I brought this back to the reflector so the folks here can get a better
idea
of what we think.
Mike N4NT
Well frankly, Mike, it irks me a little that you made my private email
public, but fair enough (I never told you not to). Please see my
comments below:
----- Original Message -----
From: "Mike Hyder -N4NT-" <n4nt_m_o_hyder@charter.net>
Compare the following two statements:
1. This only happens when you put the attenuator in front of a non-linear
amplifier. In this case, the attenuator will appear to reduce IMD products
more than the fundamental signal is, but what is really happening is that
you are reducing the drive to the non-linear amplifier stage so that it
doesn't produce as much IMD.
2.If you drop the input signal tones levels by 10dB, the 3rd order
products
will drop 30dB for a net change in relative IMD level of 20dB...
Both those statements were in your first email chastising me. In my
opinion
it is impossible for both of your statements to be true at the same time.
First you say that the attenuator only reduces the drive so it doesn't
produce as much IMD (but just "appears to"), then you say that a change in
the input level is mathematically related to the change in IMD in the
output.
Let me try to explain again, Mike. In statement (1), I am trying
to convey the notion (perhaps not very well) that if you place
an attenuator in an RF circuit, the level of the IMD at the output
of the passive attenuator (relative to the fundamental signal) will be
essentially unchanged compared with level of the IMD products
present at the input to the passive attenuator (again relative to the
fundamental signal at that same point). The only exception to this
is in the case of passive intermodulation at very high signal levels.
In that case, a passive device such as an attenuator can actually
add IMD distortion, but that sort of thing generally only comes
into play in high power multi-carrier systems like cellular telephone
base stations.
Now, I hope you will agree that if I place a 10dB passive
attenuator in front of an active amplifier stage, that it will drop
the level of the RF input signal reaching that amplifier by 10dB.
If I am operating that amplifier at a point where it is still
reasonably linear (but, of course, not perfectly linear), the level
(of the fundamental signal) measured at the active amplifier's
output will also drop by approximately 10dB. What the 3 for
1 rule mentioned in (2) says is that the when I reduce the input
signal level by 10 dB to an active stage which is producing some
level of IMD distortion at its output port, the level of that IMD
distortion will drop by 30dB (or in relative terms 20dB). In this
sense, adding the attenuator reduces the IMD, but the attenuator
is not the direct cause of the IMD reduction, it is the change in
drive level to the active amplifier that causes the IMD to change
(the less hard you drive the amplifier, the less IMD the amplifier
produces). If, for instance, I place that same passive 10dB
attenuator at the output of the active amplifier stage instead of
the input, then the level reduction of the fundamental signal as
measured at the output of the passive attenuator will still be
10dB. In this case, however, the level of the IMD distortion
present at the output of the passive attenuator will only change
by 10dB (0dB relative change). Thus, we see that while the
insertion of the 10dB passive attenuator causes the level of
the signal in the RF chain to be reduced by 10dB regardless
of whether the attenuator is placed before or after the active
amplifier stage, the effect on IMD distortion is vastly different
depending on whether we attenuate before or after the active
amplifier stage that is producing the IMD.
Again I respectfully suggest that you make your own statements without
regard to what others say. Everyone is entitled to be wrong. It is not
your job to police everyone for accuracy.
Yes, Mike, everyone is entitled to be wrong, but when you make
a statement in a public forum which I believe to be misinformative
or incorrect, don't you think I have some obligation (or at least the
right) to respond? I wasn't trying to attack you personally (heck, I
don't even know you, how could it be personal). You even said
at the end of your post "I have just exceeded my knowledge", which
implied to me that you might not be sure of what you were saying.
I was just trying to clarify what you said so that the uniformed
wouldn't come away with the impression that passive attenuators
reduce IMD. High power multi-carrier systems where PIM (passive
intermodulation) comes into play non-withstanding, all a passive
attenuator does is to change the signal level reaching downstream
stages. If those downstream stages are in some measure non-linear
(read most amplifiers), this change in signal level (brought about by
insertion of the passive linear attenuator) will cause the intermodulation
distortion products produced by that downstream non-linear stage
to change. For the 3rd order IMD product (typically the dominant
product), the change will roughly follow the 3 to 1 rule that I
mentioned previously. Read any book on RF design, or sit down
at the bench and actually measure it yourself (as I have done
countless times) and you will see that what I am saying is
correct. When it comes to IMD where you put an attenuator can
make all the difference in the world.
