3DR "Simple" Delta Printer

can you tell me what you have changed in your pins.h file?

Well, you are using Mega2560/Ramps and I am using Printrboard so my pins.h is using 81 for pins.h not 33.

Wait, If I share my Mini Kossel firmware with you......maybe.....here is my Mini Kossel pins.h that uses Mega2560/Ramps1.4 with NO auto bed leveling.

pinsMiniKossel.pdf

And here is my configuration.h for my Mini Kossel
ConfigurationMiniKossel.pdf

Thanks!

I'm going to go back to the beginning. I must have screwed my firmware up at some point. I'm going to go back to Marlin-Deltabot from Git and restart.

I really do appreciate your help, Richard!

the first thing I notice from your pins.h is that you have your end-stops connected to the min pins and you disabled the max pins? On my 3dr, they're connected to the max pins and the min pins are disabled. Is that backwards? I wonder if that's why I'm having compile issues.

I also see that your diagonal rod lengths are 181mm vs on mine they are 200mm. Do you think I should shorten those up to 181? I used John's jig so I know mine are all the same length but they would appear to be long. Perhaps I used Rich's length instead of John's.

Edited 2 time(s). Last edit at 05/24/2014 01:16PM by umdpru.

"...the first thing I notice from your pins.h is that you have your end-stops connected to the min pins and you disabled the max pins?"

That is for board #77, the pins.h is very difficult to read visually because there is so much overlap between the different boards and the logic is messy.
The max pins are setup in pins.h that work.

In configuration.h there is the possibility to mess things up with "1" versus "-1" and using "true" versus "false" for the end stops and pull-up resistors.

Yes, I have messed up my firmware so bad I had to start over from the beginning more than once.

I have started over and I can't even get Johann's deltabot Marlin to comile without changing a thing. It gives me this error "qr_solve.cpp:7:18: error: time.h: No such file or directory
"

From what I can tell, I don't have a time.h library. I didn't have the issue before, though.

I'm going to take a break and come back to it later.

EDIT: I'm back. Richard, are you using your max endstops for the endstops or the minimums? Also, what firmware did you start with as your base? The firmware that I can't turn autoleveling off on is Johann's branch Marlin-Deltabot. I can turn off autoleveling on the regular Marlin-Marlin V1 (delta version). I extracted a fresh copy of Johann's deltabot branch and ONLY commented out auto bed leveling and it won't compile and gives the same "accurate_bed_leveling_points not declared in this scope" error.

I fixed the "time.h" compile error by commenting out the #include in qr_solve.cpp. I wonder if it will present a problem. I did that on my original firmware too and I have printed some items well. Just have to nail down the first layer issues I'm having.

Edited 4 time(s). Last edit at 05/24/2014 03:00PM by umdpru.

So I downloaded the Marlin-Deltabot from Erik's git and it's different from the version I had. I was working from an old version. The auto level define is not in the configuration.h now....

this is quite the journey!

Yes, a journey. Until a few months ago, although I was aware of what firmware was, I had never uploaded or modified firmware.
Now, I am volunteering with the technology teacher at a local high school and he wants me to help him have a project for the students in Fall 2014 to use twenty Arduino Uno kits that he has.
So, I bought a book "Arduino In Action" that has all the info I need.....my summer reading and projects.

Richard,

What are the chances that I can get you to upload a copy of your Firmware, in it's entirety, to an online location like my dropbox or google drive?

I'm at my wits end here.

OK, working on uploading them now.

[www.dropbox.com]

That is the firmware for my 3DR Simple. However I use a Printrboard on the Simple.

I will next upload my Mini Kossel firmware that uses Mega2560/Ramps1.4

Dropbox link for Mini Kossel firmware

[www.dropbox.com]

I don't know what version of Marlin_Marlin you downloaded (I assume March timeframe) but it's working! I can now adjust the z endstop offsets, I don't have a Z_min endstop showing triggered in M119 and I can turn off autolevel and compile.

