XT/AT/PS2/Terminal to USB Converter with NKRO

[Soarer]

Yeah, I know what you mean.  The docs are a great reference once you know what questions to ask.  People learn things differently.  I learn the most quickly by working backwards from an example.  Unfortunately, I missed Halver's.  Most folks are going to have a specific project in mind when using this.  I imagine that IBM XT and terminal keyboard owners are probably some of your biggest users.  It may be worthwhile to provide walk-throughs for those and provide boilerplate configs for those projects' FAQs?

Hmm, I dunno, each person's project has different goals! I think it's a better approach for me to fill in the gaps in the documentation, since there's really very little that's unique to any keyboard type. Or at least to do that first, otherwise the answer to someone's AT-related question might be hiding in the 122-key tutorial, for example, and they could easily not find it.

Also, I think making an index of the example configs will help, since I'll try to highlight the notable features of each rather than just say "Halvar's config for 122-key".

[wcass]

...
I forgot to mention, the program has a few odd quirks (who doesn't?)

It ignores "Tab", it locks up after "F10", and it respects "CapLock"

If it goes funny during or after "F10", "Alt" or "Spacebar" you should use your mouse to highlight whatever is in the dialog box, then when you press a key it will go back to normal.

These are not bugs, they are "idiosyncrasies" - HAHA

ROTFL
i recomend "Aqua Key Test" - much easier to search for. download from OTD or GH links to make sure you get the right one.
http://geekhack.org/index.php?topic=34670.0
 

[fohat.digs]

...
I forgot to mention, the program has a few odd quirks (who doesn't?)

It ignores "Tab", it locks up after "F10", and it respects "CapLock"

If it goes funny during or after "F10", "Alt" or "Spacebar" you should use your mouse to highlight whatever is in the dialog box, then when you press a key it will go back to normal.

These are not bugs, they are "idiosyncrasies" - HAHA

ROTFL
i recomend "Aqua Key Test" - much easier to search for. download from OTD or GH links to make sure you get the right one.
http://geekhack.org/index.php?topic=34670.0
 

I don't get it.

Aqua Key Test would seem to be OK if you are testing an ANSI keyboard to see whether the keys are working, but since I am accustomed to keyboardtest.exe, all the others are just annoying. I don't remember what I used before, since I am happy now.

I really want the ability to identify keys not standard to an ANSI board. How do I use it to find out which key is F22? What about "Media Pause" or "Numpad *"? Aqua does show the keycode in tiny print at the bottom, but without identifying the key for a simpleton like me.

keyboardtest.exe works great for all this, too, and seems more user-friendly. I can easily live with a few known glitches if it solves my problems.

[Soarer]

Aqua's is great if you want bypass autohotkey etc., but I also (most of the time, actually) like to use a tester that doesn't hook so deeply and just listens to Windows messages. OTOH, one that doesn't go funny after F10, Alt or Space is good!

[metalliqaz]

So which do you use, Soarer? 

[Krogenar]

Also, I think making an index of the example configs will help, since I'll try to highlight the notable features of each rather than just say "Halvar's config for 122-key".

If you're going to be updating the documentation anyway, I think it would be a great idea to give a list of some places to find the materials necessary -- it was a big help to me when I made my converter having someone point out what the right connector was, and where it could be found. You couldn't post a link itself, but listing some electronics companies might help others.

[Soarer]

So which do you use, Soarer? 

Well my own, of course!

I started writing it before I knew of Aqua's (perhaps before Aqua's was available, I dunno). I might release it at some point, but it's only ever as good as I need at any given time, as a tool for programming.

[wcass]

at one time, i dreamt of writing a GUI for Soarer's that would be as easy to use as the Aikon UI. the app use hid_listen to identify what key you just hit and you use the mouse to select what to assign that key to from 4 tables (non-printed characters, common printed characters, uncommon printed characters, predefined macros). alas, i have not the time or skill.

someone might find this useful. it is a big table of HID id's and PS2 make/break codes. iirc i got if from Microsoft.
 USB-PS2.xls (63 kB - downloaded 402 times.)

