Mhackney's Rostock Max

[mhackney]

I devised a series of tests to try to home in on the incomplete infill. The idea is to minimize movement associated with one tower/axis and make a print at 3 locations with the same part orientation and infill with respect to the tower. Here's the idea:

http://mhackney.zenfolio.com/img/s8/v80 ... 7748-4.jpg

I used a combination of slic3r and Repetier host settings to layout the thing to print.

After watching this thing print several dozen times and seeing the consistency of the infill gaps I am starting to lean towards a tight U joint somewhere. I'm hoping this experiment will help me predict which set of joints (by tower/axis).

It's been a little tricky printing these out at the perimeter of the hot bed. Partly due to the temperature out there being much lower so I am having sticking problems and also because of this gradient, the hot bed itself is warped with the center higher. I currently do not have a glass build surface since mine "blew up"! My borosilicate surface has been shipped. When it comes in I will put a layer of aluminum down to dissipate heat more evenly and hold the Onyx flat with the glass plate on top. I suspect that mine broke because of the Onyx warp and the glass being held very securely around its perimeter.

I'm doing the prints now.

[Polygonhell]

FWIW I just ran a print with Slic3r generated GCode (wanted to test a brim with Nylon), and my infill reaches the edges pretty much identical to my KISSlicer settings.
So I'm going to say the issue is either mechanical, or related to calibration.
On the calibration front If you're motion is convex, the rise towards the outside of the bed can cause the filament to not stick to the bed and the Hotend to pull it away from the side as it reverses direction.
On the mechanical side it took me a while to figure out where the UJoints were binding when I sanded mine, and I probably took off more than I needed to.

What layer height are you using, I just looked at the new nozzle that was shipped to me and it doesn't have a 0.5mm diameter hole, I'm assuming it's 0.35, which would mean you want to be printing with a layer height of 0.25mm or less.

[mhackney]

I'm printing .3mm layers with a .5mm nozzle. I've completed the 3 test prints and I am pretty sure it is mechanical and should be able to predict which tower is the culprit. The print at the Z tower is near perfect. At this position, the Z axis movement is much more limited along the line to the center of the bed but there is maximal "left to right" movement.

The Y axis print is the worse of the 3 with gaps that align with the line to the center of the bed.

The X axis print is a little worse than perfect, again in the direction to the center of the bed.

I'm off for a family function so I'll have to think about what this means and post photos!

[Eaglezsoar]

FWIW I just ran a print with Slic3r generated GCode (wanted to test a brim with Nylon), and my infill reaches the edges pretty much identical to my KISSlicer settings.
So I'm going to say the issue is either mechanical, or related to calibration.
On the calibration front If you're motion is convex, the rise towards the outside of the bed can cause the filament to not stick to the bed and the Hotend to pull it away from the side as it reverses direction.
On the mechanical side it took me a while to figure out where the UJoints were binding when I sanded mine, and I probably took off more than I needed to.

What layer height are you using, I just looked at the new nozzle that was shipped to me and it doesn't have a 0.5mm diameter hole, I'm assuming it's 0.35, which would mean you want to be printing with a layer height of 0.25mm or less.

This is something I often wondered about, can the .35mm hole be drilled out to .5mm without any detriments?

[Polygonhell]

This is something I often wondered about, can the .35mm hole be drilled out to .5mm without any detriments?

I don't see why not, I'd imagine the other dimensions are all the same.
FWIW I now believe the new nozzles are very close to 0.5mm, I measured the free air extrusion width at 0.8mm which is wide even for a 0.5mm nozzle, it implies a very short nozzle orifice.

[Eaglezsoar]

This is something I often wondered about, can the .35mm hole be drilled out to .5mm without any detriments?

I don't see why not, I'd imagine the other dimensions are all the same.
FWIW I now believe the new nozzles are very close to 0.5mm, I measured the free air extrusion width at 0.8mm which is wide even for a 0.5mm nozzle, it implies a very short nozzle orifice.

Thanks for the answer, I am going to try it.
On another note, I have been waiting to use your version of repetier firmware but I need to know if you've had the time to get the LCD to work.
You seem to be one of the most knowledgeable on this forum, that's why I've been waiting. Does it really make that much difference from Marlin to make all the changes to Repetier? The Pronterface
front end looks nice but I doubt if it will work with the repetier firmware.