I believe my issues before were related to the fact that autolevel was enabled and expecting to use the Z_min endstop for the probe. I don't know why that would affect Z-tower endstop offset settings, though.

THANKS!!!!

Now, on to a proper calibration!

Edited 1 time(s). Last edit at 05/24/2014 09:26PM by umdpru.

Great, glad it's working for you.

Well, working is a relative term. I can calibrate the Z tower endstops now.

I think I have a fundamental movement problem. I calibrate the towers and am working on the center. I have gotten it close (~.1-.2 below Z0), however, I have an issue where the center is too low but points between the X-Y tower points track above the build plate and points between the Y-Z tower track REALLY low, ~1mm below Z0.

I have quadruple checked the physical build measurements now and everything is right, as far as I can tell.

the offsets being quite different at the Z tower should mean something but I can't quite figure it out? Are the diagonal rods hosed? I'm going to have to pull those and measure. Will probably shorten them to 180 while I'm at it.

Now I'm stumped...

Edited 2 time(s). Last edit at 05/25/2014 02:20PM by umdpru.

1. One of the M666 endstop adjustments might be more than the other two when the physical endstop is tightened down at a different height.
I would not worry about it because the end result you want is a flat build plane and you can get that with M666, and adjusting delta radius, plus one other adjustment below.
The M666 value gets you close to a flat plane. Then adjusting delta radius makes the print plane either a right side up bowl or an upside down bowl shape.
The purpose of the M666 values in front of the towers and then adjusting delta radius is to push the intersection of the bowl out away from the center of the printer as far as possible.
Imagine using M666 and delta radius like a blacksmith hammering out a bowl to make the diameter larger and the sides lower. First you have it upside down then it flips and you have it right side up.
You use M666 and delta radius to hammer it as flat as possible but it is never going to be perfect.
The next step is to then work on getting the plastic to stick to compensate for the imperfect flatness of the print plane.

2. Besides M666 and delta radius, there is one other adjustment in firmware that you can make. You can increase or decrease in Firmware the DELTA_DIAGONAL_ROD value.
This value is used in calculations buried in the code I don't remember where. Anyway, as you can see in Configuration.h the angle assumed is exactly 60 degrees.
However the actual angle is never going to be exactly 60 unless you custom make your rods and everything else is perfect. So that's where this adjustment helps.
Unfortunately, this is not a well used nor well documented adjustment, I don't know which way to adjust to get what result yet.

Oh, I see now. What you need to do is lower in firmware your MANUAL_Z_HOME_POS. This gives you the flexibility in the values of M666 and delta radius to get the print plane more level.
I have to keep reminding myself that physical values are what they are. The values in the firmware are different in order to get a level print plane, even though the glass build surface is flat and level.

Thanks, RTurnock. I understand the theory behind using M666 and then adjusting Delta_Radius (adjusted via smooth rod offset). I also figured out that z home had to be set low to give the headroom to use M666 since it can't be positive.

What I'm struggling with now is the inconsistent movement. I would expect movement from x-y (if concave or convex after M666) would be the same as the movement from y-z and from z-x. I'm not seeing that. I'm seeing convex from x-tower to y-tower and concave from y-z towers and then convex from z-x towers.

I took my machine apart yesterday and found some issues that need to be addressed. I also don't think my towers are parallel to the central axis of the bot which needs to be fixed. My Z carriage diagonal arm mounting lugs were also cracked. I also found my Z tower endstop was higher by 1/32" than the other two endstops. I also found that my diagonal rods are 201mm not 200mm. I imagine that would have an effect on the printing. I don't see how it would create the onconsistent movements, though.

I might print up some parts to make the frame stiffer and allow me to pull it square. I'm really having a difficult time figuring out a good repeatable way to measure the square between the towers.


what I can say is that I have printed a few vases and they came out REALLY great if you ignore the aweful first few layers. The positional repeatability of the nozzle movement is on par with my Replicator 2x which is my current benchmark. My Prusa I3 and my highly modificed MBI Cupcake CNC can't get the same repeatable positioning of the 3DR simple!

