relays

[Jasen Betts]

Good point! Weren't relays the _very_first_ gain element?

the first electrical one

before (artificial) electricity there were mechanical gain devices like capstans
(the tension on the rope is multiplied by the modulus of friction and
some (exponential?) expression involving the number of turns)

but the first device with gain was probably the goad applied to a
beast of burden


bye.

 

[Jasen Betts]

Its power gain is infinite, because once you turn the valve, the water
flows forever. It's a power integrator, not an amplifier.

does that mean a mosfet is not an amplifier either?

once the gate is charged it stays on...


bye.

 

[IanM]

Has no power gain.

John

The *first* gain element was the rudder. Plane surfice on the end of a
lever using water flow to permit one man to turn many tons of ship. You
can call that hydraulic if you like

 

[IanM]

12V 180mA relay coil switching 240V 10A

10x voltage gain, 22x current gain.
an a schmitt trigger-like transfer function.

with a little negative feedback you could have a class-d amplifier
with very poor bandwidth




absolutely, Ive run old telephone mics and speakers off 12v and got sufficient
gain for feedback-driven oscillations

wikipedia furnished this link:

http://www.comtronicsindustrial.com/vxi/tuffsetcarbon3.html




probably, magnetic amplifiers, and spark gaps, do too.


bye.

I nearly lost my coffee when I saw what actually appears to be a carbon
amplifier offered for sale over 100 years after their first commercial
application.

For completeness, there is a very nice page about the hitoric ones here:
<http://www.dself.dsl.pipex.com/MUSEUM/COMMS/mechamp/mechamp.htm>

 

[Eric]

Those don't have power gain, and cannot be configured to
self oscillate. It's like a transformer or a lever, not
like a transistor, relay or vacuum tube.


That *does* have power gain, and could be configured to self oscillate.

I think that there were single-celled organisms before the first
multicellular life that had power gain, but I need to think of
a mechanism that would cause them to self-oscillate.

Well, they reproduce. Maybe a sort of oscillation. This is getting
pretty esoteric..

 

[Eric]

Eeyore wrote:

A carbon mic connected in series with some headphones and a battery can
be made to produce a feedback squeal. Sounds like gain to me.

Cheers,

Phil Hobbs

Then a relay has gain. It can be made to oscillate by wiring the coil
in series with it's nc contact.

 

[James Arthur]

---
Typical Larkinese subterfuge.

If you prefer,


Wout
dB = 10 (log10) ------
Win


Where the electrical energy W, in joules converted into heat in a
conductor of R ohms carrying a current of I amperes is I²Rt.

Which doesn't alter the fact that a latching relay doesn't have infinite
gain, that being the point of the article.

JF

The gain isn't infinite in that Win > 0, and Wout < infinity.

True, BUT, we being engineers, it's close enough.

Cheers,
James Arthur

 

[James Arthur]

But that water will come out forever, with no further effort exerted
on the valve. The longer you wait, the more output power, without
limit. The output power is the *integral* of the input power.

John

Hey, we could do (and do do) this with MOSFETs too--load up the
gate, then leave it on.

James Arthur

 

[James Arthur]

The gate leakage is finite and measurable, so the power gain is
finite. Typical low-frequency power gain for a medium-size power fet
is probably in the ballpark of 1e16. Spec-sheet (guaranteed) gain will
be a lot less.

John

Sure--I meant it as fun thing we do, not as a contribution to
the "what is infinity" question.

Our pal Dan's got a 2-fet 2-LED flip-flop "IN/OUT" indicator
on his office door, toggled with static electricity somehow.
Cool.

Cheers,
James Arthur

 

[Jasen Betts]

Those don't have power gain,

they do while the rope is slipping.

and cannot be configured to
self oscillate.

getting the phase shift and, and avoiding latch-up seems to be hard.
but the right combination of springs, weights, and pulleys should do it

It's like a transformer or a lever, not
like a transistor, relay or vacuum tube.

it seems very similar to a common base amplifier to me
(if coulombs are inches and volts are pounds, and beta is the friction
term)

 

[Jasen Betts]

It has some gate leakage current, so no, its power gain is finite. If
the valve eventually fails for some reason, its power gain would be
finite, too... depending on the failure mode.

