Chinese USB Hall Effect Keyboard - Review and Impressions

[XMIT]

Thats really interesting. Too bad its linear only, I wonder if we could replicate a bump adding another magnet to the sides.

Btw, a guy I know bought a similar board from aliexpress that has a switch that looks pretty much the same. BUT, looking closely, his does not appears to have a magnet, but two springs instead. Well, at least the two spare ones that came with the package, as I couldnt convince him to open his keyboard ;_;
Slider color was also brown, not black.

Here some pics and link to where he bought
http://www.aliexpress.com/store/product ... 55036.html

http://imgur.com/a/U4wqX

I just checked with the manufacturer about this.

The sliders are from the same manufacturer but the switches are not Hall effect based. I think they are simple contact closing mechanisms: the smaller conical spring bridges two connectors on the underlying PCB. It is cheaper to manufacture but not as awesome as Hall switches!

Mystery solved.

[andrewjoy]

Do we know when they will be available ?

[Chyros]

I'm really liking these 50g springs quite a lot. I don't usually go for linear switches but these are some of my new favorites.

50-55 g is the perfect weighting for linear switches IMO. Right on the sweet spot.

[Virtureal]

Great to hear that the ABS and acrylic cases are less problematic. Did anyone order a white/transparent acrylic case, I'd like to see more pictures of that, other than that I'd like to echo the opinion that I'd love to see smaller bezels (and therefore smaller T nuts or another mechanism to keep the case together).

[XMIT]

No one ordered anything but a black ABS case. Black and white are the options at the moment. I'm thinking that, for an initial group buy, we'll focus on ABS cases since they work so well.

I've got a number of e-mails from the manufacturer to work through and I'll post updates here later this week. But for now: the firmware problems are understood and we're working to fix them.

Also, the key slider itself is POM, and the slider housing is polycarbonate. The POM slider is the most important thing for good switch feel.

[matt3o]

As promised this is the detail of the solder job (zoom in)



Soldering is pretty amateurish (to say the least, you can't really defend it) and glue doesn't even cover the solder completely. I totally get this is a prototype but be sure the get their QC together XMIT!

[XMIT]

Yeah, that's pretty tacky. I confirmed that these boards were hand built, and that the production run will be done in an SMT reflow oven. But that doesn't address the cables.

They are willing to do detachable connectors so long as we accept that they won't be as waterproof.

The solder job on my 87-key ABS board wasn't nearly as bad. It is by all metrics a much more refined board. I really need to work on photos of that one.

While we're here matt3o, I've asked the manufacturer for an updated firmware for our 60% boards. I used the same layout as the KBparadise V60 and found places to put all of their existing LED controls as documented in the "manual". This place outsources the firmware development and is requesting a MOQ or an additional charge to bankroll things like a programming tool, extra features, etc. But the firmware already supports one Fn layer so simply doing these key mappings should be trivial. Let's see what they come back with.

Updated 60% layout. See http://bit.ly/2bY6f4C.

keyboard-layout (1).jpg (66.34 KiB) Viewed 7335 times

I'd love for the entire firmware to just be open - like keyboard.io's firmware or TMK or Kiibohd or whatever - but I don't think the manufacturer is willing to make that jump yet. I'll see what I can do.

[the_marsbar]

That looks quite good, but why not move up to Fn+{ (where LED- currently is)? That way, it would be similar to the HHKB. This would be essential for me at least.

Detachable cable would be a huge plus. I don't need to use my keyboards under water (I am really bad at swimming, and I try not to spill tea of coffee in my keyboards).

[Phenix]

Full-waterproof would make me really overthink a detached cable. A model where one can solder in the jack oneself could be interesting

[XMIT]

It's an arbitrary choice, but, I chose to mimic the V60 instead of the HHKB becase that's what I tend to use more often as a 60% board. Also, the V60 has the same layout, and, the layout I proposed doesn't require any key changes.

I could perhaps come up with another mapping that is more HHKB-like. But right now I'm trying to see if I can get a new firmware at all. If that is easy then I may also request another HHKB-like firmware and give the option of flashing one or the other.

