Teardown - new Ellipse beamspring switches!

[Chyros]

Today we look at an awesome development; a new beamspring switch heavily inspired by IBM's original design. These come from Joe over at modelfkeyboards.com and are the type that feature in the first round of beamspring keyboards.

[Muirium]

Not even telling us if it’s brilliant or shite? You tease!

[Chyros]

Not even telling us if it’s brilliant or shite? You tease!

What's the point, off of just three switches that aren't even mounted? :p

[del20nd]

I'm very excited for the verdict on these. They might be too rich for my blood either way, but... man, what a dream.

[XMIT]

What sort of longevity is he quoting for these switches? IBM spent big bucks on a fairly exotic steel alloy for the beams in their original boards for longevity. Input Club and Kailua gave up their cloning effort because they could not source an appropriate steel. I’m worried that this is a toy that will fail after 100k actuations. To put that in perspective, that’s a couple of weeks of typing for some commonly used keys such as the space bar and ‘E’.

[webwit]

Wait, the originals were made from mithril? Should Ellipse get the help of Sauron?

[Findecanor]

Input Club and Kailh tried to squeeze the beam spring mechanism into the same size as a Cherry MX switch, and without giving up the LED window or requiring high-profile keycaps.
I would think that the beam springs would have needed to be much smaller, and bend more than in these switches.

[Obin]

Input Club and Kailua gave up their cloning effort because they could not source an appropriate steel.

I thought the primary issue was the investment cost for the tooling and almost everything else was ready to go.

[Touch_It]

Wait, the originals were made from mithril? Should Ellipse get the help of Sauron?

Most definitely. He is master of alloys.

[headphone_jack]

Ellipse himself admitted that these switches have reliability issues, specifically with the flyplate detaching from the beam spring. And when Ellipse admits something is wrong, you can bet that the actual problem is far more prolific. I can't recall if he was using the exotic alloys IBM used or not, but if not then you can bet the support thread will double in length

[Muirium]

that’s a couple of weeks of typing for some commonly used keys

Long enough to post first impressions and shower them with praise. Job done. Then hush and flip the board “immaculate condition, real shame to sell, make me an offer I can’t refuse” to some other punter, like a pro.

Whether it’s enough to cover that after badgering us about all its other issues when it first arrives, though…

[TNT]

I mean, all of these wild assumptions based on some kind of mystical spring steel IBM allegedly used and Ellipse admitting there were issues with the switch at some point (these could have been fixed by now, who knows) ... Could we wait with shitting on this thing until it's actually out?

[Muirium]

Must we? This is way more fun.

Besides, all the more scope for him to impress us if expectations remain where they are.

[del20nd]

The Input Club's reasons for abandoning their version of switches is listed here in their own words:

https://www.kickstarter.com/projects/le ... s/3485667

It was combination of tooling expense and premature failure of metal parts.

Their design was also very different, and was attempting to fit into the general height of a Cherry switch from what I read. That would mean major compromises in how much metal they could use for the mech. The Eclipse switches seem to make less of a compromise in height, so could use more material.

Still, I'll be shocked if he can do it at his projected $600 per board price point. I don't know if we have the original price of a beam spring keyboard back when IBM sold them in the 1970s, but if the Models F were the cost cut variety and retailed for over $400 unadjusted, then the beamsprings were expensive beasts indeed. Like, "only possible because a giant business was expensing a mainframe" type expensive. Building anything to that standard is going to be an insane undertaking.

[XMIT]

To be clear my comments came from conversations with Input Club directly. That Kickstarter link is authoritative.

[Chyros]

I've been in several hour-long conversations with Jacob during the development of his beamspring switches, I'm fully aware of all the difficulties he encountered, including but not limited to the metal alloys people are referring to (I shortly alluded to this in this Teardown video as well). Naturally I'll be covering this issue during my review as well.

[darkcruix]

I just want to zoom out for a second and say that it is a pleasure to see these re-productions and re-inventions have been picked up. Especially by fellow keyboard enthusiasts.

[Ellipse]

Project coordinator here. Many thanks to Chyros for taking a look at the new beam spring modules and doing a teardown that shows how the module works. The second half of the round 1 batch with the factory replaced beam springs arrived, so now the entire first round of beam spring keyboards is in stock. I will be replacing the other 150 or so keyboards' springs that arrived a while back, before the spring issue was discovered.

Orders for the first round should start next month after the remaining Model F backlog is completed (about 100 keyboards left).

As a note, there are no reliability issues detected from the beam spring switches that are any different from the IBM originals. My comments on the beam flippers detaching is no more or less frequent than my experience with the detachments of the IBM originals, with a note that they may detach as the keyboards are bounced around in shipping, so folks will need to set up and check their beam spring keyboard before they can start using it. The new ones do not detach more or less frequently, and can be reattached in the same manner as the originals. I did not mean to imply that the new modules are any different from the originals in terms of reliability. The only difference as Chryos pointed out is that the new modules have a slightly different sound to them, with the installed module sound being more similar with the round 2 boards than with round 1 (matching the weighting, key travel, etc. was the main focus for the beam spring project, not an exact reproduction of the sound, as mentioned in the first post on the beam spring project thread).

In my testing over the past couple years, I do not recall modules failing due to fatigue with usage, though to be safe I do expect to include a first aid kit of 3 to 10 modules with each keyboard at no additional cost (haven't decided yet).

Also the pricing range of $500 to $600 has been finalized and the first round already finished production and is in stock, so pricing is not an unknown variable at this point.

[Claybie]

Fantastic news!

Ellipse: Are you able to share what the board will look like and/or what the layout(s) will be?

[Ellipse]

Yes, please take a look at the Beam Spring reproduction thread here: viewtopic.php?f=50&t=26174