I share the relatively painless process of turning a junky 3D printer into a tool to help you place tricky components or to automate your PCB assembly. The minimum feature size of most electronic packages is 0.1mm. Align to that point and surface tension will pull it down to a perfect placement. This is a big deal for when you've got a BGA-style chip where joint inspection is gonna be a hassle. The modern 3D printers are putting out 10 micron positioning precision and old junk can still put in 0.1mm or better! That's precision and not accuracy, so I'll also be using a cheap usb microscope to do optical position alignment. With the unused extruder stepper motor, I simply route that to the platform and have a component rotator for 45 or 90 degree parts. The goal is manual placement of tiny components that I can't trust my zoomer hands with.
The printer you use for this has a single requirement. If it's good at printing, let it do that and don't bother it. There's no point not letting a 3D printer do what it's good at. In my case, my 4 year old Monoprice Select Mini V2 decided to stop printing reliably, probably due to the bed heater being shorted as a result of the zip-ties used for strain "relief" and cable "management". Regardless, my monkey brain couldn't figure that out in the middle of classes and so I just YOLO'ed it and got another one. To all you who read Hackaday- seriously don't take too many classes or schedule too many things or work too much. As soon as you have free time, your mind is able to do great things. Something about stress and sleep, etc.
The first step is to remove the coating on the bed. In my case, Monoprice uses an adhesive to bond the plastic print surface to the aluminum bed. I used a scraper to peel the bed off, but then the adhesive remained. Being primarily an organic goo with the gumminess of a butyl compound, I used isopropyl on the glue to cause it to swell. I don't know what the glue is though. After it swelled, I was able to use a scraper razor blade to peel it clean. DO NOT use these things lightly. They are sharpened and their edges are kept pristine and it took me many milliliters of my blood, sweat and tears to learn that lesson.
After your bed is clean, you can rest assured that your bed isn't dirty.
Now, you need to find a way to route an air-pump from a heavy mass portion of your printer up to the Bowden tube or filament feed-point. In my case, the base of the unit is chunky, so the vibrations of the motor will be distributed across a larger mass and avoid any components dislodging. You can also disconnect heater and fan cables from your control board. You'll use one of the heaters to turn on and off your pump, and since they'll probably be using PID control with a thermistor as feedback, you'll need to keep the thermistor wire connected. I chose the bed to free up the nozzle.
Depending on your printer's design, you may have a column, or some cable management system which leads to the extruder assembly. You should hijack this for your air pump's tube. I was able to feed mine through the column that the Z-axis leadscrew is in. While you have access, lube up your leadscrews because why not. The silicone and oil based lubes will probably kill each other, but YOLO.
Now, you can spruce up your old extruder assembly for its new life. The thing that really kills me is printer nozzles are usually 0.5mm or 0.4mm. That's the prefect size to handle the most common annoyingly small component, the 0402. Let's not speak of the excruciatingly small 01005, but I do have a solution for that too, actually. You'll want to clean your nozzle the best you can since any remaining organic residues will result in a chemical process called "gumming" where radicalization events from oxygen and cosmic neutrinos causes double bonds and subsequent cross-linking (seriously, cosmic neutrinos are the reason we need fat soluble vitamins D and E in our cellular membranes to keep those radicals out of our nucleii. Eat some carrots! I used the scraper to get the kapton and ceramic insulation off of the heater block and undid the set screw for the heater element.
With everything apart and after repeatedly sticking a nozzle cleaning pin through the nozzle hole, I was ready to toss in the pump.
I ended up going for a different, smaller pump since this one was way too noisy. Notice however the rubber band technique. Mounting the rubber bands between the contact interface of the housing (at the vent holes) creates a lossy medium that effectively prevents "impedance matching" of the vibration source to the printer body load. An alternative is to suspend it to avoid any contact so it "sees a high impedance" and just vibrates at itself, but that would have been too much effort. The new motor simply used a clamp with threaded inserts, modeled in ImNotARealEngineerCAD and...
I'm planning on tossing together some G-code after manually doing a few boards. The fact that most reels have 4mm spacing on the holes means I can create jigs to hold multiple components, calculate the next component position, and go there to pick it up. However, with the USB microscope I got, I'll need to write a little OpenCV script to recognize common shapes and place accordingly. I would recommend buying this particular one since that's what I got. https://www.ebay.com/itm/1600X-USB-Digital-Microscope-Handheld-Zoom-Biological-Endoscope-Camera-Pocket/123963687752 very "non-4K" quality, but perfect for some machine vision!
Super cool! What does the software side look like?