Building my Delta-Pi - Journal/Q&A/lessons learned
Posted by beamsaber
|
Building my Delta-Pi - Journal/Q&A/lessons learned December 10, 2013 06:24PM |
Just a running journal of my progress, hopefully some of you can learn from my mistakes.
I bought a Printrbot Simple ($300) with the explicit intention of using it to build a more advanced reprap 3D printer. I was at first completely enamored of Nicholas Seward's geared Simpson reprap (still am to be honest, it just looks so steam-punkishly awesome!) but the designs had not yet been released when I started this, and now that they are I find that there is a great deal of complexity in that model which I don't quite feel up to tackling just yet.
Looking through the current models being built I became intrigued by the Delta-Pi given it's elimination of expensive linear bearings and rods and instead using roller skate bearings on cheap and readily available square tube. Given I had limited funds for this project, I set out to acquire all of the parts I would need.
BOARD:
I bought a printrboard Rev E (which I felt safe with having the rev D on the simple) in hindsight, as convenient as it is to have all the stepper drivers integrated on the board, I will likely buy more modular fare in the future. The board I ordered from reprap.me was either delivered damaged, or burned out before I could finish attaching power and motors, because the Z axis motor refused to work.
I bought a second printrboard Rev D, this time from a more local shop (the Rev E was from denmark and I am in California)
MOTORS:
I found three NEMA-17 motors on ebay for $15. On receiving them I found they were fitted with a gear already and not a timing belt pulley. A bit of research informed me that trying to remove them by brute force would likely render the motor useless, and some techniques (using a chisel, metal plates, a vice, and hammer) could do the same, so I purchased a small gear puller which arrived just recently. The small gear puller actually did not fit under the gear itself. Thankfully the gear was plastic and a rather soft metal, so I used a drill bit to grind down the underside until I could fit the gear puller under it, once I was able to fit the puller on, it was like magic. They are 5 wire two phase motors, so I also had to learn how to re-wire them to work as a 4 wire motor. I then ordered a larger NEMA-17 motor for the extruder.
EXTRUDER:
I was able to find a jHead already fitted as a bowden extruder with thermistor, which was very handy.
I then printed and assembled the airtripper's bowden extruder and assembled it. I had to substitute some parts rather creatively because I was missing some screws etc, but in the end I got a serviceable bowden extruder assembled.
Right now I am using a custom printed part I built on tinkercad to hold the jHead, it is printed in PLA, so I am concerned about it melting despite active cooling. Will report back how that goes.
SURFACE:
I salvaged a glass plate from an old scanner, as well as endstops from a couple of old printers (people had left them near the trash or on the curb so I figured why not help myself to some parts!)
RODS:
I ordered aluminum square tube, which is not what the BOM calls for, and I may regret as I am already seeing how easily this aluminum tube is to scratch/dent, but it was a lot cheaper, and I can always replace it later on.
WOOD:
I purchased two oak plywood boards to use for the wooden portions. The thickness of these I was later to learn is not in line with the specs in the design (designated screws are too short) thankfully I had a lot of spare longer screws which I was able to employ instead.
PRINTED PARTS:
I purchased $20 of PLA 1.75 filament from printrbot.com off the printrbargains section (1k) I won't be doing that again as this filament was murder to work with. It jammed repeatedly (looks like they were trying to add some shine to the filament and the batch went bad, so the metallic elements in the filament kept clogging the hotend). I also ran into some serious skewing of the print on my Printrbot Simple, partially owing to voltage/acceleration issues, but primarily due to the fact that the printer, assembled as intended, has an unsupported Y axis which tilts inside the bearings about a degree and a half. I found a part on thingiverse that corrects for this, but not until I had printed all of the parts for the printer, and found that they were all faulty, especially the hotend tri and the carriages. I have since re-printed the parts, which was a little more than frustrating, but the quality is a lot better now, and I can't see any skewing in the new parts with the adjustments I made to the printer, and the correction parts attached to fix the Y-sag.
The hub tri that holds the hotend was also a problematic part. The STL that is on thingiverse is just over 100mm no matter how I rotated it, so it would not print on my Simple. I had to pull it into tinkercad and split it into multiple pieces using holes, and then add on fastening points to bolt the whole thing back together. Once I've tested and confirmed this new design works well, I'll likely post it on thingiverse for others with a small print-bed that would like to build a delta-pi.
FASTENERS:
I purchased the collection of screws from trimcraftaviationrc.com, can't recommend them highly enough. They contacted me immediately after I placed the order to let me know one size was out of stock, and allowed me to substitute a different cap.
MISC:
I scrounged up the remaining parts (pulleys etc) from various small shops online. I think I had the most trouble finding the belts.
While all of the parts were printing I went ahead and manually drafted the drilling and cutting points for the top and bottom board. If I had it to do again I would prefer to have the extra cash to get laser-cut acrylic or aluminum pre-built instead.
