(Model MF) Remodeling the Model M (aka.. the Mara)

gianni

02 Aug 2020, 18:52

zrrion wrote:
02 Aug 2020, 18:20
I imagine you could adhere some foil to the flipper to make it capacitive without having to mess with custom plastic flavors
As a test or as a definitive measure? That would bring the reliability from 100 millions actuations to 100.

gianni

02 Aug 2020, 18:53

Hypersphere wrote:
02 Aug 2020, 16:35
@SneakyRobb: The shape of the flipper looks good. However, does it have capactitive properties? The model F flippers have either carbon black or short carbon fibers added to the plastic to impart electrical capacitance. Have you added either of these ingredients to your resin?
He said that he was going to add some graphite to the mix. About 20%

User avatar
DMA

03 Aug 2020, 03:56

gianni wrote:
31 Jul 2020, 09:42
I don't know how and why you think that you'll be able to succeed where other people have found big difficulties that make a project like this a really big task that involves engineering, programming, computer science, while requiring a lot of passion and dedication. And a good mood too, to deal with people that treat you like a clerk in an convenience store. The best that you can do is join this project, and when it will be successful, making all the possible layout variations will be the obvious next step.
What a way to discourage people from doing things. Erisie, demonstrate the guy yer middle finger and proceed. Just plan for several iterations of PCB layout. CommonSense is much more forgiving - but no QMK, ever. Make your choice :D
gianni wrote:
31 Jul 2020, 09:42
For example making a curved plate isn't as simple as it seems. People did it but it required a lot of effort, since the two plates need a different curvature radius in order to keep contact with the foam.
except..
* plates don't need to be curved (see wcass' gorgeous CNCed beast of a keyboard)
* plates don't need to be metal - just strong enough to withstand the force of the springs. Acrylic is A-OK.
* foam is not indispensable - it only affects the sound. Note the word. Not "enhances". Not "improves". Just "affects". Nothing more, nothing less.
* Ground plane is still required under the pads. Tin foil is perfectly adequate for the task, just put a mylar sheet on top of it. Should you want sometihng more substantial - again, wcass has shown us the way of 4-layer PCB (borderless, too).

kmnov2017

03 Aug 2020, 08:00

Fully agree with DMA on this.

Troy Fletcher also made it out of acrylic and had a flat plate.

I don't have foam on one of my boards - I think it sounds even better. But that's a matter of personal opinion.

Also, it's worth mentioning - one can use conductive paint on the those 3D printed flippers.

gianni

03 Aug 2020, 12:48

DMA wrote:
03 Aug 2020, 03:56
gianni wrote:
31 Jul 2020, 09:42
I don't know how and why you think that you'll be able to succeed where other people have found big difficulties that make a project like this a really big task that involves engineering, programming, computer science, while requiring a lot of passion and dedication. And a good mood too, to deal with people that treat you like a clerk in an convenience store. The best that you can do is join this project, and when it will be successful, making all the possible layout variations will be the obvious next step.
What a way to discourage people from doing things. Erisie, demonstrate the guy yer middle finger and proceed. Just plan for several iterations of PCB layout. CommonSense is much more forgiving - but no QMK, ever. Make your choice :D
gianni wrote:
31 Jul 2020, 09:42
For example making a curved plate isn't as simple as it seems. People did it but it required a lot of effort, since the two plates need a different curvature radius in order to keep contact with the foam.
except..
* plates don't need to be curved (see wcass' gorgeous CNCed beast of a keyboard)
* plates don't need to be metal - just strong enough to withstand the force of the springs. Acrylic is A-OK.
* foam is not indispensable - it only affects the sound. Note the word. Not "enhances". Not "improves". Just "affects". Nothing more, nothing less.
* Ground plane is still required under the pads. Tin foil is perfectly adequate for the task, just put a mylar sheet on top of it. Should you want sometihng more substantial - again, wcass has shown us the way of 4-layer PCB (borderless, too).
You're just unpolite and detached from reality. I'm trying to be realistic and to save a guy with good intentions some delusions. He doesn't even know how to make a pcb, which is the simplest task. He pretends to do alone what the most experienced guys here have failed to do collectively. Surely if he spends a lot of time trying he'll succeed, but I wonder why he shouldn't just contribute to this project and then adapt the outcome to his layout. He's just trying to fork this project into a new one without knowing anything. Since I respect this guy and I don't like to make people waste their time, I've politely pointed out that his task is very difficult. On the other hand you're having fun looking at a guy that tries to do completely alone a very complicated project, forcing people that are working toward another objective to help him, hijacking a thread that his devoted to make a model m with a standard layout.