[sleepy916]

I haven't posted in here yet but I certainly have read this thread more than a few times. I would have been lost without it, thank you Soarer.

AKIMbO mentioned to put this here so I will.

I made this little picture looking at my post it notes. Not sure if it is useful for anybody but I'll post it here anyways. My paint skills leave a lot of room for improvement. This is specifically for my AT Model F.

(╯°□°）╯︵ ┻━┻)



Fixed...

[Soarer]

Oh, that's cool 

But something's wrong somewhere because it should have clock -> pd1 and data -> pd0 in both cases 

[sleepy916]

Oh, that's cool 

But something's wrong somewhere because it should have clock -> pd1 and data -> pd0 in both cases 

Ahh...my copy and paste skills.. 

[Krogenar]

I juuuuust unpacked a Model F 122-key (my first Model F) and I notice it has the same 240-degree DIN plug as my Model M 122-keyboard. Is it safe for me to plug this thing into my existing Soarer converter? Any thoughts? My first reactions are:

1. Frickin' HEAVY. This is the keyboard that will survive the apocalypse, for sure. It's just so solidly built. You could use it to work your upper body, like that guy who used his typewriter in Misery.
2. Yeah... it's damnably *pingy* -- "This is the machine that goes *PiNg!*" <--- Monty Python reference. It may grow on me. The sound somehow makes it sound ... cheaper? Though I know that's not the case. I'll consider floss modding eventually.
3. The feet. How do the feet engage? Just pushing on them with fingers doesn't work. The circles on the edges of the keyboard don't have any kind of fingerhold that would allow me to rotate them into position. See potato pics below.






Yes! Press the buttons on the edge, and the flippers (they're pinball-sized flipper feet) slide down. The angle at which the keyboard is set when the feet are engaged is... feels steep! Anyway, thought I'd share the unwrapping.

[rknize]

There are two sets of feet.  The inner ones are a bit tricky to coax out of the outer ones, but it gives a more typical angle.

[fohat.digs]

Those double feet area trick if you want to take the case apart - pay VERY close attention to how they are assembled. I never use them.

The sound is really too much without the floss mod. Somebody said that the board "sings" and I think that is a good description.

As to the weight, people keep talking about "6 pounds" I don't know where that comes from. All mine weight 8 and a quarter.

[Krogenar]

There are two sets of feet.  The inner ones are a bit tricky to coax out of the outer ones, but it gives a more typical angle.

Really? Gotta give those a try. Do you guys think my Soarer converter will work with this Model F? The one I made for my Model M 122-key has the exact same connector... is it safe to just plug this guy into it same as the Model M?

[Krogenar]

As to the weight, people keep talking about "6 pounds" I don't know where that comes from. All mine weight 8 and a quarter.

Yeah, this is way more than 6 pounds.

[rknize]

There are two sets of feet.  The inner ones are a bit tricky to coax out of the outer ones, but it gives a more typical angle.

Really? Gotta give those a try. Do you guys think my Soarer converter will work with this Model F? The one I made for my Model M 122-key has the exact same connector... is it safe to just plug this guy into it same as the Model M?

If it's the same connector and the converter is setup for the terminal key map, it should "just work".

[Soarer]

Model M 122-key (square label and hat key vintage): 6lbs 3.5oz / 2.8kg 

Model F 122-key: 8lbs 12.75oz / 4kg

Filco TKL: ~0.9kg (from memory)  )

[Krogenar]

Well, it "just works" -- except one key won't work. Just ONE, lol. So far I've tried removing both parts of the keycap, and when the spring itself is brushed even slightly it will register on AquaKeytest as having been pressed, but once the two pieces are replaced over the spring, nothing. Any idea what could cause that behavior? Here's hoping I don't have to take it apart.