Once I had them drafted I cut/sanded/stained/varnished them to a nice dark mahogany. Right now I have only treated the top and bottom board, but I have more wood that I intend to use to build support walls and doors for the printer. Ideally I would like to bolt the power source (a PC power supply) up under the main body of the printer with a wooden face plate over it to provide a main power switch for the whole machine, so I am going to have the support walls extend about half a foot under it and put rubber feet on it to reduce vibration transfer to whatever I place it on.
First things first of course, I want to make sure the damn thing works
.
I just finished printing the last part of the new hub-tri so I will be able to hook up the belts and run some tests probably tomorrow.
Pics as soon as I take them.
I bought a Printrbot Simple ($300) with the explicit intention of using it to build a more advanced reprap 3D printer. I was at first completely enamored of Nicholas Seward's geared Simpson reprap (still am to be honest, it just looks so steam-punkishly awesome!) but the designs had not yet been released when I started this, and now that they are I find that there is a great deal of complexity in that model which I don't quite feel up to tackling just yet.
Looking through the current models being built I became intrigued by the Delta-Pi given it's elimination of expensive linear bearings and rods and instead using roller skate bearings on cheap and readily available square tube. Given I had limited funds for this project, I set out to acquire all of the parts I would need.
BOARD:
I bought a printrboard Rev E (which I felt safe with having the rev D on the simple) in hindsight, as convenient as it is to have all the stepper drivers integrated on the board, I will likely buy more modular fare in the future. The board I ordered from reprap.me was either delivered damaged, or burned out before I could finish attaching power and motors, because the Z axis motor refused to work.
I bought a second printrboard Rev D, this time from a more local shop (the Rev E was from denmark and I am in California)
MOTORS:
I found three NEMA-17 motors on ebay for $15. On receiving them I found they were fitted with a gear already and not a timing belt pulley. A bit of research informed me that trying to remove them by brute force would likely render the motor useless, and some techniques (using a chisel, metal plates, a vice, and hammer) could do the same, so I purchased a small gear puller which arrived just recently. The small gear puller actually did not fit under the gear itself. Thankfully the gear was plastic and a rather soft metal, so I used a drill bit to grind down the underside until I could fit the gear puller under it, once I was able to fit the puller on, it was like magic. They are 5 wire two phase motors, so I also had to learn how to re-wire them to work as a 4 wire motor. I then ordered a larger NEMA-17 motor for the extruder.
EXTRUDER:
I was able to find a jHead already fitted as a bowden extruder with thermistor, which was very handy.
I then printed and assembled the airtripper's bowden extruder and assembled it. I had to substitute some parts rather creatively because I was missing some screws etc, but in the end I got a serviceable bowden extruder assembled.
Right now I am using a custom printed part I built on tinkercad to hold the jHead, it is printed in PLA, so I am concerned about it melting despite active cooling. Will report back how that goes.
SURFACE:
I salvaged a glass plate from an old scanner, as well as endstops from a couple of old printers (people had left them near the trash or on the curb so I figured why not help myself to some parts!)
RODS:
I ordered aluminum square tube, which is not what the BOM calls for, and I may regret as I am already seeing how easily this aluminum tube is to scratch/dent, but it was a lot cheaper, and I can always replace it later on.
WOOD:
I purchased two oak plywood boards to use for the wooden portions. The thickness of these I was later to learn is not in line with the specs in the design (designated screws are too short) thankfully I had a lot of spare longer screws which I was able to employ instead.
PRINTED PARTS:
I purchased $20 of PLA 1.75 filament from printrbot.com off the printrbargains section (1k) I won't be doing that again as this filament was murder to work with. It jammed repeatedly (looks like they were trying to add some shine to the filament and the batch went bad, so the metallic elements in the filament kept clogging the hotend). I also ran into some serious skewing of the print on my Printrbot Simple, partially owing to voltage/acceleration issues, but primarily due to the fact that the printer, assembled as intended, has an unsupported Y axis which tilts inside the bearings about a degree and a half. I found a part on thingiverse that corrects for this, but not until I had printed all of the parts for the printer, and found that they were all faulty, especially the hotend tri and the carriages. I have since re-printed the parts, which was a little more than frustrating, but the quality is a lot better now, and I can't see any skewing in the new parts with the adjustments I made to the printer, and the correction parts attached to fix the Y-sag.
The hub tri that holds the hotend was also a problematic part. The STL that is on thingiverse is just over 100mm no matter how I rotated it, so it would not print on my Simple. I had to pull it into tinkercad and split it into multiple pieces using holes, and then add on fastening points to bolt the whole thing back together. Once I've tested and confirmed this new design works well, I'll likely post it on thingiverse for others with a small print-bed that would like to build a delta-pi.
FASTENERS:
I purchased the collection of screws from trimcraftaviationrc.com, can't recommend them highly enough. They contacted me immediately after I placed the order to let me know one size was out of stock, and allowed me to substitute a different cap.
