F104+SSK+122+62+77+50+Ergo orders now open! New Kishsaver+Industrial Model F Keyboards

[modology]

Damn, at this rate Ellipse gonna put Unicomp in shame

[RoastPotatoes]

Is the tray they're mounted on from the prototype, or is it the one from your F77?

A note - the bottom inner assembly pictured with the new PCBs is from an original F77 - the prototypes have not yet gone into production.

Thanks! Didn't read carefully enough, was too excited

This is some nice dedication to the project.

[andrewjoy]

Of course the pads are not covered in soldermask.

dont IBMs have mask over the pads ? As far as i am aware they do but i could be wrong.

[Ellipse]

Sorry I meant the ground pads and controller holes do not have soldermask. The key pads and rest of the PCB have soldermask.

[XMIT]

If you want to see an example of RAL 7030 on an F107 please have a look here:

http://deskthority.net/photos-f62/ibm-4 ... 11432.html

RAL 7030 is not quite Industrial M gray.

Is "unpainted" an option for those of us with local powder coating facilities?

[Ellipse]

XMIT - yes I will try to make unpainted an option.

[MECHGAMER]

where did you get the pcbs from? can you give me their email?

[0100010]

Yes clear soldermask everywhere so it is protected and you can see through the PCB and see the copper traces on both sides. The ground pads and controller holes do not have soldermask.

010-allow me to quote the PCB designer to answer your question:
Xwhatsit's controller has a trace (and two resistor arrays) on the top of the board that go under the matrix "rows". Because of this, the rows have to come to the controller from the bottom of the board (where there is no obstruction).

So the answer is the row traces are routed through vias to the back of the PCB, due to some potential interference from the controller? How are you planning on attaching the controller to the PCB? It seems to be no issue with the original PCBs.

[Ellipse]

Yes. The new PCBs will be integrated with the xwhatsit controllers on the same PCB, not as separate PCBs for most designs (maybe not the F77 HHKB style shift for example). Due to the layout of the controllers, those described changes needed to be made. The new design fortunately works fine as I tested the prototype PCBs yesterday.

[Ellipse]

An update - the top and bottom inner assembly, outside Stryofoam, and box samples have been ordered! I also ordered the F77 split right shift and F62 split right shift top inner assembly samples.

The F77 and F62 case prototypes were ordered earlier this week.

Tomorrow I hope to pay for the barrel and flipper molds and enough samples for an entire F77 and F62. I decided against CNC milled plastic samples (way too expensive due to the intricate detail in each barrel), and I decided against 3D printing due to the tolerances not being good enough for this project.

The spring factory has a spring force testing machine and they will be testing each of the thousands of springs to make sure they are within the original 1980s spring force spec tolerance range. The spring tolerances I'm aiming for are two times tighter than the currently produced Model M springs, meaning the buckling springs will hopefully be slightly more even than from a new production Model M that you could buy today. Attached is a photo of original Model F and M springs and flippers (pivot plates) for reference, with some blue Industrial SSK keys in the background

20150725_225756.jpg (538.41 KiB) Viewed 5929 times

[Ellipse]

The barrel and flipper (pivot plate) molds and samples have been ordered! The complete prototypes are expected within 5-6 weeks, and then orders will open up (assuming the units test well).

[Redmaus]

Ellipse, is it too late to change my F62 order to a F77?

[Ellipse]

Nope I will note it. You can change anything up until you fill out the order form and send payment in October.

[idollar]

Nope I will note it. You can change anything up until you fill out the order form and send payment in October.

Note me down also, please.

[idollar]

Tomorrow I hope to pay for the barrel and flipper molds and enough samples for an entire F77 and F62. I decided against CNC milled plastic samples (way too expensive due to the intricate detail in each barrel), and I decided against 3D printing due to the tolerances not being good enough for this project.

I am curious, which material are you aiming for the 3D printing ? It shall be able to change the capacitance of in the PCB.
Have you checked that this material will be within the limits of allowed by the xwhatsit controller ?

These flippers may be useful for other purposes, but if the capacitance change is not similar to the original Fs:

1.- individual replacement will not be possible in regular Fs with adapter (F122 + Soarer for instance). Any difference in the closed capacitance value, due to a deviation from the nominal-built value or error, will lead to a non-working key press.
2.- individual replacement will not be possible with xwhatsit. One will need to replace all of them provided that the xwhatsit can sense the new flippers. I am assuming that this will be the case as if not, this project will not work. Note that this option is also function of the new PCB to be created.

Am I wrong ?

[Muirium]

Sounds like he's using the ultimate 3D print material: none at all!

I wholeheartedly agree with this position. 3D printing is fine for prototyping duties like checking geometry, but not acceptable for anything final that needs to stand the test of time.

