Peachy Printer ideas

Quote
koopa
Hey guys,

Like many of you, I was fascinated and enamoured with the $100 photo-setting resin printer made in a very Macgyver fashion on Kickstarter. In the spirit of collaborative work, I sent some ideas to the creator to further refine/innovate the design on low-budget, which he appreciated and told me to post here. Feel free to build/add/utilise anything here, chuck in an idea or two yourself, or let us know how to refine these ideas further:

1) Since you are already using the microphone input to sense the impedance change as drips go past, you could also run two wires up the sides of the printing flask, and use the reduction in impedance (as more salt water bridges the gap) to help work out water level. You could make it somewhat digital by only having some very small points up insulated wire be conductive, and you could increase the delineation of the points by having the wires far apart at the bottom, and close at the top.

2) It seems as though you sometimes need time to move the laser. I would take a leaf out of old-school camera technology, and have a rotating shutter. It basically could just be a spinning disk with a section (perhaps even a movable section) letting the laser through, and a small part to block it out. A $0.20 CCD could tell you if the laser was 'reflecting back' (ie not currently passing through to the fluid), and it could have a fixed high frequency sent through it (very easy to generate), so that it would only pass this high frequency through when illuminated (very easy to differentiate from the drip signal and/or water impedance.

3) To account for various issues in choosing a flask that is not perfectly rectilinear, you could allow a 'calibration routine' where effectively it prints out a 3D ruler, and you just put against an ordinary ruler and answer some questions about the X,Y,Z axis discrepancies . To save resin, you could even provide a Blender function where easy-to-file-off millimeter markings get added to certain straight axes, and then you can use even your first prototype to calibrate the machine (I've assumed that varying level-outputs are going to cause big size differences in the printed output).

Quote
CharlesNeville
Turning the laser on and off with a minimum of cost.

There are many bands of audio in the speaker output that aren't useful for controlling the mirrors. They can be used for other control signals.

To turn the laser on and off, you could inject a low amplitude, high frequency signal; eg 18kHz, -20db into the output of both L & R outputs. This could be filtered out with a very simple notch filter on the line controlling the mirrors but chances are you wouldn't actually need to.

Using another pair of notch pass filters, you could detect with high speed and accuracy whether the signals were in or out of phase. The electronics to do this would be very simple. The binary output from this detector could be used to turn the laser on and off.

Turn the laser on and off with the existing interface, two lines of code and a dozen analog components in the printer.

Quote
Spelljammer
I wanted to mention the mirrors and their odd warping of edges on a a straight line. The mirrors are a second surface type which will eat up about 10% of the laser power going in and another 10% coming out and even more at more severe angles. So, two mirrors eat up about 40 to 60% of the laser's power. They also have a ghosted reflection artifact because some of the light bounces off the surface. And, the laser changes angle when it enters the glass or plexi material, like a flashlight beam does when it enters water. The difference of densities between air and the glass cause this. Also, aiming the laser at the center of the mirror doesn't work well with the pivot point on the back side of the mirror. If the pivot is moved to the front and a first surface mirror is used, then all those complex trig problems go away.

I made a couple of pictures to show the difference between first and second surface mirrors

28stq4z.jpg

2m5b71x.png

Quote
PaulP
@Spelljammer
This is along the lines of what I was thinking for potential causes of the warping. I think the pivot point comment is particularly on the mark.

It's hard to tell without seeing the designs; but I think part of the problem is that the position of the center point is in the middle of the thread - not actually at the mirror (front or back). This would change the distance from the laser to the mirror (and mirror to mirror) based on how far it's rotated; changing where the laser hits on the next object in the chain.

I can't draw anything up at the moment; but a rough (albeit extreme) approximation would be putting a baseball card on a wheel.