[andrewjoy]

That soldering is not that bad for hand soldering without paste.

Try soldering SMD caps that are twice the size of the pads to a PCB by hand without paste , you will then know the true meaning of frustration ( i wanted tantalum SMD caps on my M2 so they cannot fail again .)

[XMIT]

I've added more photos of the ABS 87-key board to the Google Photos album:
https://goo.gl/photos/5ijmyf9RmHd9RYgE7

(Typing this back on the bamboo 87-key 70g board. I really do prefer the 50g springs, I'm going to switch the springs on this board sooner or later. The bamboo board feels much more substantial as does the acrylic board.)

Here are a few choice photos of the 87-key black on black ABS board.


Three-quarters view of ABS board with feet raised.


87 key ABS board overview - top view.


87-key ABS rear view showing feet.


Detail of feet on ABS board.


Disassembly of ABS board. It is held together with four screws and fourteen plastic clips. I lubricated the clips for easier disassembly next time.


Cable detail. The cable housing is glued to the shell. Simpler than the bamboo board.


Different stabilizers. I don't know why there are two types. The Cherry style stabilizers work great, they need to stick with those.

[matt3o]

thanks for the hard work XMIT.

honestly the whole "waterproof" thing seems a bit silly to me, not sure why I would need to scuba dive with my keyboard.

I had a look at the PCB, I believe it would be possible to remove the current controller and put a teensy (no backlight of course). I'm a bit busy these days, but it's definitely in my to-do.

[andrewjoy]

Thanks for the cool photos , as good as the ABS one looks i think i prefer the bamboo one , i also thing they need to do a premium aluminium one as well , to give it at 70s feel .

I do think the soldering needs work but as its pre production the wave / reflow soldering of mass production would solve that.

A No caps option would be good.

as for removable cable , i honestly don't care about full waterproof, just splash proof is good enough.

If its removable please push the HELL out of USB type C and failing that stay as far away from mini as you can , micro is ok , full sized B is better.

as for the firmware , can they not just leave that on TMK or something so we can mod away.

[Virtureal]

Yeah, that's pretty tacky. I confirmed that these boards were hand built, and that the production run will be done in an SMT reflow oven. But that doesn't address the cables.

They are willing to do detachable connectors so long as we accept that they won't be as waterproof.

The solder job on my 87-key ABS board wasn't nearly as bad. It is by all metrics a much more refined board. I really need to work on photos of that one.

While we're here matt3o, I've asked the manufacturer for an updated firmware for our 60% boards. I used the same layout as the KBparadise V60 and found places to put all of their existing LED controls as documented in the "manual". This place outsources the firmware development and is requesting a MOQ or an additional charge to bankroll things like a programming tool, extra features, etc. But the firmware already supports one Fn layer so simply doing these key mappings should be trivial. Let's see what they come back with.

keyboard-layout (1).jpg

I'd love for the entire firmware to just be open - like keyboard.io's firmware or TMK or Kiibohd or whatever - but I don't think the manufacturer is willing to make that jump yet. I'll see what I can do.

I think focusing on one type of case would be good, I don't mind either way for the cables.
For the firmware / keymapping have you sent that picture to the manufacturer? I have no idea what many of those legends mean and I doubt the manufacturer does either:
WASD
FPS
CF
COD
LOL
Car Race
Design
etc.

Any idea what microcontroller is being used in the boards at the moment?

[XMIT]

Virtureal, have a look at the photo album linked earlier and/or in the original post.

Those legends are from the manual. The keyboard firmware offers a bunch of gaming oriented LED controls that are described in this way. There are some photos of the manual in the photo album.

WASD - enter or leave a lock mode where the WASD keys are mapped to arrows.
FPS - illumate key caps in a manner helpful for many first person shooters
CF - same, for Crossfire
COD - same, for Call of Duty
LOL - same, for League of Legends
Car Race - same, for many car racing games

As for the microcontroller, I'm not sure, but I did take a really high quality macro photo of it. It has this text embossed upon it:

BYK870
NC837
1601

I have no clue what this is. But it has been in other keyboards before so perhaps we should try to figure that out. For example, someone on Reddit was curious too, as a quick Google search found:

https://m.reddit.com/r/AskElectronics/c ... r_is_this/

[Findecanor]

You do get more friction with off center key presses.

