[Amps] Plate modulation from power supply ripple?

Fri Oct 4 02:28:42 EDT 2013

Gary, ALL
As far as I understand modulation, mixing and multiplication are just about the same thing.  So, if there is an input port to an amplifier that changes the gain or in other words the output level when all other things remain the same, the outcome is a signal which is AM modulated.  An amplifier plate "port" causes gain to vary up and down, this constitutes multiplication between the input RF and the hum or AF signal.  As far as I understand, there is no need to run a low level active mixer in class C UNLESS you want to mix two signals coming in from the same port, not from two ports that that both control the "gate" or the gain of the amplifier.
Whether theory allows it or not, my thinking is that power supply ripple does cause some AM on all high level tube amplifiers.  However how much is the question, if 10% ripple gives 10% modulation that would mean side-bands at -20 (power) just 60Hz from the carrier.  And if the modulation 10% is not directly related to the ripple level, then even less noticeable side-bands would come out.  I think this gets interesting with 3 phase power when here in EU where the ripple voltage is either 150 or 300 Hz depending on the rectifiers.  As you may know, 3 phase 380/240 power to a private residences is the norm in OH at least.
So, if these side-bands are there, what is the level? Has anybody noticed them in real life?  
BR.
MarkkuFrom: garyschafer at comcast.net
To: TexasRF at aol.com; ww1c at outlook.com; amps at contesting.com
Subject: RE: [Amps] Plate modulation from power supply ripple?
Date: Thu, 3 Oct 2013 18:43:58 -0500

A diode detector would if the hum signal
was present. But if no mixing occurs in the transmitter, the low frequency
(hum) signal is eliminated in the transmitters output circuit as the audio
frequency never got mixed up to rf.

However in the case of the linear amp
plate circuit ripple voltage there is some modulation happening due to the varying
amplitude of the plate voltage causing the wanted signal to vary slightly in
amplitude but it will be no where near what it would be if the amplifier was
operating class C or other non linear mode.

Think about the carriers in the AM days on
the bands. If two carriers  come into a diode detector spaced a 1 KHz
apart you would hear a 1 KHz beat note between them. The modulation got done in
the detector. Switching over to a product detector the 1 KHz beat would disappear
or be attenuated a great amount. The product detector only detects the product
(multiplication) of the BFO and the incoming signal. It will not detect the sum
and difference (the mix between them) of the two carriers. 

The product detector operates as a switch
that is turned on and off by the BFO signal which allows signals to be mixed
only with it. The incoming carriers are two low in level to cause any switching
action of the product detector so there is no modulation of one carrier by the
other to cause a beat between the two. In a diode detector all signals are
strong enough to cause mixing between any other signal making it to the
detector so you get everything mixing with everything and giving all kinds of
mixes.

Another way to look at receiver mixers, as
a receiver receives more than one signal off the air the off the air signals do
not mix together and generate yet more unwanted signals or the sum and
difference between them. That is if the receiver mixer can handle the levels of
those signals coming in (it remains linear). The local oscillator is strong
enough in the mixer to cause it to do the switching (multiplication) action and
not the off air signals.

If the off air signals get to strong they
will start to cause the mixer to switch and become non linear and act as the local
oscillator which creates unwanted products, often called intermod.

The same thing happens in the linear SSB
amplifier. The SSB signal consists of multiple frequencies (from the audio)
going into it to be amplified. If some of those signals drive the amplifier too
hard (into compression) the amplifier starts to act as a mixer and while amplifying
the signals it also turns into a mixer and multiplies those various signals
together to create yet more signals as we commonly know as intermod signals. Intermod
is the process of mixing two or more signals together to create other signals,
unwanted in this case. Being close together frequency wise the new products
will fall close to our wanted signals also.

We most often think of modulation and
mixing as sums and differences of the signals that result in the new signal but
It is really not sum and difference but multiplication action that happens. It just
so happens that the end result of the products work out to the same result as
the sum and difference.

