March Update: Dye sublimation, shipping timeline, technical tips, third party video reviews, QC notes

Dye sublimation update:

I have successfully completed the dye sublimation setup! Below is a photo of the first sublimated key with the new setup. These are among the best results in any tests so far.

Upon close evaluation with my Meiji EMT stereo microscope, the sublimation is close to that of my favorite F122 from 1984. But I do not think we are yet at the point to do the comparison “can you tell the difference between which is new and which is the original?” as with the keyboard case finishes a while back (I think some people can tell the difference below). Therefore, the below photo is not indicative of the production legends.  So far there is no consensus on the forums which one is which and a number of people are on each side (I am glad to hear that!).

As a recap, the decision to dye sublimate the keys myself was because of the lower quality and high unit cost of the sublimated samples I received as well as the industry professionals telling me they no longer have the equipment they used for key sublimation (in some cases) or are unwilling to sublimate non MX keys (in other cases). The goal was XT quality sublimation for one piece Model F keys. I failed to achieve consistent sublimation results over the past 10 months of testing and had been circling back and forth with a number of sublimation industry professionals, engineers, and one of the original Model F keyboard production supervisory engineers who has been a significant help to the project.

I’ve had to redesign and switch out much of the setup. I built some of it myself and had some of it built in the US while the CNC milled jigs were done in China. In the end I had to apply advice from different industries and online forums discussing a few industrial processes to get it to work. Even the smallest things caused trouble (one example: I thought one of the sensors was bad but it just didn’t like switching power supplies so I had to get a linear regulated PS). My initial control panel setup involved analog devices only but that wasn’t sufficient, so I had to design and build a digital control panel.

Hopefully I am not announcing success too soon (!); the next steps are additional sublimation tests to evaluate consistency of results, fine tuning the sublimation parameters, and then tests of a full key set and front printed keys. If all goes well I can order the sublimation transfer sheets and begin mass sublimation. Thanks again to everyone for being patient with this final part of the project.


Shipping status update and timeline:

So far 140 Brand New Model F Keyboards have shipped out of about 2,050 keyboards ordered so far – every complete order that can ship, has shipped (not all unprinted keys have shipped as many are in the second shipment of the early bird round, expected in a couple months, or they have printed items in the order).

If your order has not yet shipped, I will note again that I cannot yet estimate when each keyboard will ship because I don’t yet know the pace of sublimation and order testing/fulfillment.


Additional technical/troubleshooting tips:

A note for those updating their function layers and keyboard layouts: the Fn key should have Fn1 selected in the drop down menu on both the Base Layer and Layer 1 of the xwhatsit GUI. If it is not selected in Layer 1 as well, you may get unwanted key presses.

I have added visual representations of the layouts and function layer to each of the four product pages to help everyone choose.

In addition to the methods described in the “QC secrets” for fixing buzzing springs, I have found that sometimes just stretching the spring and then reinstalling the key reduces the buzz – the buzz may have to do with a length of spring slightly too short for a particular key, as well as a spring end that should be a bit closer to the flipper nub.


One of the first third party video reviews has been posted on YouTube:

For this reviewer, Joc’s updated firmware with the alternative debounce threshold 11 was optimal while debounce 6 had some issues with pressed keys. The F62 was connected to a 2016 Dell laptop for this test. I don’t know the reason for the need to use threshold 6 vs. 11 but my guess is that the capacitive sensing is more sensitive to minute differences in voltage output or USB power management and that affects a small, but significant, percentage of xwhatsit controller users based on the connected equipment. If anyone has more details on this please do share!


A note on quality control, especially for those not as familiar with hardware manufacturing:

There is a QC process with dozens of steps that I go through for each new Model F keyboard – it is one reason it takes more time than expected for each one to go out. The factory also goes through this QC process before me on each keyboard and I have asked them to improve on this for the final round.

Everyone’s keyboard ships only after I test each key and fix any and needed springs or other issues.

And again my apologies that orders only started shipping last year. I know it is a long time to wait for anything.

For those on the fence about ordering a new Model F, I am getting many emails and PM’s from those who are happy with their new Model F keyboards – if you like the original Model F you’ll like the new ones! For those who have received their new Model F please do post on the forums Deskthority, geekhack, and reddit, even if you are happy so far with your Model F!

Even IBM had to perform quality control on their Model F keyboards – a 1980s IBM newsletter article posted on one of the forums a year or so ago describes how employees would use a microphone and headphones to listen to the click of a key and fix the keys that don’t sound right. You are correct that the Model F requires a higher level of quality control than other mechanical switches and rubber dome keyboards, but it is well worth the effort on my part to get the QC right, for the ability to type on a new Model F as a daily driver keyboard.

I recommend listening to the video in one of my earlier blog posts comparing a new Model F after my QC to 3 original Model F keyboards. Having restored dozens of these keyboards over the years I can say that the new Model F performs and sounds on par with many original Model F’s.