User avatar
DMA

04 Aug 2020, 21:37

gianni wrote:
03 Aug 2020, 12:48
You're just unpolite
Guilty as charged.
gianni wrote:
03 Aug 2020, 12:48
and detached from reality.
Do you have a psychiatrist license to diagnose psychosis? If not - care to apologize?
gianni wrote:
03 Aug 2020, 12:48
I'm trying to be realistic and to save a guy with good intentions some delusions.
Suuuure.
gianni wrote:
03 Aug 2020, 12:48
He doesn't even know how to make a pcb, which is the simplest task.
Neither did I, four years ago.
gianni wrote:
03 Aug 2020, 12:48
He pretends to do alone what the most experienced guys here have failed to do collectively.
And what is wrong with that, exactly?
Remote collaboration bears a huge overhead cost. I vividly remember frustration of laying out CS-MF not knowing how much vertical clearance I have between me and CONDUCTIVE bottom plate. Fifteen-minute discussion with everything on hand. Impossible otherwise.
gianni wrote:
03 Aug 2020, 12:48
Surely if he spends a lot of time trying he'll succeed,
Yeah. I ended up spending $3k on a single keyboard. But it's on my desk and it gives me joy every day.
If I listened to likes of you back then (there's an ample supply of your kind at all times, 2016 wasn't an exception), I'd have a mitsumi(or dell) rubber dome instead - because I'm not crazy to pay $400 for a keyboard. And there would be no wingnut on my desk, that's for sure.
gianni wrote:
03 Aug 2020, 12:48
but I wonder why he shouldn't just contribute to this project and then adapt the outcome to his layout.
Becase his is a _completely_ different project, he won't be able to reuse ANY parts. The only common parts are barrels, flippers, and keycaps. All of which are available commercially as of today, in virtually unlimited quantity.

Besides, you just accused him of being useless a sentence ago. What can he contribute if he doesn't even know how to do "the simplest task"? Money?
gianni wrote:
03 Aug 2020, 12:48
He's just trying to fork this project into a new one
Fork? Read the message, he's sanity-checking his plans.
gianni wrote:
03 Aug 2020, 12:48
without knowing anything.
What exactly is wrong with that, again? Everybody needs to start somewhere.
gianni wrote:
03 Aug 2020, 12:48
Since I respect this guy and I don't like to make people waste their time, I've politely pointed out that his task is very difficult.
No, you were just being a jerk. "I don't know how and why you think that you'll be able to succeed" is far from polite and a light years away from respectful.
gianni wrote:
03 Aug 2020, 12:48
On the other hand you're having fun looking at a guy that tries to do completely alone a very complicated project
It's not "very complicated".
He has the PCB already (well, he still needs to export to gerbers and place an order, but that's half an hour tops).
If he won't make a mistake of using xwhatsit for the prototype, all that remains is an acrylic plate with like 120 holes, another one with like 10, dozen bolts with nuts and a controller.
I don't know how to cut acrylic, but we can always ask wcass how he did his shiny battleship and I'm pretty sure he'll be glad to explain (knowing him - he may even produce the draft model for the cutting shop).
And if he decides to use CS and will want it in custom form factor - I'll gladly design the PCB and even assemble the controller if he ships that PCB and components to me.

It would definitely not be cheap due to Ellipse gouging $1 per switch out of the community. But complicated? Hell no.
gianni wrote:
03 Aug 2020, 12:48
forcing people that are working toward another objective to help him
Waitwhat? forcing? I don't see any guns pointed to anybody's head.
gianni wrote:
03 Aug 2020, 12:48
hijacking a thread that his devoted to make a model m with a standard layout.
It was, indeed. Four years ago. Even two.
But this year it's a discussion of how to make flippers - not because the technology is lost, but because Ellipse refuses to sell them at a reasonable price.

