Electrical Engineering - Impedance Matching

Author

Impedance Matching

karlfahy@gmail.com

2006-03-20, 10:21 am

I'm looking for help/advice designing an impedance matching circuit. I
am trying to measure noise (PSD) across a resistor, I have a 500 Ohm
resistor that i need to match to the input of a preamplifier with an
input impedance of 100 MOhm.

This would be quite easy if I only wanted it to be matched for one
frequency but I need it to be matched for a wide range of frequencies,
specifically 0-100kHz.

I have found some information which states that broadband matching
requires 4 reactance networks, unfortunately it doesn't give any
information on how to calculate their magnitudes.

Any advice on this matter would be most appreciated.

Thanks,

Karl

Salmon Egg

2006-03-20, 4:21 pm

On 3/20/06 5:16 AM, in article
1142860610.589106.141870@t31g2000cwb.googlegroups.com, "karlfahy@gmail.com"
<karlfahy@gmail.com> wrote:

> I'm looking for help/advice designing an impedance matching circuit. I
> am trying to measure noise (PSD) across a resistor, I have a 500 Ohm
> resistor that i need to match to the input of a preamplifier with an
> input impedance of 100 MOhm.
>
> This would be quite easy if I only wanted it to be matched for one
> frequency but I need it to be matched for a wide range of frequencies,
> specifically 0-100kHz.
>
> I have found some information which states that broadband matching
> requires 4 reactance networks, unfortunately it doesn't give any
> information on how to calculate their magnitudes.
>
> Any advice on this matter would be most appreciated.
>
> Thanks,
>
> Karl
>
First, I have no idea what PSD stands for.

You do not explain why you uses such a high impedance preamplifier. There
are many made to work with low impedance sources such as magnetic phonograph
pickups or dynamic microphones. To truly match the two devices you need a
circuit between the components that make your 500 ohm resistor look like 100
Megohms. When you do that, less than a picoFarad stray capacitance will
screw up your broadband match.

My conclusion is 1) You do not know what you are doing; 2) You do not really
want to match; 3) You do not know what you want.

Bill
-- Ferme le Bush

karlfahy@gmail.com

2006-03-21, 9:21 am

PSD = power spectral density

I use a high impedance preamplifier because that is what is available
to me, it is a Stanford low noise preamplifier, model SR560.

Your conclusions are fairly accurate, I don't really know what I am
doing. I want to measure noise across a resistor and have been using a
preamplifier + spectrum analyser, I should be getting white noise (a
flat PSD) but I am not getting anything close to this and have been
told that this is due to a lack of impedance matching in the circuit
(maybe this is incorrect?). Now I have looked into impedance matching
and found that an inductor capacitor combination will do the trick but
will only match at one frequency, for broadband matching i need
2capcitors and 2 inductors but I do not know how to calculate their
values for my specific case.

As I said earlier any advice would be appreciated.

Karl

TimPerry

2006-03-21, 9:21 pm

<karlfahy@gmail.com> wrote in message
news:1142944858.199965.316130@e56g2000cwe.googlegroups.com...
> PSD = power spectral density
>
> I use a high impedance preamplifier because that is what is available
> to me, it is a Stanford low noise preamplifier, model SR560.
>
> Your conclusions are fairly accurate, I don't really know what I am
> doing. I want to measure noise across a resistor and have been using a
> preamplifier + spectrum analyser, I should be getting white noise (a
> flat PSD) but I am not getting anything close to this and have been
> told that this is due to a lack of impedance matching in the circuit
> (maybe this is incorrect?). Now I have looked into impedance matching
> and found that an inductor capacitor combination will do the trick but
> will only match at one frequency, for broadband matching i need
> 2capcitors and 2 inductors but I do not know how to calculate their
> values for my specific case.
>
> As I said earlier any advice would be appreciated.
>
> Karl
>

you might want to read this article
http://www.reed-electronics.com/tmw...e/CA187331.html

then this one
http://www.gamry.com/App_Notes/SmallEchemSignals.htm

here is almost a book one the subject
http://www.rose-hulman.edu/~black/Noise/ppframe.htm

there is quite a bit more out there.

the question i have is why do you want to do it?

karlfahy@gmail.com

2006-03-23, 8:21 am

Thank you for the articles, I will give them a look.

I want to do this because I am part of a research group that is working
on thin film devices and we need to carry out noise measurements on
them. Measuring white noise across a resistor is only the first step,
my ultimate goal is to measure 1/f noise in a tunnel junction but I
need to be able to do the basics first.

Salmon Egg

2006-03-23, 4:21 pm

On 3/23/06 3:50 AM, in article
1143114652.773656.324400@v46g2000cwv.googlegroups.com, "karlfahy@gmail.com"
<karlfahy@gmail.com> wrote:

> Thank you for the articles, I will give them a look.
>
> I want to do this because I am part of a research group that is working
> on thin film devices and we need to carry out noise measurements on
> them. Measuring white noise across a resistor is only the first step,
> my ultimate goal is to measure 1/f noise in a tunnel junction but I
> need to be able to do the basics first.
>
It looks like you are trying to look at resistor noise voltage. Instead, try
looking at the noise current into a short circuit. Emulate the short circuit
with a transimpedance amplifier with much feedback. That feedback allows the
amplifier to cool its feedback resistor (active refrigeration) so that you
can measure the Norton equivalent noise current of the resistor.

To me, it seems that you are in way over your head. You really should invest
the time NECESSARY to understand noise, of components like resistors and
amplifiers with feedback. Otherwise, you will never understand what you are
doing.

Bill
-- Ferme le Bush

Joshua K. Drumeller

2006-03-23, 5:21 pm

karlfahy@gmail.com wrote:
> Thank you for the articles, I will give them a look.
>
> I want to do this because I am part of a research group that is working
> on thin film devices and we need to carry out noise measurements on
> them. Measuring white noise across a resistor is only the first step,
> my ultimate goal is to measure 1/f noise in a tunnel junction but I
> need to be able to do the basics first.
>

A lot of the noise that a receiver or other instrument detects is caused
internally. Thus would be different with each detector. Or you might be
looking for noise at the output from some device and what you actually
are looking at is noise generated in your test equipment. Keep that in
mind before you spend a lot of time chasing your tail.

Josh

LinkBot