Noiseless Damping?
Vladimir Vassilevsky
There is a loop receive antenna. It is the LC circuit with the Q at
the order of 100. It has a high resonant peak. The frequency and the
amplitude of the peak are sensitive to the component values,
temperature, particular parts, etc. I would like to damp the Q down to
something like 10, so the parameters would be much more stable and
predictable. This can be done by adding a series or a parallel resistor
to LC.
However the SNR of the antenna drops approx. proportionally to sqrt(Q)
due to the thermal noise of the damping resistor. I wonder if it could
be possible to implement the noiseless damping. Cooling down the
resistor is not an option.
One idea is to switch the damping resistor on-off with a high frequency
and a low duty ratio. Or to switch or vary the value of the capacitor so
the resonance peak will be dithered. It has many side effects, though.
Any ideas, suggestions?
Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com
Paul Hovnanian P.E.
--
Paul Hovnanian pa...@hovnanian.com
-----------------------------------------------------------------------
Procrastinators: The leaders for tomorrow.
John Larkin
What's the frequency?
1. If this is below, say 30 MHz, ambient noise could well be high
enough that the resistor doesn't do much additional harm, especially
if you stop at Q=10.
2. You can synthesize a "cold" resistor using a super-low-noise
semiconductor, if you get the impedances right. Numbers like 40K might
be possible. But a perfect resistor still eats signal energy.
3. Or you could turn the loop into a bandpass filter; add a small
additional LC resonator to make a double-hump resonance. Aren't you a
filter guy?
4... Oops, latte is gone, back to work.
John
Joop
<antispa...@hotmail.com> wrote:
If possible you can lower the L and increase the C both by a factor of
5 or so.
Cheers,
Joop
mpm
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> 3. Or you could turn the loop into a bandpass filter; add a small
> additional LC resonator to make a double-hump resonance. Aren't you a
> filter guy?
>
I have it on good authority and experience that "Antennas make poor
Filters..."
-mpm
Vladimir Vassilevsky
John Larkin wrote:
> On Wed, 12 Mar 2008 12:25:35 -0500, Vladimir Vassilevsky
> <antispa...@hotmail.com> wrote:
>
>
>> There is a loop receive antenna. It is the LC circuit with the Q at
>>the order of 100. It has a high resonant peak. The frequency and the
>>amplitude of the peak are sensitive to the component values,
>>temperature, particular parts, etc. I would like to damp the Q down to
>>something like 10, so the parameters would be much more stable and
>>predictable. This can be done by adding a series or a parallel resistor
>>to LC.
>>
>>However the SNR of the antenna drops approx. proportionally to sqrt(Q)
>>due to the thermal noise of the damping resistor. I wonder if it could
>>be possible to implement the noiseless damping. Cooling down the
>>resistor is not an option.
>>
>
> What's the frequency?
>
> 1. If this is below, say 30 MHz, ambient noise could well be high
> enough that the resistor doesn't do much additional harm, especially
> if you stop at Q=10.
I thought that, too. Unfortunately, the antenna efficiency is too low so
the resistor noise is dominating at Q = 10.
> 2. You can synthesize a "cold" resistor using a super-low-noise
> semiconductor, if you get the impedances right. Numbers like 40K might
> be possible. But a perfect resistor still eats signal energy.
Do you mean something like a small value resistor bootstrapped by the
active circuit? Thank you for the good idea! That's what I am going to try.
> 3. Or you could turn the loop into a bandpass filter; add a small
> additional LC resonator to make a double-hump resonance. Aren't you a
> filter guy?
The values appear to be inconvenient for the passive inductors. GIC
cound be the option, however got to watch for the noise.
>
> 4... Oops, latte is gone, back to work.
Thank you for your advice, John.
Mark
wrote:
>
> - Show quoted text -
Couple the loop more tightly to your receiver input?
Mark
Wimpie
wrote:
Hello Vladimir,
Your information is very limited but is it an option to generate the
resistance by active parallel Feedback? As the feedback resistor is
large enough, the impact on noise performance is minimal, but the
effect is larger because of the gain of the input amplifier.