[mhackney]

Read my posts a few back, polygonhell's repetier firmware does indeed support the LCD and the SD card reader.

[Eaglezsoar]

Read my posts a few back, polygonhell's repetier firmware does indeed support the LCD and the SD card reader.

This is confusing. I see where you showed how to hook up the LCD and you state you were using firmware from Halopend (sp), I assume the
Halopend firmware was created by polygonhell and everything works. There is too many subforums too keep track of.
Thanks for the help.

[mhackney]

I know it's difficult to follow an entire thread! I did indeed start with halopend's firmware. Then I said that I switched to polygonhell's firmware for its LCD and card reader support in polygonhell's repetier thread. So I actually switched but the reason was to get the card reader support. Halopend's firmware does support the LCD too.

[Eaglezsoar]

I know it's difficult to follow an entire thread! I did indeed start with halopend's firmware. Then I said that I switched to polygonhell's firmware for its LCD and card reader support in polygonhell's repetier thread. So I actually switched but the reason was to get the card reader support. Halopend's firmware does support the LCD too.

Okay, so that I can get my tiny little mind to understand all of this, I should get the developmental version of Polygonhell's repetier which does support LCD and SD cards.
Is this correct, for at least today?

[mhackney]

That's it! https://github.com/polygonhell/Repetier ... evelopment

Read the 2/14 comment - enabled the LCD and SD card.

[Eaglezsoar]

That's it! https://github.com/polygonhell/Repetier ... evelopment

Read the 2/14 comment - enabled the LCD and SD card.

GOT IT! Man does my head hurt from all of this...Thanks to you and Polygonhell, you guys are the heart of this forum.

[mhackney]

I posted this in another topic but wanted to capture here in my topic too.

Z, Max_LENGTH and Bed Level Calibration

I'm using Repetier firmware and host (on a Mac). I'm assuming you have a heated bed - ignore the pre-heating step if not!

1) Install your build surface (glass plate and/or blue tape, etc) and make sure the heated bed is up to temperature. I print ABS at 100°C bed, so that's the temp I used. The Onyx warps as it heats up so unless you have a flat surface on top of it, you will not have a planar surface to calibrate against. If you are printing on tape directly on the Onyx, do the following steps cold and deal with the warping issue depending on the location of your print, etc.

2) Calibrate Z at 0,0,0 in firmware. These are listed as the following in the firmware:

#define X_MAX_LENGTH 369.0
#define Y_MAX_LENGTH 369.0
#define Z_MAX_LENGTH 369.0

Here's how I did it: I set these to 375 so they are longer than my actual travel. Then I homed and brought the nozzle down to Z=20mm (G0 Z20 F2000). From there I used the buttons in Repetier host to slowly lower the nozzle (decrease Z) to the build surface. Start with the 10mm and then 1mm and finally .1mm for fine adjustment. I used a piece of cigarette paper (an old machinist's trick, it's about .001" thick) between the bed and nozzle (make sure your nozzle is clean!) to test the "fit" you can use a .001" (or metric equivalent) feeler gauge too. You want to bring the nozzle down until it just "snags" the paper or gauge. At this point, read the Z height on the Z axis display in Repetier Host. Subtract this number from 375mm to get your actual max lengths. For me this measurement was exactly 6.0mm, so 375.0mm - 6.0mm = 369.0mm. If you are using EEPROM settings, you can simply change this in the EEPROM without needing to recompile and upload your firmware. That makes it a lot faster to do and test. Check this at least 3 times to make sure it is reproducible using this process:

Home All - move to G0 Z10 - use buttons to lower 1mm nine times - place gauge - lower by .1mm 10 times - nozzle should just graze gauge

3) With your Z = 0 set you can adjust the planarity of the bed by adjusting the stops at the homing switches at the top of the towers. This is done one tower at a time with the nozzle positioned as close to the tower as reasonable. The process is described in the manual but use the "true" circumference points. These are:

X: G0 X-77.94 Y-45 F2000
Y: G0 X77.94 Y-45 F2000
Z: G0 X0 Y90 F2000

(note F2000 is feed rate to speed things up) These locations came from http://minow.blogspot.com (a MUST read for calibrating!)

