After 2 1/2 hours of polishing by balling over the switches, I washed and reassembled the switches. I don’t think more than two washing bins are needed.
While polishing, I set aside a switch every 30 minutes. So switches below represents 30min, 1hr, 1 1/2 hrs, 2 hrs, and 2 1/2 hrs of polishing with 0.5µm diamond paste. Intention was to find the optimal duration. Note that this was active polishing, not causal polishing-in-use while typing. I basically rolled a ball across switches constantly.
Test result was inconclusive because even the switch polished for just 30min was pretty smooth and smoothness increased ever so incrementally every extra 30min. Improvement seem to slow down after 2hrs but I’m not sure. I can say this though: even just 30min of polishing felt impressively smooth.
Encouraged with the result, I ordered a new batch of Cherry MX Blacks from NovelKeys. When it arrives, I’ll be making a full board of 2-hr polished MX blacks to see how they actually feel in-use.
Retooled blacks feel smooth when pressed dead center and straight down, but off-center is a different story.
Do you notice a difference when pressing the extreme top (north) edge of the keycap? That is usually where you feel the most grinding/binding on cherry switches.
There were some differences at higher grit in the last batch but not this batch.
Balling does result in side pushing so that may have helped.
I’ve decided not to try finer grits (0.25µm and lower) because, with 0.5µm grit, I’m already near the limits of my ability to discern smoothness which is to say, it’s up there with the smoothest switches I’ve personally experienced. And that’s without lube. Kinda crazy so I’d welcome second opinions from other tinkers.
lol. that may very well be. while this polishing method needs nothing more than diamond paste and a brush, most tedious part is cleaning which equipment can expedite and I am reasonably sure demand for polished MX blacks alone will be high enough let alone general polishing.
That reminds me, I got bunch of Halo housings and Panda stems to recycle back into service by polishing.
so could I just apply the diamond polish in a 3oz tupperware container like how I tub lube my switches? And then ultrasonic clean all the parts after I’m finished? That would be an acceptable amount of time for some smooth mx switches in my book.
Consistency of the 0.5µm grit paste is a bit thicker than 205g0 (close to 205g2) so I think same kind of water-soluble oil used in the paste needs to be added. Regular oil will complicate the cleaning process.
I’ve observed that finer the grit, less thick the paste is so 0.25µm grit may be tubbable/baggable although I don’t know if the grit is too fine or not.
WARNING: Working with submicron particles, like 0.5µm diamond paste, needs some care. When dry, submicron particles are small enough to float in the air. In fact, aerosols are particles smaller than 5µms which is 10 times the diamond particle in use here. I think this maybe one of the reasons why binding agents are used. Problem is particles are dissolved into water after use. If the water is just flushed into the water system, who knows where it’ll end up.
Filaments in coffee filters are about 20µms apart so layers of such filters if not N95 mask may be needed to recapture the particles after polishing then disposed carefully. We don’t want these particles floating around in the air.
UPDATE: On second thought, I think just storing the water used to remove the diamond paste in a plastic tub makes better sense. After the water dries up, pour some sticky oil to rebind the particles then wiped them in a paper towel should work in theory.
Digging more into diamond paste itself. There are also water-based, oil-soluble, and water-soluble pastes. It’s not clear if -based is same as -soluble.
Turns out the paste I got is oil-based. I could’ve sworn I read it but revisited product description doesn’t say it’s water-soluble. It came off easily enough with water and brushing but it could mean a thin film of oil remained.
This is not a popular subject so useful info beyond what’s on product pages is hard to come by but I did read that water-based paste applies more evenly than oil-based and can be easily washed off (which I assume means no brushing necessary).
One would assume that water-based paste could be thinned further using water but one of the manufacturer’s home page say this:
Can be thinned using our water soluble lubricant (see “Related Products” below).
And one of the water-based paste product description for 0.25µm grit say:
Grain size from 0.25 to 60 microns
It’s not clear whether this rather loose range is just specific to the product or just how grit is determined (by smallest particle).
Anyway, I think I know enough for messy and tedious but sufficient for DIY polishing at this point so I’m going to defer chasing these loose endpoints further until needs to scale up production arises.
I’m thinking water-based implies it’d also be water-soluble, maybe. But oil-based, for sure, would be a whole other thing. I’m sure some form of mild solvent would be needed. Do you think water-based diamond paste is the “new and improved” formula, an upgrade to oil-based?
Based & soluble are basically interchangeable in this case, just like paints. Water based/borne paints clean up with just water, while oil based & others need to be cleaned up with a solvent (anywhere from mineral spirits to methyl ethyl ketone). So the water based diamond paste should clean up much easier, but the particle size thing needs to be clarified before anyone tries it TBH.
Now in the garage better for everyone! (Also workbench is easier to clamp onto)
Added padding under the saw to reduce vibrations
Wedged some padding and a screwdriver under the trigger to dial in the speed and go hands free.
Including more photos more photos now that it is hands free so you can see what the whole setup looks like. If anyone is interested in the parts: Reciprocating Saw, Reciprocating Saw Brush Attachment, Bar Clamp, acrylic switch holder (from Zeal Sample Pack) & double sided tape
Test:
Ran a Cherry Silent Red & Boba U4 switch for 1 hour at slightly less then medium speed.
Results:
Silent Red Cherry has a slight popping sound. It still works but is worse off. After looking at this for a bit I see the saw was slightly too close to the switches)
Boba U4 no noticeable change. (this surprised me)
Next Test:
I believe trying with silent switches is not the ideal case, ordered some Cherry Black Switches from Novel Keys and the diamond paste that Don Park used. Will try two switches one by hand for 30 minutes and one by the automated method (maybe run for three minutes on medium speed or run the switch for a minute and rotate it four times in the switch holder to simulate off center pushes).
(Did run a test at a higher speed but the saw gets too hot to maintain that speed. Also it bounces to erratically to safely actuate the switch)
Let us know how it goes. Silent switches are more complex with nooks and crannies for particles to get stuck in and not removed during cleaning. I stayed away from polishing already lubed switches for similar reason.
For now, I’m focusing on just stock retooled blacks because I’ve always wanted a good set of vintage blacks but tales of inconsistencies in harvested switches kept me away. I think I already have what I set out to accomplish but need to build and use a full board for a while to confirm.
I don’t have an ultrasonic cleaner but I’m pretty sure it’ll do a better job at cleaning. I’m using plain water because plain water + hand brushing works well enough for simple linear switches.
I had only ~20 unlubed Cherry MX Blacks on hand so couldn’t polish enough for a board. The rest came in today so now the board is ready to be stepped on for leisurely polishing while coding. The case I chose is plastic Poker board, light enough for keys to be fully extended.
Can anyone recommend a model of ultrasonic cleaner that they use and like for something like this project? I have a lot of vintage blacks that I never used bc they are dirty.
