# touch lamp, the real story

In http://www.straightdope.com/mailbag/mtouchlamp.html, we get the SDAS report on touch lamps.

While the report is unlikely to lead to any casualties, it was rather off-the-wall in may respects. Like highly misleading and basically incorrect in many respects.

“Nearly all touch lamps work because of a property of the human body called capacitance, which is the ability to store electrical charge.”

Misleading: Capacitance is not a property specific to the human body, in fact any object that conducts electricity has an intrinsic capacitance.
“Most touch lamps employ an oscillator, which is a type of tuned-circuit amplifier that produces an AC current.”
Quite a bit off: I don’t know were they got that definition of an oscillator, it sounds like seomthing from a bad glossary in the Boys Book Of Science. In this case it’s neither correct nor helpful.

One can make an oscillator from a tuned-circuit and an amplifier, sure. That’s how it is described in some of your more pedantic texts, which are trying to explain the two major requirements for
an oscillator-- a frequency-determing subcircuit (the tuned-circuit), and a gain subcircuit (the amplifier).

But this description isnt helpful here, for three reasons:

#1: A typical touch-lamp oscillator doesnt use a tuned-circuit,.
#2: Very few oscilators use an explicit amplifier.
#2: The touch-lamp oscillator, as 98% of oscillators do, supplies an AC voltage, not a current.

" When you touch the lamp’s metal housing, you introduce your capacitance into the circuit. Now the oscillator has to pump charge into a much larger surface area-- yours plus the lamp’s. That causes the oscillator to detune, or change frequency.
The lamp detects this change and toggles the light on or off."

Basically NO: Well, that scheme would work just fine. Only problem, nobody would design it that way as there are much more reliable and cheaper ways to do it.

Most touch lamps use a simple R/C oscillator, which is about 90%
cheaper than a tuned-circuit oscillator.

The detection method you presented, capacitance detuning and detuning detection would also work, but again it would be way too expensive to do it that way. That would require some sort of frequency discriminator, maybe another tuned-circuit (\$\$) or a phase locked loop (\$).

Why do it that way when there are simpler methods?

Here’s a link to an actual schematic of what’s inside a touch lamp controller.
http://www.arrl.org/tis/info/pdf/59378.pdf

Note: no tuned circuits, no frequency discriminator. It probably works by applying some of the oscillator voltage to the lamp body, then noting the change in current when you touch the lamp.

I’m sure the answer writer meant well, but this answer just wasnt up to the typical Cecil standard.

You can’t make a list with only two items.

## Peace.

Sometimes, you can say ‘Hi’ to Opal.

I never said or even implied that capacitance was a property only of the human body. I suppose I could have gone into dxetail about what capacitance is, but that really wasn’t the point of the report.

Do you even know what a tuned circuit is? That schematic you linked to has one connected to Pin 6 of the IC (hint: that RC network is one). The rest of the oscillator is inside the IC, and I don’t know for sure what type it is, and neither do you. You can’t make an oscillator out of just an RC network. You need at least one active component.

As far as detection, the circuit just monitors the current flow, as you said, but note that that change in current is due to the frequency shift. I said what I said for the sake of simplicity since most people reading the report have little or no background in electronics. Seems yours is a bit lacking, in fact.

>I never said or even implied that capacitance was a property only of the human body.

If I write “Life is a property of the human body”, it suggests to some that this is a property unique to the human body. Scientists need to be clear and precise in their writing, otherwise the masses may misinterpret what is written.

>Do you even know what a tuned circuit is?
>That schematic you linked to has one connected to Pin 6 of the IC (hint: that RC network is one).

No, a tuned circuit is one that is “tuned” to a particular frequency.

An RC network cannot be tuned to a frequency.

In that light switch they use what’s known as a phase-shift oscillator, which does not require a tuned circuit. To be brief, there’s two ways to “tune” an oscillator to a particular frequency.

#1: you can use some sort of circuit that has gain > 1 at a particular frequency. This is usually a “tuned circuit” comprising of a coil and a capacitor. A coil and capacitor have opposite impedance curves, one goes up with frequency, the other goes down. Where the two curves cross, that’s the tuned circuit’s “sweet spot”, or “resonance”, where the tuned circuit will have a unique and high gain. Your typical radio uses a coil-capacitor oscillator to bring in the stations. They choose this kind of oscillator as it can be built to be very stable, drifting in frequency just a few parts per million.