Quote
JamesCooper
These are very good points, both of which we were aware of. Rylan's early prototypes used a second surface mirror (he once used old CDs), but I believe he is using first surface mirror for the recent prototypes and will definitely be using that in the final version. As for the pivot point, it gets a bit tricky. To make this simple to build as a kit, we are currently using a tensioned thread as the pivot for the mirrors. Keep in mind, this is something we are actively iterating on, so that may not be the final design, but it's simple and works fairly well right now. The problem is that we need a continuous piece of thread in order to easily attach it, so it's on the back of the mirror. I have considered building mounts for the ends of the mirror that would hold the center of the front face in the axis of rotation, but that would require a different way to mount the pivots for the mirrors. We'll see what options we can come up with that are cheap, easy to assemble, and have the physical properties that we desire. However, I have modeled the effects of this in Blender and it doesn't seem to cause a major problem. I can still draw a circle and a square without any unexpected warping.

As for the warping, I've actually narrowed it down to be caused by the orientation of the laser to the mirrors. Because the laser isn't orthogonal to the first mirror, the rotation of the mirror actually causes the light to follow an arc rather than a straight path. It's hard to describe it in words. Imagine a ray of light leaving the mirror at a shallow angle from the axis. When you turn the mirror, the start of the ray stays on the axis, but the direction of the ray rotates around the axis of rotation of the mirror. When it strikes the second mirror, it doesn't always land on the axis of the mirror. This is easily fixed by making the mirrors and laser all orthogonal, as is done in commercial 2D laser scanners. When everything is orthogonal, the light still follows a curve, but the curve runs along this axis, so it doesn't affect the apparent angle of the second mirror. Again, it's easier to draw than to explain, but hopefully that helps to understand. There is still the ever-present problem that we are controlling the angle, but drawing against a flat surface (a spherical projection on a Cartesian space), but that's easily accounted for with trig.

Quote
Spelljammer
I too have been working on a single mirror design. This is just a mirror held balanced on a center fulcrum by 4 small rubber bands. There are 4 coils and 4 magnets. Each opposite coil and magnet work together. For instance, to move left, the X coil 1 and coil 2 energize, the magnets under each is set to the opposite polarity, the right one could be North pointing up and the left one pointing down. So, one coil pushes and the other pulls for one direction, and the opposite for the other direction. I could have used just 2 coils, but I thought it would be more balanced with 4. I'll have to make a prototype.

29pfaqf.jpg

wt80p2.png

Quote
Dino
@tag16c
thanks a lot !!! I have understand your circuit . You can move the mirrors 20,000 times per second ( if the galvos/hardware can handle it) thats right?
And JamesCooper and the peachy team are do'nt working with 20kHz audio out, instead they are using a lower frequency sine wave and handle the amplitude peak with cap and diode. thats right ? For the printer they don't need that fast mirrors, for now. They are charging the peak voltage from the amplitude into the cap and lead it to the coil of the mirror, then discharging the cap, charging the next peak of the amplitude and so on, thats right ??? ( it's saying with easy words)

@ JamesCooper
you convert the 3d Modell into g-code and the g-code into audio signals. what max. frequence you use for the audio?
BTW: when it was planed to make the code of the peachy public? After the rewards for the backers are delieverd ? Or when the beta-tester received his peachy?

Quote
timogiles
I've been working on a single mirror galvanometer design also. I just put my design up on thingiverse and I figured I'd share with you guys. http://www.thingiverse.com/thing:179041. It's a work in progress but I'm making pretty good headway on it. I've ditched driving it straight from a headphone jack, if I'm going to build a printer based on this method i'd rather have it dedicated to a microcontroller. Right now I'm using an arduino and 4 FDS8858 to drive the coils.

Tim

IMG_20131106_212716_preview_featured.jpg

Quote
Peachy Printer KS Update
Below is a list of areas that need further testing and issues that we are still working to solve in order to reach first viable product:
- Memory caused by the dampening system (we’re working on a new magnetic damping system that may have very little memory)
- Resin Shrinkage
- Software usability
- Large prints
- Adding calibration processes for things like laser spot size at wide deletions, refraction, reflection, various non linearities, and offsets.