Friday, March 21, 2014

3D DLP Printer


I finally got around to taking some pictures of my 3D DLP Printer build.  I would like to give many thanks to Tristram Budel for his Instructables article detailing a DLP printer build.  This article was used as a great starting point for me.  It is full of information, and is a recommended read (http://www.instructables.com/id/DIY-high-resolution-3D-DLP-printer-3D-printer/)





These two photos are the overall of the build with the panels removed.  The panels are needed so the exhaust fan and pull the resin vapors out of the house without letting any escape.  The Bucktown polymer's resin I am currently using (which is VERY high strength once cured) has flammable vapors. I'm working on a carbon air filter with a stronger blower fan for something that doesn't need to be piped to the outside.

Some Videos: (the fluorescent lights do not cure the resin, I can keep them on all day long without issue.)
video
Here is a side view of the mirror and projector, you can also see the stepper motor the controls the Y axis slide.  This has a home limit switch in addition to the Z axis limit switches.

video

The Y-axis limit switch.
video
Z-axis limit switch. I home it, and then drop the build plate to the VAT surface whenever I switch VATs.
video
Demonstration of a rapid Z axis move from home.  The borosilicate glass is VERY strong. The suction force of the whole build plate is overcome with the NEMA 23 motor and Z stage. video

Z-axis homing operation
video

Z moves... video



I used 1" T-slot aluminum to create a "box".  I have the Y-axis vat slide on two linear rods with bearings.  There is a single NEMA 17 stepper that pulls the vat parallel to the Y-axis.  While the Z is lifting, this Y axis motion helps counter any suction forces encountered.  I have printed relatively large prints, and this was very successful.  No tilting or twisting needed.

The Z-axis is a linear ball screw stage I got off e-bay.  It has 14" of travel.  There is a NEMA 23 stepper attached to this axis. The whole thing is controlled by an Arduino Mega paired with a RAMPS 1.4 board with Pololu stepper drivers.  An ATX supply provides the power.




 The build plate is a simple cut piece of aluminum.  104mm x 204mm.  I sanded with a 120 grit sand paper and cleaned it.  The prints stick to this very well.

 The 5/8" linear rods and bearings support the vat and Y-stage.  I used some more T-slot aluminum and REPRAP printed brackets to secure it all together.

 Here is the side view.  My 1080p DLP projector (Acer H6510BD DLP) and a single surface mirror I got off ebay.  No modifications were made to the projector, I didn't remove the color wheel or anything.  I cut some acrylic to make a 45 degree angle and epoxied an adjustable mount.  I can adjust the distance to I can achieve higher resolution prints that are smaller, or larger lower resolution prints for big objects.  Still playing with the cure times to get the build times lower.

 Here is the back with the NEMA 17 stepper for the Y axis.  It has a simple threaded rod.  There is no need for high accuracy on the Y axis.  It's simply used to help detach the printed layer with a lateral force.  I have found 5mm of deflection for the complete Z axis cycle is enough.
 Here is one of my VATs that needs to be recoated.  I used a 300mm x 300mm borosilicate glass (https://www.lulzbot.com/products/borosilicate-glass-bed-300mm-x300mm) with aluminum for the walls.  Black silicone to stick it all together, and Sylgard 184 for the coating.  The borosilicate glass is "VERY" strong.  I doesn't deflect at all with a fully exposed 104mmx200mm print.  The Z axis speed has to be reduced, but they seperate with 0 missed steps...
Here is the peristaltic pump I got off e-bay to get the resin out of the VAT.  It works great, just learned the hard way that you need to flush it with isopropyl alcohol after each use....

Here are some prints I did of a compound bow stabilizer and sight mount I designed for the Hoyt Carbon Matrix:  The stabilizer was a large print, I had to make sure the Z axis was almost perfectly perpendicular to the VAT.










This is an AR-15 Quad Rail.  This was an extreme test of the full 14" travel.  The only issue was the "bleed through" of layers, I need to add more graphite powder to the resin to prevent this. 



