# Touch Lamps, principle thereof?

Discussion in 'Electronic Design' started by George R. Gonzalez, Sep 25, 2003.

1. ### George R. GonzalezGuest

Just how DOES one of those infernal touch lamps work?

Is it picking up the AC signal on my body,
or does it sense a change in capacitance, or what?

Anybody know?

Thanks,

George

2. ### Mike HarrisonGuest

It's sensing your capacitance to ground.

The switch is effectively at AC mains potential. Of course there is a large-value resistor between
the switch innards and the touch plate to limit the current when you touch it.
When you touch the plate, your capacitance causes a small current to flow.
From the switches point of view, your touch looks like someone has connected mains to the plate,
resulting in an easily detected voltage which does the switching.

3. ### Walter HarleyGuest

I'm not sure I quite buy that. I think it's close but not exactly right.

Why my "capacitance to ground"? Why not my capacitance to, say, the mains?
Or to the high-voltage lines overhead? After all, when I touch a touch
switch, assuming I'm wearing my sneakers and I'm on the second floor of my
wood-frame house, I'm not particularly closer to ground than to any of those
other things.

4. ### John JardineGuest

Touch the end of a 1Meg scope probe with your finger and the scope' can
easily show a couple of volts of mains 'pickup'. This is equivalant to a
couple of AC uA continuously flowing through your finger tip and through the
scopes' 1Meg input and finally to ground potential.
This is due to a small electrostatic potential developed across your body
from the fields generated by nearby electrical appliances and cabling. This
'pickup' for example is particularly noticable if you are near to working
florescent light fittings
These touch switches work very badly (or not at all) if there is no other
electrical equipment in the vincinity.
regards
john

5. ### Terry PinnellGuest

[snip]
Not sure I agree with you there John. I've built quite a few based on
the capacitance-to-ground principle (i.e causing a small change in
frequency of an oscillator, and hence driving a relay), such as
http://www.terrypin.dial.pipex.com/Images/Proximity.gif

It was a decade or two ago, but I recall experimenting successfully
with a battery supply and a short section of metal-wire garden
fencing, at least 10 yards from any mains.

6. ### Walter HarleyGuest

Certainly. But: why does touching the probe make a difference? After all,
the probe itself is in the same electrical field. The capacitance from the
probe tip to the nearest mains wire is not much different than the
capacitance from me to that wire, is it?

My body is basically a big conductive blob, insulated from everything around
it (but of course subject to electrostatic fields, that is, capacitively
connected to everything around it, including mains and ground). Replace me
with a similar-sized metal sphere, for the sake of science; now, touch the
probe to the metal sphere. Why does more current flow than before touching
it? What's the difference between the probe tip and the metal sphere - is
it diameter, surface area, or volume?

I think the answer has something to do with my body behaving like an
antenna: I have a lot of free charge carriers in me, and the electrical
field around me causes those charges to slosh back and forth, which in turn
causes the potential at any given point on me (relative to a fixed
reference) to change.

I think that is different than saying that I'm capacitively coupling the
probe to the mains. Am I right?

7. ### Stepan NovotillGuest

Touch dimmers etc use the most simple crappy principle you can
imagine. The circuitry itself is usually tied to the hot of the AC
line. The sensor is simply the base or gate of a transistor (IC)
coupled to the touch plate through a 4.7Mohm resistor. When you touch
the plate you are shunting a small current to ambient (ground) and
this triggers the circuit. There are no fancy proximity switch
techniqes such as you would find in the Amway doorknob alarm or
Studfinder. This is boring basic stuff for a mass produced penny
pinching consumer item.
....Stepan

8. ### Walter HarleyGuest

Okay, I thought about it a bit more (though I still haven't done the math).