Maybe you can do something with the size of the loop. When you can
extend the loop in a third dimension, inductance reduces; hence
current noise of amplifier will generate less noise. Try to use as
much as volume you have available.
Best regards,
Wim
PA3DJS
www.tetech.nl (Dutch).
Fred Bloggs
see US2787704 .......
Vladimir Vassilevsky
Wimpie wrote:
> On 12 mar, 18:25, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
> wrote:
>
>> There is a loop receive antenna. It is the LC circuit with the Q at
>>the order of 100. It has a high resonant peak. The frequency and the
>>amplitude of the peak are sensitive to the component values,
>>temperature, particular parts, etc. I would like to damp the Q down to
>>something like 10, so the parameters would be much more stable and
>>predictable. This can be done by adding a series or a parallel resistor
>>to LC.
>>
>>However the SNR of the antenna drops approx. proportionally to sqrt(Q)
>>due to the thermal noise of the damping resistor. I wonder if it could
>>be possible to implement the noiseless damping.
> Your information is very limited but is it an option to generate the
> resistance by active parallel Feedback? As the feedback resistor is
> large enough, the impact on noise performance is minimal, but the
> effect is larger because of the gain of the input amplifier.
Actually I have to go the opposite direction: generate a large
resistance from the small resistance using the positive feedback
(bootstraping or something like that).
For this application, the noise of the input amplifier is not an issue;
the dominant component is the thermal noise of R. The absolute values
of the signal and the noise are not critical. The goal is to optimize
the S/N of antenna while keeping the Q at 10.
> Maybe you can do something with the size of the loop.
Understood. The energy caught in the loop is approximately proportional
to the volume of the loop. However we have what we have.
> When you can
> extend the loop in a third dimension, inductance reduces; hence
> current noise of amplifier will generate less noise. Try to use as
> much as volume you have available.
>
> Best regards,
>
> Wim
> PA3DJS
> www.tetech.nl (Dutch).
Thank you, Wim
VLV
John Larkin
If you had a high-impedance noiseless amplifier with a high inverting
gain G, and use a negative feedback resistor R, the composite input
impedance looks like R/(G+1), but has the current noise of the
high-value R, so the apparent resistor noise temperature is below room
temp. I've seen nuclear-sensor amps that use feedback transformers, to
make a preamp that looks like a 50 ohm load to the detector but has
much less Johnson noise; I may have a paper around here somewhere.
You may as well combine the fake "cold" resistor with the preamp,
namely design a preamp that kills your Q to the desired extent and
looks like a cold resistor and has a basically low noise figure.
Some of the phemts have noise figures of 0.4 dB, equivalent to about
28 Kelvins, when properly matched. They tend to get noisier at low
frequencies, for certain values of "low."
>
>> 3. Or you could turn the loop into a bandpass filter; add a small
>> additional LC resonator to make a double-hump resonance. Aren't you a
>> filter guy?
>
>The values appear to be inconvenient for the passive inductors. GIC
>cound be the option, however got to watch for the noise.
>
>>
>> 4... Oops, latte is gone, back to work.
>
>Thank you for your advice, John.
>
>
Hey, how about this?
>http://www.google.com/patents?id=Y1sfAAAAEBAJ&printsec=abstract&zoom=4&dq=5051700
John
mi...@sushi.com
wrote:
Would a parallel loop that intercepted the flux of the main loop lower
the Q if this parallel loop were terminated?
John Larkin
Sure, but the termination will be a noise source.
John
Tim Wescott
going to do yourself any good at all.
Even if you had a dewar full of liquid nitrogen handy to keep your
loading resistor in, all the resistor is going to do is burn up signal
that would have otherwise reached the detector. At no point in the
spectrum will the amount of energy reaching your detector be greater
with loading than without -- on the contrary, the resistive loading will
just lower the response of the antenna where it had previously been more
sensitive.
The only place that I could see such resistive loading being an
advantage is if you are attempting to receive a signal so broad that it
is filtered by the antenna -- then you may gain something.
