Alright, here are my preliminary findings.
Basically, the goal of flame-polishing is to liquefy the plastic, so that surface tension will cause the molten plastic to "self-level," eliminating the bumps left there by the well-used Cherry molds.
First thing I noticed when I popped open one of my spare MX browns, is how tiny the stems are. The tactile "legs" are only about 1mm thick. I knew it wouldn't take much to mess them up.
I don't have a blowtorch handy, but I do have a Bic lighter (I'm not a smoker, BTW, but everyone needs a way to make fire). Because I only have two hands, I couldn't get any pics of the flame-polishing process. But basically I held the stem upside-down by the cruciform with a pair of needlenose pliers in one hand, while passing the yellow flame of the lighter back and forth across the legs at a nice, even speed. I initially took 4-5 passes, but didn't really notice a difference so I let the stem cool and tried 7-10 passes (I probably should have counted, but it's pretty intuitive, really). This produced a visibly smoother surface. I found that it didn't take much to overdo it, as seen it the picture below:
Flame-Polishing MX Brown Stems by
Nick Bair, on Flickr
From left to right: stock, flame-polished, and overcooked.
On the middle stem, you can notice a slight rounding of the corners of the legs compared to the stock stem. You can see that the legs have been completely deformed on the right stem. I flame-polished three stems. Two came out good and one was overcooked. I figured two was enough for some testing.
While I was in switch-opening mode, I went ahead and lubed a stock switch with thick Krytox from TechKeys. I lubed both the legs and the side sliders. I marked the leads with different colors so I could tell them apart during my testing.
Testing
Blind-Testing Cherry Switches by
Nick Bair, on Flickr
The leftmost switch is unmarked, and the stem is stock. The black switch has been lubed. The two blue switches have been flame-polished.
For testing, I decided to do 10 rounds of blind tests and log the results. For each round, I shuffled the switches and laid them out in a row without looking. Then I went down the row, one by one, doing my best to guess which switch it was and logging the results in a spreadsheet. Here are my test results:
| Round | Switch #1 | Switch #2 | Switch #3 | Switch #4 | Notes |
| 1 | Lubed | Flamed | Stock | Flamed | Guessed right on every one. Initially thought #2 was stock until I felt #3. |
| 2 | ? | ? | ? | ? | Inconclusive. I was really confused for some reason during this round. Sophomore jinx? |
| 3 | Flamed | Lubed | Flamed | Stock | Guessed right on every one. |
| 4 | Flamed | Flamed | Stock | Lubed | Was unsure about my initial guesses, so I took a second pass before looking and guessed every one right on the second pass. |
| 5 | Flamed | Flamed | Lubed | Stock | Confused flamed with both lubed and stock on the first pass before settling on the correct guesses on a second pass. |
| 6 | Flamed | Flamed | Lubed | Stock | Guessed right on the first pass. |
| 7 | Stock | Flamed | Lubed | Flamed | Guessed all on first pass except #1 (I thought it was flamed at first until I felt #3 and #4). |
| 8 | Flamed | Stock | Lubed | Flamed | Guessed all on first pass except #2 (I thought it was lubed at first until I felt #3). |
| 9 | Flamed | Stock | Flamed | Lubed | Mixed up stock and lubed on this round. |
| 10 | Flamed | Flamed | Stock | Lubed | Mixed up flamed (#2) and stock on this round. |
FindingsSome clear findings emerged during my testing that I thought were really interesting:
- The differences between all three types are pretty subtle, at least when tested as bare switches rather than in an actual keyboard.
- Despite the relative subtlety, it was pretty easy to guess which switch I was using most of the time. This proves that there is indeed a difference.
- Lubed switches are much smoother than stock, but have much less tactility.
- Flame-polished switches are somewhat smoother than stock, and have a much more pronounced tactility. They almost feel like ergo clears.
- Interestingly, lubed and flame-polished switches are easier to tell apart than either type compared to stock, because they are both on opposite ends of the tactility spectrum.
- When I became confused during testing, it was usually due to the relative tactility rather than the smoothness.
ConclusionFlame polishing works, but it's more risky than lube because it's easy to destroy a stem. The overcooked stem pictured above is basically useless now. It feels weird and the stem gets stuck on its way back up. I believe I could master the art and nail down the precise timing to be able to produce a whole batch of correctly flame-polished switches, but there would probably be more than a few casualties along the way. I'm wondering if the higher heat of a torch might do better than the lighter, since it can superheat the surface very quickly without allowing time for that heat to conduct its way deeper into the plastic and cause deformities to occur. If I ever end up getting one of those butane mini-torches, I might give that a try. But right now I don't have very many spare browns, so I'm not going to try and make any more with the lighter.
I've got some Gateron Browns coming as part of a Planck kit from Massdrop. I'm going to bag up these switches I used for testing, and when the Gaterons get here I might do another round of testing with one or two of those.
So I would say the jury's still out on flame-polishing. But it is definitely a worthwhile idea to explore if you have some browns to spare. And if you have a torch, I'd be really interested to hear how that goes.