I can see that I've got a bigger surface area than a scope probe. So
presumably I've got a bigger capacitance to any nearby conductor than the
probe would. Let's say I've got capacitances to both 120v mains and to
ground lines in the area. Then I'll form a capacitive divider; the
situation is like this:

120VAC
|
C ---
---
|
o------.
| |
| .-.
C --- | | scope
--- | | 1MEG
| '-'
| |
o------'
|
===
GND

Where "C" is larger when it's my body than when it's just the scope probe.
Insert other C's to other sources as desired; in general the weighted
average will still be nonzero.

9. ### John JardineGuest

Yes, I suspect you're right. I've come across lots of simple human body
models from a health and safety perspective but nothing in the way of
electrical field pickup. For easier visualisation, my own homespun body
model is a 6' x 1' thin walled cylinder of slightly conductive plastic
filled with a low resistance salty electrolyte. The whole lot hovering about
1/8th inch above the floor .
It would also seem to imply or infer, internal circulating currents (eddy
currents?), skin effects! and preferential frequencies of resonance. Both
for reception and maybe transmision?.
regards
john

10. ### Terry PinnellGuest

Yes, which squares with my reply to John a couple of days ago.

11. ### AlanGuest

Having used only a 10pf cap and two pins on a PIC to make a very
good touch sensor (so good it could detect my hand was within 6 inches
reliably), I think you're all pretty far off on this.

Has nothing to do with your capacitance to ground, would work if you
were near no other conductor at all in space. Has entirely to do with
your body being a relatively large conducting body compared to what's
being charged, you ARE the ground in these circuits.

Think of this, a small cap one side to AC one side open:

AC----| |----- open

NO charging current. So if you had a very sensitive low load probe
and put it on the open lead, you'd see the AC voltage. No charging
current means no voltage across the cap, so of course the open lead
follows the AC voltage just like an open resistor. You touch that
open lead, and there is no way the AC side can charge and discharge
your body to 120V AC through that tiny capacitor, so your body holds
it at the middle while the AC side swings around you. There is a very
small current going into and out of you that balances, your body is
huge in comparison so it barely moves you as a reference point. You
are the EARTH.

An Osc probe tip has AC on it, from noise inside the scope it's just
balanced so you see no signal. Touch the tip and you damp that the
tip side, the signal is relative and from inside the scope.

Put a small cap on two microcontroller pins. One pin is drive, one
is sense. Make both pins output to ground, to set the charge on the
cap to zero. Make sense an input so it's high Z. Now raise the drive
pin side to 5V. If nothing is touching sense, you will read 5V on it
shortly after raising drive, a few cycles. If you touch sense, there
is no way the 5V rise can pull your whole body up 5V through the cap,
it will read zero. More impressive than that, even if you don't touch
the pin you affect it. Even if it still rises to 5V because you're
not touching it, your body being nearby will SLOW the rise. By
checking for it being high within a very short time and using a small
capacitor you can tell if there is a large conductor within quite a
few inches, it's relatively large charge is capacitively coupled to
the sense pin even when not touching. Even when your not touching or
close enough to keep it from going to 5V, your body being near changes
the charge slope of the sense pin side. Charge time slope is a much
more reliable non-contact way to sense than the end voltage.

The lamps surface is very lightly coupled to AC through the resistor
then the FET's internal capacitance. Your body acts as a relative
ground (big charge pool), the tiny current going into and out of you
as the other side goes up and down activates the circuit.

Note that you're not a good earth, with a large capacitance or
touching a bare conductor your body will go up and down the AC
assuming you're not really grounded and getting electrocuted. But
your body has plenty of charge capacity to hold one side of a tiny cap
fairly steady when the other side is moving around.

Some clowns (Q something in the name) have a patent on it. I think
anyway, there description of what they're doing is so obtuse you can't
tell if they even know what they're doing. If so I think it would be
easily bustable, I have code from years before their patent and
consider it obvious. I was looking at other touch circuits and it
struck me to get rid of the other components, and swing a small cap up
and down with the other side open just like an open resistor.
Actually it works nearly the same with a very large resistor, but you
will eventually charge up through the resistor if you're touching it.

Alan