Introduction: SLAM - SLA Machine, a Custom Built SLA LCD Printer
So after a lot of research I decided to pull the trigger on building my own DIY SLA LCD based printer.
I will share a couple of posts with you to illustrate how I got to where I am now. The build is not complete yet but I am in the last steps before my first print.
First I want to thank all who have taken their time to answer my questions or in other ways contributed to my progress in this project.
My first thought was to buy a Wanhao D7 but I wanted a bigger printer. I don’t like compromises and so I decided to build my third 3D printer.Previously I have built two FDM printers which are modified kit printers, one Delta and one cartesian.
My main inspiration was the videos from Ionel Ciobanuc on Youtube and Ion Gurguta Thank you! I decided to use a LCD UV-LEDs and NanoDLP with DirectDrive. What an effort by Nano DLP devs! The last is not yet set in stone and I have both a RAMPs and DirectDrive ready to implement. I also want to send some cred to Thinkering On Steroids and the entire Google Plus OpenSLA group for great inspiration.
Now to my project!:
Step 1: LCD Hookup With PC and RPi / Nano DLP
First I Ordered the LCD:
www.aliexpress.com : 13.3 Inch 3840*2160 4K 100% NTSC New Original UHD IPS Display DisplayPort DP Driver Board LCD Module Screen Monitor Laptop HDMI
This screen is a spare part for a Lenovo Yoga.
When I received the LCD I tested it first with my computer and it was fine out of the box. Everything working as expected. I then hooked it up to my RPI which had Raspbian Jessie preinstalled. Now this took some tinkering and tweaking to get going (3 hours, google and a candybar ). At first i saw only fractions of the screen or it worked partially in portrait mode. At last i found the right config for the screen. After the hallelujah moment of seeing the console in full resolution I fired up the install of NanoDLP. No issues after following some guides and also setting the parameters for the display!
Here are the parameters i used:
gpu_mem_1024=256 framebuffer_width=3840 framebuffer_height=2160 max_framebuffer_width=3840 max_framebuffer_height=2160 hdmi_ignore_edid=0xa5000080 hdmi_drive=2 hdmi_pixel_freq_limit=400000000 hdmi_group=2 hdmi_mode=87 hdmi_timings=3840 1 48 32 80 2160 1 3 5 54 0 0 0 25 0 220430000 3 ## The Screen is a replacement screen for Lenovo Yoga 720-13 ## Productnumber: B133ZAN01.1 ## Resolution: 3840X2160 ## Connector EDP40 pin
Step 2: Prepare to Separate the Backlight From the LCD
Why do i strip the backlight you might ask?
Since i want to use UV curing resins i need UV Light instead of the regular backlight which comes with the LCD Panel. It is possible to use the backlight as is with daylight curing resins but this is normally a slow process and it would limit the number of resins i can use.
Why do people strip the polarizer away from the LCD?
Normally this is not needed but some LCD screens are assembled with polarizers which block UV light. This is a showstopper if you want to use UV for curing the resin. Therefore you can strip the polarizer away from the LCD and replace it with a more suitable polarizer which has better properties for this purpose.
At a later time i intend to use two polarizers. When the polarizer is hit with light it blocks 50% of the light which translates into heat. In order to protect the LCD I want to remove the heatload from the LCD panel. This is the reason for the first polarizer.
The second polarizer will be mounted directly in contact with the screen oriented in the same direction as the first polarizero enhance contrast as much as possible. Maybe this second polarizer is not needed but as it is already glued to the LCD I have ordered i will try keeping it.
I am now ready to strip the LCD backlight!
This was a process I spent some time with but using a magnifying glass it is not to hard to find all hooks behind the tape around the border. Notice that there are hooks also behind the metal tape.
To make your life easier you can slightly bend the aluminum back outward where the hoocks are.
Step 3: Finalize Backlight Separation
After removin as much tape as i could and freed all hooks I was able to successfully remove the backlight. You can now also inspect all layers of filters which come loose. I also scarified a corner of the LCD to explore the polarizer film and other films attached to the display. In the end I kept everything that was glued and removed everything that was loose or came loose with the backlight. Now I wanted to test the LCD and it fired up perfectly and using a flashlight I could see the NanoDLP test patternsthroug the LCD.
Step 4: Lightsource UV LED
Next step was a light source. UV LEDs have been used successfully so I ordered a 100W UV LED405nm with driver and lens and also some 3W UV LEDs 395nm, 405nm from Ali Express and also some quality ones from some local vendors, some resistors and 15 and 30 deg lenses for the 3W LEDs also fron AliExpress.
[ Before proceding... USE CORRECT PROTECTIVE EYEWEAR with strong lightsources!!! ]
The 100W led was useless if you want a uniform light for a 13.3 inch screen with a reasonable optic setup (lens fov / distance to LCD also the lenses i got where crap which further pushed me towards the smaller leds. The 3W LEDs proved to be a good choice and the 15 deg lenses where a good pick. The other lenses where not so good and produced a poor ununiform projection. I also tested the heat conductive adhesive (Arctic Alumina) i intend to use for attaching LED to some cooling and it preformed well with a very small stepper driver heatsink and a LED.