Current settings:

First 5 layers - 13 seconds
Each 100micron layer - 6.25 seconds

I use Creation Workshop for the prints with the code modified to move the Y axis instead of X for the "tilt" setting.http://www.envisionlabs.net/

Source Branch:
https://github.com/Pacmanfan/UVDLPSlicerController

Some of the STL files I made for this build:

This is the build plate mount. It secures to 1" Aluminum T-slot.  I used epoxy to secure to the aluminum plate.  
https://drive.google.com/file/d/0B8S25d-5M3YyWTlmVWd4RWlKTUk/edit?usp=sharing

This is the retainer that secures the vat to the Y-axis slide.  It's simple but very effective.
https://drive.google.com/file/d/0B8S25d-5M3YyM0FhSGMxb3VRYUk/edit?usp=sharing

Y-carriage slide stem:
This takes a simple hex nut and secures to the T-slot aluminum to control the Y-axis slide.
https://drive.google.com/file/d/0B8S25d-5M3YyenlRVDBwVVpKbHc/edit?usp=sharing

Z-arm mount - A reinforced right angle mount for the Z-arm. Not currently being used in the pictures, but does work well when 4 are printed.
https://drive.google.com/file/d/0B8S25d-5M3YyZ21lOUZGMWdqQk0/edit?usp=sharing


ADDITIONAL FINDINGS SUMMARIZED FROM COMMENTS/EMAILS and GENERAL QUESTIONS I'VE HAD SINCE HACKADAY POSTED A LINK TO MY BLOG:

** Z-stage - "THOMSON LINEAR MOTION 2RBM120DMHL QUICKSLIDE 14-3/8" TRAVEL" The pitch is large, which surprised me that 100 micron layers are no problem for it. The ballscrew has no measurable backlash, atleast not with my budget calipers. I've tried down to 50 micron without issue. I would need to do some more experiments to see what it's capable of. I'm using 1/16 microstepping on the stepper drivers with a big NEMA 23 stepper on this axis.

When I get some time I'll try some experiments to see what layer thickness I'm able to get reliably. The 1080p projector is a dream for X, Y.

** I chose this projector simply because it was tried and tested with the instructable article by Tristram. The amount of UV that comes out of it without a single modification is wonderful.

** Bucktown polymer ZVE500-V420  UV resin:
   12 second initial layers (5 layers)
   5.25 seconds per 100micron layer
  Some prints with overhang geometries need some graphite powder added to prevent bleedthough from layer to layer. 

** Makerjuice SubG+ resins.
   First 5 layers - 13 seconds
   Each 100micron layer - 6.25 seconds

More updates to come. If you would like to donate to my projects fund, my paypal is: zothar@zothar.com  All donations will contribute to open source efforts and documentation to make others lives easier.  It's a pain trailblazing a new or dirty path, especially with limited funds...

33 comments:

  1. Awesome work! How much did you spent to make this printer?

    ReplyDelete
    Replies
    1. Thanks. The z axis stage was the most expensive at about 180 off ebay, other than the projector. Probably close to 1800-2000 if someone were to keep an eye on ebay for good deals.

      Delete
    2. You choose this projetor for some reason? Because what i saw is that it can be the most expensive part of the printer.

      Delete
  2. Could you post a link to the auction for the Z-Axis stage? It would be helpful to see so I could look out for something similar..

    Do you have any specs for that?
    From the photos the pitch seems very large, what is the min layer thickness you can get?

    Awesome project!

    Cheers!

    ReplyDelete
  3. The Z-stage link is no longer there, but my ebay archive has this as the description - "THOMSON LINEAR MOTION 2RBM120DMHL QUICKSLIDE 14-3/8" TRAVEL" The pitch is large, which surprised me that 100 micron layers are no problem for it. The ballscrew has no measurable backlash, atleast not with my budget calipers. I've tried down to 50 micron without issue. I would need to do some more experiments to see what it's capable of. I'm using 1/16 microstepping on the stepper drivers with a big NEMA 23 stepper on this axis.

    When I get some time I'll try some experiments to see what layer thickness I'm able to get reliably. The 1080p projector is a dream for X, Y.