#2 way: You can use some combination of components that have a 180 degree phase shift. Oscillators need “positive feedback” around the amplifier stage. The amplifier usually has a 180 degree phase shift in itself, so the feedback components must add another 180 degrees of phase shift. It turns out a R/C network, just a resistor and capacitor, DO provide a variable phase shift versus frequency. So at some frequency the R/C combo will provide just the right amount of phase shift, and the oscillator will run at that frequency. But this is not in any way a “tuned circuit”. RC oscillators are not very stable frequency-wise, as the slope of the phase shift vs frequerncy curve is very shallow, allowing the oscillation frequency to drift many percent. So they’re relegated to designs that don’t need much intrinsic frequency stability, like say touch lamps.
> The rest of the oscillator is inside the IC, and I don’t know for sure what type it is, and neither do you.

See above.
>You can’t make an oscillator out of just an RC network. You need at least one active component.]

Never said you could.
>As far as detection, the circuit just monitors the current flow, as you said, but note that that change in current is due to the frequency shift.

No, it is not. There is no connection between the frequency-determining components, the R/C network, and the touch input.

Even if there were, you’ll note there’s no frequency discriminating components, so even if there was a frequency shift (which there isnt), there’s no components there to detect the frequency shift.

> I said what I said for the sake of simplicity since most people reading the report have little or no background in electronics.

That’s exactly why one should write as clearly and correctly as possible, so those nice folks don’t go off on tangents. That’s how many an urban legend has started.
>Seems yours is a bit lacking, in fact.

Well, let’s see: I’ve been designing electronics for, um, 34 yrs now.

I’ve designed from scratch several weather satellite signal demodulators, several modems, several PC sound cards, built transmitters, receivers, oscilloscopes. My father and brother are both Electrical Engineers, one a PhD.

I hate to be such a nit-picker, but it doesnt take that much more work to get the facts straight. And this still is the “STRAIGHT DOPE” isnt it? Or has it been renamed “WHATEVER HALF-BOTTOMED HALF-TRUTHS WE FIND LYING AROUND” ?

Grumpily but respectfully yrs,

grg88

Well, it doesn’t have a real frequency, but there’s nothing wrong with an imaginary frequency, is there? And i/(RC) is the frequency of an RC circuit (or j/(RC) , for you EEs).

>but there’s nothing wrong with an imaginary frequency, is there?

That was the point in my official EE training that I threw in the towel. All that pure theory stuff full of wj’s gave me the willies.

Luckily one can still go far on the practical side of EE with Ohms law, a reactance slide rule, Winfield Hill’s book, a bunch of IC application notes, and a Digi-Key catalog.

I have a plasma ball, (a glass globe with lightning) thats sound activated. when I pluged it in in the same room as mt touchlamp they interacted. when a sound caused the ball to light up the light would turn on or off. can anyone tell me why.

Oh, I forgot:
>That schematic you linked to has one connected to Pin 6 of the IC (hint: that RC network is one).

That’s the wrong R/C network. There’s only two R/C networks that are not part of the power supply, and you picked the wrong one. You can’t have an R/C network that’s part of an oscillator when one end of the “R” is floating, unconnected to anything!

That’s the touch sensor input, which needs a simple R/C filter to keep out the ambient electrical noise.

The oscillator R/C pair is the one connected to pins 13 and 14.

Regards,
George

I’d just like to say:

“AC current”?
Is that anything like “ATM machine”?
Or “HIV virus”?
Or “PIN number”?

Or, to stay up to date, “Navy NCIS”?

I never expected all this argument about my silly touch lamp question. You guys take these things seriously huh? Anyway, thanks for all the responses, my cat and I feel much better now.

I hate to hijack a hijack, but I can’t let this just sit out there.

All these examples are completely correct. English idiom demands a following (or preceding as is Navy NCIS) noun. And there is a proper way of referring to an acronym and an improper way.

OPEC countries is correct. OPEC nations is not.

So when you see examples of this, think to yourself: “this is following standard English convention,” not: “this is stupid.”

A bit grumpy myself.

So you’re saying that when a news report says something like, “The CIA led an invasion of Cuba . . .” the writer should be chastised for using incorrect English? That the report should instead read, “The CIA agency led an invasion of Cuba . . .”?

I’d like a cite for the contention that English “demands” a noun. I contend that simply “OPEC” is more correct than either “OPEC countries” or “OPEC nations”, as in, “OPEC raised crude prices today…” Similarly, when I use the ATM, I enter my PIN. Saying “ATM machine” and “PIN number” is unnecessarily redundant.

How come OPEC or CIA aren’t nouns?

To my recollection, Proper Nouns are, in fact, nouns. That means names of people, places, and organizations. Thus, “George Washington” is a noun. So is “Central Intelligence Agency”.

Just because the formal name has been shortened to an acronym doesn’t make it any less the formal name. Unless you also think that “Mr.” in “Mr Magoo” should also be followed by the proper noun: “Mr Mister Magoo”?

Mr. Mister Magoo? Didn’t they have a hit with “Bwoken Wings”?