Bongofish
October 19, 2019, 09:19:52 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
News: You can attach photos hosted by the forum rather than using an external image hosting site, this means they will stay forever and not disappear after a year or two.
 
  Home Help Search Login Register  
  Show Posts
Pages: [1] 2 3
1  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: March 18, 2011, 07:01:56 PM
thanks tpope for pulling it out.

SU-040-X01 --- darn!  yet another wacom model number not in my little list.  They have a *lot* of different models, it is crazy!! http://wiki.bongofish.co.uk/doku.php?id=bongofish:penenabled  I will add this model.  Could you measure the "active area" (or the "apparent" active area), thanks.  You said 12 inch lcd -- Would you know if that is 12.0 inches or more?  I assume that is not a wide screen aspect ratio, right?

It's hard to tell from the wires on the sensor board, but I do know the effective active area is pretty close to exactly 12" and that the sensor board fit exactly under the LCD, if you include the LCD bezel in the measurements.  Not widescreen, but the standard 12" XGA netbook/tablet display of that era.  In fact, it looks like the dimensions were the same as on my LE1700.
2  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: March 17, 2011, 10:03:47 PM
I am going to interface penenabled tabletPC sensor boards -- the serial-based ones -- and convert them to appear as USB Wacom device.  I am looking for the pinouts of the different models.  In all your laptop ripping, have you encountered such sensor?  

(Well, I know you have a tablet pc -- but a functional one, I am not asking that you risk-open it for me -- but it would be very cool that you do so if you can put a multimeter on the running device Smiley )

If you have any info regarding Penenabled boards -- I would like to "list" the models that are out there and know a bit more about them. (and which sensor is inside which laptop, etc.)

I've only torn apart one so far, a Thinkpad X40T (I believe, it was a while ago).  The tablet part is right here pinned to my "wall of electronic bits" so I'll take a few photos...  

Here is the full board (front side, backside is totally covered by a red metallic film, shielding probably.  The digitizer itself (leaving aside protuberances) is 10.125" by 7.625" roughly, and fit behind a standard 12" LCD.  I tossed the LCD when I found that I couldn't drive it (wrong connector) from my NJYTouch controller.  It's also honestly smaller than I wanted to bother with even if I could have gotten it working.
[attachment=3]

Here is the PCB itself.  As you can see, it's a 14 pin connector back to the computer, but that's all I could tell you about it.  There is nothing useful on the back but circuit traces.  No printing or pinout definitions or anything like that.
[attachment=1]

...and here is the back label with all of the various part numbers (the most important probably is the Wacom SU-040-X01)
[attachment=2]

While I'm not opposed to running some current through this (somehow) and testing leads with a multimeter (again, somehow) I also don't have any idea where to start on something like that.  I mean, I could easily get a +5V signal from a random USB cable but I wouldn't know where to try and wire that and the ground, nor what I'd be looking for on the other pins.

Sadly, when I pulled this apart, I wasn't really thinking about getting it work again myself, or I would have at least cut off the cable at the connector to make testing easier.

3  Screen Tablet malarky / Heyaaaalpppp / Re: Intous 2 A4 build advice needed on: March 14, 2011, 03:20:52 PM
One option would be to pick the screen you want, based on a successful build here, or at least a conformation that the folks with the controller can drive it.  Then order the preprogrammed controller first.  Don't order the screen until you have the controller in your hands, and if it doesn't work, return it for another one of the same type.  This means you're committing to a given screen type, but short of buying a programmer (another $60+ to the controller price) that's pretty much a given anyway.


4  Screen Tablet malarky / The Trading post / For Trade: Have LCDs - want Wacom on: March 11, 2011, 07:47:45 PM
Now that I've run through and tested my stack of screens, I have a lot left over that I don't really need (most of the ones I thought were bad were from dead laptops, and the screen was rarely the problem) but I am still without a tablet. Thus, I would like to propose a trade.  I am looking for a 9x12 Intuos tablet of any vintage, as long as it works, comes with a pen and is USB, not Serial.  In exchange, I will offer a to-be-negotiated number of the following LCD panels:

2x   BOE Hydis   HT14P12-100   14.1"   1400x1050
1x   Samsung   LTN141P4-L01   14.1"   1400x1050
8x   Samsung   LTN141P4-L02   14.1"   1400x1050
1x   HannStar   HSD150PX14   15"   1024x768 (2 PCB Model)
1x   LG.Phillips   LP150x05 (A2)   15"   1024x768

All of these I have confirmed work with the NJYtouch controller (just tested them all yesterday) and I'll supply as much information as I can about what driver program to use, datasheets, etc.

The disadvantage of the 14.1" screens is that they are smaller than the working area of a 9x12 tablet, though at least their resolution somewhat makes up for that. The advantage is that I have a ton of them, which means one or two can be broken in the process with little harm done.  As an added bonus, all of them ran for me on the exact same controller program (makes sense, they were all out of similar generation Thinkpads).

The 15" screens are of course a perfect fit for the most common tablets, but there's only one each, and one of them is a 2 PCB model that would require an extension cable to unfold the boards.  Either of them could work, but since the controller is only programmed for one screen at a time in China, it's more of a risk.

Anyway, while I'm happy with any USB Intuos 9x12 tablet, I'll obviously give preference (and would consider trading larger numbers of panels) to newer models if available. To keep shipping from getting insane, this is probably a USA-only trade, but I'm willing to discuss any offer.  Basically I have screens I don't need, and I'm hoping someone out there has a tablet they don't need...

Heck, if someone by some miracle has an Intuos4 Large and wants half of these or more, I'll find a way to ship them to you.  :-)
5  Screen Tablet malarky / Monitors / Re: NJYtouch controllers on: March 11, 2011, 07:25:00 PM
Yeah.  ADJ is defined in their datasheet as "Luminance Adjustment" but unfortunately there's no further information about how the signal works in the documentation I have.  I couldn't tell you if the OSD controlled it I'm afraid, because of the corruption.  The only reason I knew where the brightness control was at all is that before I reprogrammed the board to test the next panel in my stack, it was working for that first one.  That said, since mine wasn't even wired to use the ADJ signal, there's no real way I can test it.

However, while the controller you are using isn't the same as mine (which has a Realtek chip) it does sound like the inverter is pretty close to identical between the two of them, which bodes well for whatever that signal might be.

I think I need to find another desktop monitor to tear down, just so I can see how they feed to the inverter boards...
6  Screen Tablet malarky / Monitors / Re: NJYtouch controllers on: March 11, 2011, 05:01:22 PM
My guess is that EN/ENA is a simple on/off Enable signal (true that it could be a PWM signal too, but I suspect not, "Enable" pins are very common in electronics and they always work the same way: ON/OFF switches. (most of them are active-low -- meaning you need to ground it to ENable, but either way, it is a ON/OFF switch).  If you see a start * next to or a line over or under the letters "EN" -- that means "Active-Low". If you play with the OSD, is there a feature of brightness/color control that does not seem to work?  If you are able to see the CCFL light going out of seams/small openings of the LCD Panel case, do you see it dimming when playing with the brightness (or whatever that would control the backlight luminance/brightness) on the OSD?

That was the test I needed to perform, and I've just done it.  None of these let me see the backlight through any seams, but I just plugged in a spare CCFL tube and adjusted the brightness on the OSD on the dark LCD panel. Between the corrupt OSD text and the darkness of the panel I can't be 100% certain, but I'm fairly convinced that the only brightness and contrast settings the OSD can affect are all in the tint of the individual pixels. The CCFL remained the same brightness throughout.

Between that anecdotal evidence and your knowledge of what ENABLE means in terms of electronics (plus their separate, unused, "ADJ" lead) I'm willing to bet that it's just a on/off like you say.  That bodes well for certain kinds of LED backlights I think, simply because that makes the connections easier.

Quote
Driving LED-based LCD Panels is, to me, a very important topic as more and more LED-based LCD Panels will become available on the after-market.

Agreed.  The closest I got to having one to play with was a dead Macbook, and they use so much double sided tape that I damaged the panel getting it out (it was also quite likely dead).  We're on a 3-5 year cycle here for computers, but they normally only get to me via friends in the IT group after 5 or more, so it'll be a while before I have a whole stack of LED backlit panels to play with.  But anything I do manage to find, I'll post here.
7  Screen Tablet malarky / Monitors / Re: NJYtouch controllers on: March 11, 2011, 04:32:11 PM
What about the connection going to the LCD Panels? They are all the same and matched? I thought there is some variety out there, no?

Of all of the laptops I tore apart (nearly 60) only a couple of the really old ones had a completely different style of connector.  I'm afraid I cant' give any more details than that because I tossed them immediately.

The 12" panels I had (which I also tossed) used what I *think* might called TTL, which looks a lot like LVDS but with a shorter connector.  NJYtouch does make a controller that can push to that kind of panel, but it's a different model.

All of the rest (mostly in the 14" to 15" range) have been LVDS.  I'm afraid that while I've ripped apart a lot of dead desktop monitors for their buttons, and the nice tiny wires they use, I haven't paid a lot of attention to their connectors.  They all looked similar, but might have not been identical.  What I think will be the difference for some of the larger monitors is that it's the same style of connector, just longer with more pins.  But I'm not even sure about that.  The connector they supplied is 30 pins (some were empty) but their header on the controller board is only 28 pins.  I know the controller board I bought can drive up to 1920x1080 or maybe higher, for TV-sized screens, so it might be that this one cable is pretty much universal.

Anyway, at least all of the laptop screens I've tried used the same physical connector, but they often ran different signals over that connector.  The two differentiating factors that the controller files seem to look for are the number of signals and whether you are running them in singles or doubles.

While not comprehensive, there's a nice large selection of LCD datasheets here:
http://www.beyondinfinite.com/library.html

I used that for a couple of the problem children of my stack of panels.  From there, what you're looking for is whether you have 3 or 4 pairs of data signals with a single clock pair, or 6 or 8 pairs of data signals with two clock pairs.  3 and 4 they call singles, while 6 and 8 they call doubles, which has something to to do with that they call "Odd and Even" data input.

Anyway, while their documentation is a bit murky at times, the end result is that you count the pairs, and count the clock signals.  If there is one clock signal, you're a SI6L if you have 3 pairs or a SI8L if you have 4 pairs (3 pairs is Red, Green and Blue, I do not know what the 4th is but none of my panels used that signal type).  If there are two clock signals, then you're looking at DO6L or DO8L for six and eight pairs respectively.  I *think* that the single or double has much to do with resolution.  All of the 1024x768 panels I tested (though I threw some out at the end because I have way more than I need) were SI6L and all of the 1400x1050 and higher panels were DO6L.

Anyway, those four types of signals seem to be pretty much a standard in LCD panels.  All you need to do at that point is find a panel in their list (and the program files all include that data in the filename) that has the same signal type and the same resolution.  From there, you keep trying them until one works, though they recommend if one is by the same company, use that first.

For reference, here are all of the panels I tried, their size, resolution and the driver that ended up working for them:
AU Optronics   B141XG09   14.1"   1024x768   B140XG10_SI6L_1024_768_5KEY_ENG_NA_R_RM5X51
HannStar   HSD150PX14   15"   1024x768   HSD150PX1A_A02_SI6L_1024_768_5KEY_ENG_NA_R_RM5X51
LG.Phillips   LP150E02   15"   1400x1050   LP150E06_DO6L_1400_1050_5KEY_ENG_NA_R_RM5X51
Sharp   LQ150U1LW22   15"   1600x1200   LQ150U1LW22_DO6L_1600_1200_5KEY_ENG_NA_R_RM5X51
LG.Phillips   LP150x05 (A2)   15"   1024x768   LQ150X1LGN7_SI6L_1024_768_5KEY_ENG_NA_R_RM5X51
Samsung   LTN154P1-L04   15.4"   1680x1050   LTN154U2_L04_DO6L_1920_1200_5KEY_ENG_NA_R_RM5X51
BOE Hydis   HT14P12-100   14.1"   1400x1050   LP141E2_B1_DO6L_1400_1050_5KEY_ENG_NA_R_RM5X51
Samsung   LTN141P4-L01   14.1"   1400x1050   LP141E2_B1_DO6L_1400_1050_5KEY_ENG_NA_R_RM5X51
Samsung   LTN141P4-L02   14.1"   1400x1050   LP141E2_B1_DO6L_1400_1050_5KEY_ENG_NA_R_RM5X51
8  Screen Tablet malarky / Monitors / Re: NJYtouch controllers on: March 11, 2011, 04:12:30 PM
Do some of the panels you have use LEDs backlight ? Maybe you could provide us with the list of the panels that works fine: that would be invaluable data to us !

I'm afraid all of the panels I use are CCFL, which is fairly standardized across the LCD industry.  I've attached a photo of the controller attached to one of the panels, though I'm sure you've seen these before. 

On the left the the controller itself, with the VGA and power cables connected, and wires to the button panel, backlight inverter and LCD.

Note that the button panel for the 3xxx and 5xxx series has only 5 buttons.  The 6xxx series has 7 buttons because that includes audio controls as well.  They are all on the same PCB, and you can see on the board where some components (notably 4 medium sized capacitors and a IC) are missing on mine because it's not the audio version.  This is relevant because the header for the button panel is 10 pins on the board, and the connector for the button panel I'm using is only 9 pins, with 8 used.  They shipped it assembled, but be careful to note this if you ever unplug the panel, as it aligns onto the leftmost 7 pins of the header. 

The inverter in the photograph is of course the narrow PCB assembly on the bottom.  Every single laptop LCD I've seen has the exact same 2-prong CCFL connection and single bulb setup, always along the bottom edge, so that's an easy one.  Desktop monitors tend to use 3-pin connectors and run in pairs or in sometimes in quads for larger screens.  This board can drive them all I think, but I only got the laptop version. 

So, the backlight header on the main controller board has 6 pins.  12V, 12V, EN, ADJ, GND, GND.  The header on the inverter I have only has 5 pins, +12V, GND, ADJ, [blank], ENA.  EN/ENA is defined in their datasheet as "Backlight Control" while ADJ is defined as "Luminance Adjustment."  All that said, on my board, there are only three wires between those two headers, +12V, GND and ENA.  Unless brightness and contrast settings are completely controlled by the color of the pixels in the LCD (which is possible, I honestly don't know) I'm thinking that with the ENA lead must be more than juston or off. 

But maybe not.  I could probably test this with one of my spare backlights from a disassembled panel.  I'll be hampered by the fact that the OSD is both broken AND I'd be running the main panel without a backlight for the experiment, but if I can test I will.

Anyway, the key point here is that the board itself appears to be able to supply a pretty standard voltage, on two rails if necessary, and couples that with at least an enable signal if not actual control over brightness.  While I'm not sure you'll get any "out of the box" compatibility with a LED backlight, this doesn't sound like it's outside the realm of possibility either.
9  Screen Tablet malarky / Monitors / NJYtouch controllers on: March 10, 2011, 10:35:06 PM
I had some money come in from a freelance job, and decided to treat myself to a controller kit from NJY Touch.  I didn't bother to go through eBay at all, just emailed Christina direct from the address on the wiki.

Because I have 32 LCD panels of 17 different types from my laptop demolition program, I knew I needed not only the controller package but also the programmer.  After much confusing cross comparison of listings, I decided to go with RM5451, as it was the higher resolution driver (x4xx) with both DVI and VGA inputs (5xxx).  I also ordered the USB programmer and a power adapter just in case nothing I had here would work.

Christina answered my email pretty much as soon as the time difference would allow, and after a few emails back and forth my order was placed.  I asked for the controller to be programmed for one of the panels at random, and didn't object to the price for DHL shipping, since I was being impatient.  All told I spent US$127.48, with $28 being shipping.  I'm assuming I could have asked for "slow boat" and gotten it in a month.

Instead, with my last email and PayPal payment sent Friday afternoon, the order shipped Monday morning and I had the controller last night (Wednesday)!

Sadly the panel that I'd programmed (one of my nicest, a 15.4" 1920x1080) looks to be dead.  The backlight goes on for a minute showing a decidedly red-tinged screen, and then the light goes out.  You can still see the screen just barely, but I'm assuming the red is still there and it's basically shot.  :-(

So, I had to wait impatiently until I could bring it in to work where all of the other panels are stashed away.  Installing the programmer was a bit of a chore.  There are instructions, but they're not always perfectly clear.  Also, they do mean it when they say this is for XP only.  Even in compatibility mode I've had no luck getting it to work under Windows 7, though it may still be possible.

Instead, what I did was to install "XP Mode" which is a Microsoft-provided virtual machine running XP.  It takes a while to set up, but once that's done you have a nice clean XP machine for this kind of thing.  You have to use the meta-menu on the top left to basically push the USB device to the XP box, and installing drivers took a while, but after maybe an hour of somewhat frustrated tinkering I got it working.

One thing not noted in the instructions is that you need to "register" the software to use it.  I got a bit worried about this, but on a whim when I opened the registration application, I just pasted the "Request SN" into the "Activation SN" box.  Thankfully that worked.

The programmer itself feeds data to the controller over a VGA cable.  The software is not user friendly, but basically all you are doing is picking from a list of precompiled monitor specs and pushing that over the wire.  The biggest problem I found is that once I disconnected from the programmer to test on my panel, the only way (and I tried EVERY combination of plugging and unplugging) to get it to see the controller again was to unplug from USB, plug back in, grab the USB device from Windows 7, then grab the USB device again (there are two logical "devices" attached to this port, and both need drivers and pass-through) and then try again.  Frustrating, but now that I know what to do it's not a big deal.

There's also this black slider switch on the side.  They tell you to slide it towards the VGA connector (the programmer has DVI and 2 VGA ports) and that either port will work.  Well...  That's not entirely true.  It looks like it's actually a 3-way switch, and sliding all the way to the right means you need to use the VGA port on the end.  I haven't tested yet with the switch in the middle position and the other port, it's been a bit of black magic to get this working anyway, and I'm not wanting to mess with it yet in that way.

[attachment=2]

With all the setup done, I've started to go through my stack of panels to see which ones work.  I already had a spreadsheet with their model numbers and resolutions, which helped.  For the first few, the included library had perfect matches, so no problem there.  Plug in the controller, push up the new code, unplug and plug in the monitor. 

One problem that I'm waiting to hear about is that every time so far it's ended with a warning that the CRC check failed.  The only noticeable effect of this error (or I presume due to this error) is that the on screen menu is completely munged up. [attachment=1]

This hasn't caused any real problem with screen performance or resolution or anything, but it's a bit annoying.  The menu still works, since I remembered the buttons to adjust the brightness, and could do that "blind" so it's just the display that's screwed up. But still annoying. 

I got through about four panels that way, lining up model numbers to match exactly.  For the next two, I picked, mostly at random, what looked like close matches based only on model number and resolution.  One worked perfectly that way, and one had some nasty looking random static (but still worked) until I switched to another guess.

At this point I'm down to the panels that have no close analogues in their database.  However, all is not lost.  Apparently as long as you're looking at a LVDS connector, there aren't actually that many different ways of sending signals to a display.  What the documentation suggests is that you find a datasheet for the panel in question and find out how it's sending data (there are only a few choices, 6 or 8 bit and single or dual channel IIRC).  Then it's just a matter of picking another file that happens to match that spec and resolution and trying it.

So that's where I am now.  I've already done one that way successfully, and am starting on the others now.  When I'm done I'll try to share more of my experiences, as well as the resources I've found.

10  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: March 01, 2011, 10:27:16 PM
Hey!!  that's very cool!  -- Using a diode was the best thing to do afterall! You've got a very good intuition there.

Yeah, I really lucked out on that one.  I also just found a work discard (out of a laser printer IIRC) with dozens of regular and zenier diodes (mostly of the glass variety) so I'm well stocked.

Quote
The "glass" diodes ones are also very popular indeed. For the "specs" all the "small" ones should have worked because their spec is more like how much voltage they can take or their "reverse breakdown voltage", their "basic" operation is pretty standard.

Good to know, thanks.  I've got a lot of them now, some zenier, some regular, some regular with yellow stripes and a few oddballs here and there.  It'll likely be trial and error next time, but at least I know what effect I'm looking for.

Quote
Laser-cut:  watch out that "laser cut" does not equates "it works" -- far from it -- one reason is that you do not have the material in your hands (not all plastics are the same) and you are not "cutting/sculpting it" yourself with a "try-it & refine-it" tight loop. Unless you already made a working prototype, you will probably need a couple of rounds before getting the thing right. (But you might get lucky or just deeply know how to work with tolerances to not miss it).  It is quite cool to design a complex shape and get it in your hands for real. Besides you seem to like to do these things, so that will be an invaluable experience.

All agreed.  My plan right now is frequent testing with cardstock mockups, careful measurement where material thickness comes into play and a bit of luck.  Ponoko has a limited selection of acrylic in the thicknesses I want, but I figure that if I end up making minor mistakes (the kind not caught by cardboard mockups) I still have the dremel to try and fix them, rather than wasting the whole pattern.  We'll see what happens.  I have two more Dell keyboards at the moment, though they are not the same model.  At a guess, one will be similar to Mk1 and the other similar to Mk2 in terms of PCB size and layout, but I won't know until I strip them. 

Quote
Ruggedness:  the LE1700 is a rugged-tablet-pc -- so I am thinking maybe there is already spots/holes to put some sort of "bracket" that would be sturdy. I mean some sort of metallic bracket that would go into "holes on both sides of the laptop" (if there's such a thing).  I see they sell some accesories like a keyboard that also doubles as some sort of "docking station".  Anyway, just a thought.

There is a docking station and a clip-on keyboard, both of which use metal kingposts that fit into metal lined holes in the frame.  The problem is that they attach to the bottom of the tablet (in landscape mode) or the right side (in portrait mode), neither of which is convenient for a right-handed pen user.  There are even programmable buttons and a d-pad on the LE1700 (it is, by far, the best artist slate on the market today, despite being 4 years old) but they end up on the top or right. 

If I could find a few spare free (instead of $50 and up) LE1700 keyboards I might be tempted to try hacking one of them anyway, and just find a 3rd party video driver to let me rotate the 270 degrees to place it on the left.  It would have the advantage of being usable in portrait mode.  An interesting thought certainly, but I think not worth the destruction of my only keyboard right now.

Quote
The LE1700 uses a Wacom driver -- so I assume this is a wacom sensor -- is this a "Penenabled"?  I guess my question is: are you getting the pen pressure data?

It is penabled, like other TabletPCs.   ...and yes, with the Wacom driver I get full pressure sensitivity across all applications.  It's really a nice machine overall.
11  Screen Tablet malarky / Tutorials and useful knowledge / Mk2 Completed on: February 28, 2011, 05:56:51 PM
Well, the glue is dry and I've had a day to play around with Mk2, as well as use it seriously for a few things.  I've decided, like Mk1, not to bother with capping any of the sides.  The wires are all protected well enough that I'm not too worried about them, and cosmetically, I kind of like the MacGyver look anyway.  :-)

One drawback of using scavenged buttons from so many sources is that, even when they look identical, there are at least two different styles in here, and you can feel it in the subtle differences in how easy it is to click them.

I also am not totally pleased with the plastic sheeting I used.  At the corners it's stuck in places to the glue, and that makes in particular the control-z key a bit harder to press than the others.  It's also just rigid enough that if I'm not careful I can sometimes accidentally trip the Alt key when I press Shift.  Though that also may be due to the fact that I think the shift key is one of the ones that got a stiffer button.

Also, it's not perfectly sturdy, given the connection point.  Now, in regular use that's not a problem, as it's not load bearing and I would have a reason to not hold onto the back of the device while I push the buttons.  But I almost had a broken device (or worse, broken USB port) when my 6-year old rambunctious son was climbing all over the bed not watching where he was going.  Sadly I'm not sure there's much of a solution to that, other than "be careful."  I've pretty much come to the conclusion that moving the buttons up onto the bezel of the tablet is just too uncomfortable to use, and that would also add 6mm to the overall height of the device.

If I wanted to make this unique the the LE1700 and worthless on other tablets, I still have the option of the double USB port, or a lip onto the bezel to make it sturdier.  I'm certainly going to consider those options when I work on Mk3, though at the moment I'm not seeing either as all that viable.

Overall though, I'm very pleased how this is working.  It's incredibly useful and makes my tablet a lot more functional when I'm doing more serious work than sketching, and it's just so pleasing to have built something like this myself.  While all of the complaints above are valid, none of them detract significantly from the value of the device for me.  It's sturdy enough, comfortable enough and usable enough that I could very well stop here with no regrets.

But I'm not going to stop here.  For one, there are still refinements I could make, and for another, I have a friend with a LE1700 and I really want to make one for him as well.  So I think the next step is first to find two (at least) identical keyboards, and then start to seriously design a full kit in lasercut plexiglass.  This I imagine will be even slower going than Mk2, and will of course involve spending some actual money, so I'm going to be a lot more careful in measuring and templating (probably in cardstock) before I buy anything.  But it's doable, and I think the process will be a good learning experience. 

Thanks to all who've replied for the advice and encouragement.  Reading this forum (and not just conversations in this thread, but all of the other build logs) has been a great source of inspiration.  I'm actually feeling a lot more confident now about my ability (someday) to take the next step and build a DIY Cintiq of my own.  ...probably with a custom button bar.  :-)


12  Screen Tablet malarky / Tutorials and useful knowledge / Diodes to the rescue! on: February 26, 2011, 05:51:51 PM
...
So one thing you could try is to cut the wire (where I painted black) -- (in case you didn't do that already to fix your CTRL key!!) and try to hook a diode in series there scavenged from another device. Try connecting one way and then the other. Nothing can break.

