RepRap: Blog

To keep the angle of the point with the slide more vertical, I've put a kink in Probe 9. This keeps the width of the deposited line even regardless of the direction the probe is travelling in. One question I has was "Will the resin still climb the probe?" and the answer is yes. As you can see here it still accumulates nicely. Unfortunately the probe was a bit close to the edge of the foil, and this resin is mostly set solid:

Note that the probe still worked to some extent in this condition. I'll get it off by soaking in acetone, which has worked previously. This illustrates the importance of placing the probe's dipping point at least 0.9mm back from the edge of the foil, and preferably with the UV LED underneath the foil to avoid UV casting up onto the side of the probe.

The slide in this case was made using adhesive aluminium tape approx 60μm thick. The assembly process is as follows:

Clean slide with a razor blade and isopropyl alcohol to remove previous experiments. Cut a piece of adhesive foil on a cutting mat with a small box cutter approximately 1/3 the length of the slide.. It is important to cut the foil shiny side down.

Stick the foil on one end of the slide, leaving a 3mm gap to the end. The gap is so that the slide can be placed in a slide storage box after experimentation. Place a piece of printer paper on top of the slide and burnish the foil flat using a cylindrical polished chrome burnisher - I use the hex bit adaptor from a set of drill bits.

Place a second glass slide flat across the first perpendicularly and drive it in short motions against the edge of the foil. This curves the edge up slightly, creating a retaining wall for the resin.Run an isopropyl-dipped cotton bud up and down this edge to remove any exposed foil adhesive. When doing this, press the cotton bud against the edge of the foil laterally to give the retaining wall a crisp edge. The slide is now ready to be used as per the previous version, but the Touch Plate Height will need to be adjusted as the foil is thicker.

 

This is quicker to construct, and less prone to the probe tip penetrating the foil. It is, however, not as consistent, and I have found that making test dots of resin with the probe is the only way to be sure of probe height. It also tends to be more wrinkly, and might benefit from burnishing before adhering. However, if the wrinkles are not near the actual reservoir they are unimportant.

I have made another prototype gripper. This one is 13x larger than the one I intend to micro-print, which in this configuration would have a jaw spacing of 230μm (example has 2mm spacing, corresponding to 150μm). Obviously a version with wider jaws that don't completely close is possible. The hypodermic body and plunger would be replaced with a 0.5mm hypodermic needle and 0.23mm wire, assembled manually (I may add alignment stops). The actual actuator is undefined at this point but I has ideas. This one is a bit prone to torsion of the plunger, the plunger anchor is too short resulting in some curl, and the jaw tips need to be taller for more gripping area. All of these issues should be fixable. As it stands, it can grasp a sheet of paper and drag it along the desk.

 With minor changes it can be modified to close when pushing or pulling the actuator.

I've got a Svbony SV205 "8Mpixel" imager. It's a YUV USB3.0 device so it plugs in to Linux and just works. Unfortunately it didn't plug straight in to the Kronus trinocular microscope, but a few hours later I'd modelled an adapter that I'm almost happy with. The plastic shell wasn't sufficiently lightproof though, so I added some aluminium tape (which, yes, I will test as a slide touch plate at some point).

 

Notice how the USB lead wants to pass between your eyeballs when the imager is in the correct orientation. 

I've not played much with the lighting and so forth yet, but I took this screenshot of the calibration slide while fiddling around. The squares are 50μm and the linear divisions are 10μm. Looks like I may have a bit of tilt in my adaptor as the left side is a bit out of focus. Unlike the phone camera there is no clever filtering and pixel-guessing going on, so it looks a little drab.

3D Files for the adaptor are here https://www.printables.com/model/1544204-svbony-imager-to-konus-microscope-adaptor

Did a quick experiment with some Dutch Guilding leaf to see just how conductive it is. I put some nail gel on a slide, placed a sheet of paper on top, and used a cylindrical burnisher to rub it down. The idea being to form a thin layer of resin, which I then set to hold the guilding in place. Looks nice and shiny:

However, Dutch Guilding is thin sheet brass, not nice malleable gold. It did not conduct at all - look close and you can see where I was poking in vain with a multimeter.

Zoom in with the microscope and we see why:

It is just a maze of little cracks in there. With real gold you'd be able to burnish it out, or meld another layer on top. Worked brass is brittle and just breaks up. Its oxide layer prevents melding.

I need gold. Gooooold!!! 

