What causes the difference in key feel between IBM Model M and F?

[sypl]

It seems most people consistently rate the Model F key feel higher than the M. I've never tried an F, but I am curious as to what exactly is it that causes the different key feel. Do people have theories? Is it the chassis? The bigger flipper? Are the springs different? The metal mount plate?

Here's a reminder of their force curves:
https://plot.ly/~haata/179/ibm-membrane-buckling-spring-1989/#/
https://plot.ly/~haata/200/ibm-capacitive-buckling-spring-f-at/#/

They're broadly similar but the F actuates later and has greater overall key travel. How does that happen, btw? The barrel lengths are the same, are they not?

Not trying to start a fight here, I'm just interested in the why.

[Techno Trousers]

The barrel dimensions are the same, but everything else is different. F springs are lighter and the flippers are longer. The F has individual barrels in a metal plate, and the M has them all molded into a single plastic plate. I think the main design driver that caused a difference in feel is the need for M flippers to hit the membrane with a particular amount of force, rather than just getting the flipper down in order to register capacitance in an F.

I'm surprised, though, that they ended up signing off on the membrane buckling spring design as it exists. It's such a step down in feel from the F. But maybe they had no choice-- it's pretty obvious that keeping the same keys, barrel dimensions, and curve was a priority.

[fohat.digs]

The Model M was designed as an extreme cost-cutting measure and could obviously be built for a fraction of the cost of the F.

I believe that the difference comes from the fact that the F has multiple steel plates in various states of tension and compression that make it feel so alive.

[sypl]

The barrel dimensions are the same, but everything else is different. F springs are lighter and the flippers are longer. The F has individual barrels in a metal plate, and the M has them all molded into a single plastic plate. I think the main design driver that caused a difference in feel is the need for M flippers to hit the membrane with a particular amount of force, rather than just getting the flipper down in order to register capacitance in an F.

Thanks for your answer. It makes sense that a lighter spring would give a later actuation point.

Regarding needing greater pressure to activate the membrane, I wonder if anyone has tried an F spring in an M keyboard to see if it fails to activate? Would be an interesting experiment.

I'm also curious as to why it had to be a membrane. Is it really that much cheaper? Are the capacitive pads used in the F not also printed on a flexible plastic sheet?

[Findecanor]

I think the main design driver that caused a difference in feel is the need for M flippers to hit the membrane with a particular amount of force, rather than just getting the flipper down in order to register capacitance in an F.

In the Model M's flippers, it is not the front of the foot that compresses the membranes but rather a cam (a protrusion) on the bottom in the middle. The cam in the middle presses down the hole in the middle membrane, whle the sides of the foot rest against where there are three layers of membrane — and for that the foot only needs to tilt forward under tension.
The front of the foot (the "flipper") is only to stop it from rolling further forward, and for that it does not need to be as long as on the Model F.

See closeups on Sandy55's page.

[fohat.digs]

You can put a Model M spring onto a Model F pivot plate and it will work.

As soon as you have taken both models apart and seen the construction, you will understand that these are completely different animals.
 

[Techno Trousers]

In the Model M's flippers, it is not the front of the foot that compresses the membranes but rather a cam (a protrusion) on the bottom in the middle. The cam in the middle presses down the hole in the middle membrane, whle the sides of the foot rest against where there are three layers of membrane — and for that the foot only needs to tilt forward under tension.

Thanks for the clarification.

[sypl]

You can put a Model M spring onto a Model F pivot plate and it will work.

As soon as you have taken both models apart and seen the construction, you will understand that these are completely different animals.

I actually meant the other way. If you put the lighter F spring in an M, will it still work fine? And how much does it change the actuation point and key feel?

[sypl]

In the Model M's flippers, it is not the front of the foot that compresses the membranes but rather a cam (a protrusion) on the bottom in the middle. The cam in the middle presses down the hole in the middle membrane, whle the sides of the foot rest against where there are three layers of membrane — and for that the foot only needs to tilt forward under tension.
The front of the foot (the "flipper") is only to stop it from rolling further forward, and for that it does not need to be as long as on the Model F.

See closeups on Sandy55's page.

That page is great! Thank you. I see I was wrong about the F PCB. It's a curved fiberglass PCB. That can't have been easy. I imagine it'd be simpler nowadays with an aluminium PCB, but how on earth did they do that back then?

[fohat.digs]

You really need to take these keyboards apart and look at them. Blind guesswork will not reveal their inner complexities to you, and describing things in words is inadequate.

The fiberglass PCB is flat but flexible and is bent into curvature in the middle of the "sandwich" that includes the other plates as they are asembled.