Signal to Noise Ratio ?

[Skywave]

This is one for the techs. amongst us!

BTW I am a little undecided about the right Section for this, since it also involves Test Equip. I trust that it will be moved accordingly if required.

Yes, I know that I should know this, but it was a long, long time ago when I was last doing this sort of thing professionally. I need a brief reminder to get me started again

I've looked around the Web for this, but no joy.
So here goes . . .

Take one typical communications receiver and an assortment of test gear - sig. gen., AF power meter, a selection of attenuators, and a noise generator -- possibly home-built using an A2087 noise diode. (This is the nearest I'm going to get to acquiring one!)

What is the technique for measuring the S/N ratio, how is it specified and what practical value does it have? What is the significance of the S/N ratio of RX 1 compared to RX 2?

C'mon you techs - let's be hearing from you!

Al / G8DLH

[jim_beacon]

The normal method is to feed in a modulated signal (usually 1KHz 30% AM), and measure the ouput level, then switch off the modulation and measure the ouput level. The two values give the signal to noise, usually expressed in dB.

A Signal to Noise is value is actually signal + noise : noise, and is normally expressed as x dB for y uV input, z% Modulation at a specified RF and AF frequency.

You will sometimes see it quoted as a maximum RF input to give a known ratio, for example 20dB S/N with 1uV or less input, and sometimes as a minimum ratio for a given input, e.g. better than 10dB S/N at 3uV

Changing any of the parameters can massivly affect the reading, so it can be difficult to compare data between manufacturers, as they all have their own favourite audio frequency and modulation depth.

It gives an indication of how good a receiver is a picking a signal out from the background rubbish.

Jim.

[Biggles]

The only other thing to add to Jim's explanation is that the higher the signal to noise ratio the better the receiver. i.e. 20 dB s/n is better than 10 dB s/n for the same signal level input. For FM receivers sinad readings are more commonly used, or alternatively, a quieting measurement is taken. An alternative to using microvolts is to use dBm as an input level. It all depends on what you were brought up on; some people use microvolts, others use dBm. I feel that I may have opened a can of worms here! You are probably going to ask me to explain this further.

Biggles.

[YC-156]

If, on the other hand, you were talking about the receiver input noise figure, the procedure is as follows:

With a 50 ohm dummy load connected to the receiver input, measure the AF output voltage of the white noise with AGC disabled and in AM mode. Any bandwidth is suitable for this. A suitable resistor should be used in parallel with the voltmeter as an AF 'dummy load' while doing these tests.

Now replace the dummy load at the antenna with the noise diode setup. If the noise diode setup is built to spec, you increase the current through the noise diode by increasing the filament voltage until the AF *power* has doubled, Ie. until the voltage across the AF dummy load resistor has increased by sqrt(2) ~ 1.41.

The receiver noise figure is now usually related directly to the saturated current through the noise diode. The exact equation will be in the datasheet, but is often 1mA per dB of noise figure.

Hope this helps a bit.

Frank N.

[XTC]

SiNAD - Signal to Noise and Distortion is the usual quoted standard.

See.

http://www.aeroflex.com/products/sig...cles/sinad.pdf

Which is what Jim and Biggles have been saying.

Somewhere, I've got a WW article on SiNAD, which I think is a bit different for FM sets.

By the looks of that link, you need a good signal generator which a modulates a pure sine wave accurately and a SiNAD meter, which looks like a distortion factor meter with extra filtering. As the measurement of noise is important, you really need true RMS meters for the voltage measurement, or a separate true RMS meter. Most distortion factor meters have sine wave average meters, and that usually doesn't matter too much, however, I think that measuring noise is quite different, and that's where the true RMS measurent becomes important.


Pete.

[Skywave]

Originally Posted by Biggles

I feel that I may have opened a can of worms here! You are probably going to ask me to explain this further.

Biggles.

Not at all! I'm quite OK with all that you've said, incl. SINAD, dBm / uV, etc. No further explanations required!

Thank you for your input: much appreciated.

Al / G8DLH

[Skywave]

Originally Posted by jim_beacon

The normal method is to feed in a modulated signal (usually 1KHz 30% AM), and measure the ouput level, then switch off the modulation and measure the ouput level. The two values give the signal to noise, usually expressed in dB.

A Signal to Noise is value is actually signal + noise : noise, and is normally expressed as x dB for y uV input, z% Modulation at a specified RF and AF frequency.

Jim.

Thanks.
Further down the page there is an comment on Input Noise Figure - this is what I was trying to recall. (It all comes back to me now!)
However, your comment is of value, since the method you give provides an easy way to determine the S/N.
The reason I ask all this is so that I can assess the performance (in this particular parameter) of a few comms. receivers I own to check performance / need for maint. by comparing the measured figures with the manufacturer's data.

Again, thank you.

Al / G8DLH

[Skywave]

Originally Posted by YC-156

If, on the other hand, you were talking about the receiver input noise figure, the procedure is as follows . . .etc . . .

Hope this helps a bit.

Frank N.

Thanks!

Ah - ha! THIS is what I was trying to recall from 30+ years ago!
Err, it helps MORE than "just a bit"! As soon as I read it, it all seemed so familiar!

Looks like I'm going to be a little busy, building my own noise generator (I have a number of A2087 somewhere!).

BTW - if I do this - and fit a mA meter to monitor the diode current- have you any thoughts on how to calibrate the finished noise gen (x mA on meter = y uV of noise)?

Again, many thanks for your comments.

Al / G8DLH

[Skywave]

Originally Posted by XTC

SiNAD - Signal to Noise and Distortion is the usual quoted standard. See.:
http://www.aeroflex.com/products/sig...cles/sinad.pdf . . . . etc . . .

Pete.

