Full Version: Pinebook 14" keyboard misses keypresses UPDATE
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3
See Post #4 for key module pics & analysis.

The 14" Pinebook keyboard misses keypresses if you're a very fast typist. 

This issue is a mechanical problem, not apparently electrical and occurs because fast keypresses will from time to time 'hit' on the corners of the keys. If the keys are pressed distinctly in the center of the key the misses are not apparent.

Also, hunt&peck typists will likely not have this issue either;  for the reason stated above.

The Pinebook prototype does not have this issue because the mechanism under the keytop is slightly larger;  the keybutton has a 'further' key-travel just by a tiny bit;  just enough that full key travel occurs regardless of where the key is pressed (center or corner) !

The downside of the prototype keyboard is that there is a noticeable 'gap' around the key;  and the keyboard is NOT sealed.  Consequently, dust can get into the keyboard and gum it up.  The newer keyboard was designed for less key-travel (and less dust) but which unfortunately resulted in missing keypresses. 

I was planning initially (for the purposes of electrical testing) of swapping out the lower case assembly (containing the prototype keyboard and trackpad) with the lower case on my pre release Pinebook.  This has shown to be not practical. The prototype keyboard is plastic riveted into the lower case, and the lower case is quite a bit larger than the 14" lower case mold-- significantly.  The lower prototype case is 349.25mm wide, and 231.78mm deep. The trackpad is white and is significantly larger too; 74.6mm x 103.19mm.  The only practical way to test the prototype keyboard with my 14" pre release Pinebook is to run an FFC cable extension assembly (with connectors) across the table to the running Pinebook.  

This test is probably better suited to the laboratory, frankly, but I am game for trying it if I can find the right adapter and extension;  will keep you posted on that !

The only other difference I can 'see' is that the prototype keyboard has the unimportant dreaded Microsoft Windows logo on the menu key (third from the left, bottom row). All things being equal, I can see why the hardware developers chose the present keyboard over the 'Microsoft' keyboard with dust openings!  Unfortunately, the released keyboard & case did not undergo sufficient pre release product testing.

Notes:   user's list:
sample case 1
(07-11-2017, 03:16 AM)MarkHaysHarris777 Wrote: [ -> ]The 14" Pinebook keyboard misses keypresses if you're a very fast typist.

I am not a very fast typist, even a fast typist and I have this issue often, but I expected it because I have read about it a long time before I got my PB Smile
(07-11-2017, 04:10 AM)Wizzard Wrote: [ -> ]
(07-11-2017, 03:16 AM)MarkHaysHarris777 Wrote: [ -> ]The 14" Pinebook keyboard misses keypresses if you're a very fast typist.

I am not a very fast typist, even a fast typist and I have this issue often, but I expected it because I have read about it a long time before I got my PB Smile

Right.   I knew about it early on... and frankly, it doesn't really bother me any longer;  I got used to it ( took about a week of use ) and its only a minor annoyance for me now;  and for the price point of the pinebook, its really a non issue ;  if it were a mac book pro ( and cost $1400 dollars! ) I would expect it to be perfect.  But its a great little community computer with some great features for a tiny amount of money , soooo
[attachment=869]  [attachment=870]  [attachment=872]

I was finally able to extract the keyboard module from the prototype pinebook base case !  This was a painstaking process made almost hair-pulling impossible due to all of the plastic rivets holding not only the keyboard module, but also the module shield !  I included a standard evo plus SD card from Samsung for relative size correlation; obviously top and verso.  This appears to be the same module as the 14" released version, with the minor exception that the menu button has the ominous Microsoft Logo ;  the mark of the beast.   The pic top right has the module laid out over a sheet of white paper so the rivet marks can be seen more clearly. 


In the pic above the height of the key-travel is highlighted.  The bubble domes are quite high, and the module is not sealed;  in other words, the bubble dome is not the typical 'sheet' of domes, rather each key button is an individual bubble dome.  Consequently, if anything goes on this keyboard (dust or moisture) its going to end up inside the unit.  There is a metallic shield under the keyboard (also plastic riveted) which will catch dust and moisture as long as the unit is left laying flat!  


The pic above shows the "Y" key button top cap removed with an LED back-light coming in from the bottom of the module.  Clearly visible is the nylon scissor jack cap top support ( which keeps the key level during transit ) as well the rubber bubble dome.  The domes on this keyboard module are fairly large and quite tall;  which is why they work after all !  Because the bubble dome is tall the scissor jack can keep the key button level for good key press contact regardless of where the key is hit, corner or center.  

The next step in this project is to connect the keyboard to my pinebook and see whether the typing improves.  I suspect that it will;  the next decision will be whether to attempt to remove the module from my pinebook and swap out the prototype keyboard;  after all I can swap the dreaded Microsoft Logo key , and assuming the layout dimensions are the same then all I have to do is find a good way to redo the plastic rivets--  at this point I'm considering tiny droplets of 30 minute epoxy resin.
I have determined that it is NOT practical to try to replace the keyboard module on the 14" pinebook, without also replacing the bottom case top bezel.