The power gain of the valve, or a latching relay, is an integral with
no upper bound. If you name any finite power gain, I can name an
interval over which the gain is 10x as much. So there is no upper
limit. As an engineer, "no upper limit" means "infinite" to me.

what we have for latching relays, faucets, and mosfets is a
gain-bandwidth product.

at some high switching rate they perform poorly, but at lower
frequencies they excel.

 

[JosephKK]

Nope. One gain nothing doing that.

A button depression controlling mucho power is by definition gain. The
problem is that there is only destructive results in that application.
A very different issue.

 

[JosephKK]

---
OK.

Pout
dB = 10 (log10) ------
Pin

Where Pout is the power dissipated in the load and the relay contacts,
Pin is the power needed to latch the contacts, and
dB is the power gain in decibels.

Clearly, no matter how minute the amount of power dissipated by the coil
while latching the relay, or how great the power dissipated by the load
and contacts after having been latched, unless Pin is zero the power
gain can never be infinite.

JF

A bit of a fine distinction, but technically correct. It can have
infinite energy gain though. As energy is the time integral of power
and the switched energy can grow without bound and the switching
energy is finite.

 

[JosephKK]

There you go again, typical sloppiness between energy and power. The
power gain is finite and large, the energy gain is unbounded and
increases indefinitely with time.

 

[JosephKK]

No, it was probably hydraulic.

John

Close, steam control valves for steam engines. At least for man made
stuff.

 

[JosephKK]

Attach lever to ox. Whip ox. Ta dah! Power gain...
/No actual oxen were whipped during this post

No, that is just using a more powerful prime mover (motor). The whip
and the earlier goad (a stick) however is another matter.

 

[JosephKK]

That is true. A transistor base (a lump of doped silicon)
in and of itself has no gain either. It's only when assembled
into the more complex "mechanism" we call a transistor that the
device exhibits gain. It's just that we tend to think of the
transistor taken as a whole as being a circuit element, and tend
to ignore the fact that it is really a composite "machine". The
power that ends up in the load originates in the external power
supply, not the transistor.

Take the relay example. It employs an electromagnet to actuate
a switch comprised of a spring-loaded lever with contacts. A
small amount of electric power for the magnet can control a much
larger amount of electric power via the contacts. If you change
the magnet for a handle, you can control a large amount of
electrical power with a small amount of mechanical power.

Greg; P(out)/P(in) is gain, no lever makes it, no matter how
configured.

 

[Mark Zenier]

It has some gate leakage current, so no, its power gain is finite. If
the valve eventually fails for some reason, its power gain would be
finite, too... depending on the failure mode.

The power gain of the valve, or a latching relay, is an integral with
no upper bound. If you name any finite power gain, I can name an
interval over which the gain is 10x as much. So there is no upper
limit. As an engineer, "no upper limit" means "infinite" to me.

Isn't power gain always specified over some freqency response?
For a latching relay, I guess it would from be 1/MTBF to
1/(on-time + off-time) Hz.

The power gain would be
(duty cycle) * (contact rating in amps) * (contact rating max volts)
/ (switching energy to turn on + switching energy to turn off) * Hz

Mark Zenier [email protected]
Googleproofaddress(account:mzenier provider:eskimo domain:com)

 

[Mike Monett]

The power gain of the valve, or a latching relay, is an integral with
no upper bound. If you name any finite power gain, I can name an
interval over which the gain is 10x as much. So there is no upper
limit. As an engineer, "no upper limit" means "infinite" to me.

John

There's always a boundary condition. The sun will expand in 5 billion years
and envelop the earth. Your valve will melt. There will be no water since
all the oceans will boil away. The entire process will stop working. This
limits the gain to a finite value.

There is no place in the universe where you can put the valve and guarantee
it will work forever. Therefore the gain cannot be and is never infinite.

The only thing that has infinte gain is a thought or an idea. It weighs
nothing, occupies no space, has no temperature to measure, and can easily
be distributed to others.

The act of processing a thought costs energy, that is true. But before it
pops into your brain, no energy is expended.

So the gain becomes infinite as soon as you are aware the thought exists.

Regards,

Mike Monett

 

[JosephKK]

But that water will come out forever, with no further effort exerted
on the valve. The longer you wait, the more output power, without
limit. The output power is the *integral* of the input power.

John

Again the confusion between power/flow and energy/volume.