Comment by howardbeware | Hacker News Reader

Comment by howardbeware | Hacker News Reader

Parent Full thread

We evaluated Cherry Reds for the project, but there were several things wrong with them. In order of our priorities:

1. It's very difficult/expensive to make custom key caps that mount to Cherry MX stems. By custom key cap, I mean a completely new shape, like this:

http://3.bp.blogspot.com/-g_Bi0FseNUk/VGBKMdshu-I/AAAAAAAAAe...

We actually did come up with a method for doing this with Cherry MX switches, but the cost was prohibitive. I'll add more detail on the campaign page to make this clear.

2. The actuation force of Cherry MX Reds was just a bit too high. We're making a stenotype keyboard, which should have a very light touch. The Cherry MX Red might have been marginally acceptable, but we're much happier with the lower actuation force of our switch.

3. As you pointed out, the availability of Cherry MX Reds isn't great.

4. The price of Cherry MX Reds is much higher.

Regarding the Topre switches, I agree that 30g is almost too low. Also, the mounting post for the Topre switches seemed challenging from the perspective of creating custom key caps, though we didn't actually try.

As for the openness of the project and there being only one manufacturer, is your concern about availability, control of the IP, or something else?

Okay. Interesting response here. There's some points I hadn't considered.

1. I see your argument here but I also wonder if there are processes that you have not considered. I don't know if you're familiar with ClickClack's keycaps, but that's an artist who does custom shaped keys. I'm 99% positive, as I've done the same process myself, that he sculpts a master, makes a mould and pours resin keycaps. You can make these in huge bulks quickly for cheap this way. Getting the Cherry MX backing is as simple as cutting the stem off a key and gluing it to your top. It makes choices of plastics/resin more difficult (you can't just mix and pour PBT...) but this process is super cheap. Cheaper than 3d printing and cheaper than machining. You can build a vacuum degasser for $50 + an air compressor.

2. Cherry switches can be spring modded to much lower actuation forces. There are places where you can buy 30g-45g switches, in bulk, and it's probably cheaper than tooling them yourself. I honestly can't imagine using less than 45g and maintaining accuracy, especially on something that's basically close to being a chording keyboard like the Stenotype (but then again, I'm not a steno).

4. I guess it doesn't matter for your use case because you have to make your own keycaps, but for someone who will be using the switches for something else this is not true. Regular keycaps for ALPS switches are either going to be substantially more expensive or someone trying to dump old/excess stock. I don't even know where you could get ALPS keycaps made in a group buy -- people have tried. The campaign page is advertising advantages that really don't exist outside of your specific use case.

That's really my only gripe. I don't see where there would be a use case outside of your specific project. I know folks sitting on huge bulks of ALPS switches that are sitting doing nothing because usually nobody wants them. Forgive my myopia.

Alps is an interesting choice for your project, especially since you don't have to worry about stabilizers at all.

1. The problem wasn't in creating the shape we wanted - that's easy with CNC. The problem was creating the cross mounting post for Cherry. The process that ended up working was somewhat as you suggested: we basically used a lathe to remove material from a standard key cap until just the post was left and then we attached the post with a good adhesive to our custom key caps. The problem with this process is that it is very expensive in terms of labor/time and requires a high degree of precision to ensure uniformity. I think I might eventually make a keyboard using Cherry MX switches using this process, but for the current project the goal is to produce at least 100 keyboards while keeping the cost down. We tried for several months to make the Cherry MX Red work, but ended up deciding Matias was the better option in this case.

2. Modding key switches works well for a single keyboard, but really causes costs to soar if used in production. This wouldn't work for us because one of the main design points of the keyboard we're working on is making it financially accessible to a wide range of people.

4. This is a fair point. Honestly, I'm not too deep into the mechanical keyboard world beyond this one project, so I don't have a good feeling for other people's use cases. My hope, though, is that people do some really strange things with these switches and start innovating in ways I wouldn't have thought of.

Thanks for all your feedback - it's been really helpful. I haven't yet updated the campaign page to reflect it all, but I will soon.

Reply on news.ycombinator.com