That's a shame ...

To be fair, this was a really aggressive, press the corner of the key cap sideways, press. I can't think of any switch - even Topre - that does well in this case.

OK. I was afraid that it was as bad as Cherry ML ...
I don't even like Alps partly because of how bad they are off-centre key presses. I am a bit sensitive about that.

Analog sensing is inherently possible with any Hall sensor but it depends on the amplifiers after the sensor outputting an analog level as opposed to a digital signal. I can ask but I believe this implementation uses digital logic coming out of the switches.

OK. I understand that it depends very much on the signal.

The reason why Topre can do analogue sensing is because the conical shape of the spring makes the signal have an almost a linear relationship to travel. If the spring is a regular coiled spring the travel/signal curve is more like a hockey puck and much less useful.

3. Stick with standard 6KRO USB for maximum BIOS/OS compatibility. If someone has a problem with this please explain to me why you need NKRO.

There are lots of threads here and elsewhere debating if NKRO is really needed or not... so I won't go into that.
But Soarer had found a way to have both. His PS/2 to USB adapter firmware uses it, so you could try it out for yourself by soldering together an adapter from a cheap Pro Micro.

Wow, some people really don't like the backlighting.

You either love it, hate it or don't care.
I personally don't think it is about backlighting itself - because it can be turned off, but how you make backlighting.

First of all, backlit keyboards often come with bad stock keycaps: Translucent ABS that have been painted black and with the legends lasered off.
The painted surface often does not feel good and these kinds of keys often wear the worst: more and more black paint is peeled away to reveal larger and larger glowing blobs of light.
If this new keyboard's caps are doubleshot, then that's great. But because doubleshot are still uncommon, backlighting still has a connotation as being cheap, because of all the keyboards witht bad keycaps.

I also don't like how backlighting is done on Cherry MX, which wasn't really designed for backlighting but for indicator LEDs in locking keys.
Because the LED is near the top edge, all symbols are scuffed into the top centre of the keycap which both looks bad and you lose the meaningful corner-based layout that has been the standard since the Model F. I wonder why no Cherry MX keyboard has had the LED on the left - then it could have used correct corner-based layout at least for ANSI.
I also find the light bleed to look like crap, especially when there is no side border around the keys.
I do like how Logitech and Omron with the Romer-G switches put the LED in the middle of the switch and without light bleed. Those are also able to illuminate the shifted symbols on the numeric row correctly, but they are still unable to do left or right-shifted illuminated legends.

I like it when backlighting is done in a smart way, though. Many gaming keyboards allow you to have different colour schemes for different games, highlighting those keys that are used in that specific game.
I also think there laptops where the light comes on automatically if the room goes dark.

[SL89]

This is my new favorite thread.

[shreebles]

Really interesting.

Replying just to subscribe

[webwit]

I don't get the fuss except as a superior and reliable activation technology. But it is not related to the implemented key feel, which is left open? Although all I've tried were butter smooth but boring (to me) linear implemented by a simple spring. I wonder how tactile will be implemented.

I'm not a fan of unicorn vomit led keyboards, but one advantage of a keyboard with full rgb led implemented is, if supported, that you could program stuff like status lights (caps lock and such) on any key you want, in white, red, pulsating vomit, whatever you like.

[XMIT]

Here's a thought experiment, motivated by my opening up my RC930 earlier to lubricate it: what would a Hall effect based RC930 look like? The sliders could have a little cylindrical indentation at the bottom for a magnet, the springs would not be there, and the aforementioned PCB could fit under it. From the top you couldn't tell the difference, and it wouldn't feel any different.

With this particular board I'm also intrigued by the construction. It lends itself nicely to easy case mods with the layered construction, and, individual switches can be removed without any soldering at all.

Stay tuned for some photos of the acrylic 60% version of this board, likely some time later this week.

[ohaimark]

I received mine yesterday evening -- way later than the usual delivery time. I feel bad for the driver. I'll get a review up this weekend.

