#sla-printer

[makerspacelt]

Vaidas shares his experience on his latest creation - SLA 3D printer. Here goes the details so dig in.. ;)

Finally, a few months ago I finished another SLA printer (LCD based). And it works great, prints really well with way faster cure times than I thought. Now I think I can share my experience.

But to begin with, my goal was not to make the smallest and the fanciest device, but it was supposed to be quite adjustable, easy to work with and upgradeable if necessary (or if something goes wrong).

I started to with frame itself. I wanted to make it out of ordinary shelf components so I chose standard aluminum profiles (40mm x 20mm) (image: Profiles). I went for “box” design with two sections: bottom part for electronics and upper part for actual printing chamber. And now, after finally completing this printer, I must say that this design is really comfortable to work with: you can access all segments of the printer from all sides, you can do that very quickly with no stress. You do not have to disassemble the whole printer, when you want to adjust something that is deep within.

Profiles

Aluminum profiles were powder coated using self-made powder coating system (images: Powder Coating, Painted profiles, Powder Machine). It uses old CRT monitor transformer (50k V), paint gun is also 3D printed and uses “Nestea” bottles to feed powder. I was amazed how expensive are powder coating paint guns… and it took as one evening to make one for us.

Powder Coating

Painted profiles

Powder Machine

Profiles were joined together using my own design 3D printed “corners” (images: Frame Assembly, Frame Assembly_2). Each profile was also fixed with M4 screws to each corner. After joining profiles together, I had fully functional frame. See images.

Frame Assembly

Frame Assembly_2

Then it was time for Z axis (images: Carriage Head, Carriage Plywood, Carriage Top). I am a fan of thick linear rails, so I used here 16mm calibrated steel rods. I like thick rods since they compensate well any wobbles that come from usually curved thread. Stainless steel M8 thread was used. Head itself was redesigned “Cristelia” version. I did not like to use only two bearings for carriage, so I made my own carriage that uses 4 linear bearings since that brings way more stability into the system. Build plate and the rest of the head is “Cristelia” design. It is quite good although a bit bulky, but it works well. All in all, Z axis itself after calibrating it, was off only by ~10um within 10cm, i.e. after moving axis from 0 to 10cm, at the top it was off only by ~10um. Since everything was made by hand – that is pretty good. Disclaimer: my hands are not very well “calibrated” I was simply lucky this time :)

Bottom part and printing chamber was separated by hand-cut plywood plate (image: Carriage Plywood). I have chosen wood, since it is less expensive and easier to work with. I coated it with lacquer and it is just fine. At least when you need additional hole, you can make one easily.

Carriage Head

Carriage Plywood

Carriage Top

Now electronics (images: Inside, Dashboard)... First of all, my goal was to design electronics to handle 100-130W LED (I did testing at the end and it handled that power well). I use two PC power supplies (since I have a bunch of those): one for LED and one for remaining electronics (RPI3, Arduino, power for fans, steppers etc.). LED is powered using 300W step-up boost converter. Since I installed powerful 1 ohm resistor, I can measure current usage of LED circuit with voltmeter and with another regular voltmeter (measuring LED voltage drop) I end up having a nice dashboard, which helps me to see actually emitted power by LED and I need that, since I change LED power for my own reasons quite often. Another dashboard screen shows voltages of RPI3 and Arduino (helps to debug any issues). Everything is cooled using a bunch of fans from PCs.

Inside

Dashboard

Now, the LED part… It was (as expected) the most complicated one, where I spent most of my time. I also thought of using an array of LEDs, but after trying a lot of alternatives I went for a single LED. Moreover, it was way easier to work with single LED than with array. I was just not able obtain better results with an array than I did with a single LED. I will add image of light uniformity with single LED (image: Light Distribution). Although I designed everything for >100W LED, I use 50W LED (400-410nm). Since I still think of making daylight device one day, I think this power reserve might be useful later.