The keyboard module on the production units is also plastic riveted to the top bezel, and the module is also covered with a metal shield which is also plastic riveted to the top bezel. Removing the keyboard module from the bottom case top bezel (for most people) would cause so much damage that the case would probably not go back together smoothly and even so the keyboard might be so warped that it would not work anyway.

In my opinion the bottom case top bezel (with new keyboard module insert) should be made available as an option. The bottom case top bezel can be replaced easily , transferring the "guts" from the old bottom case along with the bottom case bottom bezel. 

I do not plan to try to replace the module in my 14" until such time as the entire top bezel is made available with the good keyboard module.
I wonder if this problem falls under the warranty as it is clearly a defect. The most annoying key not working is the space key on my model.
The rest of the keys work pretty fine.
If someone has a tip on repairing the space key itself (it only reacts pressing in the middle and not thefar rightorleft) I would be pretty happy.
As even under warranty it would cost a bunch sending the device back to china.
Try this hack, it helped me a llttle.
[attachment=943]   [attachment=944]

Ok,  I finally got to plug the old style keyboard module into the Pinebook for power-on testing;  as expected the old style keyboard module ( with longer key travel ) completely eliminates the missing keypress issue !  

Big Grin

So, on the one hand,  the missing keypress issue is entirely hardware related to the new style keyboard module assembly;  it is NOT firmware or software related and its not about placing extra "stiffiness" into the case ( padding hacks will only damage the Pinebook ! ).  Its all about the superior bubble domes and larger key travel in the old style module !

Having said that,  I got a BIG FAT surprise too;  which is apparently firmware related !

On my original bad keyboard module the Alt keys are completely dead;  both of them, left & right;  and, this from day one-- they didn't die, they just never worked.  On the old style keyboard module ( swapped in today for power-on testing )  BOTH Alt keys are ALSO dead !   That blew me away.

So the dead keys issue is NOT the module, matrix, ffc cable, fpc connector, or defective bubble domes;  this is all about either a defect in the 8051 ( MCU for key scan ) or the firmware that produces scancodes for the kernel from the 8051 !  

I do intend to swap out the keyboard module because then I'll be able to type on it !  The keyboard can be remapped for Alt key functionality until we figure out how to flash the firmware to the 8051.

The left pic clearly shows the fpc connector;  these are very fragile !  when lifting that snap bar, be very careful.

The pic on the right is the old style keyboard module being connected for subsequent power-on testing.  I used xev, evtest, and showkey to test every key on the keyboard at center, edges, and corners;  both Alt keys are still dead, but the missing keypress issue is resolved !

[attachment=950]   [attachment=951]

This is the tear-down at this point;  this is very tedious and exacting work because it all has to go back together without being damaged. Everything is taped and riveted. All the connectors are very fragile and easily broken, and many of the wires are delicate.

The left pic is my bag of guts !  The right pic has the keyboard module ffc cable peeled back ( it is stuck to the shield board with some tacky 'stuff' that is almost impossible to remove, but I managed. All of the white dots are actually styrene plastic rivets ( fifty-eight [58] of them ) which hold the shield in place, as well the module under it !  Also, the module itself is also  riveted to the bottom case top bezel;  I think you can clearly see why the issue is NOT firmness of the keyboard !  Just the opposite, that keyboard module is so firmly attached to the bottom case top bezel that its damn near impossible to remove it without damaging something !  Stuffing things into the case is not only a waste of time, its silly.

The Plan

If you look carefully in the top right corner of the right pic you will see that the robot missed one of the 58 styrene rivets !  yay   ... I only have to remove 57 of them !

The plan at this point is to carefully remove all 57 of the remaining rivets and save them.  After cleanup and install of the replacement module, I will use a special tip on my solder station and refuse the rivets ( this also will be tedious but should work fine since the styrene material is inject molded and recycleable ).

Of course the keyboard module must be removed in exactly the same way.  I plan to use aluminum tape sparingly when I remount the module and shield. 

stay tuned, its about to get messy



Note:  The pic above is the tool I'm using to do the work, and the first two ( of 57 ) rivets removed !  haha ;  only 55 rivets to go!   " 55 rivets of beer on the wall, 55 rivets of beer,  take one down and pass it around, 54 rivets of beer on the wall ! "    Wink

[attachment=953]   [attachment=954]   [attachment=955

Pics above show the keyboard module coming out;  I did have to remove the display in order that I don't damage it while trying to wrestle the replacement keyboard module into place.  The issue is that each rivet has to be pushed into position from the other side and then "heat-squished" into place.  Too much work to risk the display -- out it came !

[attachment=956]   [attachment=957]

Left pic above are the top rivets;  normally they would protrude through both the keyboard module and the shield. I have decided to NOT do that in reverse -- I will be heat-squishing the replacement module to the top bezel 'alone' and then aluminum taping the shield in place above ( below ) that.  

The keyboard module base metal is 1mm wider and 1mm longer than the original keyboard module !  Not sure yet if this is going to be a problem or not ( at this point, I'm not going to worry about it ). Otherwise the keyboard replacement module fits perfectly in the top bezel, something I am thankful for and frankly fairly surprised to see !
Pages: 1 2 3