Edit: Or on Monday. I always forget about my crappy home internet.

[Hypersphere]

@XMIT: I haven't followed this thread closely. It's interesting that someone is reviving Hall Effect switches. I have some questions -- apologies if these have already been asked and answered:

1. Would it be possible to put a contact leaf in the switch housing to produce a tactile or tactile-clicky switch similar to an Alps switch but retaining the Hall Effect sensing mechanism?

2. Could the case be designed to permit a detachable cable and still be waterproof?

3. Could the keyboard be made fully programmable?

If I were to go for one of these, I would probably be most interested in a 60% form factor. It would also be nice if it were fully programmable or in a HHKB layout. A compromise would be DIP-switch options for converting a standard layout to something close to HHKB.

[Findecanor]

If the shape of the switch housing and slider had been a bit different then it might have been possible to put buckling rubber sleeves in-between the housing and the keycap.
Rubber domes and sleeves are not always bad for tactility and they don't induce any friction as tactile leaves would.
Many Foam-and-foil and similar switches (such as Mitsumi "hybrid switch") have tactile versions that use them, and while those switches are often disliked, they are often disliked more for the mushy landing - which is because of another element of the switch.
If you take a Key Tronic foam-and-foil keyboard apart, place the switch plate on plinths and type on it with the foam having nothing to reach, it does actually not feel half bad.
I have long asked for a switch that would combine a rubber dome or sleeve together with a spring (in parallel) because I think it would be possible to produce a switch that would feel like a smoother Cherry MX Clear.

[XMIT]

The answers to all of these questions are buried somewhere in the thread but I'm happy to answer them again.

1. Would it be possible to put a contact leaf in the switch housing to produce a tactile or tactile-clicky switch similar to an Alps switch but retaining the Hall Effect sensing mechanism?

Yes. Sensing - a magnet moving up and down - is completely separate from tactility in this design. A click or tactile leaf, Alps style, could come in the future. However, it would result in designing a new slider, cutting molds for it, testing, etc. and is not something that will happen in the near term. The same is true for damping at the top and bottom of the stroke.

2. Could the case be designed to permit a detachable cable and still be waterproof?

Yes, to the extent that water could seep into the connector and possibly short USB power and ground if the whole thing were submersed. Resisting coffee, water, beer, etc. is one thing, being immersion proof is another. The consensus is moving toward detachable cables - perhaps even USB C. The manufacturer is open to this so long as we accept that it may not be immersion proof.

3. Could the keyboard be made fully programmable?

Anything is possible in the long term. The current situation is that the manufacturer doesn't really "own" the firmware - they license/purchase it from some other group that makes the firmware. This is why I'm having some trouble getting a better 60% firmware out. The other group needs to pay a developer and wants money to make a programming tool, etc. My preference would be to "liberate" the board with an open source firmware. We'll see what happens here.

If I were to go for one of these, I would probably be most interested in a 60% form factor. It would also be nice if it were fully programmable or in a HHKB layout. A compromise would be DIP-switch options for converting a standard layout to something close to HHKB.

Perhaps. DIP switches also affect waterproofing, and, the manufacturer requests certain minimum orders for substantial reworks such as board changes. I've communicated to the manufacturer that the current firmware offering on the 60% board is not acceptable - I don't want to do a group buy with missing keys from the Fn layer and the unreliable buttons on the right hand side.

I hope we get this sorted out because these are really nice keyboards with a lot of potential. At this point I expect that the initial group buy will be with 87 key ABS boards in ANSI and ISO, and that we'll use this as a stepping stone for further work.

[Techno Trousers]

I don't know if anyone's brought this up yet, but will these have media key functions? It's kind of lame, I know, but I've grown accustomed and rather fond if the way my QFRs have them arranged a function layer on the F keys. Obviously, full programmability would be the bees knees, though.

[XMIT]

Right now, no, media keys are not present. I'm still working to see what we can do for programmability at all. I may know more tomorrow.

[XMIT]

The saga continues.