MISC:
I scrounged up the remaining parts (pulleys etc) from various small shops online. I think I had the most trouble finding the belts.
While all of the parts were printing I went ahead and manually drafted the drilling and cutting points for the top and bottom board. If I had it to do again I would prefer to have the extra cash to get laser-cut acrylic or aluminum pre-built instead.
Once I had them drafted I cut/sanded/stained/varnished them to a nice dark mahogany. Right now I have only treated the top and bottom board, but I have more wood that I intend to use to build support walls and doors for the printer. Ideally I would like to bolt the power source (a PC power supply) up under the main body of the printer with a wooden face plate over it to provide a main power switch for the whole machine, so I am going to have the support walls extend about half a foot under it and put rubber feet on it to reduce vibration transfer to whatever I place it on.
First things first of course, I want to make sure the damn thing works
.I just finished printing the last part of the new hub-tri so I will be able to hook up the belts and run some tests probably tomorrow.
Pics as soon as I take them.
|
Re: Building my Delta-Pi - Journal/Q&A/lessons learned December 11, 2013 07:30AM |
Great job! Looking forward to see these pictures 
Have you got your jhead for the official store or from a different website?
Have you got your jhead for the official store or from a different website?
|
Re: Building my Delta-Pi - Journal/Q&A/lessons learned December 12, 2013 01:49PM |
See pics attached Sorry for the bad quality, only had my camera phone handy. Will take better pics soon.
Currently trying to get the aluminum square pipe smoothed out, as the company I ordered it from thought it would be a great idea to plaster a very sticky bar-code sticker on each rod, which is using some kind of glue which refuses to be removed.
Sorry for the bad quality, only had my camera phone handy. Will take better pics soon.Currently trying to get the aluminum square pipe smoothed out, as the company I ordered it from thought it would be a great idea to plaster a very sticky bar-code sticker on each rod, which is using some kind of glue which refuses to be removed.
|
Re: Building my Delta-Pi - Journal/Q&A/lessons learned December 12, 2013 11:46PM |
I was able to start a dry run print for the first time today. Still a lot of calibration to do, and I am noticing that the motors are having a bit of trouble providing sufficient force. I think if I can oil some of the hinges and the aluminum pipe that will help a lot.
The endstops had me mystified for a bit. Apparently the right settings are to have the pullup active, the min endstops deactivated, the max endstops activated and not reversed, and the pins.h had to be updated to swap the pins for the min and max endstops.
Also the home direction needs to be 1 rather than -1.
Thinking I may want to add active cooling to the motors as well, as they are getting rather hot when operating only a short while.
Still.. it moves!!
The endstops had me mystified for a bit. Apparently the right settings are to have the pullup active, the min endstops deactivated, the max endstops activated and not reversed, and the pins.h had to be updated to swap the pins for the min and max endstops.
Also the home direction needs to be 1 rather than -1.
Thinking I may want to add active cooling to the motors as well, as they are getting rather hot when operating only a short while.
Still.. it moves!!
|
Re: Building my Delta-Pi - Journal/Q&A/lessons learned January 13, 2014 03:03PM |
OK, so, the problem is becoming more clear.
I used a multimeter to confirm the winding set I was using and found that my 6 to 4 wire mapping was wrong. I was following the assumption that the 6 wire pins were in sets of three EG:
AAABBB
The truth of the matter however is that the pins mapped more like this:
BAAABB
With those wires mapped properly it greatly reduced the vibrations I was getting from the stepper motors (YAY!) I also found an issue with one of the belts needing to be tensioned, and some flaws in the print of one carriage that were scraping against the belt (Dremmel to the rescue!)
Even with all of this done, however, I am still seeing the following problem:
As I run a dry run print, over time, the hotend actually lowers rather than raising with each layer. By layer 3 the printhead has dropped about a half millimeter (almost a perfect inversion of what should be happening)
I have tested homing and moving to XYZ000 with different speeds and I am not seeing any difference in where it lands (tried values from 1000 to 5000) so I don't think it is slippage.
The good news is, except for this problem, everything is working
I used a multimeter to confirm the winding set I was using and found that my 6 to 4 wire mapping was wrong. I was following the assumption that the 6 wire pins were in sets of three EG:
AAABBB
The truth of the matter however is that the pins mapped more like this:
BAAABB
With those wires mapped properly it greatly reduced the vibrations I was getting from the stepper motors (YAY!) I also found an issue with one of the belts needing to be tensioned, and some flaws in the print of one carriage that were scraping against the belt (Dremmel to the rescue!)
Even with all of this done, however, I am still seeing the following problem:
As I run a dry run print, over time, the hotend actually lowers rather than raising with each layer. By layer 3 the printhead has dropped about a half millimeter (almost a perfect inversion of what should be happening)
I have tested homing and moving to XYZ000 with different speeds and I am not seeing any difference in where it lands (tried values from 1000 to 5000) so I don't think it is slippage.
The good news is, except for this problem, everything is working
Sorry, only registered users may post in this forum.