At each one of these locations, you move down (using the Z buttons) until you get to Z=0. If you are lucky, the gauge will just graze the nozzle and you are done at that tower. If not, the nozzle will either be above the bed (case A) or the display will show that Z has not been reached (case B).

Case A Measure the gap and set the stop screw at the top of the tower (see a few lines below).

Case B Write down the extra distance the Z could go and set the stop screw as below.

Setting the stop screw: These screws are 6-32. That means they will advance .03125" per revolution. Converted to metric that is 0.79mm per revolution.

Case A - let's say that you measured a gap of .4mm. This means that the firmware thinks the tower length is .4mm shorter than it actually is! So, you need to "raise" the stop screw (lefty-loosey or counter clockwise looking from the top) .4mm - which is about 1/2 of a turn.

Case B - let's say that Repetier host shows (Z axis display) that you still have .2mm to go before reaching Z=0. This means the firmware thinks the tower length is .2mm longer than it actually is, so you need to "lower" the stop screw .2mm or about 1/8 of a turn, in this case righty-thighty or screw clockwise from the top.

That's it! Now you should go back and recheck everything 1 or 2 times to be sure.

Note that there is an advanced planarity adjustment described in the link above. I have not done this yet since my glass plate broke and I'm waiting for the replacement. If your machine is really "off" the re-check step above will show that Z=0 at X=Y=0 will not be correct. I'd adjust one more time through the entire process and if you still have this problem, then do the planarity adjustment. Hopefully I'll be there by Tuesday or Wednesday when my plate arrives!

There is an "auto calibrate" mode/command in the LCD controller firmware in Repetier. I have not had a chance to investigate it. In theory, all of the above could be accomplished in soft settings with the right math. It probably does not hurt to try to calibrate mechanically first though!

DISCLAIMER: there is some probability that there are mistakes or better ways to do the above. This is my first delta printer and I'm only a few weeks in and don't have a lot of experience yet!

NOTE/WARNING - since posting this on the original topic I tried playing around with the auto calibrate in the firmware exposed in the LCD Controller. Not intuitive and I hosed my configuration! I ultimately had to reupload the firmware with EEPROM disabled to get back to a good state. All sorts of weird things were happening!

[mhackney]

Following up on my post yesterday on the controlled experiment to determine why I am getting gaps in the infill on my prints. Here are the 3 test prints laid out to show how they were printed.

[img]http://mhackney.zenfolio.com/img/s8/v77 ... 8494-4.jpg[/img]

Note that I oriented the fill to go from the center of the build plate to the tower. You can read the details in the earlier post.

I know how this story ends so hang in there!

First, look at the part at the Z tower. It actually looks pretty reasonable and is the best infill I've had to date. At this position, there is very little movement along the axis (from the tower to the center of the table) and a lot from left to right.

Now look at the part at X. It may be hard to see but I have gaps at the arrows. This is what I typically have seen and the problem I'm trying to fix.

Then look at the part at Y, it's the worse one of the three. Gaps along multiple sides.

After thinking on this overnight, I hypothesized that one side of the Z delta arm U joints may be stiff or have backlash. I couldn't predict which side but it just seemed like Z was the culprit. I disassembled the delta arms from the Z carriage and immediately discovered that the delta arm on the X axis side was very stiff compared to the others. I took it apart and found 2 problems: 1) I had used a small screwdriver to install the arm on the aluminum bearing - BAD IDEA. I had several gouges in the face of the bearing that had burrs large enough to create a lot of friction. I also discovered that I had not "defuzzed" the delta arm U joint when I rounded the top. The fuzz had caught between the arm and joint, again adding friction. I cleaned all of this up, regreased with lithium grease, reinstalled and prepared for a test print to see if I found the problem. And here is the result - this piece was printed at the Y position so it was in the same spot as the worst of the three prints in the photo above:

[img]http://mhackney.zenfolio.com/img/s8/v85 ... 0018-3.jpg[/img]

This part is very nice! No infill gaps and nicely filled overall.
So I am ready to proclaim that stiff delta rod U joints can lead to odd printing behavior! Given the number of folks with infill gap issues, I think re-investigating the U joint fit will eliminate those Delta Arm Blues!

Now I can go on to more fun things!

Cheers,
Michael

[Polygonhell]

Just a point for the calibration, DONOT believe the Z Height reading in repetier host, use M114.
Repetier host and all the othe host software I've used tracks the position separately and accumulates errors while jogging. M114 tells you exactly where the firmware thinks it is.