But for modulation to occur you can not
sum signals together, they must be multiplied.

73

Gary 
K4FMX

From: TexasRF at aol.com
[mailto:TexasRF at aol.com] 

Sent: Thursday, October 03, 2013
4:07 PM

To: garyschafer at comcast.net;
ww1c at outlook.com; amps at contesting.com

Subject: Re: [Amps] Plate
modulation from power supply ripple?

Seems that the receiver mixers and
demodulator would provide this mixing/multiplicaton action and introduce
apparent hum modulation of the signal.

This is an interesting aspect of rf
design that I have not previously considered.

73,

Gerald K5GW

In a message dated 10/3/2013 10:49:25
A.M. Pacific Daylight Time, garyschafer at comcast.net
writes:

AM modulation is a multiplication (mixing) process. Without a
switch (amplifier driven into cutoff) there is no multiplication of the
signals. 

The variation in plate voltage due to power supply ripple will be
seen in the output as addition and subtraction of the main signal and ripple
voltage but at a very low level if the amplifier is linear.  No side bands
are created as they are in an AM signal.

73

Gary  K4FMX

From:
TexasRF at aol.com [mailto:TexasRF at aol.com] 

Sent: Thursday, October 03, 2013
10:44 AM

To: ww1c at outlook.com; garyschafer at comcast.net; amps at contesting.com

Subject: Re: [Amps] Plate
modulation from power supply ripple?

Hi All, even if the plate current did not change, wouldn't the
changing plate voltage result in changing power input and corresponding changes
in power output?

That sounds like modulation as well.

Taking this to an extreme, removing the filter C altogether, any
hum modulation then? How much?

73,

Gerald K5GW

In a message dated 10/3/2013 8:35:17 A.M. Pacific Daylight Time,
ww1c at outlook.com writes:

Hi Gary, All

I was thinking about your statement and looking a typical tube anode characteristics
(curves).If I read this correctly, change in plate voltage does cause a change
in plate current.Is this not plate modulation?  I am not sure if the
operating class changes this.Perhaps I need to simulate this too...

Thanks

MarkkuWW1C

> From: garyschafer at comcast.net

> To: ww1c at outlook.com; amps at contesting.com

> Subject: RE: [Amps] Plate modulation from power supply ripple?

> Date: Wed, 2 Oct 2013 12:36:06 -0500

> 

> You don't hear the ripple because it does not modulate the linear
amplifier.

> If you were to run the amplifier in a non-linear mode such as class C then

> it would plate modulate it.

> 

> 73

> Gary 
K4FMX

> 

> > -----Original Message-----

> > From: Amps [mailto:amps-bounces at contesting.com] On Behalf Of Markku

> > Oksanen

> > Sent: Wednesday, October 02, 2013 12:40 AM

> > To: amps at contesting.com

> > Subject: [Amps] Plate modulation from power supply ripple?

> > 

> > 

> > All

> > I was simulating (with LTspice) a three phase power supply where the

> > secondary is in star configuration with full wave rectification. You
can

> > get also a second voltage from this by connecting to the center of
the

> > star, this voltage is half of the full wave rectified voltage and is

> > only half wave rectified (per phase), 3 pulse in stead of 6 pulse for

> > the higher voltage.

> > It turns out that even relatively high value (tens of uF) filter

> > capacitors would leave some 10% of ripple on the lower voltage. 
Now the

> > question is: How much plate modulation this make?  The ripple
frequency

> > is 150 Hz and to me this looks like your regular high level, plate

> > modulated AM transmitter.  Why do we not hear this on a typical
signal

> > if it is there?

> > 

> > MarkkuWW1C

> > _______________________________________________

> > Amps mailing list

> > Amps at contesting.com

> > http://lists.contesting.com/mailman/listinfo/amps

> 

_______________________________________________

Amps mailing list

Amps at contesting.com

http://lists.contesting.com/mailman/listinfo/amps