The only two choices that I can suggest are to use an RF amplifier that
itself loads the antenna down, or take the suggestion of one of your
other respondents and change the antenna to more efficiently capture
energy over a broader band.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html
Wimpie
wrote:
> Wimpie wrote:
> > On 12 mar, 18:25, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
> > wrote:
>
> >> There is a loop receive antenna. It is the LC circuit with the Q at
> >>the order of 100. It has a high resonant peak. The frequency and the
> >>amplitude of the peak are sensitive to the component values,
> >>temperature, particular parts, etc. I would like to damp the Q down to
> >>something like 10, so the parameters would be much more stable and
> >>predictable. This can be done by adding a series or a parallel resistor
> >>to LC.
>
> >>However the SNR of the antenna drops approx. proportionally to sqrt(Q)
> >>due to the thermal noise of the damping resistor. I wonder if it could
> >>be possible to implement the noiseless damping.
> > Your information is very limited but is it an option to generate the
> > resistance by active parallel Feedback? As the feedback resistor is
> > large enough, the impact on noise performance is minimal, but the
> > effect is larger because of the gain of the input amplifier.
>
> Actually I have to go the opposite direction: generate a large
> resistance from the small resistance using the positive feedback
> (bootstraping or something like that).
Based on your information, that is not obliged. Generating a certain
low noise input impedance/resistance (to damp or load a circuit) can
be done with or a low value resistor with series feedback or a large
resistor with parallel feedback.
I did the same in a 8 MHz circuit to generate a prediscribed load for
a 2 resonator LC filter (to avoid resonance and keep the curve).
Vladimir Vassilevsky
Tim Wescott wrote:
> Vladimir Vassilevsky wrote:
>
>>
>> There is a loop receive antenna. It is the LC circuit with the Q at
>> the order of 100. It has a high resonant peak. The frequency and the
>> amplitude of the peak are sensitive to the component values,
>> temperature, particular parts, etc. I would like to damp the Q down to
>> something like 10, so the parameters would be much more stable and
>> predictable. This can be done by adding a series or a parallel
>> resistor to LC.
>>
>> However the SNR of the antenna drops approx. proportionally to sqrt(Q)
>> due to the thermal noise of the damping resistor. I wonder if it could
>> be possible to implement the noiseless damping. Cooling down the
>> resistor is not an option.
>>
>> One idea is to switch the damping resistor on-off with a high
>> frequency and a low duty ratio. Or to switch or vary the value of the
>> capacitor so the resonance peak will be dithered. It has many side
>> effects, though.
>>
>> Any ideas, suggestions?
>>
> I suggest you do some basic analysis, because I'm not sure that you're
> going to do yourself any good at all.
You are mistaking me for some other idiot, pehaps. Of course, I have
accounted for the tradeoffs.
>
> Even if you had a dewar full of liquid nitrogen handy to keep your
> loading resistor in, all the resistor is going to do is burn up signal
> that would have otherwise reached the detector.
So what? The noise of the electronics is well under the other sources of
noise. I trade off a tiny part of signal to get the robust operation.
> At no point in the
> spectrum will the amount of energy reaching your detector be greater
> with loading than without
And this is actually bad. The sharp resonance peak can slip off the
frequency of interest and pick up some interference. Hence the dynamic
range has to be increased by Q times, which is problematic.
BTW, I have tried the automatic tuning with the GIC, too. The simple
straightforward solutions are too noisy, the good solutions take too
many parts. Varactors are inapplicable.
-- on the contrary, the resistive loading will
> just lower the response of the antenna where it had previously been more
> sensitive.
The absolute sensitivity is not a problem. The S/N and the dynamic range
is what matters.
> The only place that I could see such resistive loading being an
> advantage is if you are attempting to receive a signal so broad that it
> is filtered by the antenna -- then you may gain something.
In the addition to the above mentioned reasons, the gain and the phase
shift of the loaded antenna are very stable and predictable.
> The only two choices that I can suggest are to use an RF amplifier that
> itself loads the antenna down,
That was considered, too. The total noise balance is going to be worse.
> or take the suggestion of one of your
> other respondents and change the antenna to more efficiently capture
> energy over a broader band.
This misses the point. There is no problem with capturing enough of energy.
mi...@sushi.com
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
Drat! Foiled again.
