buttons that don't register well

If you're having problems with the hardware of a DM42, post about them here.
gduncan
Posts: 10
Joined: Sat Mar 03, 2018 2:18 pm

Re: buttons that don't register well

Post by gduncan » Thu Mar 29, 2018 3:55 pm

Hi Meino,

I agree that the knobs on the underside of the keys should be centered to contact the center part of the domes. When I swapped the domes as previously mentioned, I took great care in their positioning. I actually tried moving the domes to optimize the contact point several times until I was satisfied. For me, it was really only the ENTER key that was problematic, and to a lesser degree the EXIT key.
Is there any suspicious "something" near the left side of the ENTER key, why may cause the "pre-trigger" of the left dome of the ENTER key?
I can't figure out what the problem is here.
PS: I have some dust under the glass of the LCD, which is clearly visible, when the sun or a lamp throws its photons with a shallow angle onto
the LCD...any idea how to remove it?
I noticed the same thing on mine. Spraying around the LCD screen with compressed air may move some of the dust pieces around. But I'm not sure how to get them out completely. For me, the dust pieces are small enough not to bother me much at all.
P.S. my HP 35s purchased new came with a huge piece of dust under the screen that did bother me.

Best,

Greg

dove
Posts: 14
Joined: Sat Mar 24, 2018 11:35 pm

Re: buttons that don't register well

Post by dove » Thu Mar 29, 2018 4:00 pm

Walter wrote:
Thu Mar 29, 2018 6:12 am
dove wrote:
Wed Mar 28, 2018 1:50 pm
would you say that this mod has made your keys more clicky all around or less clicky?
Is registration a function of clickiness? :? I thought it's more a matter of proper alignment (or call it centralisation) ... and the absence of dirt (or fuzz) under the domes ...
I believe they are correlated, in two senses:

1) The most clicky response is generated when pressing the dome dead center. This is also probably the most reliable actuation point for registration. So, if all the keys have gotten clickier, the alignment is almost certainly better, and I'd expect registration to improve.

2) The "false activation" I discussed is possible because with a mushy key you can overcome the weakened tactile bump without putting enough force into the key to bottom it out and activate the PCB (at least, that's my theory). With a clicky key, as soon as you cross that tactile bump you have loaded enough force onto the key that you are essentially guaranteed to bottom it out. It would be much more difficult and unnatural to get your finger to stop applying force after the buckling point because it is so dramatic. This is similar to static vs. dynamic friction. If you're pushing a box from standstill that has a small difference between the coefficients of static and dynamic friction with the floor, you can carefully start it moving at near zero velocity. If on the other hand the difference between the coefficients is large, you're going to have to load a lot of force into it and then suddenly it will jolt forward. That jolt is exactly what we want in the behavior of these metallic dome switches: either the key doesn't move at all, or it jolts into bottoming out.

dove
Posts: 14
Joined: Sat Mar 24, 2018 11:35 pm

Re: buttons that don't register well

Post by dove » Fri Mar 30, 2018 12:31 am

I just tried sanding off the PCB sprue and pressing the PCB downward from the top edge while tightening. It didn't make a noticeable difference.

Aside, I think I read someone suggesting aligning the PCB by holding onto the beeper. Be careful...the adhesive connecting it to the PCB is not too strong and I've loosened mine a bit in the process of doing that.

mcc
Posts: 167
Joined: Fri Jun 23, 2017 3:10 am

Re: buttons that don't register well

Post by mcc » Fri Mar 30, 2018 5:15 am

Hi Dove,

for me it works to a certain amount (not meant as "after that it works perfectly").

For pushing down the PCB I use the battery holder since it hase some thick soldering point.
The beeper is only glued (as you alread mentoined) and the USB port has only some really tiny
solder points.

Has SwissMicros some hints here, since Michael has put all the DM42 together?

Cheers
Meino
DM 42 - SN: 00373, Firmware v.:3.9.1 / 3.9.1 as compiled by SwissMicros

dove
Posts: 14
Joined: Sat Mar 24, 2018 11:35 pm

Re: buttons that don't register well

Post by dove » Sun Apr 01, 2018 7:00 pm

I just took a closer look at things because the bulging of the case caused by the beeper, which leaves the bottom left corner loose, has continued to bother me.