Hope that clarifies what I was trying to say. Sorry if I ruffled your
feathers in the process.
73 de Mike, W4EF............
----- Original Message -----
From: "Michael Tope" <W4EF@dellroy.com>
To: <n4nt_m_o_hyder@charter.net>
Sent: Saturday, July 30, 2005 1:20 PM
Subject: Re: [TenTec] Filters, amps and IMD, etc
Mike,
I'll take this off the reflector. There is no need to make this
a public pissing contest (there is already too much of that
on this reflector). In any case, I assumed you said something
you did not, because your statements were NOT clear, and
in my opinion they could easily be misinterpreted. I have
reread your words in the "clear light of day" several times
and each time I do, I come to the same conclusion. That is:
a) This guy doesn't understand what he is talking about.
or
b) This guy does understand what he is talking about, but he
has explained things so poorly that I can't tell for sure.
Whatever the case, the only reason I chimed in was I
was worried that others might misinterpret what you said
and thus come away misinformed. Tell me that you
wouldn't do the same if you believed someone was being
provided misinformation? Of course you would, Mike.
As far as your statement about being "short-circuited by
fear" I guess you are accusing me in a roundabout way
of jumping to conclusions based on some neurotic fear
of new knowledge or some other personal insecurity.
Well, I suppose that I have my share of personal
insecurities, but I can assure you knowledge of basic
IMD theory isn't among them.
I do appreciate your advice, however, about guarding
against neurotic tendencies. Your are correct that when
left unchecked they can be an impediment to personal
growth and learning. I do try to stay vigilant in that regard.
73 de Mike, W4EF..............................................
----- Original Message -----
From: "Mike Hyder -N4NT-" <n4nt_m_o_hyder@charter.net>
To: "Discussion of Ten-Tec Equipment" <tentec@contesting.com>
Sent: Saturday, July 30, 2005 6:27 AM
Subject: Re: [TenTec] Filters, amps and IMD, etc
In the clear light of day, reread all the words below. What you've done
is
take my fairly clear and correct statements, assumed that I said
something
I
did not or assume I did not know something I did know and then take issue
with your assumptions. I respectfully suggest that in the future you
concentrate more on helping the original requestor instead of taking
issue
with someone else who is trying to help.
Almost all of what we learn we learn by analogy. In some people that
learning is short-circuited by their fear -- I don't know if it's fear of
being wrong or ignorant or what -- so that when they hear the first part
of
the analogy, their internal response is, "I know that!" and the new
knowledge is lost.
Guard against that.
73, Mike N4NT
----- Original Message -----
From: "Michael Tope" <W4EF@dellroy.com>
To: "Discussion of Ten-Tec Equipment" <tentec@contesting.com>
Sent: Saturday, July 30, 2005 2:07 AM
Subject: Re: [TenTec] Filters, amps and IMD, etc
----- Original Message -----
From: "Mike Hyder -N4NT-" <n4nt_m_o_hyder@charter.net>
Hi Mike--
As I said, I was over my head. Please help me understand something you
said
in your last paragraph, to wit: "If you drop the input signal tones
levels
by 10dB, the 3rd order products will drop 30dB for a net change in
relative
IMD level of 20dB (e.g. IMD changes from -20dBc to -40dBc as you drop
the
tone levels 10dB from the amplifiers 1dB gain compression point)."
Please compare that with what I said: "For this reason, a 10dB
attenuation
in the line will attenuate the desired signal by that much but will
attenuate the intermodulation distortion products by much more" and
explain
where my thinking differs from yours.
73, Mike N4NT
Perhaps it is just the way you worded your statement, Mike, but I
took what you said to mean that an attenuator actually reduces IMD
regardless of where it is in the RF chain. An attenuator can't reduce
IMD products that are already present at its input any more than
it can reduce the fundamental signal. It will reduce the IMD at its
input by an amount equal to the amount that it reduces the
fundamental. If the IMD is down 40dB relative to the fundamental
signal at the input of an attenuator, it will still be down 40dB relative
to the fundamental at the output of the attenuator. An attenuator can,
however, reduce the fundamental signal level hitting a downstream
non-linear stage thereby reducing the amount of IMD which that
stage produces. In that case (non-linear stage downstream from the
attenuator) the amount of IMD produced by the offending non-linear
stage will be reduced in greater proportion than reduction in the
level of the fundamental signal (~3 for 1 proportion in the case of
the 3rd order IMD products).