[mhackney]

Good tip Polygonhell! I'll start using that from now on.

[geneb]

mhackney, will the calibration points you list fit on a Phebe I bed area? If not, I'll adjust them until they do fit and then add both sets of points to a new rev of the manual.

Thanks for researching this so thoroughly!

g.

[mhackney]

Hi Gene, I'm not sure if they'll fit - probably not. I have a phoebe on my H-1.1 so I'll try to measure it and see.

Cheers,
Michael

[Polygonhell]

No they won't fit, just multiply all the numbers by 7/9 that should fit.
Any scale will retain the points in a circle which is what matters.

[mhackney]

I installed some ultra bright LEDs on my Max. After playing around with a few ideas, I decided to drill 3 small holes at a 45° angle between the U joints on the delta carriage.

[img]http://mhackney.zenfolio.com/img/s8/v75 ... 7916-4.jpg[/img]

I supported the carriage with a roll of tape. I started the hole vertically and then once started, angled the drill to complete. It's a 13/64" bit so the wires from the LEDs fit through. I bought a bunch of these 12V LEDs for this and other projects:

[img]http://mhackney.zenfolio.com/img/s11/v3 ... 6570-4.jpg[/img]

I wired them up with a switch mounted in the LCD Controller door. There is a hole conveniently labeled "Lights"!

[img]http://mhackney.zenfolio.com/img/s8/v83 ... 6382-4.jpg[/img]


Wire routing:

[img]http://mhackney.zenfolio.com/img/s8/v80 ... 6166-4.jpg[/img]

Here's a video of the Rostock taking a trial run with its new lights.

[youtube]http://www.youtube.com/watch?v=5aTc7oI4-k8[/youtube]

It would be nice to have 6 LEDs but 3 fit and they are angled in towards the nozzle so they do cast a nice beam. I am thinking about shortening the length of the 3 aluminum posts that attach the barrel assembly so the nozzle protrudes more below the delta carriage, maybe another 1/2 or so. That would give me the room for a fan duct like highcooley's.

[img]http://mhackney.zenfolio.com/img/s1/v49 ... 6310-4.jpg[/img]

[dbarrans]

I had the very same thought last night, about using shorter spacers to lower the hot end, but it was to let me see the workpiece better during a print.

- dan

[mhackney]

Dan, that was one of the reasons I want to lower the nozzle too, it is tough to see under there!

[mhackney]

My new borosilicate glass is due to arrive from SeeMeCNC in a couple of days. So, I'm doing some prep work for it. The Onyx is a great heated bed but like all heated beds, it isn't perfect. I have been plagued by 2 problems.

1) The center is a lot hotter than the perimeter. Even after letting the bed warm up for extended time, there is a pronounced temperature gradient. Many heated beds have this issue. The nichrome bed I made for my H-1 especially since it has a fairly short coil of nichrome. I solved the even heat problem with an aluminum (1/16" on the H-1) plate between the heated bed and glass.

2) My Onyx bed warps at high temp. The center pops up because the edges are held in place and thermal expansion wins. This may have contributed to the demise of my original window glass plate. In any case, it makes it a bugger to get the first layer or even 2 right - especially on larger prints.

So, I am going to kill 2 problems with one aluminum plate! The idea is to fasten a 1/8" thick aluminum plate over the heated bed. I decided to go with thicker aluminum for its structural strength over the 1/16". Here's a picture ready to install:

[img]http://mhackney.zenfolio.com/img/s3/v7/ ... 7498-4.jpg[/img]

It is 12 3/16" diameter with a flat and cutout for the electrical connections at the back.

[img]http://mhackney.zenfolio.com/img/s8/v81 ... 4304-4.jpg[/img]

This cutout will also allow the full circle borosilicate plate to rest on top and not interfere with the electronics (I can bend the LED back out of the way).

I need to get some longer 4-40 flat head screws to mount through the aluminum, Onyx, melamine, and spacers!

[Eaglezsoar]

May I ask how you cut the aluminum so perfect?

[mhackney]

The outer perimeter was cut on my 12" metal bandsaw. I cut A LOT of metal so I'm pretty good at free handing! the cutout for the back was done on one of my milling machines by hand. It's pretty simple.