Repairing Wiper Arm

The rear wiper arm of our Hyundai Santa Fe ends in a clip which snaps over the shaft of the wiper blade assembly. It's made of a very tough plastic but ours broke during a blade change.

The local dealer wanted $80 for the part. Auto parts stores don't stock it. No Internet vendors carry that exact part, and they wanted $50-$60 for similar parts for other models. So I decided to try to fix it.

What broke was a kind of claw that went under the shaft, from left to right in this picture. I did not think I could replicate that exactly with the required strength. 

But I noticed there are two holes in the arm, quite parallel, one T-shaped on the left and one D-shaped on the right. I could make a part that would go into both holes and fully trap the shaft. With the right dimensions, the friction fit should hold it securely. Chamfers on the edges of the post help avoid excessive friction in the corners.

PLA wouldn't do because it's not terribly strong and does not tolerate heat. Nylon is tougher and more heat tolerant, and is somewhat bendable and resilient. (See my bicycle part.) I have a spool of Taulman Bridge Nylon which should do well. Here's an article about printing with nylon. 

I used a 0.25mm nozzle because these dimensions are pretty small. The key point was to make the T-post and the D-post just the right size to fit tightly, but be removable for future blade replacement. I designed this parametrically in Fusion 360, and I defined a "tolerance" added to the dimensions of each post. That makes it easy to tweak the size. I printed just one half of first unit so I could clearly see how the posts lined up. Then I printed four iterations getting the sizes just right so they would fit with just the right friction. The filament cost per part was about 8 cents, and it took 30 to 40 minutes to print, so no big deal to try a few times getting it right.

The picture shows a lot of stringing, typical of nylon in my experience. That really does not matter for this application so I did not work hard at cleaning them off. Usually I use a metal bit in a Dremel tool to clean up, but it really doesn't work on nylon. So I just snipped the worst and left the rest.

Printing it as shown, the posts were prone to layer separation. I did not want to try to print it on its side and deal with support. I counteracted the layer issue with three methods:

1. 100% infill. For a tiny part like this there's no reason to make it hollow.
2. Higher print temperature to try to bond the layers.
3. Heat treating the finished part. I had never tried this before. See this article. I did this on just the final iteration, and the posts seemed really strong after this.

Here's how it looks installed inside the wiper blade, holding the shaft in place.  

And here it is installed on the car. The base is only a few millimeters thick, so it doesn't get in the way of anything.

The total material cost was about 35 cents. Much better than $80 for a whole new arm!

UPDATE: It didn't work. The next time she drove through the car wash, the force of the brush mechanism popped it right out of the arm. I ended up buying a whole new arm.

Telescope Adapter (Eyepiece Projection) for 1.25" Eyepieces and Micro Four Thirds Mirrorless Camera

Previously I developed an adapter which connects the camera directly to the telescope focuser, so the telescope is the camera's lens. This is called the Prime Focus method. In my case, the focal length is 910 mm. While this works, the images of planets are very small. They can be scaled up (cropped), but because they do not cover many pixels of the camera sensor, they don't preserve much detail.

Another approach is to take a picture of the image produced by an eyepiece, which magnifies the image. The amount of magnification depends on the focal length of the eyepiece and the distance in front of and behind it. I have not yet determined this for my adaptor, but probably will at some point.

There are various focal length and magnification calculators available, but they seem to vary and will depend on measuring the actual distances resulting from the adapters. (I'm starting with a 15mm Plössl eyepiece, and I also have a 6mm to work with... they are not the same physical size.)

As mentioned in my previous post, it starts with a part connected to the telescope's focusing tube which accepts a standard 1.25" eyepiece or other part. 

Next is a tube which inserts into the 1.25" opening and holds the eyepiece at some distance up from it. So that tube is slightly larger than the eyepiece and contains a shelf so the eyepiece rests at the right height. The outside dimension therefore depends on the specific eyepiece to be used. I'm using an SVBony eyepiece, so I measured that carefully. Maybe in the future I can make it larger to accept others. But since the design is all parametric, other users could resize it if needed.

Next is a tube that slips over the eyepiece holder and can slide up and down. This is important in order to "project" the image onto the camera sensor at the desired size. More on that later. I made some estimates of the minimum and maximum lengths. This part includes threads for a set screw to lock this outer tube onto the inner tube at the desired length, without screwing into the actual eyepiece. I went through a couple of iterations of exterior shape to get one that prints an overhang with no drooping. I got some knurled screws from Home Depot that work fine.

It ends with a part connecting to the camera's bayonet connector. I found the Micro Four Thirds connector as a STL file. (It's pretty complicated and works well so there's no reason to redevelop or change it.) The "standard" way to connect this to telescope adapters is called a T-mount, which is a big ring, 42mm in diameter if I recall correctly. I had some issues with the standard T-mount threads so I'm using a somewhat larger custom thread for now. I may go to a standard thread at some point. This part has male "T-mount" threads, and the outer eyepiece tube has female threads, so it all fits together.

Depending on the distance to the object being imaged, the telescope focus travel may be extreme. To accommodate a wide range, I developed an extension tube that can be inserted between the camera connector and the outer tube. This extension is a 14mm long hollow tube with male "T-mount" threads on one end and female on the other. I think the tube is needed for celestial objects, and not needed for terrestrial objects. More on this later.

The inner and outer tubes need to slide snugly but smoothly over each other, so they need to fit well. Due to shrinkage of the PLA plastic I used, it took me several iterations of adjusting a tolerance variable before these fit properly. Here is how all the parts appear together, in cross section:

And visualized:

And in real life:

Because the M43 bayonet connector has a lot of fine details, I used my smallest nozzle on the previous design. It's 0.15mm. Unfortunately after printing a few parts it became hopelessly clogged. I went down a rabbit hole trying to unclog it with solvents. Since that was taking a long time, I tried printing with a 0.25mm nozzle, and it worked fine too on the M43 as well as the threads. It's maybe a little more rough, but it works. In the meantime I ordered another 0.15mm nozzle and some 0.15mm cleaning wires.

If the tubes are not set to the right length, the image is highly vignetted, i.e. a small round image in the middle of the camera sensor, highly distorted. By extending the tube the image gets larger and less distorted, until it becomes a usable image. I'll need to do some tests to see if the distortion completely straightens out or remains an issue to deal with in post-processing.

This tube will bear the weight of the camera and needs to be robust enough to not allow it to bend or fall. The thicknesses of the inner and outer tubes are parameters so I can thicken them if necessary. It seems to me that on the one night I have tested it, there was one part that I felt needed to be thicker so that a set screw would not distort it. I think it was the insert on the lower tube, that goes into the telescope focuser. I'll have to check on that again.

The good news is that the adapter did work as intended. The bad news is that I had tracking problems with my telescope's computer, so I was not able to get any useful images. I've taken some steps to improve that, and have some more testing and calibration to do. I also need clear skies and decent temperatures to do some further tests. I'll update this article as I get more results.

These parts and the Fusion 360 design file are now published at https://www.thingiverse.com/thing:4834778

Another step will be to figure out the length needed to make a lower tube for my 6mm eyepiece, which is considerably longer than the 15mm. The length of the adapter may increase the leverage that the weight of the camera places on the telescope tube, which may change the power that the telescope drive motors need to exert to move the whole assembly. We'll see how that goes.