Worth noting is some optics theory. Distacnce between the bed and the LCD is less sensitive with a fresnel lens and a point lightsource at the fresnel focal length than with a even distribution of multiple lightsources. A point lightsource is the optimal for a projector which is what we actually would want here if we where only looking for contrast but the benefit in even light distribution possibillity for cooling etc makes the choice of even lightdistribution with multiple lightsources very apealing. The later option results in light spreading in all directions on the other side of the LCD which is why we ideally want zero distance between the lcd and print surface. a projector based solution is not sensitive to this other than setting correct focus.
Step 5: Designing the Frame
Given the above I had the info I needed to design a frame. All I knew is I wanted a rigid one so I started the design process with the following requirements: -Big aluminum plate for mounting LED:s -4040 Aluminum profiles -20mm Linear rails with 4 blocks -1604 Ballscrew -300x300x3mm glass surface -5mm aluminum sheet for cutout parts. After maximizing the number of LEDs I could fit to the surface underneath the LCD I counted 242 LEDs. Which just happens to match perfectly with a connection schema, 11 series 22 parallel. I intend to use a 40v PSU to power the LEDs and using the LED array Calc (‘http://led.linear1.org’) this is perfect if I use 40v to supply my 700mah LEDs which have 3.6v forward current with 1ohm 1W resistors on each series. The result is 10W resistor heat to 610W used by diodes for heat and light. So the machine will have plenty of Lumen :P
When it comes to cooling I don’t know how much cooling I will actually need since I don’t know the ratio between idle time and exposure time. I have three options: Water cooling which can be integrated into the plate, Passive cooling using up to 50 southbridge cooling fins or milling cooling fins into the platform. If airflow is needed I have planned for having the option to integrate 4pcs Noctua 160mm fans mounted to the bottom plate sucking air into the printer through a thin filter. Here is the design I came up with! Rigid is the keyword ? To be continiued…
Step 6: Mill the Led Panel
I made my panel fron 15mm 6061-T6 aluminum Stock sheet. This is a time consuming task as there are many details and it is a challence to get a nice part within a resonable time and cheap CAM software. Id recomment you to order it prefabricated if you dont want to spend unnessesary time.
After the part was done it was spraypainted with heat tolerant matte black spraypaint and sanded at the cooling contactpoints.
Step 7: Mount LEDs to LED Cooling Panel
To be able to mout the LEDs to form a precise grid you will need a template so go ahead and make that one to. It takes some extra time but it is worth the effort! Since the adhesive i used, Arctic cooling two component silver/ epoxy mix, hardens quite fast i recomment mounting two rows of 2x11 leds at a time. When this is done you need to solder the leds, another time consuming task which took me about 3 hrs to complete and another hour to fix bad solderpoints and add resistors and connecting the series.
The math behing this is done with an online Led array calculator using values of my Leds as a baseline and fiddeling with it until i all adds up and produces as little heat in the resistors as possible. i ended up with 22x11 Leds (242 of them...) connected to a 40v 1000W PSU loosing 10 W at the resistors total.
In the alst picture you can see some lenses added to the leds.
Step 8: Manufacture Mainplate, Top Plate, Bottom Plate and Cut Extrusions to Length
This really does not require any detailed explanation. All plates are made of 5mm Aluminum Sheet and the extrusions are 40x40 R90 Corner profile with 8mm slots. The corner profiles are easily cut if you use a blade made for aluminum and give it some cutting oil half way through.
Step 9: 3D Printed Parts
While these parts can be milled i found it easier to just print them... Eventuallt i might upgrade the printer in the future but for the sake of putting it together in a resonable time i fount it better to print these parts. First the ballnut housing and also the mount for the stepper motor driving the x-axis. I dont expect the stepper to get hot so pla will likely be fine but given the care taken on other parts it feels cheap to print the stepper mount but hey - It´s a work in progress...
Step 10: The VAT
While this can have many solutions i descided to go with a simple solution with no real fancy twists to it. This is the first vat I make and also the first one i see first hand so I can onlt relate to what is working for others. There are many 3Dprinted vats out there i dont believe they are up for the task of keeping tension and withstand the tearing forces when the printarea exceeds the standard 5.5". While many have sucess with bigger vats I decided to mill mine. To be able to do larger prints without refilling the vat i wanted a height of 50mm. This makes for a expensive piece of aluminium which mostly will be transformed into chips i started looking towards alternate manufacturing methods. I descided to try welding.....
The vat consists of a main frame, a fixation ring and a raiser which will build the height. The tensioning ring fixates the FEP to the main frame in a classical manner. Researching other vats it is common to use about 32-35mm spacing between the screws so i took the safe route and used 30mm. When the Fep is secured in the frame the raiser is inserted into the frame and desired tension is built by securing the raiser using its mounts.
Beginning the manufacturing i made the hard part first. Checking some videos on YT i descided to actually give welding a try. First i used a crap piece to determine meltingpoint for the stock and welding wire I was using and when I was comfortable I moved on to the real part. The raiser consists of one piece of sheet aluminium bent into shape using a rig and then the ends are welded together. To finish up I used the samding tools i had at hand. a belt sander would have been preferable but I had to settle for a oscillating sander and a dremel...