Thank you.
As I said elsewhere, the real reason behind all of this is to check performance of comms. receivers (that go to 30 Mc) - especially post-repair.
I shall refer to that Site for further guidance!

Al / G8DLH

[oldeurope]

I am doing it in the same way like Frank N toled it.
My noise generator makes a high level and I use attenuators to reduce its output level. So the output of the noise generator is exactly known.

Kind regards
Darius

[Skywave]

Originally Posted by oldeurope

I am doing it in the same way like Frank N toled it.
My noise generator makes a high level and I use attenuators to reduce its output level. So the output of the noise generator is exactly known.

Kind regards
Darius

Thanks.
Err, perhaps I'm missing something here, but if you have an attenuator of specified loss - say 10dB - you'll need to know the power in to it to calculate the power out. No? (I assume everything is Z- matched.)

Al / G8DLH

[GMB]

The normal method is to feed in a modulated signal (usually 1KHz 30% AM), and measure the ouput level, then switch off the modulation and measure the ouput level. The two values give the signal to noise, usually expressed in dB.

Is that dB of the voltage ratio, or the power ratio? (The standard confusion over dB usage).

[XTC]

Originally Posted by G8DLH

Thank you.
As I said elsewhere, the real reason behind all of this is to check performance of comms. receivers (that go to 30 Mc) - especially post-repair.
I shall refer to that Site for further guidance!

Al / G8DLH

I just found the EW&WW article. July 1999.

There were a few sorts of sensitivity measurements used by different makers of communication receivers, such as minimum discerible signal sensitivity (input needed to produce a signal 3dB above the noise floor), standard o/p power sensitivity (all different) etc.

SiNAD is really intended for FM as another poster said.

Generally, for an older comms receiver the maker quotes something like:

For S/N ratio of 15db @50mW output
AM modulated 30% @ 400Hz better than 5uV all ranges.
CW on ranges 1,2,3,4 .........


So, you can check it with a good RF signal generator (no radiated RF) and an AC millvoltmeter, prefererably true RMS.

If you are working on older comms receivers, you are really checking to see that they are back up to scratch as per the maker's spec, not cross comparing them, so look for the sensitivity measure quoted in the data and aim for that as a bill of health.

Pete.

[Skywave]

Originally Posted by GMB

Is that dB of the voltage ratio, or the power ratio? (The standard confusion over dB usage).

Huh? dBs are dBs my friend!

Al / G8DLH

[Skywave]

Originally Posted by XTC

If you are working on older comms receivers, you are really checking to see that they are back up to scratch as per the maker's spec, not cross comparing them, so look for the sensitivity measure quoted in the data and aim for that as a bill of health.

Pete.

Exactly. That's what I intend to do - which was why I raised this Thread in the first place. I threw in the "cross - compare" notion simply to develop the thread and stimulate responses.

Al / G8DLH

[Sam]

Originally Posted by GMB

Is that dB of the voltage ratio, or the power ratio? (The standard confusion over dB usage).

Originally Posted by G8DLH

Huh? dBs are dBs my friend!

I agree a dB is a dB, but the way they have to be calculated is different dependant on if you have a Voltage or Power ratio.

For voltage,
[dB]=20log(Vout/Vin).

For power,
[dB]=10log(Pout/Pin).

This is because power is proportional to the square of the voltage.

Our lecturer used to like this trick to make sure you were reading the question properly in Exams!

Sam

(After editing - I was initially wrong: got the 10 and 20 transposed. It was stuff I last used in the first-year though and thought I had gotten it wrong as soon as I posted )

[Skywave]

Originally Posted by Sam

I agree a dB is a dB, but the way they have to be calculated is different dependant on if you have a Voltage or Power ratio.

Sam

Oh dear! Here we go again!

I have an amplifier that has a gain of 3dB. It doesn't matter if I say its 3dB "voltage" gain or 3dB "power" gain. 3 dB is 3 dB!

This amp. has an input R and a load R of 50 ohms.
Therefore, 1 watt in gives 2 watts out in the load,
since Gain = 10 log (P2/P1) = 10 log (2/1) = 10 log 0.3010 = 3 dB

The input voltage will be sq.rt of (PR) = sq. rt.(50) = 7.07 volts.
The output voltage will be sq. rt of (PR) = sq. rt (2x50) =10 volts.
Gain = 20 log (V2/V1) = 20 log (10/7.07) = 20 log (1.41) = 20 x 0.15 = 3dB

Whether your referring to a voltage ratio or a power ratio, in either case, the "Gain" is 3dB.

As I say, dBs are dBs !

Therefore, to get bacck on Thread, when the original contributor said "until the output increases by 3dB" there is no confusion!

Al / G8DLH

[YC-156]

Originally Posted by G8DLH

TBTW - if I do this - and fit a mA meter to monitor the diode current- have you any thoughts on how to calibrate the finished noise gen (x mA on meter = y uV of noise)?

Nope, sorry, I don't.

The noise generated is by definition wideband, so AFAIK you cannot assign a definite voltage level without knowing the bandwidth of the particular piece of receiving equipment being tested.

Originally Posted by G8DLH

Again, many thanks for your comments.

Al / G8DLH

You are quite welcome.

Best regards

Frank N.

[Sam]

I agree Al, the gain will be the same for a situation wether or not you have a Power or Voltage ratio. A dB is something you cannot measure - only calculate. A measured quantity has to be compared to a reference to be able to calculate a dB gain. The way this gain is calculated depends upon the quantity measured.

Sam

[GMB]

I have an amplifier that has a gain of 3dB. It doesn't matter if I say its 3dB "voltage" gain or 3dB "power" gain. 3 dB is 3 dB!

Mmm. So if I have an amplifier with a voltage gain of 2 and a power gain of 1000,000 then what dB gain do you say it has??