer.  Quite a bit of difference, isn't it.

My point is, measuring one layer only, doesn't give you proper feedback, because it's not as accurate as the layers that go on top.

NOTE:  Depending on how perfectly level your bed it, the first layer may actually show differences in the extruded filament width.  That's because in the area were the bed is closer to the head, the extruded filament will be squished more than the areas where the bed is farther.

Now, regarding the pen test.  It probably wouldn't be a good test, because it doesn't take in consideration the extrusion (or over-extrusion) width, it lays down a perfect line with the same width all the time.  This would be especially true on the first layer where the "squishing" mentioned above occurs.

I'm quite aware of what the main thread is about, but you missed the point of my post.  Here is what I said:

The first layer is supposed to be thicker and somewhat squished in order to promote adhesion, for this reason, the first layer usually looks more "filled in" than the rest, and the filament looks like it was a little over extruded.

What that means is that a single layer test, such as printing one layer only of the Velociraptor posted by n2ri, will NOT tell you if you have backlash or not, because the infill will appear to be connected to the perimeters due to the aforementioned squishing.

The rest of the post is an explanation of why this happens, and how to test this for yourself.

Take  look over here:
http://reprap.org/wiki/G-code#M99:_Set_ … teresis_mm

Also note that these commands (M98 & M99) have been deprecated as they clash with the current implementation.

Nevertheless, you can implement the hysteresis codes as they were previously done, as long as you don't have, or try to implement any additional macros.

As for whether they need to be implemented the values inn the firmware, the answer is you can but you don't have to be.  You can simply add them to your custom G-Code in Slic3er, Cura, or whatever slicer you're using. 

If you want to store the values in your firmware you should be able to, by using the M500 command.  Basically send the M99 Xmm Ymm Zmm Emm command (where mm is the actual value you want), then send M500.  That will store the commands in the firmware.

Got it, thanks.
I'll try this and report back.
BTW, to restore to defaults I should use "M99 X0 Y0 Z0 E0" right?

Sorry, what does deprecated mean here?
The newer firmwares which Solidoodle might use in the future don't have them anymore?
Or/and there is a new command which replace it and do the same? Which one is it if so?

Not necessarily.  If you already have some settings built in, you should go back to those, as they would be your "factory settings".  Use the M98 command to get the current setting values, and write them down for future use in case you want to go back to them

Deprecated means obsolete, no longer in use.  The newest versions of the Marlin firmware are reserving these commands for use with macros.  Basically, you can run a macro in the middle of you Gcode, by simply sending the M98 command followed by the macro.  Once the macro is executed, the Gcode continues from where it left off.

So, you can create a macro out of your Hysteresis commands, save it as a My_Macro.g file, then send the command: M98 PMy_Macro.g, which would execute your macro, then return.

I don't know what version of Marlin is currently in the latest Solidoodles, so I don't know if they're implementing the M98/M99 for hysteresis, or for Macros, or at all.

Do a M98 command and see what you get.  If you get X, Y & Z values, it probably has built in Hysteresis values, If it comes back with nothing or an error, then it might not even have support for it, which means you'll have to compile your own modified version of the Firmware and upload it to your board.

yes thats what I thought. so when I upgrade to an E3Dv6 hot end with .4mm nozzle it may do it. I also intend to buy extra nozzles both smaller and larger for different materials and models to work better with other slicer profiles to meet special needs. thanks for testing this.

"n2ri,for the thickness of the sheet metal, in Inches system you use gauge xx,
In Metric system you use mm, we can convert them each other."

OK then what MM is 16 gauge? and what MM is 24 gauge? I dont think its just a matter of converting to inches. I was a sheet metal tin smith for many years and we never dealt with either for thickness of the material for duct work thinner than 14 gauge.

never saw galvanized sheet metal in odd numbered GA. always even so 20ga like the corner angles used on dry wall at edge of walls on corners but without the inset and nail holes. but that would be only the upper side angles, as the rest is stamped/formed of one piece and riveted at corners giving it enough strength for Sam to stand on it so not quite as flimsy as suggested. and sufficient for intended use.