Edited 2 time(s). Last edit at 05/27/2014 09:27AM by umdpru.

Wow, you have four 3D printers! And they are all different!
Yes, the Delta design is different in terms of movement, calibration, leveling, alignment.
Wish I could see what you have going on, but I am not any help with hands on things from this distance.

My 3DR is really stiff. The 2020 extrusions plus the two smooth rods on each column work well and they fit very tight into the plastic frame parts.
I have a problem that I go to print and it starts and moves down then stops......no error message, nothing to tell me why.
Might be a loose connection.

Yes, 4 printers. I'm no noob when it comes to these but this 3dr is kicking my butt. I'm putting serious thought into turning it into a kossel mini. I like the frame on that machine better. I'm going to reprint the Z tower carriage. I hope that one being messed up was the cause of my issues.

Edited 1 time(s). Last edit at 05/27/2014 04:17PM by umdpru.

Yes, I built a Mini Kossel before I built the 3DR and really liked it, except for the E3D hot end was very sensitive and so I printed another effector assembly from the 3DR Simple and bought a Ubis hot end from Prinrbot.
Then I changed the 240mm horizontals to 350mm and made 340mm diagonal rods. Ready to test it out tomorrow and try printing.
I want to print parts to build an Eggbot and the Cyborg Beast - a prosthetic hand.

I've been researching my options to convert what I have purchased for the 3DR simple into a Kossel. Ultimately, if I can't get it to work right then I'll convert it. I recently purchased a MakerfarM I3v upgrade kit for my MF I3, though, so anymore printer purchases will hae to wait.

I found one interesting problem with my 3DR Simple. The Z-carrriage mounting lugs for the diagonal rod were cracked. The crack was such that it allowed the through bolts to sit lower on the carriage. This could account for the LOW nozzle height when the Z-carriage was raised (bolts/arms) move down slightly as would happen when the machine is moving the nozzle from y to Z (Z carriage moves up). I dind't have any more spectra line so I superglues the carriage lugs back together and reinforced them a bit. I'll print some more at a higher infill when I get more spectra line. I might get belts/pulleys anyways although it would be a shame to waste the money I spent on the machine spools.

Your experience and mine run in parallel paths.
Yesterday I took the Airtripper Extruder off my Mini Kossel and replaced it with the Greg's Wade Extruder off my 3DR.
Earlier I had bought a Ubis hotend to replace the E3D. Last night it finally worked reliably for a 5mm pyramid test print.
The 3DR Ubis hot end will replace the Ubis on my Simple that had a broken thermistor that I replaced but is now unreliable due to wire shorting intermittently so I need to take it off and rebuild it to insulate the wires better, or stop them from bending.
Hoping to print more parts for Eggbot, also planing on building Watercolorbot and Mirobot (as seen on kickstarter).
Take Care.

Very nice.

I believe I may have found my issue.

[forums.reprap.org]

Using [kreature.org] as a reference.

I think one of my towers is slightly non-symmetrically located.

Quote
JohnSL
It's starting to come together. Yesterday I printed the three parts that make up a third of the base. I went through several iterations before the parts you see here, which are themselves not even final. I added alignment dimples and indentations between the three parts, which makes it really easy to assemble these parts perfectly. I also added nut indentations in so you don't need a wrench to assembly the parts. I'm pretty happy with how these turned out, so I'm printing the next generation of parts in order to assembly the second third of the base.

hi just a question how many hours it total did it take to finish the base and the top pieces when it was split up into smaller segments?