SO. Stop being a jerk, apologize to Erizie, and NEVER AGAIN discourage people from trying a new hobby. It's a friggin HOBBY, the journey is as important as the destination, if not more.
It's not your money, it's not your time. If you have something positive/supportive to say - say it (preferably in a non-condescending manner, and avoiding any commitments or setting any time frames.). If not - JUST DON'T. Not your business.
If _you_ _personally_ walked that particular path and want to share your findings - make a post (you should've done it back then!) and refer to it.

A friendly reminder to all of you
Please post about things you tinker with and post as you go, not when it's all 100% complete (unless you're planning to make money of it, hehe). Those posts are proverbial Shoulders of Giants. You - you personally - are The Giant. Do your part.
No matter your experience level, no matter how stupid/trivial things you're doing - post the progress. If you don't - that knowledge you gained today is lost in a year. Like tears in the rain and all that.

Erisie

05 Aug 2020, 00:46

I just want to say this: thank you, thank you, DMA.
DMA wrote:
04 Aug 2020, 21:37
And if he decides to use CS and will want it in custom form factor - I'll gladly design the PCB and even assemble the controller if he ships that PCB and components to me.
See? I wasn't even aware of the existence of the CommonSense controller until I read it here. I thought my options were limited to QMK and xwhatsit.

And yes, I am perfectly aware that the only thing I would be able to reuse from my existing Model M is the case: everything else would have to be created (PCB, controller, top and bottom frames) or bought (hammers, keycaps, barrels).

gianni wrote:
03 Aug 2020, 12:48
He pretends to do alone what the most experienced guys here have failed to do collectively.
Gianni, question: if I wanted to "do [it] alone", would I use an existing design as a template? Would I come to a forum to ask for help on the project?
gianni wrote:
03 Aug 2020, 12:48
forcing people that are working toward another objective to help him, hijacking a thread that his devoted to make a model m with a standard layout.
I have never done any such thing. All I did was asking whether my modified PCB layout design could work, and, if not, what changes were needed. At any point I have "forced" or "hijacked" anything at all. So stop saying that.

User avatar
wcass

05 Aug 2020, 03:47

A basic principal of capacitive and contact keyboard matrix design; rows and columns should intersect exactly once. If they intersect more than once, you will have two switches that produce the same result (at best) or two that do nothing (at worst). Your 121 key matrix will likely be 6 by 21 or 22. If you have a spreadsheet app, lay it out ... rows 1 through 6, lettered columns A through U or V... it is OK to have unused cells, but you can't have two switches in one cell. You can have two functions for one switch via layers, but the switches themselves must be unique.

That is a good starting place.

Erisie

05 Aug 2020, 04:20

wcass wrote:
05 Aug 2020, 03:47
A basic principal of capacitive and contact keyboard matrix design; rows and columns should intersect exactly once. If they intersect more than once, you will have two switches that produce the same result (at best) or two that do nothing (at worst). Your 121 key matrix will likely be 6 by 21 or 22. If you have a spreadsheet app, lay it out ... rows 1 through 6, lettered columns A through U or V... it is OK to have unused cells, but you can't have two switches in one cell. You can have two functions for one switch via layers, but the switches themselves must be unique.
That is a good starting place.
Gotcha.

I was checking out i$'s FEXT design again, and I noticed that the reason that you see contiguous column traces is because the row traces are a total clusterf**k (obviously, a necessity in order to work around the original M mounting holes): a single PCB row in his design spans more than one "physical" row, moving up and down the board in order to fit the design. Since I'm using PCB rows that are parallel to the physical key rows, I can't have that workaround.

I had made the dummy mistake of thinking that you could share more than one row contact in a single column (which, for what little I know about these things, didn't make sense to me while working with the design). It's an easy correction from there. I'll post my corrected design as soon as it is done.