You also said the following:
"Amplification works in the opposite way. It brings up one
signal but brings up the IMD products more than it does the
one signal. In addition, if the amplifier is not perfectly linear, it
serves to accentuate the harmonic products and worsen the
problem (see paragraph 1 above)."
The wording of this statement implies that IMD products are
created even in perfectly linear amplifiers (e.g. your wording
implies that problem is always there, and just gets worse when
the amplifier is non-linear). In point of fact a pefectly linear
amplifier (if you could find one), doesn't generate any IMD
products. Any IMD already present at the input of a linear
amplifier will be amplified by the same amount as the fundamental
signal. Of course, any real amplifier will have some non-linearity.
If you increase the signal level to a typical amplifier that has some
inherent non-linearity, the IMD products will grow by an amount
greater than the increase in input signal level (~3 for 1 in the case
of the 3rd order product). This is where the concept of intercept
point comes from.
If you keep increasing the signal level to a real amplifier (e.g. any
real amplifier will have some inherent non-linearity), since the 3rd
order IMD products grow faster than the increase in fundamental
signal level, eventually the level of the 3rd order products present
at the amplifier output catch up or "intercept" the level of the
fundamental at the output. Actually, however, a typical amplifier
doesn't ever get there as its output will saturate just before it
reaches the 3rd order intercept point, so generally the intercept
point is a mathematical extrapolation from the slopes of
fundamental signal level vs. input power and the 3rd order
product level vs. input power curves.
In any event, perhaps this is what you meant to say, but it
just wasn't clear to me from your wording.
73 de Mike, W4EF................................
----- Original Message -----
From: "Michael Tope" <W4EF@dellroy.com>
To: "Discussion of Ten-Tec Equipment" <tentec@contesting.com>;
<k4qo@earthlink.net>
Sent: Friday, July 29, 2005 10:34 AM
Subject: Re: [TenTec] Filters, amps and IMD, etc
----- Original Message -----
From: "Mike Hyder -N4NT-" <n4nt_m_o_hyder@charter.net>
IMD is the product of two or more signals. 3rd order IMD is a signal's
fundamental mixing with the second harmonic of another signal...
For this reason, a 10dB attenuation in the line will attenuate the
desired
signal by that much but will attenuate the intermodulation distortion
products by much more.
That isn't quite right, Mike. This only happens when you put the
attenuator in front of a non-linear amplifier. In this case, the
attenuator
will appear to reduce IMD products more than the fundamental signal
is, but what is really happening is that you are reducing the drive to
the
non-linear amplifier stage so that it doesn't produce as much IMD.
If you put a 10dB attenuator after the stage which produces the IMD
distortion, it will knock the signal down and the distortion products
by the same amount, 10dB. In other words, if the IMD products are
already present, the attenuator will knock them down by the same
amount as the desired signal (the attenuator can't tell the difference
between the desired signal and the distortion products).
Amplification works in the opposite way. It brings up one signal but
brings
up the IMD products more than it does the one signal. In addition, if
the
amplifier is not perfectly linear, it serves to accentuate the harmonic
products and worsen the problem (see paragraph 1 above).
A perfectly linear amplifier by definition will amplify distortion
products
by the same amount as the fundamental signal (amplifiers can't tell the
difference between the signal we want and the signal we don't want).
Most
amplifier, are somewhat non-linear so they typically add IMD products.
The amount of IMD that they add depends on how hard you drive the
amplifier. At the amplifiers 1dB gain compression point, an amplifier
will typically add 3rd order IMD products that are approximately 20dB
down from the fundamental tones. If you drop the input signal tones
levels
by 10dB, the 3rd order products will drop 30dB for a net change in
relative IMD level of 20dB (e.g. IMD changes from -20dBc to -40dBc
as you drop the tone levels 10dB from the amplifiers 1dB gain
compression
point).
73 de Mike, W4EF.......................
I have just exceeded my knowledge.
73, Mike N4NT
----- Original Message -----
From: "Randy K4QO" <k4qo@earthlink.net>
To: <tentec@contesting.com>
Sent: Thursday, July 28, 2005 11:50 PM
Subject: [TenTec] Filters, amps and IMD, etc
If the amp that kicks in at the tighter bandwidths is the problem, why
not just adjust the gain of that amp down a bit. Wouldn't that improve
the IMD without resorting to 3rd party filters? Or is it that ANY
additional amp stages, regardless of gain causes the IMD to get worse?
Iquiring minds wanna know!
73,
Randy
K4QO
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|