Light Distribution

LED assembly (images: Collimator, Collimator Top) consists of a large piece of aluminum plate (image: Heatsink For LED) with additional heatsinks attached at the bottom. PC fan is attached at the bottom as well. LED is screwed on the plate. Then there is first part of the whole assembly (everything is 3D printed, ABS) which holds Fresnel lens on top (see image: Collimator). Then there is second part – mid part, which had a purpose to be short and easily changeable if I needed to adjust distance between LED and LCD, so bear with me, but I did take a lot of precautions. It was easier to print another short part than entire assembly. Last part is top part, which is attached to a plate of black PMMA plate with a rectangular cut, which holds LCD display. There are a few fans to get cooler air inside, I am not really sure now that they are making some sort of impact, so you can say that they are useless, but… oh well: they live there and it seems they are happy. Interior of this collimator is covered with aluminum foil. Foil itself also did not make any serious changes/improvements since Fresnel itself collimated light, but, I would say, it was a little bit better. I did not observe some sort of light unevenness due to the foil (image: Light Distribution). LCD itself was placed into the rectangular cut of black PMMA plate – thus closing system. LCD is the same as used for YHD-101 (KLD). With this setup, I did not notice any heat issues at all. LCD is warm, but that is all.

Collimator

Collimator Top

Heatsink For LED

This LED assembly was fixed to plywood plate that separates bottom part from printing chamber (image: Inner Chamber). I used springs to fix that in place and springs also allow me some sort of adjustments. Another feature is that it helps me to avoid crashes if Z axis accidentally goes to low – springs will compensate that.

Inner Chamber

Frame was covered with milled orange PMMA, which is very easy to remove, when you need to access some parts of the printer. Doors are made out of piano hinge (image: Hinges) and door lock (image: Door Lock) is made using simple magnets and 3D printed parts. At the bottom of the frame, there are adjustable legs (image: Legs), which you can screw up and down to adjust the level of the printer.

Hinges

Door Lock

Legs

VAT I made was a mix of various designs, since I made a lot of them in recent years, this was just a rough assembly which appears to be working really well (image: VAT)

VAT

I have attached image of printed calibrations parts  (images: First Print Calibration, Motor) 

First Print Calibration

Motor 

And at the end, I must say, that everything turned out quite OK and I cannot say that I would do something in a completely different way if I had to start over again. It works OK, maybe it is possible to play again with array version or to simplify collimator assembly, but all in all it does the job.

[makerspacelt]

OpenSLA

..and again Vaidas has a story about 3D printer ;)

A few months ago I finished another SLA printer. This one is pure DLP printer based on Acer H6518BD beam projector. It works great, prints extremely well and has one quite unique feature that saves a lot of my time and adds more versatility. Therefore, I think I will share my experience on how this build was made.

But to begin with, I would like to start with initial goal of this machine and what was the main purpose of it. I used to have one DLP printer before, which was based on Acer H6510BD beam projector. That printer was made entirely from scrap and lacked versatility, was quite difficult to use and to adjust, especially when it comes to XY resolution. Each XY resolution change was very time consuming process, prone to errors and required extensive calibration procedures in order to meet dimensional requirements. Firstly, I tried to look for DLP printers that had a feature of automatic adjustment of XY resolution. However, I failed to find a reasonable one. Therefore, I thought that my goal with this one will be to make DLP printer that would have an ability to adjust its XY resolution automatically via software as needed. It was also important to support small and larger VATs for various XY resolutions.

I started to with the frame itself. It was made out of ordinary components, so I chose standard aluminum profiles (40mm x 20mm). I went for “box” design with two sections: bottom part for electronics and upper part for actual printing chamber. This design as a platform was also used for my previous LCD printer that was described several months ago. This design works very well, allows easy access to all parts of the printer and maintenance is also extremely easy.

Profiles were powder coated and were joined together using my own design 3D printed “corners”. Each profile was also fixed with M4 screws to each corner. After joining all parts together, I finally had fully functional frame.

Then it was time for Z axis. I like thick linear rails, so I used here 16mm calibrated steel rods since they compensate any wobbles that come from Z axis due to usually curved threaded rod. Acme M8 thread was used with pitch and lead of 2mm.