I am now of the opinion that it is the domes themselves that are inconsistent. This has been mentioned elsewhere by some people including one of the only reports of someone succeeding at fixing their keyboard woes. Pressing on the individual domes carefully and in the center, I can now feel that the keys I find snappier do have snappier domes and vice versa. Why the bottom row in particular should have mushy domes and why there is some consistency in complaints with this row between customers remains a mystery. But, I no longer feel that alignment alone can fix this, and having busted out a ruler, I don't think the alignment is too bad. Now, I'm kind of expecting that with use all of the domes will revert to the feel of the mushier ones, but we'll see. I won't really be disappointed if that happens.

Separately, I found I could get the left side of the ENTER key to be clicky, at the expense of the right side, if I completely loosened that PCB screw. Not an option, but an interesting observation. The bottom line here is that the PCB tension is preloading the left ENTER dome, I'm pretty convinced now. Moreover, this is the same edge of the PCB as the beeper, and it seems to me that the PCB can't be pushed against the housing quite as far on that edge as on the other edge (CLX edge). I wanted to blame the plastic ledge/cliff on the housing immediately behind the LCD, or some component on the dome side of the PCB, but I can't really see any asymmetries with either of these things or anything else. I'm stumped.

There's something funky going on with the ENTER side of the case. I think this is the cause of the ENTER key problems and the bulging induced by the beeper. I'm a little surprised I haven't heard anyone else mention the latter so far. Does anyone else have that problem? In my case, I've now resorted to "fixing" it by putting a small cushion of masking tape on the back side of the PCB behind the EXIT key. Not ideal, but it feels much better now, and I think I'll finally be able to use the machine peacefully going forward :)


ETA: I initially tried repositioning the beeper a little to see if it could sit in a spot with more room in the metal backing, but this didn't work.

mcc
Posts: 167
Joined: Fri Jun 23, 2017 3:10 am

Re: buttons that don't register well

Post by mcc » Sun Apr 01, 2018 7:43 pm

H dove,

I also thought about something to get the space between PCB and the rest of the world more defined.
What about a kind of "washer" (not tested so far):
WIth a paper puncher created little disks from a thin, non comressable material -- at least
non compressible with the small force the screw create under normal conditions of usage.
Then drill a hole into the midth of such a disk (temporary fixing the disk with a piece of sticky tape
may help).
Use the result as a washer underneath the PCB screws.

As material for such washer came the folio into my mind, which is used for overhead projectors (the
old, non-digital once).

HTH!
Cheers
Meino

EDIT:
PS: Doesn't work...forget about that idea. Sorry...
DM 42 - SN: 00373, Firmware v.:3.9.1 / 3.9.1 as compiled by SwissMicros

dove
Posts: 14
Joined: Sat Mar 24, 2018 11:35 pm

Re: buttons that don't register well

Post by dove » Mon Apr 02, 2018 4:22 pm

You know what they say about great minds....;)

I was working a little on this idea after my post yesterday. I didn't have a rigid enough material to act as a washer, so that specific idea remains to be tested and I'd be curious to hear if anyone does. Instead, I tried the opposite, stacking up some small support material on the dome side of the ENTER edge of the PCB to relieve the space there. I cut up some old plastic packaging into shims to try. In general, I found I could get the left side of the ENTER key to work but then the right side loses tactileness (I think the right side of the button might end up too far away from the dome), or in some configurations the whole button would lose tactileness or only the center would retain it. Basically the same result as toying with the screw tightness, which isn't surprising. What I do find surprising however is that at no point did the response of any other button get noticeably modified, even though in some cases I was stacking pretty thick shims directly next to them instead of next to the ENTER key. The ENTER key is unique and I don't quite understand why, beyond the obvious fact that it is the only double wide key. But it's not even like it's particularly close to the screw...

As an alternative to toying with screw tightness, this idea has the benefit that you can keep the screw tight so it wont get lost, but I prefer to use the right side of the ENTER key anyway so I abandoned it. distr

I would be very curious to hear the results of a washer experiment. Unfortunately, I don't think I'll have a suitable material to try this lying around anytime soon, and it isn't a high enough priority for me to go seek some out. Also, I suspect that even if we get the whole switch to be tactile, the ENTER key will still be stiffer overall, and that's the more pressing (:P) issue for me.