    Miguel - I chose this projector simply because it was tried and tested with the instructable article by Tristram. The amount of UV that comes out of it without a single modification is wonderful. It works excellent with both Bucktown polymer ZVE500-V420 and Makerjuice SubG+ resins. The bucktown has a slightly smaller cure time per layer (500mSec), but seems to need more graphite powder to prevent bleedthrough.

    ReplyDelete
  4. Hi, Dan Beaven

    how do you control projector with ramps1.4. I mean open- closed projector in each layers. what about pin of ramps for connect with rs232 pin in projector. can you show me the pic of ramps boards.
    regards
    sontaya (spansupa@hotmail.com)

    ReplyDelete
    Replies
    1. Luckily, creation workshop controls the Projector blanking through software. The RAMPS 1.4 is only used for the Z and Y axis movement. However, you can use the FAN output on the RAMPS with a TTL level converter and custom M code (don't know off hand which one) to turn the FAN on and FAN off to control projector blanking. What I did was set my windows background to Black and use Creation Workshop to handle it. Hope this helps.

      Delete
    2. thank you very much Dan. sorry i don't come to see comment long time

      Delete
  5. Hi! I came to your 3D printer page while I was searching for DIY DLP printer. Your work is awesome! I'm planning to build my own soon.
    If you don't mind could you please tell me why you chose linear ball screw+THOMSON LINEAR MOTION not to use trapezium spindle from Damencnc.com as discribed at Instractable.com by TristramBudel? does linear ball screw+Thomson more reliable than Trapezium spindle?

    Also why did you use additional aluminum plate between the build plate and purple plastic mount?
    and I'm very curious that you use different type of two NEMA.Both of them seem like same stepper axis and similar torque...

    ReplyDelete
  6. The THOMPSON linear motion setup was used because I got a good deal on ebay. :) I didn't choose it specifically, but it turned out to work better then I could have expected. This linear ball screw has no significant backlash. I'm able to easily get 50 micron layers with it.

    The NEMA 23 stepper motor on the Z axis is 425 oz-in. The Z axis is capable of exerting a lot of force to help overcome suction. Probably too much force. This is definitely over specc'd for the job. The NEMA 17 motor on the Y axis is for the sliding motion to help release the print on each layer. It's from my extra reprap parts. I had two of these for the Y axis, but this wasn't needed. I move 3mm per +6mm Z axis lift and -5.9mm cycle. You definately need to find the sweet spot on feed rates for both axis.

    The extra build plate is a result of my experimentation with a smaller surface area. This plate had a textured surface and I was hoping the resin would stick better. Turned out it didn't make a difference. So I epoxied a larger build plate on so that I could make larger prints.

    ReplyDelete
    Replies
    1. Thank you for your reply! It is great advice and conclusion. I will try soon.

      Delete
  7. Hi, Dan Beaven

    As I follow your advice I've got NEMA23 and some equipments:)
    I'm working for my DLP printer design now and one of question coming up.
    May I ask you why did you choose sliding vat system instead of tilt one?
    I'm concerning tilt system can reduce width of printer also vat's cubic volume rather than sliding one.

    Regards,

    ReplyDelete
    Replies
    1. I decided on the sliding vat approach because of a research paper I read (http://www-bcf.usc.edu/~yongchen/Research/Digital_Material_2012.pdf) It demonstrated the lower forces needed for layer separation with a sliding approach.

      Many people have had a lot of success with a tilt mechanism. With the 14 inches of travel on mine, I wanted to minimize the separation forces even further and gave it a try. I believe both approaches are effective. For my large surface area prints, I believe the sliding approach is marginally better.

      Thanks,
      Dan

      Delete
    2. Thank you for your quick response and sharing a nice reference.
      I will check it out.

      Regards,
      Masa

      Delete
  8. This comment has been removed by a blog administrator.

    ReplyDelete
    Replies
    1. Thanks for the compliment Ann. I would like to discuss this more and answer any questions you may have. However the chinese to english translator did not do a good job. Please try a different translator and i would be happy to answer any questions.