Head itself was redesigned “Cristelia” version. I did not like to use only two bearings for carriage, so I made my own carriage that uses 4 linear bearings since that brings way more stability into the system. Acme M8 thread was used with pitch and lead of 2mm. Build plate and the rest of the head is “Cristelia” design, but I had to make additional modifications to support sliding build plate. This allows me to use small VAT with multiple resolutions, because when projector is moved up and down, center of image also moves accordingly. Although a bit bulky, it works quite well for my needs.

Bottom part and printing chamber was separated by milled plywood plate. I have chosen wood, since it is less expensive and easier to work with. I coated it with lacquer and it is just fine. At least when you need additional hole, you can make one easily.

When it comes to electronics, I use single laptop power supply: step-down converters were used to produce 5V for RPI3, Arduino and actual 12V were used for fans. Everything is cooled using two high-power fans: one sucks air out and sits next to projector and second one pushes cool air from the outside.

XY adjustment (i.e. build area adjustment) system was the most intriguing one. In order to adjust XY resolution, projector must be able to move up and down in a controlled way. 

The easiest way was to exploit similar motion solution as it is for Z axis: threaded rod with Nema stepper motor. 

For this purpose I designed and 3D printed special plate (in white plastic) that was attached to projector and also housed M8 stainless steel nut for threaded rod. 

Since extreme accuracy was not mandatory here, I used drawer rails to move projector up and down. For my own surprise, those worked really well. It is also worth mentioning, that projector itself will not be moved extremely often thus this movement system does not have to be over-engineered and simply has to do its job when needed. At the bottom position, end-stop switch was installed. This marked the lowest resolution I was able to obtain and in this case it was ~75um XY. The highest resolution that was supported is 37um (which is the highest one that this projector actually supports due to limitations of optical system). For those, who do not know, adjustment of XY resolution also adjusts maximum build area available. So adjustment of XY resolution also provides a feature to have larger or smaller build area when needed.

Frame was covered with milled orange and clear PMMA, which is very easy to remove, when you need to access some parts of the printer. Doors are made out of piano hinge and door lock is made using magnets and 3D printed parts. At the bottom of the frame - adjustable legs.

I have also made an adjustable VAT system. VAT itself is mounted on aluminum profiles that work like rails. This allows me to adjust position of the VAT based on XY resolution since image of projector moves up and down, when resolution is adjusted.

What is important to say, that this build plate and VAT adjustment is only necessary, when small VAT is used. When larger VAT is used (that supports ~75um XY resolution), no adjustments are required and XY resolution can be changed in seconds, because the system is designed in such way that bottom line of the projector image always stays in the same position and only upper part of the image moves (that’s how optics of this projector actually work). Thus if you use such XY adjustment design on a daily basis and you are OK with larger VAT – simple and non-adjustable build plate and VAT is completely acceptable and that would save you a lot of engineering work.

Software I use, is nanoDLP with raspberry PI 3. During the first use, I decided that ~37um, ~50um and ~75um XY resolutions are OK for me (later I added one more as seen in picture), thus I have empirically figured out correct projector positions in relation to end-stop switch position. Then each position was selected, adjusted and dimensional calibration of printed parts was executed (this had to be done only once). After all of this, I had precise projector distances from end-stop to corresponding and precise XY resolutions. NanoDLP allows custom buttons at Z axis page, therefore I just added three of those with fixed G code: homing and going to a specific location for appropriate XY resolution. So the only time I had to make some sort of calibration was only at the beginning, when I built my machine. It’s been 4 months already and it still works perfectly. The only negligible issue was that when “Floor” button was clicked, firmware homed ALL axes (thus, Z AND projector), therefore I had to make separate button to home Z axis only.

I have attached some images of printed models. It prints really well and this DLP is perhaps better than my LCD printer. The only downside of such DLP printers with beam projectors is uneven UV light output across entire build area and optical aberrations, which is possible to correct with masks and other tools if needed. Apart from that – works great and DLP projector is awesome heater for our laboratory during cold winter days!

And at the end, I must say, that everything turned out OK and I cannot say that I would do something in a completely different way if I had to start over again. If some of you decided to make something similar, do not hesitate to contact me if you need any help.

Materials used for printing models in pictures are AmeraLabs modelling 3D printing resins. You can get those at our website www.ameralabs.com