John Larkin
Thermodynamics proves that the universe is cruel.
John
Tim Wescott
> Tim Wescott wrote:
>
>> Vladimir Vassilevsky wrote:
>>
>>
>>> There is a loop receive antenna. It is the LC circuit with the Q at
>>> the order of 100. It has a high resonant peak. The frequency and the
>>> amplitude of the peak are sensitive to the component values,
>>> temperature, particular parts, etc. I would like to damp the Q down to
>>> something like 10, so the parameters would be much more stable and
>>> predictable. This can be done by adding a series or a parallel
>>> resistor to LC.
>>>
>>> However the SNR of the antenna drops approx. proportionally to sqrt(Q)
>>> due to the thermal noise of the damping resistor. I wonder if it could
>>> be possible to implement the noiseless damping. Cooling down the
>>> resistor is not an option.
>>>
>>> One idea is to switch the damping resistor on-off with a high
>>> frequency and a low duty ratio. Or to switch or vary the value of the
>>> capacitor so the resonance peak will be dithered. It has many side
>>> effects, though.
>>>
>>> Any ideas, suggestions?
>>>
>> I suggest you do some basic analysis, because I'm not sure that you're
>> going to do yourself any good at all.
>
> You are mistaking me for some other idiot, pehaps. Of course, I have
> accounted for the tradeoffs.
>
No, Mr. arrogant-but-human, I'm mistaking you for someone who is asking
for assistance. Just because you're Russian doesn't mean you have to act
like the stereotype.
But it's good that you're thinking.
I'm not sure how this is consistent with your other statements about the
total noise. Is this one of the things that you've analyzed so
thoroughly that you can diss me for suggesting that you analyze it, or is
it one of the things that you know you don't have to do your homework on
because Everything is Bigger in Russia?
>
>> or take the suggestion of one of your other respondents and change the
>> antenna to more efficiently capture energy over a broader band.
>
> This misses the point. There is no problem with capturing enough of
> energy.
Well, why didn't you say so in your original post? You're perfect, so it
can't be that you forgot to mention it.
Knowing your frequency of interest would help, too, and whether you're
interested in sky waves, ground waves, or whatever happens to impinge
upon your antenna.
At 300kHz, a one-meter capacitive probe connected to the gate of a JFET
works nicely -- it receives enough atmospheric noise that any energy loss
to inefficiencies is negligible, and it's quite small compared to a
wavelength.
--
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com
Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html
Vladimir Vassilevsky
What a shame.
I used to be of better oppinion about you, Mr. Wescott. Never mind.
VLV
Vladimir Vassilevsky
John Larkin wrote:
> On Wed, 12 Mar 2008 20:59:05 GMT, Vladimir Vassilevsky
> <antispa...@hotmail.com> wrote:
>>
>>John Larkin wrote:
>>
>>>On Wed, 12 Mar 2008 12:25:35 -0500, Vladimir Vassilevsky
>>><antispa...@hotmail.com> wrote:
>>>
>>>>There is a loop receive antenna. It is the LC circuit with the Q at
>>>>the order of 100. It has a high resonant peak. The frequency and the
>>>>amplitude of the peak are sensitive to the component values,
>>>>temperature, particular parts, etc. I would like to damp the Q down to
>>>>something like 10, so the parameters would be much more stable and
>>>>predictable. This can be done by adding a series or a parallel resistor
>>>>to LC.
>>>>
>>>>However the SNR of the antenna drops approx. proportionally to sqrt(Q)
>>>>due to the thermal noise of the damping resistor. I wonder if it could
>>>>be possible to implement the noiseless damping.
>>>2. You can synthesize a "cold" resistor using a super-low-noise
>>>semiconductor, if you get the impedances right. Numbers like 40K might
>>>be possible. But a perfect resistor still eats signal energy.
>>
>>Do you mean something like a small value resistor bootstrapped by the
>>active circuit? Thank you for the good idea! That's what I am going to try.
>
>
> If you had a high-impedance noiseless amplifier with a high inverting
> gain G, and use a negative feedback resistor R, the composite input
> impedance looks like R/(G+1), but has the current noise of the
> high-value R, so the apparent resistor noise temperature is below room
> temp.
