By the way....

I worked out your method of Noise Figure measurements of two receivers
at a mathematical way and it seemed right.
I am glad the same statement where made by the professionals under us.
They are a bit "rusty" but we need them !


"bviel" schreef in bericht
...

"Steve Kavanagh" schreef in bericht
om...
"bviel" wrote in message
...

Thanks for the suggestions. I'll have a look at the web references.
Just a couple of notes on things that won't work.

Noise performance of Mar's MMIC's should be noted on datasheets.

Yes, but they never show the range of noise figures, only a typical or
worst case numbers. Of course if the noise figure is very low (PHEMT)
then the variability is pretty small as long as the circuit is done
right.
The noise figure of a MMIC is flat from DC to Ghz, the one I worked with.
Did not test the Mar's.

Terminate receiver with 50 Ohm resistor, measure output and label it
zero dB.

This doesn't work, because the noise is dominated by the receiver
noise, not the resistor thermal noise (unless the receiver is very
very good !).
The method is from JT44 EME software where you can measure noise figures.
The Help with the program should explain everything.
I myself am in the position to measure noise with my scope for a "second"
opinion.
The noise of a high frequency transistor b-e diode should deliver the
right
sort of noise
also called white noise.
Or high frequency diode microwave type.
The transistor has because of his junction noise behavior the most real
white noise,
or general white noise.




73,
Steve VE3SMA

By the way....

I worked out your method of Noise Figure measurements of two receivers
at a mathematical way and it seemed right.
I am glad the same statement was made by the professionals under us.
They are a bit "rusty" but we need them !
http://www.mth.msu.edu/~maccluer/Lna/noisetemp.html
Amplifier Noise Measurements layout linked, by a German company
that sells LNA's.


"Steve Kavanagh" schreef in bericht
om...
"bviel" wrote in message
...

Some assorted comments and follow-on questions on your suggestions:

- (from bviel) The Elecraft noise generator is not a calibrated unit.
They give a typical ENR but specifically state that it varies from
unit to unit. Their low-level signal generator is calibrated (though
perhaps not quite accurately enough) but this would bring in the issue
of measurement errors between noise and sine-wave powers. And it only
works at one frequency.

- (from bviel) Which MMIC did you find had flat noise figure to 1 GHz
? My experience is that MMIC NF specs are usually at a frequency
above the flat part of the spectrum, as this is more indicative of the
high frequency performance. Are there any with noise figure also
specified, or at least well characterized, at low frequencies (within
the spectral region where the NF is flat) ? But the other question
with MMICs is the unit-to-unit variation in noise figure which I don't
believe is ever specified. If all units of a given part have, say,
between 2 and 3 dB noise figure then a MMIC amp could be a fairly good
standard for amateur use, but if the variation is from 1 to 4 dB then
the usefulness is limited, in my opinion.

- (from bviel) I don't have the JT-44 software but I did look at the
(full) manual. On page 35 it describes the "Measure Sub-Mode" which
allows noise levels to be compared. However it does not appear to
support absolute noise figure measurements (unless you use a
calibrated noise source). Is there more in the online help ?

- (from Mike Andrews) Terry Ritter's stuff seems mostly concerned with
the degree of randomness rather than absolute output power, which
makes sense as he is concerned with cryptography. But I did find,
elsewhere on his web page,

http://www.ciphersbyritter.com/REALRAND/92102201.HTM

a low frequency noise generator based on a diode source followed by a
limiting amp. I imagine this concept could be extended to HF by
suitable choice of amplifier. I am not sure of the significance of
the very different output waveform (pulses of random width and fixed
amplitude) compared to the true random (Gaussian) noise from the diode
itself. I am not sure if I am willing to tackle the math needed to
understand this or to calculate the ENR - can anyone help ?

73,
Steve VE3SMA

Steve,

There is another method of measuring noise figure that does not need a noise
source.

The reason that it is not used as often is that it is harder to measure low
noise figures accurately. Since there are still a few layers of rust on my
brain I will not geive the details here, since i do not want to say anything
that is too far off.

This other method involves measuring the gain of the device under test and
then measuring the noise power output with the input terminated properly.
The noise output is a combination of the DUT noise and the 'ambient' noise
from the input termination, times the gain of the receiver. Working
backwards you can then determine the DUT noise. As you can probably guess,
the reason it doesn't work so good on small noise figures is that any errors
in the measurement can easily outnumber the actual noise figure you are
trying to measure! And I have actually had to deal with this problem. That
is where half my gray hairs came from (the other half from being laid off).