The existing microcontroller and firmware vendor didn't want to open things up (so no customizations). The solution? Pick a new vendor! They're moving to a new 32-bit 8051 based MCU for new boards. I don't yet know if it is pin compatible with the old one, or if it will require a new PCB.

Some folks reported stabilizer binding issues on the bamboo boards. We've root caused this to the problem of fitting a square peg in a round hole. Seriously: the stabilizer pieces have sharp square corners, but since the holes for the switches are CNC cut, the end mill size dictates the minimum radius. We're exploring two solutions: new stabilizers (their idea) and cutting past the corner to fit the existing stabilizers (my idea).

At this point I'll probably order another 60% board to prototype with these two fixes. The ABS and acrylic 87 key boards are nearly perfect as is and will likely be the offering for the first group buy.

[lot_lizard]

I received my Bamboo TKL on Tuesday, and have been giving it a couple of hours a day the last couple. I want to thank XMIT for letting me participate, and also for driving the initiative in the first place. I have Hall Effect sensors all over the house and shop, but I have never tried one in a stand-alone keyboard. A unique experience for sure. My thoughts...

• Bamboo body - I am a huge fan of bamboo in general. The grain is always very clean, seldom warps, and prefer the carbonized look over natural. They did a nice job laser cutting/engraving the 3 layers, but fell pretty far short when it came to polishing the end result. I can tell the production was a bit rushed when I first opened the case to a strong varnish odor. The edges are quite rough, and there is a pretty uneven transition between the layers. You can tell that very little sanding was done once the plates were cut. Surfaces all the way around are fairly rough. Definitely has potential though with 5 minutes of extra attention to detail.

  

  IMG_8451.jpg (655.67 KiB) Viewed 6945 times

• Fasteners - as others have mentioned, the sex bolts are a bit excessive and could be a higher grade. Something a little smaller that was countersunk on both the screw and nut sides would be ideal.
• USB cable - as others have pointed out, something detachable would be ideal. Storing attached cables is messy, and the waterproofing need is an odd requirement (not sure what their target market is)
• LED backlighting - I realize I am in minority here, but I appreciated the numerous light effects. Not for myself, but my 4 year old daughter is having a blast with it, and I like her taking interest in keyboards. If the cost built-in isn't too much, and it can be turned off, it is a nice feature... and behaves well (I am not having the quality issues that others addressed)
• Caps - this one has already been beat to death, but I would second what everyone else has said. Cap quality itself is fine, but the legends are not ideal (my daughter couldn't find 'A' even though she is very aware what it looks like and where it is located on a keyboard).
• Switch - The whole point of this exercise... I consider the Hall Effect sensor and the mechanical switch to be two completely different things (though they obviously work together). You slap the right magnetic field in almost any physical switch mechanism, and this PCB in theory could pick it up (hence the waterproof nature of the board). The Hall Effect sensing is working very well. On something that behaves like a linear switch with no-tactile bump, I tend to touch type even more... and it took me a bit to get the actuation point down. It seems lower than an Alps, but I didn't verify that. It could be the 70g switch I am using as well (which makes touch typing that much easier). I would like to see the point of actuation raised a hair (at least for the 70g switch), or have a "bump" along the way.
  
  The physical switch is very smooth, and I am enjoying the 70g springs. I think it is actually a little loose though, and if you press laterally on the key at the base of the stroke, it grabs slightly. This occurs on every key for me, but in some cases, can get stuck altogether.

  

  IMG_8450.jpg (991.5 KiB) Viewed 6945 times

Overall impressions are really high though. It is a very enjoyable typing experience. Upgradable firmware is not a big deal to me on a TKL, but would be for pretty much every other layout. Well done finding this, and putting together the initiative. I would happily pay 50 additional dollars for this same board if we could raise the quality level in several key areas. Those need to be improved though before it would be board that I would really embrace.

[XMIT]

Thanks, lot_lizard!

For key caps: we'll offer a bare bones option since the current caps aren't popular. But, they are double shot, and PBT, and inexpensive.

For "sex bolts": future bamboo boards could, as an example, have the top two layers glued together with attachment bolts at the back. We'll probably head in this direction.

For detachable cables: we're exploring USB C as an option here.