2 Feb
2014
2 Feb
'14
1:26 p.m.
Hi,
I've just got (another!) AT&T UNIX PC keyboard to replace the one with
missing keycaps and -- woe and behold -- this one has a few dead keys!
For bonus points, it doesn't use the same switches as the other
keyboards: it uses ITW "Magnetic Valve" keyswitches instead --
Details here:
http://deskthority.net/wiki/ITW_magnetic_valve
Photos here:
http://deskthority.net/keyboards-f2/reverse-rubber-dome-switch-t449.html
http://park8.wakwak.com/~kaineko2/keyboard/album20110107.html
These switches were apparently used in Decision Data, Aydin and Devlin
keyboards, and consist of two metal pins and a tiny ferrite in the top
of the black keyswitch base, and a pink plastic key which houses a large
chunk of ferrite. A small chunk of rubber inside the keycap forms the
spring which returns the key to the "not activated" position.
The problem I have is that the ferrites on several of the keys have
broken in transit... with these being delightfully obscure and
apparently obsolete, my chances of repairing this thing (without parts
at least) are slim.
Does anyone have any of these keyswitches kicking around in a spares
box, or a scrap keyboard which uses these keyswitches?
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
3 Feb
3 Feb
1:20 p.m.
Hi,
I've just got (another!) AT&T UNIX PC keyboard to replace the one with
missing keycaps and -- woe and behold -- this one has a few dead keys!
For bonus points, it doesn't use the same switches as the other
keyboards: it uses ITW "Magnetic Valve" keyswitches instead --
Are those the ones with a core that's sturated by a permanent magnet, and
which is desaturated when the key is pressed so that there is increased
coupling between 2 'widnigns' on said core?
HP used thsoe in the HP9845 (at least). Really odd to test, you had 2
pairs of pins that were dead short irrespective of the state of the key.
I had to take one apart to figure out what was going on. Acutally HP also
had some 3-state ones (up, pressed, pressed hard) with 2 cores, and 6
connections (there was one pair of pins that went to a winding on both
cores, then separate windings, one on each core, to the other 2 pairs of
pins).
-tony
4 Feb
4 Feb
5:10 p.m.
On 03/02/14 21:20, Tony Duell wrote:
Are those the ones with a core that's sturated by
a permanent magnet, and
which is desaturated when the key is pressed so that there is increased
coupling between 2 'widnigns' on said core?
Yep, that seems to be the design principle. There are two U-shaped leads
(four terminals) -- drive and sense -- and a core-memory-sized ferrite
ring in the hoops, like this:
___ ___
/ \ / \
| | | |
| XXXXXX |
| | | |
The "X" is the ferrite ring.
I might be able to fix the switch if I can remove the wire hoops and
find a suitable replacement ferrite, though that seems at least as
difficult (possibly moreso) as finding replacement keyswitches.
I'm actually tempted to get a Nd magnet and a few small ferrites and try
this on a breadboard. Seems like it might be a fun experiment to do in
an evening and then write up (or demonstrate in a video).
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
5 Feb
5 Feb
6:32 a.m.
I might be able to fix the switch if I can remove the wire hoops and
find a suitable replacement ferrite, though that seems at least as
difficult (possibly moreso) as finding replacement keyswitches.
I'm actually tempted to get a Nd magnet and a few small ferrites and try
this on a breadboard. Seems like it might be a fun experiment to do in
an evening and then write up (or demonstrate in a video).
Look for RF cores, not ones for shielding. These are usually easy
to saturate.
Dwight
1:45 p.m.
On 03/02/14 21:20, Tony Duell wrote:
Yes, that looks like the innards of the HP ones.
I think HP just had one magnet on the key plunger, no fixed magnet. The
magnet is normally near the ferrite, wehen you press the key it moves
down out of the way.
[...]
Are those the ones with a core that's
sturated by a permanent magnet, and
which is desaturated when the key is pressed so that there is increased
coupling between 2 'widnigns' on said core?
Yep, that seems to be the design principle. There are two U-shaped leads
(four terminals) -- drive and sense -- and a core-memory-sized ferrite
ring in the hoops, like this:
___ ___
/ \ / \
| | | |
| XXXXXX |
| | | |
The "X" is the ferrite ring.
I'm actually tempted to get a Nd magnet and a few small ferrites and try
this on a breadboard. Seems like it might be a fun experiment to do in
an evening and then write up (or demonstrate in a video).
Theorigianl magnets are nto that strong. I don't think you need an Nd magnet.
The circuitry that HP used should be in 'my' HP9845B schematics. From
what I rmmemebb the actual key drive/sense circuitry doesn't involve any
custom or hard-to-find parts. There is a single-chip keyboard encoder,
but you wouldn't need that if you were jsut making one key as a
demonstration.
-tony
6:01 p.m.
On 05/02/14 21:45, Tony Duell wrote:
This one is the same - magnet on the plunger, but just a ferrite on the
key base.
I noticed -- but I have Nd magnets in stock from a 3D printing project
(and some smaller ones from an aborted micro-actuator project), anything
else I'd have to order in.
Yep, that
seems to be the design principle. There are two U-shaped leads
(four terminals) -- drive and sense -- and a core-memory-sized ferrite
ring in the hoops, like this:
___ ___
/ \ / \
| | | |
| XXXXXX |
| | | |
The "X" is the ferrite ring.
Yes, that looks like the innards of the HP ones.
I think HP just had one magnet on the key plunger, no fixed magnet. The
magnet is normally near the ferrite, wehen you press the key it moves
down out of the way. I'm
actually tempted to get a Nd magnet and a few small ferrites and try
this on a breadboard. Seems like it might be a fun experiment to do in
an evening and then write up (or demonstrate in a video).
Theorigianl magnets are nto that strong. I don't think you need an Nd magnet. The circuitry that HP used should be in 'my'
HP9845B schematics. From
what I rmmemebb the actual key drive/sense circuitry doesn't involve any
custom or hard-to-find parts. There is a single-chip keyboard encoder,
but you wouldn't need that if you were jsut making one key as a
demonstration.
I'll see if there are any schematics floating around... otherwise a
simple description will likely be enough to reproduce it.
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
6 Feb
6 Feb
12:44 p.m.
This one is the same - magnet on the plunger, but just
a ferrite on the
key base.
Posibly even the asme part as HP used. I don't think HP made them
themselves.
As I mentioned, HP had a few wiht 2 ferrites and 3 loops of wire. In the
up state, both ferrites were saturated by the magnet. When you pressed the
eky, it moved the magent far enough from the upper ferrite to allow it to
couple the 2 loop,s but the lower one rmained saturated. Press the key
hard and rthe magnet was moved dwon far enough to allow both ferrites
top[ couple their respective loops. This was used for the 4
cursor-movement keys in the hP9845, which auto-rpeat if pressed hard.
I noticed -- but I have Nd magnets in stock from a 3D
printing project
(and some smaller ones from an aborted micro-actuator project), anything
else I'd have to order in.
You mean you don't have a box of assorted magnets?
The circuitry that HP used should be in
'my' HP9845B schematics. From
what I rmmemebb the actual key drive/sense circuitry doesn't involve any
custom or hard-to-find parts. There is a single-chip keyboard encoder,
but you wouldn't need that if you were jsut making one key as a
demonstration.
I'll see if there are any schematics floating around... otherwise a
simple description will likely be enough to reproduce it.
3924
days inactive
3928
days old
6 comments
3 participants
participants (3)
-
ard＠p850ug1.demon.co.uk -
classiccmp＠philpem.me.uk -
dkelvey＠hotmail.com