[idollar]

Sounds like he's using the ultimate 3D print material: none at all!

I wholeheartedly agree with this position. 3D printing is fine for prototyping duties like checking geometry, but not acceptable for anything final that needs to stand the test of time.

But I am going even further, I am not just asking about the quality of the 3D print. It may be sufficient for prototyping as you said. My point is that the material should be analysed. It shall be able to change the capacitance. Pure plastic does not. so I wonder what he is thinking to do about it. The solution shall combine physical and electromagnetic properties.

[Ellipse]

No 3D printing will be used, not even for the prototypes. I will be using a specialized resin for injection molding that contains a good amount of carbon so that it is capacitive. I will use another material for the production run if the prototype resin is not satisfactory. The prototypes will be made with the full production molds I ordered. The prototypes should be done in 5-6 weeks.

[Techno Trousers]

Idollar has a good point though... Will you be testing one of new pivot plates in an old school Model F to ensure they are interchangeable and work together? That would be really useful for spare parts.

[Ellipse]

Yes they will be tested in an original and new model f.

[fishcola]

This is incredible! I *was* saving up for my first Topre board as a birthday gift for myself, but that's just can't happen anymore. I'm torn between the 62 and the 77... The 77's numpad/arrow cluster can be reprogrammed with a xwhatsit or soarer? I need to familiarize myself with those controllers...

[Ellipse]

Yep the xwhatsit controller being used here is fully programmable.

[fishcola]

I'm curious, some pictures in the kishsaver photo thread show ANSI/ISO enters... I'm assuming this is for ANSI ver., but I noticed the R5 backspace is sometimes replaced by 2 x 1u buttons, remapping backspace to where it is on the HHKB.. How will the modifiers be handled on these boards? I'm probably missing something, but I'm at least aware that wider modifiers have two prongs, are these cases of switch swapping?

[chzel]

In their infinite wisdom, IBM decided to make the barrel plates almost universal, helping us along the way!
All mods larger then 2u actually have two independent barrels. With contacts underneath. So to split the Backspace, you open it up, install the extra flipper, and close it back up!
ANSI to ISO is the same process, with the ANSI "\|" becoming the top of the ISO Enter and the left side of the ANSI Enter becoming the key next to the ISO Enter.

[Elrick]

Signed on for ONE F77 only because I heard such nice things about Model-F's in general and would love to own one keyboard, built to Ellipse's standard.

Plus Ellipse seems to be a serious individual here, so anything connected to him looks to be a 100% certainty.

You are the ONLY one doing this in a sea of CherryMX and ALPs like keyboards, so it's nice to see someone with a real fetish for ancient IBM switches.

[Techno Trousers]

^^ I totally agree. This is a really special project. I've tried pretty much every type of mechanical switch ever created, and if I were forced to only use one switch type for the rest of my life, it would be the Model F capacitive buckling spring.

So far Model F users have been limited to people who are willing to find an F AT and put up with that super heavy space bar, or put in a fair amount of work to swap controllers or put in a Soarer's. This will be a pretty simple way of getting an easily reconfigurable layout, and it should make those Korean forums envious!

[Ellipse]

Thanks Elrick for the vote of confidence! Now we are just waiting for the prototypes I ordered to arrive in hopefully 5-6 weeks (maybe a little longer).

119 keyboards' worth of interest so far

Given this project is likely going to involve ~$40,000+ I am going to be very careful that this lives up to the F62 / F77 name. The flippers are the big question mark. I have discussed recommended barrel and flipper materials with one of the Unicomp old timers. One of the early decisions I had to make was whether or not to electroplate the flippers to get them to be capacitive; they confirmed that the original flippers were not electroplated.

At this point I expect to do some testing for a week or so on the prototypes when they arrive (as well as some HD video and sound recordings with some nice XLR microphones), and then open up orders for a month or so.

[Elrick]

At this point I expect to do some testing for a week or so on the prototypes when they arrive (as well as some HD video and sound recordings with some nice XLR microphones), and then open up orders for a month or so.

I think you'll hit the 200+ mark easily as everyone swarms in to witness true history here on DT and on GH.

Because this might be the ONLY time a real Model-F like keyboard is being manufactured for keyboard fanatics everywhere, then do expect a lot of interest to occur. You actually might be the ONE to turn me into a rabid IBM switch zealot and that is what scares me the most.

[romevi]

Bah. Can't decide between a 62 or 77. I plan on getting an F AT, so I'm not sure whether I should get the 77 for that right cluster or get the more niche 62.

[macmakkara]

HMM... This idea just come to my mind... Would it be possible that in future we could see model f AT compstible metal case? :p