To be clear, when I say "washer idea" I'm thinking about a rigid washer between the screw head and the back side of the PCB, with the intent of better distributing the force of the screw across the surface area of the PCB. If you mean a washer on the front (dome side) of the PCB, than that would be equivalent to some of the "shim" approaches I tried. I didn't try everything in that approach though of course, so I encourage others to give it a shot and report back.

gduncan
Posts: 10
Joined: Sat Mar 03, 2018 2:18 pm

Re: buttons that don't register well

Post by gduncan » Mon Apr 02, 2018 5:09 pm

I also considered adding a washer of some kind for the screws holding the PCB in place. I've noticed that the keys in general and especially the ENTER key feels better when the screws are looser. But loosening the screws too much results in them unscrewing during use and falling out. A washer would help to get a better spacing while maintaining a tight hold in the screws.

mcc
Posts: 167
Joined: Fri Jun 23, 2017 3:10 am

Re: buttons that don't register well

Post by mcc » Mon Apr 02, 2018 5:44 pm

Hi,

The first "washer idea" - which is complete invalid, since I had the wrong mechanical image of the screws, the PCB and ... --
was to put some kind of a ... washer (you guessed it already...) between the PCB and what the PCB is laying on.

The mechanical reality is different:

The counterpart of PCB (mechanical-vise) is the plastic housing of the LCD or some slightly thicker/higher plastic ridge between LCD
and the field of the keys.

Another reason for a different tactile feeling of the left and right side of the ENTER key may be that the ENTER key is the only key, which
has to operate two domes. The slightes difference in height and/or point where the ENTER key is pressed may cause one dome act differently
than the other.

What do we know so far?
1 The keyboard is prone to misalignments of all sort.
2 The keyboard is prone to a wrong tightening of the PCB screws.
3 The "0" key (and in my case the "6" also) seem to be the first which
fail completely, when point 1 or 2 are "established"/"current" (oh damn,
words, words, words...especially ENGLISH words...TRINITY - HELP! )
4 In all cases, when the ENTER key is affected, the left side of that key
seems to feel mushy/swampy/non-clicky. I didn't read about a non-functioning
ENTER-key so far.
5 All other keys seem no to be affected that much/at all.

Open questions:
What is the reason for the black foam-thingy on the battery-holder...it is compressed by the steel back
so it has a mechanical influence on the PCB alignment/twist/tilt?
Since the foam thingy is black...is this electrical conductive (high impedance) to avoid static electricity
to charge up? Should we leave that in place in any case?

Does the speaker has mechanical contact to the back? If "TRUE" then same questions as for the foam thingy
are valid...

Man, we do a lot of research here and I wish someone of SwissMicros would chime in and give us a hint...we do
some sort of design fixing here...

Cheers
Meino
DM 42 - SN: 00373, Firmware v.:3.9.1 / 3.9.1 as compiled by SwissMicros

dove
Posts: 14
Joined: Sat Mar 24, 2018 11:35 pm

Re: buttons that don't register well

Post by dove » Mon Apr 02, 2018 8:50 pm

On my DM42, the speaker absolutely contacts the metal back. In fact, I would say it is THE main contact point, and it is absolutely responsible for depressing the top portion of the PCB when fully assembled. This is tied to a nitpick of mine. It doesn't have enough room to do so (maybe the LCD ledge at fault? or the back case needs more room). When I disassemble the case, I can see the speaker leaves a crescent shaped mark on the inside of the steel case from where its bottom edge makes contact. When fully assembled, that portion of the case and housing do not stay flush to eachother (the other side, with the battery, is fine), it bulges out and leaves a gap between the metal back and the plastic housing right where the speaker is, and this gap propagates all the way down to the EXIT key corner where it leaves an annoying looseness I had to fix with a small cushion.

I don't know if this is related to the ENTER key problems, but I do find it a curious coincidence that both problems seem to revolve around issues with space in the assembly of the same side of the device.

Post Reply