Thank you!

gianni

05 Aug 2020, 10:13

DMA wrote:
04 Aug 2020, 21:37

I won't read what you wrote, this is just pure madness. OMG these two guys :lol: :lol: :lol: You're just off topic and you're trying to fork a project forcing people to look at your very wrong stuff. Good night.

User avatar
wobbled

05 Aug 2020, 14:58

could anyone sum up this drama so I don't have to read the equivalent length of the bible? Thanks

gianni

05 Aug 2020, 15:29

wobbled wrote:
05 Aug 2020, 14:58
could anyone sum up this drama so I don't have to read the equivalent length of the bible? Thanks
Just two guys trying to divert the thread from remodeling the model f to making from zero a model f with an ortholinear layout and asking for help here, publishing pcbs made with photoshop, with so many errors that nobody wanted to look at those drawings.

This is offtopic and should stop.

User avatar
ppCircle

05 Aug 2020, 19:00

zrrion wrote:
31 Jul 2020, 00:21
I would think that something like that, but with plastic studs for the pivot wouldn't bee too difficult to do, and would hopefully smoother than an all metal part.
It's actually interesting, because they are insanely smooth. I need model f to compare but it's definitely better than model m.

Erisie

05 Aug 2020, 22:33

gianni wrote:
05 Aug 2020, 10:13
you're trying to fork a project forcing people to look at your very wrong stuff.
I am not forcing anyone to do anything. Do you even understand what "forcing" means?
gianni wrote:
05 Aug 2020, 15:29
with so many errors that nobody wanted to look at those drawings.
Unless they are "nobody" to you, I already got feedback from wcass and DMA on my idea.

Has anyone ever told you that you are a very impolite individual, gianni?

gianni

06 Aug 2020, 10:10

Erisie wrote:
05 Aug 2020, 22:33
gianni wrote:
05 Aug 2020, 10:13
you're trying to fork a project forcing people to look at your very wrong stuff.
I am not forcing anyone to do anything. Do you even understand what "forcing" means?
gianni wrote:
05 Aug 2020, 15:29
with so many errors that nobody wanted to look at those drawings.
Unless they are "nobody" to you, I already got feedback from wcass and DMA on my idea.

Has anyone ever told you that you are a very impolite individual, gianni?
stop with this spam show, nobody cares, write to me with a pm if you have something else to say

This is what someone else said about you. He says what I said, but that's ok for you.
"...There are actually too many things wrong for me to go into it. ....

The idea behind this thread is to take an off-the-shelf M and make it an F by swapping out just membrane, flippers, and controller. Changing the key layout to ortholiniear would require that AND a new barrel frame and probably back plate."

Really, stop this shit show, you're diminishing yourself.

Erisie

07 Aug 2020, 02:19

wcass wrote:
05 Aug 2020, 03:47
A basic principal of capacitive and contact keyboard matrix design; rows and columns should intersect exactly once. If they intersect more than once, you will have two switches that produce the same result (at best) or two that do nothing (at worst). Your 121 key matrix will likely be 6 by 21 or 22. If you have a spreadsheet app, lay it out ... rows 1 through 6, lettered columns A through U or V... it is OK to have unused cells, but you can't have two switches in one cell. You can have two functions for one switch via layers, but the switches themselves must be unique.
I followed your advice, wcass. I redesigned i$'s layout to make it work according to the specifications (six rows, 22 columns, for a total of 113 potential keys) and this is the PCB result:

Top/Bottom
Image

Top
Image

Bottom
Image

What else would it need to be done for this design to become viable?

Thanks again for your help.
Attachments
Top.pdf
(589.08 KiB) Downloaded 10 times
Bottom.pdf
(586.53 KiB) Downloaded 8 times
Top-Bottom.pdf
(592.37 KiB) Downloaded 10 times

Hanslau

24 Sep 2020, 09:57

It's unfortunate this project never came to be.

pandrew

30 Sep 2020, 10:37

Fricked is doing something interesting here, make sure to fill in his interest check survey:
viewtopic.php?f=7&t=24559

Post Reply

Return to “Workshop”