[rknize]

Sounds like the spring is bent or not seating in the key cap correctly. If its the latter, leaning the board back while inserting the cap sometimes helps

[poxeclipse]

rknize is right, it happens during shipping, some hammers move and the spring doesn't sit properly in the key cap.
The key cap has, sometimes, a small nib, which should sit inside the spring; if not, then the spring doesn't bent and doesn't move the hammer, so there is no change in capacitance and no signal sent. The sound is different, with no 'ping'.

[sleepy916]

I noticed that the Model Fs were harder to get the keycap on correctly than the Model Ms.

[Krogenar]

rknize is right, it happens during shipping, some hammers move and the spring doesn't sit properly in the key cap.
The key cap has, sometimes, a small nib, which should sit inside the spring; if not, then the spring doesn't bent and doesn't move the hammer, so there is no change in capacitance and no signal sent. The sound is different, with no 'ping'.

I'm going to give that a try tonight. Pregnant Mrs. Krogenar was very needy last night, so I didn't have a chance to test this. The fact that the spring registered any kind of keypress at all made me feel that there's got to be a way to fix it. The keyboard itself arrived very well-packed by the sender. The spring did not appear or feel as though it was bent. It had the same 'springiness' as a neighboring, working spring.

[simkev]

If you have a USB keboard that has NKRO over a usb to PS2 adapter could you make that board NKRO via the teensy without the PS2 adapter?Solder the ubs cables to the teensy?

[alaricljs]

If the teensy is configured with a firmware that interfaces to a PS/2 board and supports NKRO then yes, it should work.  You'd need to attach the keyboard via PS/2 protocol pinout.  If for you that means cutting the cable and soldering it direct instead of using an adapter or some other method, then yes it will work if you get the pinout correct.

[Soarer]

If it's a keyboard with a captive cable (i.e. no USB socket on the keyboard) it might be easier to remove the cable and wire the Teensy inside the keyboard, since the cable is quite likely to have a connector where it joins the PCB. Of course, you'd still have to work out the pinout.

[Krogenar]

rkinize, fohat, I recognize your expertise with the Model F. I leaned the 'board backwards, and inserted the keycap sections, and all is well. It took a few tries, but now it's working! Thanks so much for your help guys, it is greatly appreciated. (tips hat)

[fohat.digs]

It can take several attempts to get it to seat properly. Sometimes they are just stubborn. Making the spring flop forward is usually helpful.

[rknize]

rkinize, fohat, I recognize your expertise with the Model F. I leaned the 'board backwards, and inserted the keycap sections, and all is well. It took a few tries, but now it's working! Thanks so much for your help guys, it is greatly appreciated. (tips hat)

Great!  The same trick works on the M, too.  If you look under the key cap, there is a little nub where the spring is supposed to seat.  Gravity makes the springs lean towards you when you insert the cap, so it has to ride the ramp and then hopefully pop over that nub.  Sometimes it just won't, so you have to guide the spring yourself.

[Soarer]

I just added a v1.11 update to the first post:

• Added support for IBM RT keyboard.
• Added display of boot mode state by making converter lock LEDs glow dimly.
• Added extra error codes to help determine where timeouts occur more precisely.
• Prevented debug and rawhid output when keyboard_protocol is not set (i.e. in BIOS mode).

The last change, to prevent debug output when in BIOS mode, is another attempt at fixing the long-standing incompatibility with certain Dell machines (M6300 (BIOS rev A14) and Inspiron 630m). If anyone has one of those, please test!

[Halvar]

Thanks Soarer for constantly pushing this forward! 

[jpeart]

Soarer, thank you for all your efforts in developing this firmware.  I spent the last few weeks buying different adapters/converters trying to get an older keyboard working to no avail.  I found your converter while googling and decided to give it a try.  I ordered a Teensy 2.0 last weekend and I was very excited to see it had arrived today.  Took me just a few minutes to hook it all up for a test run and it works perfectly!  With the other products I had purchased I was getting a lot of key repeats, but I have not seen a single problem using your firmware.

If anyone is interested I'm tracking the progress of my little project here http://geekhack.org/index.php?topic=42150.0
Basically I'm trying to put a Raspberry Pi inside a Keycat keyboard and make my own kind of Commodore 64.