Precisely. However I am more interested in the opposite trick: if a
small resistor is bootstrapped by a low noise amplifier with the gain
close to unity, the input impeadance is multiplied while retaining the
voltage noise of the small resistor.
> I've seen nuclear-sensor amps that use feedback transformers, to
> make a preamp that looks like a 50 ohm load to the detector but has
> much less Johnson noise; I may have a paper around here somewhere.
Yes. The simple math shows that it is possible to reduce the noise
temperature by about of the order of magnitude.
> You may as well combine the fake "cold" resistor with the preamp,
> namely design a preamp that kills your Q to the desired extent and
> looks like a cold resistor and has a basically low noise figure.
It works, although there are some pitfalls because of the phase shifts
and the stability concerns.
> Some of the phemts have noise figures of 0.4 dB, equivalent to about
> 28 Kelvins, when properly matched. They tend to get noisier at low
> frequencies, for certain values of "low."
>
>
> Hey, how about this?
>
>>http://www.google.com/patents?id=Y1sfAAAAEBAJ&printsec=abstract&zoom=4&dq=5051700
They basically do the same thing plus the C1/C2 transformer feedback
coupling. This allows for the additional flexibility, not sure about the
stability margin though.
JosephKK
<antispa...@hotmail.com> wrote:
My quickest thoughts were using a PIN diode to limit the "Q" or a
quiet JFET to reduce "Q" as needed. Both could be dynamically
adjustable.
mi...@sushi.com
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
Wait a minute. Does this mean you can't cut taxes, build up the
military and balance the budget at the same time?
Vladimir Vassilevsky
> Wait a minute. Does this mean you can't cut taxes, build up the
> military and balance the budget at the same time?
Of couse you can. Just make debts and print money. What an interesting
time! I lived through the disintegration of the USSR, now I experience
the demolishion of the USA, I wonder if I will see the demise of China!
VLV
John Larkin
The US has one of these stock market bubble-meltdowns about every 20
years on average, although modern technology may allow them to get
closer together, as in the dot-com bust and the latest mortgage thing;
both, as it turns out, being Clinton legacies. The relaxation time
constant seems to be about equal to the greed/fear decay time of
amateur investors. Some money changes hands, some people get hurt,
lots of press gets generated, Congress makes a lot of speeches and a
few new laws, and life goes on. The US still has pretty good basics:
lots of land and food, a young and vigorous population, a good
environment for science and technology and business.
John
mi...@sushi.com
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Sun, 16 Mar 2008 13:49:48 GMT, Vladimir Vassilevsky
>
>
>
Do you think there will be a meltdown when the powers that be realize
nobody is looking at their stupid adverts? Yahoo spends $850 million
for something like Facebook. Hell, I wouldn't spend a nickel on
MySpace, Facebook, or number 3 in the same genre. Everytime I see a
Myspace page I think what a bunch of f-ing arseholes.
I'm looking at the Google "sponsored" ads at the side of the page as I
type. They are all from Taiwan. Only question marks appear since I
guess I don't have Asian fonts loaded. Oh, but these fools will get a
bill for their advert being "displayed"
Michael A. Terrell
mi...@sushi.com wrote:
>
> Do you think there will be a meltdown when the powers that be realize
> nobody is looking at their stupid adverts? Yahoo spends $850 million
> for something like Facebook. Hell, I wouldn't spend a nickel on
> MySpace, Facebook, or number 3 in the same genre. Everytime I see a
> Myspace page I think what a bunch of f-ing arseholes.
>
> I'm looking at the Google "sponsored" ads at the side of the page as I
> type. They are all from Taiwan. Only question marks appear since I
> guess I don't have Asian fonts loaded. Oh, but these fools will get a
> bill for their advert being "displayed"
Yahoo bought Geocities a few years ago, then limited the websites to
20 MB of traffic per 24 hours, which was stupid because the more people
seeing a website with their ads means more revenue.
Yahoo bought Broadcast.com, and shut it down. It was home to a lot
of radio stations online, and had very good audio quality. It went down
without notice to the people who used it to hear their favorite
stations.