understanding key calibration

[cobalt sail]

Reading in a few other posts / messages in this server that a calibration process is executed every time the keyboard is power-cycled.

I'm noticing that after that initial calibration a majority of keys on my board register 100% actuation at around 85% actuation physically. However, after a key is fully pressed, the key seems to register correctly throughout the entire range with every subsequent press.

It would seem that keys are calibrated in some way every time they are pressed?

A bit of a nuisance that full calibration requires pressing every key first, especially if you have the keyboard connected to a USB switch. Obviously I don't understand the inner workings, but I think I'd prefer that calibration be stored manually in the profile (store current calibration button or something) with a way to manually recalibrate when needed.

[haughty monolith]

As far as I know they are working on a different method for the calibration. Most likely to avoid the two second phase after the board got power. At the moment only the resting position of the key should be calibrated while end point is fixed, as far as I know. The end point is likely determined by the sensor itself during production. The thickness of the PCB is known and constant. The sensor is kind of reversed, the voltage gets lower the stronger the magnetic field is. Meaning there is a limit where it reaches zero Volts and that is always 100%.

[cobalt sail]

My end point is definitely changing... upon boot my Fn key registers 100% actuation at 85% actuation physically. After pressing it down fully and releasing the key I get 85% actuation registered at 85% as it should be.

I'm verifying this with "Touch" lighting effect as well as the "Test your bindings" box on the Advanced Keys page.

I am waiting a full 5-10 seconds before touching any keys upon powering up.

[haughty monolith]

It would make sense that the highest value which can be archived would be used as the max. But so far I never had the case that a key was registered as 100% even it wasn't bottomed out.

[cobalt sail]

Here's a video example.

As can be seen around 85% actuation physically the lights jump to full, you can see them flicker from 85% percent straight to 100% as I bounce the key at that actuation. After pressing the key fully, that upper range (85-100%) has a smooth transition.

[haughty monolith]

What modifications have you done?
Is it the same when you use the module in the original case with the original foam?

[haughty monolith]

On which firmware version you are?

[rugged night]

I’m confused what’s the issue?

Is the keys not fully actuating or are they actuating before you hit them

[cobalt sail]

1. modified new case to fit board (used Dremel to remove extra standoffs)
2. swapped case
3. reused oem foam + added some dampening material
4. lubed and installed Zeal stabs
5. taped board
6. burger mounted o-rings
7. lubed switches
8. swapped keycaps
9. added dampening material to larger keys

[cobalt sail]

They fully actuate before the key bottoms out. After bottoming out the key once it looks like the detection range is readjusted. Seems like it can tell I went passed it's assumed max travel, and adjusts.

Basically initial calibration is poor, after pressing every key once calibration is adjusted and is good. But this resets every time the board is powered down, ie every time I use my USB switch.

I would assume this could be "fixed" by allowing users to store calibration values and re-calibrate manually when needed.

[haughty monolith]

Quiet some work. I guess the module itself is pretty much unmodded or? Like you haven't changed the foam between the mounting plate and PCB, right?

The tape mod can causes problems if there isn't enough room for the legs of the switches. It can be that the switch isn't fully clicked in and so would provide some flex which can cause input problems and a weird feeling.

Maybe @empty crystal can tell more about how the scanning calibration works and why you can reach 100% early and it seems fine after the first press.

[cobalt sail]

Didn't consider that, I guess I could make a bunch of holes lol

I'd be surprised if that was causing too much trouble though.

[cobalt sail]

Yeah stock foam between the board and plate.

[haughty monolith]

There was a user which had such issues due to the tape mod.

[haughty monolith]

So as long that isn't more compressed than the original assembly it should be fine.

[cobalt sail]

I only did the tape mod because I didn't want to short the board in the case, but maybe I can just tape parts of the case instead of the board.

[haughty monolith]

The foam in the case should already prevent any possible shorts.

[cobalt sail]

But this case has some extra stand offs that I couldn't Dremel out easily at the bottom, and little elevated metal lips around the outer edge.

[haughty monolith]

Right, they can be problematic.

[cobalt sail]

But yeah I doubt the tape mod is helping acoustically. I'll just tape off anything I'm worried about in the case.

But You'd think that tape on the back of the board would push the switch away from the PCB, which you'd think would have the opposite affect. Like impossible to achieve full actuation, in my case full actuation is occurring before the switch is bottoming out.

Either way it's worth a try.

[haughty monolith]

The plate can help even it may only changes the sound character. Just make sure the switches are fully inserted and that the tape isn't pushing the switch up via the legs.

[cobalt sail]

At the same time I still think an option to store your calibration would be useful seeing as people are doing mods and such.

[haughty monolith]

It just depends how well you have done the tape mod. I only inform you so that you are aware of potential issues.

[haughty monolith]

As mentioned as far as I know they are working on a better way of doing the calibration. Mainly to reduce the two seconds it takes after each power cycle.

This causes issues because people aren't aware of it.

[cobalt sail]

Yeah I can see why people get confused. It's all very interesting, especially that they built in dynamic calibration during use. But it makes sense why'd you want that.

[empty crystal]

This sounds correct. The first time a key is pressed after calibration it figures out the end-point, as it doesn't know for sure if the last x% is reachable.
After the first time the key is fully pressed it will remember what the lowest point was and use that as the new end value.

I haven't read the rest of the thread, and what I just shared may not be completely accurate as I am simplifying what was already simplified before to me.
But all you should know, the calibration process is technically incomplete until you fully press down the key to the lowest point.

So assuming you have modifications such as o-rings, stoppers, anything else that blocks out full-switch range, those should be calibrated for after you press down on the key

[rugged night]

What issues

[cobalt sail]

Cool so it sounds like exactly what I was thinking.

I guess the unfortunate part with this system is that calibration is reset every time the board is powered down. For me this is a lot because I keep it connected to a USB switch. Would be cool if we could store our calibration settings manually. Sounds like you guys are looking into different options already tho.

[haughty monolith]

I would expect that due to the USB switch you would have less issues. As long the switch is active (which it should be due to the power demand of the Wooting).

[cobalt sail]

Every USB switch I've used cuts connection completely each time the device is switch. Evident by the lights turning on and off between switches. Switch is externally powered, but yeah that's always been the behavior on every one I've tried as well.

I don't think the USB protocol would allow for host switching without power loss unless the device itself is externally powered.

[haughty monolith]

My USB switch does the same thing and my LEDs also shortly turn off. But the device still gets power, the LEDs only went off for me because between the switch the device is put into sleep. The calibration takes around two seconds. So if you switch the output to an active host and the LEDs are only shortly off it would be fine. Only if it takes two seconds again it would mean there was a power loss.

It's not the USB data connection which causes the calibration but the power loss. So as long there is power provided by the switch it's fine.

[empty crystal]

I believe you can keep your USB devices in sleep mode even after shutdown
Probably somewhere in the windows power management settings

[haughty monolith]

Only works if the system is still on, meaning in a power state where USB is still powered. Otherwise some mainboards allow to keep the USB ports powered on even if the system is shut down.

But this is unrelated to the USB switch.

So far a USB hub should not handle the power state of their downstream ports depending on the power state of the upstream port.

[empty crystal]

I know mine stays powered since it stays on the same profile etc..

Plus the LED on my microphone stays on

[haughty monolith]

I guess such hubs rely on the USB specification what a connected device don't draw much power when in sleep.

[cobalt sail]

Interesting what switch do you have?

[haughty monolith]

https://www.aten.com/us/en/products/usb-solutions/docks-and-switches/us424/

Due to the power demand of the connected device I use it in an active configuration.