Beogram CDX 2, 1987, touch sensors

[N Cook]

As a block of 4 non functioning probably a problem with the MAB8441 I2C
encoder, otherwise plays CDs in their entirety but cannot preselect tracks.
But I was just wondering how these sensors worked (reliably ?), presumably
just relying on stray domestic rf induiced in humans. There are no metal
pads to touch on the surface but are under the plastic cover and each
"aerial" pad connected to the unbiased base of a transistor with a gain of
100 or so.

Incidently the keyboard + encoder of this one is much like the Beogram CD
4500

 

[N Cook]

As a block of 4 non functioning probably a problem with the MAB8441 I2C
encoder, otherwise plays CDs in their entirety but cannot preselect tracks.
But I was just wondering how these sensors worked (reliably ?), presumably
just relying on stray domestic rf induiced in humans. There are no metal
pads to touch on the surface but are under the plastic cover and each
"aerial" pad connected to the unbiased base of a transistor with a gain of
100 or so.

Incidently the keyboard + encoder of this one is much like the Beogram CD
4500

mystery deepens
The key action has to pass through 3mm of glass
Removing from the cover and touching the pins that connect through to the
bases activates that switch.
Any thickness/thinness of plastic covering those pins and touching then no
action.
There is a part loop of aluminium that is part of the overall casing that
goes around the edge of this key pannel , connected to ground but I don't
see how that is involved

CDX 2 , type 5161 if that is relevant

 

[N Cook]

mystery deepens
The key action has to pass through 3mm of glass
Removing from the cover and touching the pins that connect through to the
bases activates that switch.
Any thickness/thinness of plastic covering those pins and touching then no
action.
There is a part loop of aluminium that is part of the overall casing that
goes around the edge of this key pannel , connected to ground but I don't
see how that is involved

CDX 2 , type 5161 if that is relevant

weirder and weirder, anyone seen the like?

For the CDX 2 in front of me, 20 "keys"
Each transistor has the base connected to a pin that touches soft conductive
plastic in a well under the glass and each base has a 8p2 capactior to
ground and also a resistor in parallel but the value of that varies between
55K , 100K and 120K for no apparent reason.

Looking at the CD4500 schematics with 8 "keys" i would expect all the same
but there, that resistor, varies between 22K,27K,33K,47K and 56K,
transistors all the same BF840 and each circuit around 4071 inputs the same.

Looks like these Rs need going up to 200K or so

 

[b]

mystery deepens
The key action has to pass through 3mm of glass
Removing from the cover and touching the pins that connect through to the
bases activates that switch.
Any thickness/thinness of plastic covering those pins and touching then no
action.

then get a hot pin and make a tiny hole above each switch thru the
outer plastic ;-)

 

[N Cook]

then get a hot pin and make a tiny hole above each switch thru the
outer plastic ;-)

The activation pin goes into this compressible conductive plastic (100 to
200 ohms or so) in plastic wells of the casing but on top of that is a 3mm
sheet of glass.
Even if it was very high lead content glass I cannot see how that would make
it conductive.
Today I will try temporary drawing pins to some of these pins to see if they
need an intercept area for rf reception. I would really like to know why
these spread of base resistors , almost as though selected on test, but are
written into the schematic.
There is not a left to right spread or, nearest the mains transformer is
lowest, or any seemingly rational explanation for the distribution of
values.
Also a bench supply through 10 or 100M to each pin to see what the trigger
level is of each "key" through the I2C encoder. Although all keys work by
direct touching as one block of 4 is not working through the glass I suspect
the state change level of one line of the I2C chip is different to the other
4.

 

[N Cook]

Beogram then

The activation pin goes into this compressible conductive plastic (100 to
200 ohms or so) in plastic wells of the casing but on top of that is a 3mm
sheet of glass.
Even if it was very high lead content glass I cannot see how that would make
it conductive.
Today I will try temporary drawing pins to some of these pins to see if they
need an intercept area for rf reception. I would really like to know why
these spread of base resistors , almost as though selected on test, but are
written into the schematic.
There is not a left to right spread or, nearest the mains transformer is
lowest, or any seemingly rational explanation for the distribution of
values.
Also a bench supply through 10 or 100M to each pin to see what the trigger
level is of each "key" through the I2C encoder. Although all keys work by
direct touching as one block of 4 is not working through the glass I suspect
the state change level of one line of the I2C chip is different to the other
4.

Pushing a drawing pin in one of the base pins then piece of 0.2mm polythene
over that and touching would trigger but not doubled thickness, or 0.5mm
ptfe or through 1.5mm microscope slip glass.

Could not connect DC to any pin without the wires on their own triggering .
Connecting me to the pins via 100M ohm then no key triggered, 10M all
triggered and in range of 33M and 50M showed up the difference of ones with
base resistors of 56K compared to 120K

I will change the 4 problem key base resistors from 120K to 220K and
reassemble and see what happens or not

 

[N Cook]

So much for all that palaver, reassembled but there were even less operative
keys.
Which left just the conductive plastic that miust be causing problems.
I prized out one of these conductive plugs from its plastic well under the
glass. There is a ledge in these wells, and now removed, the conductive
cylinders are not solid all the way down but open to a hollow cylinder at
the ledge which is very weak and splits at that point, even before my
getting to it.
The glass has some metalisation , perhaps nickel spray, on the rear of the
glass.
Screwing up some aluminium foil in these wells to make contact with the
frail and resistive , order 1K across 5mm, and then fixing compression
springs to the base pins to touch the aluminium has solved the problem.

Also soldering a piece of 26mm square brass to the 10mm diameter drawing pin
then stuck in a base pin, shows that the variable base pins probably vary
according to the (unseen as glass is glued to the plastic) variable areas of
metalisation under each legend.
With larger contact area (higher C ?) then finger injected rf would trigger
easily through 1.5mm of glass or even 4mm of perspex.

As a refinement can anyone think of phospher bronze springs or something non
corrodable for this application. I used ordinary steel springs. I tried fine
heater element , whatever metal that is, but although it looks springy it is
not reliably so - it relaxes. Or even where to find phosphor bronze wire
from a non specialist supplier, to make some springs.

In the archives its obvious that a lot have people have been defeated by
these B&O touch sensors - but the solution is now in the public domain.