[fohat.digs]

Oh, no! It's that pesky fohat.digs again!

A couple of months ago I had a big problem after upgrading a motherboard from Asus to Gigabyte. I keep a powered USB hub at the back of the desk, and my keyboards and mice plug into it. I have done this for a long time, but the Gigabyte motherboard refused to boot (not even a peep or POST) with the powered USB 2.0 hub connected. When I went to a non-powered hub, the problem disappeared.

There are USB 2.0 and 3.0 sockets on the back, and since I have a 3.0 external hard drive, I wanted to upgrade the USB hub. I bought a nice SIIG 4-port powered USB 3.0 hub, and plugged it into the (previously vacant) 3.0 port.

Now, the computer boots fine and recognizes both of my mice, but will not recognize my F-122/Teensy (Teensy 2.0 with v1.10 Soarer's Converter) until I unplug it and hot-plug it. I have been through several power-up/power-down cycles and it is still happening.

I know that motherboards are funny about these things, sometimes, and the stock answer is: "plug the keyboard directly into the motherboard" but I would prefer to make the hub work if I could.

Are there any new answers since last time?

Thanks! Harry

[rknize]

Look for a BIOS upgrade?  This weird boot behavior with keyboards usually has to do with legacy keyboard emulation that the BIOS tries to do with USB keyboards.  Look in your BIOS options and try disabling legacy keyboard emulation if such an option exists.

[vipjun]

Just wanted to show my appreciation for this project. thanks soarer!

[AKIMbO]

Successfully modded one of my AT F's to a traditional ANSI layout (using a function block where Fn + 1 is F1, Fn + 2 is F2, etc). 

Thanks again Soarer!

[Halvar]

That's beautiful.

I wish Unicomp's European resellers had those media keycaps to use it with my M-122.

And that there were any Model Fs to be found here in Europe...

What key do you use for the Fn key?

[AKIMbO]

That's beautiful.

I wish Unicomp's European resellers had those media keycaps to use it with my M-122.

And that there were any Model Fs to be found here in Europe...

What key do you use for the Fn key?

Capslock + 1= F1...all the way through Caplock + = which is F12

[sleepy916]

Successfully modded one of my AT F's to a traditional ANSI layout (using a function block where Fn + 1 is F1, Fn + 2 is F2, etc). 

Thanks again Soarer!

Show Image

That looks to be in pristine condition! Now just send it to my house, I'll pm you my address shortly. ;]

[AKIMbO]

Successfully modded one of my AT F's to a traditional ANSI layout (using a function block where Fn + 1 is F1, Fn + 2 is F2, etc). 

Thanks again Soarer!

Show Image

That looks to be in pristine condition! Now just send it to my house, I'll pm you my address shortly. ;]

It is...I actually struggled with whether I wanted to mod it or keep it in stock condition for quite some months.  Today I said fugg it and modded it.

[Halvar]

After looking at other Model F pictures: wow, how much did you have to mod here? 

Did it look like this one with the big-ass return key, small backspace and all?
http://www.clickykeyboards.com/index.cfm/fa/items.main/parentcat/9934/subcatid/0/id/529210

That's no small feat then ... I feel dumb for even talking about keycaps first...

[Krogenar]

Akimbo, I'd love to do what you did (more ANSI-like layout) with my Model F-122, but frankly I don't want to risk ruining my 'board. Looks awesome. Do those additional keys feel particularly different from the original keys? I mean the actuation, sound, etc.?

[fohat.digs]

The mod is totally easy, assuming that you can take the board apart and put it back together.

It is a great opportunity to tidy everything up, and swapping 2 spring locations is utterly trivial.

Of course, putting it back together is the hardest and most challenging part.

[AKIMbO]

After looking at other Model F pictures: wow, how much did you have to mod here? 

Did it look like this one with the big-ass return key, small backspace and all?
http://www.clickykeyboards.com/index.cfm/fa/items.main/parentcat/9934/subcatid/0/id/529210

That's no small feat then ... I feel dumb for even talking about keycaps first...