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

chris,

The time required to print the pieces is highly dependent on the printer your using and it's capabilities. John had a very fast MendelMax to make his parts. I'm using a highly modified Makerbot Cupcake CNC which prints at 70-80mm/s. I believe John said his MM was printing at 150mm/s earlier in the thread. My bottom corner pieces take 7 hours a piece on my Cupcake and the wings take approximately 1.5-2 hours per unit.


As for my 3DR, I have decided to reprint my frame pieces. There are some inconsistencies that I can't live with. I also need to rule out the frame as a calibration problem. I printed my first frame components on my MBI Cupcake before upgrading to a better Z-stage and a J-Head. I previously had a belt driven z-stage and a MK 5 Makerbot extruder which wasn't terribly precise.

So far, I have the three bottom corners printed and they are absolutely much better than my first attempt. I'm hoping my printed frame was my problem.

I have tried calibrating a bit further. I have been able to level Points X, Y, Z, and C. But points L, F are high and R is REALLY low (under the build plate). I think the problem is a radial positioning error in the X-tower.

I'm also going to build a mini-kossel since it has much better "factory" support in the community right now.

It would appear that this thread has died.

Has anyone gotten their 3dr simple to print well? I see many have them "printing" but how many are dialed in?


An update on mine:

1. I have finished re-printing the bottom frame pieces. It seems the first round of parts weren't accurate. The new bottom frame places one tower approximately 1mm further away from the others. 2 always align but the third is always out by about a mm. have to assume this affected the angular placement of the towers too. I'm hoping this was the cause of my major calibration issues.

2. I will be using a three bearing carriage to gain better linearity in the carriage movment. Right now, the linear force applied by the spectra line is not centered on the carriage and induces a moment and torque loading on the bearings which can cause binding. A three bearing carriage will fix that. I may just glue two carriages together and have 4 bearings.

3. The motor mounting boss on the bottom corners could have been fused to the corner sides in the model. There is still available flex in the corner pieces because of a lack of bracing that transfers loades from the sides. tying the motor boss to the sides will substantially strengthen this piece. I won't repritn them but will add a glued in filler piece (glued with abs slurry).

4. I am going to glue all frame pieces together and then screw them together. This will substantially increase the rigidity of the frame.

5. I am going to print out flat pieces to glue to the "wings" to fully box them in after the frame pieces are glued together. This will also substantially increase the rigidity of the frame.


If I fail to get this machine calibrated then I'm going to sell the frame (minus Ramps and motors and move on).

Chris,

My bottom frame took a total of 34 hours of print time for all 9 pieces.

I estimate my upper frame will take approximately 28-29 hours of print time (less involved upper corner pieces).

The carriages took 4.5 hours to print all three.

The end-effector will take approximately 4-5 hours to print the pieces.

These times are for my machine running at 80mm/s and 500mm/s^2 acceleration settings.

Edited 1 time(s). Last edit at 06/05/2014 08:48AM by umdpru.

Quote
umdpru
Has anyone gotten their 3dr simple to print well? I see many have them "printing" but how many are dialed in?

Mine is printing parts which I'm quite happy with:

Very nice!

I'm quite jealous, actually. Did you have any difficulties calibrating it? What did you use to print the parts in the first place?

I had a thought last night that my steps/mm for the three axes could be causing my problems ( I left them at 55.5 default). I calibrated those and everything is very close so that wasn't it. I'm back to waiting to finish printing a new frame.

I'm seriously hoping that the frame is the issue.

Quote
umdpru
Very nice!

I'm quite jealous, actually. Did you have any difficulties calibrating it? What did you use to print the parts in the first place?

I had a thought last night that my steps/mm for the three axes could be causing my problems ( I left them at 55.5 default). I calibrated those and everything is very close so that wasn't it. I'm back to waiting to finish printing a new frame.

I'm seriously hoping that the frame is the issue.

Took a bit of tinkering to figure out how to calibrate it (described in the previous pages of this thread, I believe). Parts were printed on a RepRapPro Mendel in ABS, although that took some effort as well because it's not an enclosed unit so I was fighting warping issues.