If you are measuring an entire receiver there are a few things you have to
be careful with. The receiver must be a linear receiver (no FM, AM diode
detector, etc.---basically just SSB). There should be a filter to pick just
one sideband. Turn the AGC off. Make sure you measure the gain in the
linear region, which also applies to a simple amplifier.

If you have a DUT with a known noise figure, I think that this would be one
way of calibrating a homebrew noise source.

Jim
N8EE

Can you measure the noise power of the noise source by comparing
the output (through the RX) with the power from a signal
generator (again through the RX) of known output power?
Presumably using some form of averaging of multiple readings
across the RX passband.

- Stewart G3YSX

Steve Kavanagh wrote:

Some assorted comments and follow-on questions on your suggestions:

- (from bviel) The Elecraft noise generator is not a calibrated unit.
They give a typical ENR but specifically state that it varies from
unit to unit. Their low-level signal generator is calibrated (though
perhaps not quite accurately enough) but this would bring in the issue
of measurement errors between noise and sine-wave powers. And it only
works at one frequency.

- (from bviel) Which MMIC did you find had flat noise figure to 1 GHz
? My experience is that MMIC NF specs are usually at a frequency
above the flat part of the spectrum, as this is more indicative of the
high frequency performance. Are there any with noise figure also
specified, or at least well characterized, at low frequencies (within
the spectral region where the NF is flat) ? But the other question
with MMICs is the unit-to-unit variation in noise figure which I don't
believe is ever specified. If all units of a given part have, say,
between 2 and 3 dB noise figure then a MMIC amp could be a fairly good
standard for amateur use, but if the variation is from 1 to 4 dB then
the usefulness is limited, in my opinion.

- (from bviel) I don't have the JT-44 software but I did look at the
(full) manual. On page 35 it describes the "Measure Sub-Mode" which
allows noise levels to be compared. However it does not appear to
support absolute noise figure measurements (unless you use a
calibrated noise source). Is there more in the online help ?

- (from Mike Andrews) Terry Ritter's stuff seems mostly concerned with
the degree of randomness rather than absolute output power, which
makes sense as he is concerned with cryptography. But I did find,
elsewhere on his web page,

http://www.ciphersbyritter.com/REALRAND/92102201.HTM

a low frequency noise generator based on a diode source followed by a
limiting amp. I imagine this concept could be extended to HF by
suitable choice of amplifier. I am not sure of the significance of
the very different output waveform (pulses of random width and fixed
amplitude) compared to the true random (Gaussian) noise from the diode
itself. I am not sure if I am willing to tackle the math needed to
understand this or to calculate the ENR - can anyone help ?

73,
Steve VE3SMA

"Jim" wrote in message ...

This other method involves measuring the gain of the device under test and
then measuring the noise power output with the input terminated
properly

....any errors
in the measurement can easily outnumber the actual noise figure

Thanks for reminding me of that one, Jim. But I see your point about
errors. For example, in measuring the gain one needs a standard. One
of the few pieces of real test gear I have is a bolometer-type RF
power meter which can measure about -13 dBm accurately. If the
measurement bandwidth is 1 MHz (suitable for VHF, perhaps) then
thermal noise is -114 dBm. So I need about 100 dB gain for a very low
noise figure DUT. To measure that I might need five 20 dB attenuators
as a standard, each with perhaps +/-0.5 dB accuracy if I am lucky...so
there's +/-2.5 dB error (well, I suppose I could cross my fingers and
RSS the numbers). Or I have a diode-type power meter that will
measure lower power, which leads to issues of how the detector
responds to noise. And then there's the problem of knowing the noise
bandwidth precisely...

is where half my gray hairs came from (the other half from being laid off).

I think about half of mine come from the latter factor too !

If you are measuring an entire receiver there are a few things you have to
be careful with. The receiver must be a linear receiver (no FM, AM diode
detector, etc.---basically just SSB). There should be a filter to pick just
one sideband. Turn the AGC off. Make sure you measure the gain in the
linear region, which also applies to a simple amplifier.

Definitely. Though, with my method the input never gets more than 3
dB above the receiver noise floor and in most cases a well designed
receiver will have no AGC response at that level. But with a preamp
in front it usually will activate the AGC, so AGC has to be switched
off when comparing preamps - which of course is impossible to do in
most ham rigs !

If you have a DUT with a known noise figure, I think that this would be one
way of calibrating a homebrew noise source.

Hence the interest in NF repeatability of MMIC amps, since they are
hard to build wrong, are well matched over a wide bandwidth and don't
require tuning for best noise figure. The known DUT can also be the
standard itself (in association with a receiver of only roughly known
NF) to avoid issues of errors in calibrating the noise source ENR.

73,
Steve VE3SMA

You also have to know the noise bandwidth of the system to use this
method. If the response is dominated by a single, fairly steep-sided
filter, this is easy. But otherwise (such as if both IF filtering and AF
amplifier response shape the overall response), some calculation and/or
measurement is required. I've gotten what I believe are reasonable
results on HF and AF amplifiers using this method. But I haven't tried
it on a really low NF system, so would heed Jim's caution.

Roy Lewallen, W7EL

Jim wrote:

Steve,

There is another method of measuring noise figure that does not need a noise
source.

The reason that it is not used as often is that it is harder to measure low
noise figures accurately. Since there are still a few layers of rust on my
brain I will not geive the details here, since i do not want to say anything
that is too far off.

This other method involves measuring the gain of the device under test and
then measuring the noise power output with the input terminated properly.
The noise output is a combination of the DUT noise and the 'ambient' noise
from the input termination, times the gain of the receiver. Working
backwards you can then determine the DUT noise. As you can probably guess,
the reason it doesn't work so good on small noise figures is that any errors
in the measurement can easily outnumber the actual noise figure you are
trying to measure! And I have actually had to deal with this problem. That
is where half my gray hairs came from (the other half from being laid off).

If you are measuring an entire receiver there are a few things you have to
be careful with. The receiver must be a linear receiver (no FM, AM diode
detector, etc.---basically just SSB). There should be a filter to pick just
one sideband. Turn the AGC off. Make sure you measure the gain in the
linear region, which also applies to a simple amplifier.

If you have a DUT with a known noise figure, I think that this would be one
way of calibrating a homebrew noise source.

Jim
N8EE

On 14 Oct 2004 16:44:35 -0700, (Steve
Kavanagh) wrote:

I haven't really been following this thread but if you are looking for
noise diodes and associated stuff, you should take a look at:

http://www.noisecom.com/

Seems to me that at one time they would calibrate a homebrew source
for a modest fee. That may have been a ham that worked there paving
the way; not sure.




|"Jim" wrote in message ...
|
| This other method involves measuring the gain of the device under test and
| then measuring the noise power output with the input terminated
| properly
|
| ....any errors
| in the measurement can easily outnumber the actual noise figure
|
|Thanks for reminding me of that one, Jim. But I see your point about
|errors. For example, in measuring the gain one needs a standard. One
|of the few pieces of real test gear I have is a bolometer-type RF
|power meter which can measure about -13 dBm accurately. If the
|measurement bandwidth is 1 MHz (suitable for VHF, perhaps) then
|thermal noise is -114 dBm. So I need about 100 dB gain for a very low
|noise figure DUT. To measure that I might need five 20 dB attenuators
|as a standard, each with perhaps +/-0.5 dB accuracy if I am lucky...so
|there's +/-2.5 dB error (well, I suppose I could cross my fingers and
|RSS the numbers). Or I have a diode-type power meter that will
|measure lower power, which leads to issues of how the detector
|responds to noise. And then there's the problem of knowing the noise
|bandwidth precisely...
|
| is where half my gray hairs came from (the other half from being laid off).
|
|I think about half of mine come from the latter factor too !
|
| If you are measuring an entire receiver there are a few things you have to
| be careful with. The receiver must be a linear receiver (no FM, AM diode
| detector, etc.---basically just SSB). There should be a filter to pick just
| one sideband. Turn the AGC off. Make sure you measure the gain in the
| linear region, which also applies to a simple amplifier.
|
|Definitely. Though, with my method the input never gets more than 3
|dB above the receiver noise floor and in most cases a well designed
|receiver will have no AGC response at that level. But with a preamp
|in front it usually will activate the AGC, so AGC has to be switched
|off when comparing preamps - which of course is impossible to do in
|most ham rigs !
|
| If you have a DUT with a known noise figure, I think that this would be one
| way of calibrating a homebrew noise source.
|
|Hence the interest in NF repeatability of MMIC amps, since they are
|hard to build wrong, are well matched over a wide bandwidth and don't
|require tuning for best noise figure. The known DUT can also be the
|standard itself (in association with a receiver of only roughly known
|NF) to avoid issues of errors in calibrating the noise source ENR.
|
|73,
|Steve VE3SMA

I am hoping that accuracies of +/- 1 to 2 dB might be achievable.
If you are satisfied with +/- 1 to 2 dB variation, the link I gave to
the Mar MMIC 35dB "calibrated" Noise generator should be
accurate enough.
That why I "forgot" the approximatly statement.

You can not compare noise power with a calibrated sine wave generator.
That's right, not directly.
You can calculate the RMS for a sine wave.
Also calculate the RMS for noise power, but its another formule.
Use the same impedances and frequency's.


If you know the amount of (milli) Ampere's through the noise diode and
know the impedance of the load, you can make power calculations like
we do with DC.
With the calculation method you don't need to have a calibrated noise head.
Someone else discussed the method, let me give the formule.
The formule is in Chris Bowick's book RF Circuit Design.
Its about Shot Noise. (Not thermal)
In^2 = 2qIdcB

In^2 = the mean square noise current
q = the electron charge (1.6 x 10e-19 coulombs)
Idc = the direct current in Ampere's
B = the bandwith in Hertz

Onces you have the mean square current, calculate the power in the load
resistor.
Did not made calculations with it, since I have the opportunity to measure
noise
right away. Will do it in future to check the formule.

The MMIC I used.... I was afraid there would come a question about it.
It's in a factory designed preamp of Japanese origin, has a forgotten
product
number, I mean National, not sure of that.
Searched hours for datasheets, because I was convinced the NF of the device
was less at 144Mhz, the manual stated approximatly 2dB 2Ghz.
EME use 144Mhz, with the knowledge of bipolar transistors that have an
increase of NF with frequency, 144Mhz would have a NF of about 1 - 1.5dB.
The manual said nothing about a lower NF at lower frequency's.
So bought it with the deal, not good money back.
Found the datasheets, the NF was 2.8dB flat from almost DC to 2Ghz.
From there comes the idea of flat NF respons of MMIC's.
I brought the device back to store and got money back. 150$.
More research learnt me that similar devices with less NF in order of 0.7dB
were avaiable for 20$, OK no box around it, no blinky leds.
I know only that the MMIC is obsolete today, forgot the type number.

A single mosfet, BF981 does 1.7dB at 100Mhz and cost about 1$.
I love homebrewing.
With that device you could also calibrate your noise generator.
It matched the input impedance of the 50 Ohm receiver.
Many EME amateurs use it in their preamps.
If they all use the same scheme it would be accurate within a variation
of say +/- 0.1dB ?

Since I can measure noise power, how to calibrate whitout a calibrated
noise head is not a hot item for me anymore.
Just practice.
Before that, I read several books, spent hours on the internet, collected
many
data.
Its only to share thoughts, nothing science, professional or years of
experience,
and yes I can be wrong.
The goal was maybe you can do something with it, for me, maybe there is a
clever
way to determine the NF without calibrated noise head, maybe even whitout
math.

The JT44 software program, I just finished the interface from receiver to
PC.
Details at the original PSK31 homepage.
Used a "Jensen" audio trafo, to prevent ground loops.
A lot to do... no practical data at this moment.

Have read Terry Ritter's stuff.
He said, no white noise at all. Just pink. See the graph's.
Read in some university books, noise is of pink behavior at the lowest
frequency's.
The graph showed the audio band.
White noise is flat (whitin 15 - 20Khz partitions) and at higher
frequency's, and random.
Whit respect to the measurements of course.

There is an increase of NF to higher frequency regions.
How came that MMIC to a flat respons ?

Sorry I was wrong about the noise figure at 100Mhz for the mosfet BF981.
It has to be 0.7dB instead of 1.7dB!

It can still better with the BF998, to get an impress of the scheme try the
following link http://lea.hamradio.si/~s53ww/4xbf998/4xbf998.htm

I think it can be reproducible whitin small variations.