Happy New Year!

The hand has healed up enough to make more slides, so I printed all the ideas I had kicking around while convalescing. I tried a few 200μm diameter micromugs, which ended up being dog bowls again, and a chain of hinged squares to test multiple folding. I'll cover that in a later post when I've origami'd it.

But the main push was to try and print a pyramid, on the grounds that they're infamous for standing up. Mine was 400μm on each of its three sides, and 300μm tall, making it the tallest and largest thing I've tried to print.

Of course, it didn't go as expected. It looks like it's spent 10,000 years in a rough desert and learned bad things from a ziggurat:

It didn't help that I probably started the probe a bit too low either. My guess is I need to get the layer size (2.5μm) down even further to make sure there's enough resin in the middle of it, maybe adjust the print width a little lower too. However, it did print all the way up even if it did sag like a poorly-baked cake! Where the print is sliced a little denser at the top and the corners you can see some facets trying very hard to happen.

The UV cured the resin from underneath too, so that answers whether the underslung tiny LED can cure relatively thick layers of resin.

Other good news, the printer stayed together throughout the entire process, and the resin was in the reservoir was still perfectly good after all the exposure to air and scattered UV. I should point out too that this was printed using short segments rather than dots of resin. That is significantly quicker to do. It would be interesting to compare the dot-by-dot approach for quality at some stage.

Why is it dark grey? It isn't. That's a refraction of the black microscope stand underneath. Doesn't it just highlight all the defects beautifully? The resin is actually quite dense and optically good, which is why I think I need to just make the layer spacing a bit thinner. Could probably print simple optics. I'll finish off with a shot taken straight down onto the frosted glass surface:

 

Working on improving the photography. The trick with propping things up on lead bullets only gives me a couple of directions of viewing, so I've designed a gimbal rig. With this I can turn the microscope slide any which way under the binocular microscope. Getting good pictures is important because saying "I am creating tiny transparent things that only I can see" is a good way of getting an express ticket to the Funny Farm. Here's a good angle of the folded squares:

 

It has two sets of pivots, and the base can be rotated to give a third degree of freedom. This is what the contraption looks like:

 

If you want to build one I've put the STL model and source here: https://www.printables.com/model/1534011-microscope-slide-rotating-gimbal-mount

Back again, been doing a bit of organization behind the scenes. The hand is still a bit wobbly so I increased the size of the hinged squares to 0.8mm on a side and two print lines thick to give me a fighting chance.

The sides are 30 or so layers thick and two print lines wide. Line thickness is dependent on direction of travel of the angled probe, so I'll make one with a vertical tip later.

The lower square is filled in with 2 solid layers. I tried depositing unset resin on the hinges, but didn't get enough on and added some manually with a hypodermic afterwards.

After gently teasing the hollow square off the slide manually under the binocular microscope, I folded it up 90 degrees, and held it in place with a "finger" - the little hinges are capable of pushing the 15g weight aside, so I'll have to make a heavier one:

Then I added resin to the hinges and cured it all with a UV lamp. The square stayed vertical, so I threaded one of my fine greying hairs through it for the top photo.

This is the current microphotography rig:

I have a new ring light that allows me to use the crappy 0.6 Mpixel microscope imager. A new 8 Mpixel one is on order. I have been planning a better gimbal rig for orienting the slides. I hope to get μRepRap under the rig too once I have joystick control software, which will allow more delicate manipulation of finer prints and better imagery. 

I have now posted the results of FPath Experiment 011 which is part of the FPath project. This is not an actual experiment per-se, rather, it is documentation of the image recognition and Graphical Stigmergy techniques applied in Experiment 010,

Experiment 010 used an interesting image recognition technique in order to automate the movement of a probe at the sub-millimeter level. The major benefit of this technique is that it is fast enough to keep up with a sequence of 640x480 images arriving at a 30 fps frame rate. 

Also, which may be of some interest, is a discussion of the mechanics of the Graphical Stigmergy algorithm. This discussion shows how complex emergent behaviors can be obtained from combination of relatively simple actions and the modification of the environment to send signals.

I just thought that some of you might be interested in how all this is done. The video explains all: https://youtu.be/be725uWk4c8

The image below shows a frame from the video - Ok, I know, it isn't the most sophisticated graphics you've ever seen. (click on the image to enlarge, watch the video for context) 

 

Labels: FPath, Graphical Stigmergy, Image Recognition, Sub-millimeter Control