Recent Posts31
Group Buys and Preorders / Re: Beam Spring 104+SSK+122+62 Reproduction Project! Round 1 now shipping
« Last post by Ellipse on Tue, 14 April 2026, 20:16:27 »Here is a summary of the Round 2 improvements over the Round 1 beam spring boards:
An improved, dampened, and more refined sound while typing, due to:
• Extra case material and foam on top of the modules (like on the originals) for added dampening. The casing is made from custom-bent metal (see below factory videos) whose dimensions were based on scaling the measurements you kindly provided me. The Round 2 keyboards were made to look like the classic appearance of the rounded case beam spring boards (basically all the boards except the angular 3278 and 3279 boards (a prototype was made of this design but most folks disliked it). It's also inspired by the process used to make the IBM Displaywriter keyboard case (the bottom was stamped and folded, not die cast; the new beam spring boards were all stamped, folded, and welded with custom-made stamping tools).
• Improvements during a long R&D process over the past few years (the injection molds for the modules was redone and the beam module metal materials were changed to make them sound closer to the originals; though as I mentioned in my updates, the goal of this project was to bring back the beam spring mechanism and not to exactly recreate the sound or feel as exactly as with the Model F project).
• Also the inner assemblies are now closed with a torque screwdriver, which allows you to specify the torque and it will let you know if the torque is too high and if the screws are too tight. The purpose of this process is to optimize how tightly the bottom and top inner assemblies are attached to each other. In our research over the past few years to get these keyboards to be in line with IBM's standards, we have found that having a slight looseness to the assembly, as shown in one of the below videos, results in a far nicer sound and typing experience. You don't need a torque screwdriver for your own maintenance of these keyboards, you can just back off the tightness a few turns by eyeing it and keep things loose for optimal sound and performance.
Additional improvements:
• These are the very first beam spring keyboards with Rico's new open source RP2040-based Leyden Jar controller (same controller as on the F122's). All the new Model F and beam spring boards still run Vial. These now have enough memory and pins for 18 instead of 16 columns, improved matrix scanning speed, as well as additional Vial features, including layers and macros, that did not fit in the ATMEGA32U2 used in the xwhatsit controllers. The other keyboards that do not require 18 columns (B62, BSSK, etc.) will continue to use the xwhatsit controllers until stock runs out, as I've noted on the forums. I stocked up during the chip shortage many years back, before the Leyden Jar was released.
• Improved QC - no one is expected to need to adjust the bend of the metal parts inside the module due to a factory QC error with some of the original modules that was disclosed at the time. The most common issue is just a flyplate (the metal part connected to the flipper) that gets separated from the rest of the module while in transit. This QC improvement was introduced with the second batch of Round 1 beam spring keyboards which have been in use for a couple years now without major reports of failures.
• Reattaching the flyplate is now even easier than both the IBM and Round 1 originals where you had to use tape and a paperclip. The modules now have a custom-molded press fit plastic washer on top so you can quickly and easily disassemble the module by hand and reattach the flyplate (or replace a damaged inner plastic part). This was also introduced with the second batch of Round 1 beam spring boards which used a glued-on metal washer for each module. Also in general the beam spring is much easier to take apart compared to a Model F; you only need to unscrew the bottom inner assembly plate instead of using pliers to separate it. You can even leave the keys installed (except for the modules that you need to remove to fix, of course). For these reasons, resolving issues requiring disassembly that come up over the years might be even easier and quicker to repair than many issues of the Model F.
• The boards still come with a mini-first aid kit of 4 modules in anticipation that some may break while in shipping, though many folks order the 12 module full first aid kit.
And as always, for folks who have been waiting a while now and want to see the reasons behind the extended production timeline (and for everyone who is curious about the Beam Spring keyboard production journey), everything is explained in detail on the updates page on the project website.
The aforementioned beam spring Round 2 videos:
An improved, dampened, and more refined sound while typing, due to:
• Extra case material and foam on top of the modules (like on the originals) for added dampening. The casing is made from custom-bent metal (see below factory videos) whose dimensions were based on scaling the measurements you kindly provided me. The Round 2 keyboards were made to look like the classic appearance of the rounded case beam spring boards (basically all the boards except the angular 3278 and 3279 boards (a prototype was made of this design but most folks disliked it). It's also inspired by the process used to make the IBM Displaywriter keyboard case (the bottom was stamped and folded, not die cast; the new beam spring boards were all stamped, folded, and welded with custom-made stamping tools).
• Improvements during a long R&D process over the past few years (the injection molds for the modules was redone and the beam module metal materials were changed to make them sound closer to the originals; though as I mentioned in my updates, the goal of this project was to bring back the beam spring mechanism and not to exactly recreate the sound or feel as exactly as with the Model F project).
• Also the inner assemblies are now closed with a torque screwdriver, which allows you to specify the torque and it will let you know if the torque is too high and if the screws are too tight. The purpose of this process is to optimize how tightly the bottom and top inner assemblies are attached to each other. In our research over the past few years to get these keyboards to be in line with IBM's standards, we have found that having a slight looseness to the assembly, as shown in one of the below videos, results in a far nicer sound and typing experience. You don't need a torque screwdriver for your own maintenance of these keyboards, you can just back off the tightness a few turns by eyeing it and keep things loose for optimal sound and performance.
Additional improvements:
• These are the very first beam spring keyboards with Rico's new open source RP2040-based Leyden Jar controller (same controller as on the F122's). All the new Model F and beam spring boards still run Vial. These now have enough memory and pins for 18 instead of 16 columns, improved matrix scanning speed, as well as additional Vial features, including layers and macros, that did not fit in the ATMEGA32U2 used in the xwhatsit controllers. The other keyboards that do not require 18 columns (B62, BSSK, etc.) will continue to use the xwhatsit controllers until stock runs out, as I've noted on the forums. I stocked up during the chip shortage many years back, before the Leyden Jar was released.
• Improved QC - no one is expected to need to adjust the bend of the metal parts inside the module due to a factory QC error with some of the original modules that was disclosed at the time. The most common issue is just a flyplate (the metal part connected to the flipper) that gets separated from the rest of the module while in transit. This QC improvement was introduced with the second batch of Round 1 beam spring keyboards which have been in use for a couple years now without major reports of failures.
• Reattaching the flyplate is now even easier than both the IBM and Round 1 originals where you had to use tape and a paperclip. The modules now have a custom-molded press fit plastic washer on top so you can quickly and easily disassemble the module by hand and reattach the flyplate (or replace a damaged inner plastic part). This was also introduced with the second batch of Round 1 beam spring boards which used a glued-on metal washer for each module. Also in general the beam spring is much easier to take apart compared to a Model F; you only need to unscrew the bottom inner assembly plate instead of using pliers to separate it. You can even leave the keys installed (except for the modules that you need to remove to fix, of course). For these reasons, resolving issues requiring disassembly that come up over the years might be even easier and quicker to repair than many issues of the Model F.
• The boards still come with a mini-first aid kit of 4 modules in anticipation that some may break while in shipping, though many folks order the 12 module full first aid kit.
And as always, for folks who have been waiting a while now and want to see the reasons behind the extended production timeline (and for everyone who is curious about the Beam Spring keyboard production journey), everything is explained in detail on the updates page on the project website.
The aforementioned beam spring Round 2 videos:
