Datahand Keyboard Resurrection

[mikeincousa]

I purchased a Datahand keyboard ca 1999. With a rebuild maybe seven years ago it served me well. I purchase it as a prosthetic device. It helped me greatly with coordination and composition. I achieved a high level of words per minute with very high accuracy.

Datahand suddenly went out of business some time after that. I am more or less a disaster with a standard QUERTY keyboard. The oversized ones help some but I do not come close to what I could achieve on the Datahand. I put a link at the bottom that will give the background. There multiple images in Googlel

A little over a year ago the left thumb switch quit working properly. Then I thought it a problem with the IR relay. That key is the most used one and it could have simply worn out. I can get it to toggle one way once but it gets stuck somehow. I can manually reset and but I can only do it once again. It is has been over a year since I tired to use it so I am a bit rusty on the details.

I later thought there could be something wrong with one of the IC's further into the circuit? I am sure they are primitive given today's technology. But I would think the device does not need the latest and greatest to function like new with a repair of some sorts?

Along the way I have thought that replacing the relay or installing a TTL gate could fix the problem?

I have not engaged in any serious bread-boarding for several decades. A few years ago I purchased an Arduino Uno but did not get past the basics. I thought that if push comes to shove the Uno or a device like it could bypass the stall.

I had reason to think about extracting firmware a few months ago and that got to again think about resurrecting the Datahand.

I need a lot of guidance and direction on this so I will appreciate any thoughts or experiences you might kindly provide. I have a lot of very technical / legal writing ahead of me. It would be very helpful to do that with the Datahand.

Thanks for any thought or suggestions.

https://en.wikipedia.org/wiki/DataHand

 

[wejn]

If you search for “lalboard” and “Jesusfreke” in your favorite search engine, you will find an opensource datahand clone.

All you need is a 3D printer and some soldering skills.

 

[mikeincousa]

I am looking for a repair,not a replacement. This is major project beyond my skills and interests and time availability.

 

[JesusFreke]

Apologies for resurrecting a dead thread, I just randomly ran across this No idea if you still need help, but I'm the lalboard guy wejn mentioned -- but I'm also quite familiar with datahand internals and could probably help diagnose/fix the issue.

 

[mikeincousa]

What a nice surprise! Not hearing anything I had given up on the post. Thank you for making the connection.

Talking off the top of my head the critical failure is that one of the thumb keys (left I think) is not engaging properly.

As I recall I can get it to toggle once from a start-state but then it goes dead. What do you suggest that I do?

Thanks for any thoughts.

 

[JesusFreke]

Is it the "shift/capslock" key that you press in a downward direction? And if so, is it the first detent for shift, or the full press for capslock that is the problem?

Does it work again (once?) after replugging the keyboard?

 

[mikeincousa]

I have the DataHand out of storage.
I have to admit that I am quite rusty with it. I went over the manuals and that quick review brought back some of the operations.

Just completed a quick first run.

It powers up.

The only functionality I have found so far is the toggle between the NAS mode and Mouse mode.
I did not get any keys to work in this mode, but I am recalling that some other operation is needed besides the toggle?

The connecting cable is tight. I have a series of adapters for creating the cabling into my USB hub. All the supplied connects are PS/2 predating the USB era. I am thinking that maybe I will need to purchase a more direct patch cord? In contract if they were impaired some how I would not be able to mover the cursor?

All the lefthand keys seem totally dead save for moving the cursor with the paddles, likely in the mouse mode?

My next step will be opening up the right and left cases and blowing them out well.

Maybe you have some thoughts on this first stall at this point?

I will add to this reply after I compete the internal cleaning.

I am recalling that when I put it away the problem was with what the manuals and training show as the "N" key. (Left thumb up). I may be wrong on that though?

So far have not been able to find any mention of the functionality of that key in the manuals.

I need to quit this for now and may not be able to get back to it until the middle of the coming week.

I use Skype and Google Talk: if you feel a need to talk in real-time I will pass back my access numbers. BTW I am at UTC-7.

 

[mikeincousa]

I had forgot your instruction so I powered-up again.

The Normal (green) light came on. I was able to do a limited amount of typing with either hand. Somehow in typing I did something that brought up the NAS light. I could not toggle back to Normal. Not sure what I did and do not have the focus to sort it out any further right now.

When I rebooted the keyboard the Normal light reappeared: I could type letters again with either hand. I did this a couple of times with the same results.

I think this is what you were looking for?

 

[JesusFreke]

It sounds like at a minimum, there is probably an issue with the left-hand "up" key (the N key you mentioned - "N" for "Normal"). That's the key that switches back to normal mode.

I think the first thing to try would be to take it apart and unplug/replug and just generally "jiggle" the cable connecting the thumb cluster to the main board in the back of the left hand unit. Loose wire connections were a fairly common problem with my datahand at least.

If that doesn't work, there are a few places that might be the problem. The only problem I can think of offhand that would cause *only* that single button to stop working would be if the LED or phototransistor for that key stopped working.

To check if the LED itself is working, you should be able to see the IR from the LED with your phone's camera. There's an LED and phototransistor on either side of the base of that key. The LED will be the one that's nearest to you, when the keyboard is in a normal typing position. You'll need to peek in at the LED from the back side, since the LED is shining towards the back. You'll need to be in dim light, and the IR will show up as a purple glow.



The only way I can think to diagnose a phototransistor failure would be to check it with a scope, or maybe a multimeter. I'm not sure if the duty cycle would be high enough to get much averaged voltage with a multimeter.

Other than that, there are 3 ICs on the main board for that hand that could fail. A buffer for driving the LEDs, a multiplexer for scanning through the keys, and a schmitt trigger for debouncing the keys. Any problems with those components would almost certainly cause problems with multiple keys. So it's probably worth checking each key, and see if that normal mode key is the only one that doesn't work, or if there are any others. In particular, looking at my notes, the other key in the same scanning matrix "row" is the return key for that thumb.

But yeah, the datahand never had proper USB functionality. They just included a ps/2 to usb adapter.

 

[mikeincousa]

Thank you so much for all your sage comments. I will start working through them late this week.

Looking for a loose/ cracked wire is certainly like the first place to start, following the classic axiom the device has to be plugged in to work!

I do not own an oscilloscope right now but I know there are on-line apps that function like one.

I have been considering trying those or purchasing a simple used one for an Arduino project. Maybe you would have an interest in it?

The essence of the project is inflating and deflating a Blood Pressure cuff and data logging the points for unfiltered/ massaged file for graphing and calculations. The commercial units do not provide such data. Instead they engage multiple smoothing routines that, to me, obfuscate critical data and trends,

Thanks again for your insights and guidance. I will report back when I have made some progress...

 

[mikeincousa]

I am slowly proceeding with your recommendations.

In the past I used a Y-adapter to change the PS/2 Data Hand cables to USB. That did not work well with my current machine and powered hubs. I moved the project to my very dated Linux box. The PS/2 sockets are native.

Like with my previous Window machine I could type but the other functions do not operate as they should. If I can get this going again I learned that there are PS/2 PCI boards that will fit into my Window machine.

I have the left hand unit open. I have located the thumb switch cable that you mentioned (pink). I am ready to start doing some assessments as you suggested.

It looks like there are 8 wires in the assembly. Is one more prone to breaking than another? Could there be multiple broken wires?

My thought is completely removing the cable and checking each wire for continuity, Plan B would be purchasing a new cable and see what happens. Do you have specifications on the cable and a source for a replacement?

Thanks for any thoughts.

PS. Given the age of the keyboard and your report of broken wires maybe it would be wise to replace all the cables in both units? Any thoughts?

 

[JesusFreke]

It's not really a matter of it *breaking*, per say, more of just something not quite making good connection. Maybe the contacts inside the connector get some corrosion on them, or something like that. That sort of issue is usually very intermittent and hard to diagnose. Like, if you have the unit opened, it's likely the wires aren't in exactly the same position with the same forces on them as when the unit is closed, so the connection may not exhibit the same problem.

I would first start with jiggling the thumb cable at both ends, maybe unplug it and replug it, etc.

A new cable is certainly a possibility as well. I don't have my datahand handy atm, so I can't comment on the number of wires, but the connectors are JST XH connectors. Although, if memory serves, the cable for the thumb is actually a Y connector, with a second branch coming off for the LEDs. It's unlikely you would be able to find exactly that cable - unless someone is selling one specifically as a datahand replacement or something.

It's certainly possible to build a cable though. You might be able to find pre-crimped wires, and then you would just need the 3 connectors (for the mainboard, LED panel and thumb). Or, you can buy a crimping tool and crimp connectors onto wires yourself. digikey, mouser, etc. should have the connectors and the contacts.

crimping tutorial on youtube

 

[mikeincousa]

It's not really a matter of it *breaking*, per say, more of just something not quite making good connection. Maybe the contacts inside the connector get some corrosion on them, or something like that. That sort of issue is usually very intermittent and hard to diagnose. Like, if you have the unit opened, it's likely the wires aren't in exactly the same position with the same forces on them as when the unit is closed, so the connection may not exhibit the same problem.

I would first start with jiggling the thumb cable at both ends, maybe unplug it and replug it, etc.

A new cable is certainly a possibility as well. I don't have my datahand handy atm, so I can't comment on the number of wires, but the connectors are JST XH connectors. Although, if memory serves, the cable for the thumb is actually a Y connector, with a second branch coming off for the LEDs. It's unlikely you would be able to find exactly that cable - unless someone is selling one specifically as a datahand replacement or something.

It's certainly possible to build a cable though. You might be able to find pre-crimped wires, and then you would just need the 3 connectors (for the mainboard, LED panel and thumb). Or, you can buy a crimping tool and crimp connectors onto wires yourself. digikey, mouser, etc. should have the connectors and the contacts.

crimping tutorial on youtube

 

[mikeincousa]

I tried the shaking from the top but did not notice any change.Your po, about a difference in force on the wires between open and closed is well taken. I will try the jiggling again but now from both ends next session ,with the case open and the unit connected. Any concerns here?

I found the cable connectors easily enough.

I will also look for the fork on the thumb.

It looks like the switches in the wells and thumb arrays are magnetic or reed. I do not have experience with either. I will need to bone up for learning how to test and clean them.

If it boils down to the thumb switch fork I would look hard for one. If not building one is not of the range of possibility.

I have a very basic breadboard for working with my Arduino and a collection of electronic tools. I recently found a caliper IC puller. Looks like it would be a good choice for safely pulling the cables. I will purchase one. Professional IC Extraction Tool

Given what you say I think that re seating all the cables is a critical first step. Comments? See below.

Do you have any suggestions for a trial circuit to replace the Y?

What is the black plastic cover for?

I have been thinking of this on and off all day. I found my last correspondence with the factory: it was in 2011. Then they went over the unit, did something that I cannot recall now, and swapped out the processor. I recall them saying that the unit was very dirty when they opened it. Mindful of that my first effort in this run was taking the covers off and and blowing out the inside the best I could. There was a fine layer of dust but far from enough to choke a horse.

The case is not sealed. I have lived in dusty environments, New Mexico and Colorado, since the 2011 rebuild. Fine particles from fires and the like are ubiquitous. The connectors could very well become slightly coated with air pollution. My conclusion from this line is that a grit particle could be gumming up the works,

Bye for now. Thanks again for your thoughts and comments and support.

 

[JesusFreke]

The switches are optical, with a magnetic key return. There are no electrical contacts there to clean. The contacts I was referring to was the contacts in the cable itself.

I don't know what you mean by replacing the Y cable with a circuit. It's just a cable. Some of the pins on the mainboard connector go to the thumb unit, and other pins go to the LED unit (again, if I'm recalling correctly. It's either the thumb connector or the finger connectors.. but I'm pretty sure it's the thumb).

Yeah, the datahand is surprisingly resistant to dirt and grime for the most part. When I switched to the lalboard, I just put my datahand on a shelf as-is - and looking at it now.... holy cow was it nasty. haha.

The cables are a bit of a pain to unseat, the first couple of times you do it. There's a little lip on the 2 small edges that you can get something under to help pull up. But I think it's generally considered "okay" to pull on the wires directly if needed as well, as long as you try and pull on all of them equally. See, e.g. https://www.reddit.com/r/electronics/comments/1mxx6h/is_there_an_easy_way_to_unplug_these_type_of/

 

[mikeincousa]

I am not following you : "magnetic key return". I have not been able to get any hits in the Web. Could you point me to an example, preferably the ones used in the DH? My only sort of hit. https://www.digikey.com/en/products/filter/magnetic-reed-switches/193

Is this the unit that you built? The one in your build website?
http://xahlee.info/kbd/lalboard.html

It looks like it would work for me but I do not think I have the resources to build one right now. But if I cannot get my DH working that might be an option.

Have you patented it? Considered going commercial?

Do you have layouts/ templates of the pcb's? They would be helpful if I have to dig that deep?

 

[JesusFreke]

They are not magnetic reed switches, they are optical. The key press is sensed by having the key stem move between an IR LED and an IR phototransistor, and sensing the voltage change of the phototransistor. Mechanically, it uses magnets to hold the key in the unpressed state, and to return to the unpressed state after being pressed. But the magnets don't play any role in the actual sensing of a key press.

Yes, that's me I don't have the time, resources, expertise or even desire to try to patent anything in it. I don't have time currently to try to make a business out of it, but I won't rule it out in the future.

You can find info on how I made the PCBs here: https://github.com/JesusFreke/lalboard/wiki/Vinyl-Cut-PCBs. Although I have been working on a new iteration of the design, which will use an actual PCB, that can be fabbed by a standard PCB fab.

 

[mikeincousa]

I sort of have the general drift on the keys, but lacking experience I can't fully grasp the operations. Are the LED, IR photo-transistor, and magnet a commercial assembly? If so where could I get a look? If not could you point me to the components / a circuit that I could mock up for gaining a better understanding? Thanks.

I will likely come back to your other topics later.

 

[JesusFreke]

No, they are discrete components in a "custom" (to to the datahand) plastic housing. The LED and PT sit on either side of each key stem, with the stem normally blocking the PT from "seeing" the light from the LED. When the key is pressed, the key stem moves out of the light path, and the PT senses the light and changes voltage.

I don't know the exact LED and PT components that the datahand uses, but they're very similar to the ones I use in the lalboard: LED, PT

And then, for example, for the N,S,E,W keys in each key cluster, there is a magnet on the key stem, and then a small metal strip in the central part of the plastic housing for the key cluster. The magnet in the stem is attracted to the metal strip, and so it keeps the key vertical and provides a nice tactile feel as the magnetic force quickly drops off as soon as the key stem starts getting pressed.

(cross sectional view)

       |(  <- (key)
       |
       |] |  <- (metal strip)
       |  |
       o  |

] is a magnet, embedded in the key stem.
o is the pivot for the key

The keys on the lalboard are similar, except that instead of a metal strip opposite the key, it uses another magnet.


Interestingly, it's a very different force profile/tactile feel from a normal mechanical switch. With a normal mechanical switch, the maximum force will be somewhere in the middle of the key travel (depending on the exact mechanism used, of course). But with the keys in the datahand, the maximum force is at the very start of the key travel, since the magnet in the key stem is right up against the metal strip. And then quickly falling off as the magnet moves away from the strip. But the key travel is short enough that there's still enough magnetic attraction when the key is fully pressed to "pull" the key back up into the vertical position.

The basic circuitry for each key is just the normal circuit for a phototransistor. See, e.g. https://www.elprocus.com/phototransistor-basics-circuit-diagram-advantages-applications/#:~:text=A phototransistor works just like,device only needs 2 pins.

 

[mikeincousa]

Resuming. Sorry for the delay. So goes life.

1. Switches. Thanks for the detailed explanations. I think I have grasped the basic concepts. In general for the optical switches. I liken them to an optical beam functioning as a door counter. For pulsing / switching I liken that action to a telegraph key.

2. Wires. I also did some homework on the failure of thin wires. That came after thinking about the longevity of POTS and doorbell circuits wires. The takeaway was that there is not any conclusive failure-modeling for these, especially for the POTS wires after decades service.

3. Potential Clue? Left Hand Thumb Assembly.

I poked and wiggled both of the red wires in middle and at the connectors the best could without taking all the assemblies out of the case. I did not see any reactions or differences on the screen.

Next. Recall that I could not run the DataHand and a flat keyboard on my Window machine. I concluded then that Windows 10 code did not like two different keyboards. Now I wonder if I just had the two PS/2 lines reveresed; or if that blockage came from the stacking of adapters, or both?

My old Linux (Ubuntu 18.04 running on P4D hardware) does not have any problems running both keyboard designs. I used both at the same time for the Left Thumb testing.

When I was testing the Left Thumb action with the DataHand like before I could switch the DH to the mouse mode but I could not get back to the typing mode. When the DH locked like this I found that I could type a few characters using the flat keyboard, then the green Left Hand light came back on, and then I could resume typing with the DataHand!

I am wondering if you see a clue in this? To me it sounds like the firmware it okay and this adds to the speculation that I am dealing with a cracked wire?

4. Given the above to you have any other suggestions for in-case testing?

Absent any new thoughts I think the next step needs to be conducting out of case continuity checks starting with the left had thumb assembly?

5. Pin Assignments. Have you mapped which wire is doing what functions in each of the connector circuits? Such an inventory might be helpful for isolating a broken wire.

Thanks again for your ongoing support.