I had a few diodes, both through mount (is that the term?) and surface mount, already scavenged from some of the earlier demolition jobs.  Most of them look like the glass ones in the image you posted.  So I took a random one, no clue what the rating was (I can hardly read the numbers on the side and don't know what they mean anyway) and wired it in as you described.

It... just... worked!  First time.  Perfectly!

By the time I finished writing my last post, I had completely talked myself out of using a diode, I just figured I was being naive and clueless.  I was all set to go with "good enough" and just clip the wire this time. So thank you, very much, for the mini diode tutorial and encouragement!

After adding some shrink tubing to either side of the connection with the diode, I glued on the frontplate.  It'll take about a day to dry completely, so by tomorrow I will hopefully be using Mk2 to start designing Mk3. :-)
13  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: February 25, 2011, 08:54:50 PM
Id like to applaud your cleverness in this keyboard hack. I don't think I'd have had the patience to figure out the terminal bridging sequences for the keys.

I missed this before, sorry.

Honestly, tracing just these five keys wasn't all that difficult.  A bit tedious, and I do think I might try a suggestion I read somewhere of scanning in the circuit and then using the fill tool in Photoshop to pick out the leads.  But even manually it wasn't bad.  Even where I screwed up the tracing (I think Alt in this one) I was only off by one lead, so playing with a bit of wire and shorting across settled that.

If I needed to trace a dozen, or the whole keyboard, I may be more tempted to make a grid map (in this case it would be 14 rows by 14 columns) just shorting every combination of leads and writing what they do, if anything, in the intersection.  The Tablet Input Panel in Windows makes this easy, since it lights up the keys you're pressing.
14  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: February 25, 2011, 08:51:10 PM
Ctrl-Z: yipes!  You could have two microswitches glued onto the same CTRL-Z cap?
Yep.  That's going to be my first try at fixing the issue in Mk3.  Luckily, I have literally dozens of switches from all of the laptop teardowns, so I can afford to experiment a bit.

Quote
Or find a microswitch that would actually make 2 parallel contacts at the same time (a lot of switches have that, but quite rare for  micro-switches me think).

I haven't even scratched the surface in what's available to buy yet.  I poked around the element 14 store a bit (they call these guys "tactile switches" for whatever reason) but I was quickly overwhelmed by options, and couldn't see any that had more than two inputs or two plus ground in my brief search.

Unfortunately I still don't know a lot of the terms so I'm going to have trouble narrowing down the search past that.

Quote
The next option would be to have a mini "relay" that the Ctrl-Z would actuate that would do the two keys in parallel.  I do not know how much "power" is going through those lines (if you could use to actuate a relay or a similar electronic switch).

Honestly, that's probably beyond my abilities right now. Simple circuits are ok, but I think anything fancier than attaching two switches together might be over my head.  Not that I'm opposed to figuring out a more clever solution, but I think I need to find myself a grounding in electrical theory that goes farther than "a switch connects two wires" before I go too much further. 

My understanding of diodes is VERY shaky, but if they limit the direction current can flow, might I be able to use one of them to avoid the feedback loop I built?  Probably not, now that I think about it.  But as I said, my actual understanding of electronics is amusingly naive at the moment.

15  Screen Tablet malarky / Tutorials and useful knowledge / Re: TabletPC Quickkeys on: February 25, 2011, 06:59:49 PM
While still cursing myself for the wiring snafu, I did at least get the faceplate finished.  The button caps were all glued onto a sanded piece of thin plastic I had lying around.  After some testing I confirmed that at least when resting on top of the tactile switches, the plastic was flexible enough so that no single button ended up pressing its neighbors.

I had some LED light guides from one of the disassembled laptops at work that happened to all be about the right height to work as standoffs, so after notching the plastic and punching a hole int he middle, all five were glued into place. Some of the glue ended up leaking under the corners, which was thankfully just enough to pin the plastic to the faceplate, while remaining just flexible enough to let the buttons press easily.

Minor measurement (or lack of measurement) issues aside, this fits pretty well onto the button tray, and as soon as I clip the wires to fix my control-z screwup, I'm going to glue the faceplate onto the unit.  The backplate was already attached, using a few more pieces of pen barrel as standoffs.
Pages: [1] 2 3
Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Valid XHTML 1.0! Valid CSS!