      Delete
  9. Very cool project! I'm also looking into DLP/resin-based printing. Have been using a ABS/PLA filament printer for a while, but I think its too much calibration, and poor quality vs resin-based printing. Is the printer easy in use when when you have built it? And how much time did you spend assembling the printer? I see Brudel used about 1000-1500 hours, but I'm guessing that is with all the research he have done.

    Cheers!

    ReplyDelete
    Replies
    1. Thanks! Thankfully, there is was alot of information online to guide my design and implementation. It took about 2 weeks to get it functional. Maybe about 60 hours. I reused alot of parts from my 2 cu ft PLA printer, so I didnt have to order much. Once it's built and calibrated it's real easy to keep using. The resin can be a bit messy and sometimes the VAT needs to be recoated (I made 2 so that I wont have any down time).

      Delete
  10. This comment has been removed by a blog administrator.

    ReplyDelete
  11. Thanks for your write up, I've found it very helpful.
    I'm currently designing a similar printer up in Solidworks CAD, are you able to answer my question
    Using the same projector as yours (Acer H6510BD) and wanting a build area of 250mm X 250mm, what is the minimum and maximum focus length?
    I.e. distance from the lense to the plate of glass

    ReplyDelete
    Replies
    1. The build area will be a 16:9 ratio I believe, to get the full 1920x1080 resolution. I'll assume you want to do 250mm x 140.625mm and take a few measurements for you tonight. You'll have to post some pics of your build, I'm always interested in looking at 3D Printer builds :)

      Delete
    2. That will be great if you can come up with those minimum and maximum focus values.
      I'll post some images of the CAD designs once its finished, I've gone for a "Keep It Simple Stupid" approach of making it taller and have the projector facing straight up to the bottom of the build area.

      Delete
  12. I’m eager to find the valuable information and for me this is the right place to get the good stuff.Informative industry news

    ReplyDelete
  13. Hey Dan

    Did you managed to measure the minimum and maximum focus distances for 250mm x 140.625mm build area?

    ReplyDelete
    Replies
    1. So sorry it's taken me so long. I've had several things come up that have distracted me. I'm putting a reminder in to get this info for you tonight.

      Delete
    2. 295mm is the closest I could get it to focus at 250mm x 140mm

      Delete
  14. Thanks meant for sharing this type of satisfying opinion, written piece is fastidious, that’s why I’ve read it completely.Click here

    ReplyDelete
  15. Thanks for finding out those measurements.
    Did you use the ramp v1.4 kit for the control side of things?

    ReplyDelete
    Replies
    1. I found a great combo deal on ebay for the Arduino Mega2560 with Ramps 1.4, pololu drivers, and LCD Screen. Seeing as how the Z axis is the most important, the screen was an extra and the extra drivers I have set aside as spares.

      Delete
    2. Thats handy, as I have a mega and HY-DIV268N-5A drivers spare, with the ramps 1.4 board, what scratch did you use for the mega?
      And can you expand on where you changed the Y-Axis tilt function to 5mm travel?

      Cheers
      Dan

      Delete
    3. I used the Erik Zalm Marlin firmware: https://github.com/ErikZalm/Marlin The Y-axis tilt function travel was specified in Creation Workshop. Depending on which version you have I believe it's located in different areas. Which version are you using?

      Delete
    4. Im using version 1.0.0.75, I see there is a "Slide / Tilt Value" under the "Lift and Sequence" in the configure slicing profile.

      How do you turn this value into Y-Axis movement of 5mm? in the Marlin sketch?

      Also what does a ramps shield do?

      Delete
    5. The RAMPS shield is the interface between the Arduino MEGA microprocessor and the stepper motors / limit switches / and temp probes etc... For this we only use it for Limit Switches, Z axis and optionally X or Y. The Arduino is the middle man between the steppers and computer that receives the Gcode commands generated by Creation Workshop and translates them to movement.
      The Slide / Tilt value I believe defaults to use the X axis on that version. Even if it is setup for the Y, you should probably hook it to the X axis on the RAMPS board. The Slide Tilt value you enter is the number of mm to move. 5 mm i've found to be good value for my setup.

      Delete