Yup, it sure did look like that (small backspace and all).  Like fohat said, it isn't a hard mod....open up the board and move a few springs around, add a couple white barrel inserts from an old Model M, and close her back up.  The hardest part is stripping and soldering the AT cable to the Teensy.

Akimbo, I'd love to do what you did (more ANSI-like layout) with my Model F-122, but frankly I don't want to risk ruining my 'board. Looks awesome. Do those additional keys feel particularly different from the original keys? I mean the actuation, sound, etc.?

Nope...the additional keys feel and sound exactly like the original keys.

[Retrete]

Nice post! I have a pseudoterminal  kevex ray 7000CM, and I want to adapt the keyboard in order to use it with a modern computer, the keyboard terminal use this port;
http://www.flickr.com/photos/95230480@N02/8683117351/in/set-72157633342730175

As I don't have a clue about this, I also took photos of the terminal w/o tape;
http://www.flickr.com/photos/95230480@N02/sets/72157633342730175/

The keyboard looks great 


Thanks in advance!

[Soarer]

Beautiful! A bit crazy in construction, with a PCB and wire-wrapping!

Not something this converter can convert though. Do you know what switches it uses? If they are simple contact switches, then the easiest way to get it working would be to remove the chips and wire in a new controller. I'm currently working on one aimed at non-programmers, or there's hasu's code which is pretty easy to configure and build.

[Retrete]

Beautiful! A bit crazy in construction, with a PCB and wire-wrapping!

Not something this converter can convert though. Do you know what switches it uses? If they are simple contact switches, then the easiest way to get it working would be to remove the chips and wire in a new controller. I'm currently working on one aimed at non-programmers, or there's hasu's code which is pretty easy to configure and build.

I don't know exactly what type of switches are, non clicky and the loudest I've seen ;
http://www.flickr.com/photos/95230480@N02/sets/72157633364166384/

Thanks again!

[Soarer]

They look rather like Cherry M7 (or M6?) with the 'T' mount stem... Catalogue 1982, pp22,23. There's even a couple of keyboards in there with a similar layout to the main alpha block of yours.

So most likely to be simple contact switches. You could check with a multimeter easily, if you have one...?

The first task in replacing the controller would be to identify the rows and columns. I did that by tracing the PCB tracks on an Ortek MCK-84. It's quite tedious It's not really necessary to go to such detail; simply knowing which are rows and which are columns would suffice, and the rest can be worked out after the controller is wired up.

Some straight-on pictures of the PCB would help, if you get to taking it apart that far. It looks like those supports could be unscrewed from underneath, leaving them attached to the PCB. Not sure if they'll block the view of anything important though!

[Retrete]

I've already disassemble the keyboard cover and I took a couple of photos, It isn't linear ;
http://www.flickr.com/photos/95230480@N02/sets/72157633350772239/

[Soarer]

Hah! I wasn't expecting it to be a double-sided PCB... at least it makes one side very easy to trace out! There's clearly 8 traces there for the rows, plus one trace for the shifts which seems to go to ground. On the other side there are probably 11 (or more) traces for the columns. All the lines presumably go to these sockets:



I'd guess that the rows all go to the one on the left, which looks like it has a resistor pack in it (acting as pull-ups or pull-downs). Testing with a multimeter - or even a cheap continuity tester - would confirm the connections.

The 'controller', such as it is, consists of all the chips (12?). None of them look big enough to be a CPU, but 12 is more than would be needed for a completely 'dumb' keyboard. So that's kind of interesting, but probably not very useful! (There's what looks to be a voltage regulator near where that ribbon cable connects, so some of the chips will simply buffer voltages).

So as controller swaps go, it looks fairly straightforward - remove all the chips, wire a Teensy or Teensy++ to some of the connections on the the sockets I marked, customise and load some code, and it should work! Of course, I don't know how comfortable you'd be doing that kind of thing...