SDR-Console Power dBm Accuracy


Larry Dodd
 

I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


Tony_AD0VC
 

Not accurate unless you do something to calibrate them. The relative readings are fine but you can't take an absolute measurement unless you have calibrated your setup with a known signal. 

Tony


From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of Larry Dodd <101science@...>
Sent: Wednesday, June 23, 2021 6:53 PM
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: [SDR-Radio] SDR-Console Power dBm Accuracy
 
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


jdow
 

That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


Larry Dodd
 

jdow,
Thanks. We use a calibrated on/off noise source to do a Y Factor Method for SDR noise figure measurements. I have found just using the spectrum dBm measurements easier than the dBm meter. It usually reads -140 dBm constantly with low noise floors. 
Thanks,
Larry

On Jun 23, 2021, at 9:49 PM, jdow <jdow@...> wrote:

 That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


jdow
 

Hot and cold noise sources are useful. Depending on the amplifier pure liquid nitrogen may be a good friend for accuracy. If you can pull in a precise calibrated step attenuator and nice hot noise source it might be interesting to see what you get as you add in attenuation. And ideally you'd work with a long term RMS average on the noise. (I think Simon does that on his spectrum.) Generally a Mark 1 human eyeball is not a very good tool to read spectrum analyzer noise levels.

An S-Meter is pretty poisonous to use with noise readings. But you might discover it can give you some useful readings when deltas are involved. If I had to do it I'd probably take some tools I already have and doctor them to do the precise job that's needed = set a precise IF bandwidth and measure and long term average the I^2 + Q^2 noise level coming out of the IF filter. I do not know that any tool already made out there caters to this. And I have spent over 3/4 century learning how to be lazy when faced with this sort of issue.

{^_-}

On 20210623 18:57:21, Larry Dodd wrote:
jdow,
Thanks. We use a calibrated on/off noise source to do a Y Factor Method for SDR noise figure measurements. I have found just using the spectrum dBm measurements easier than the dBm meter. It usually reads -140 dBm constantly with low noise floors. 
Thanks,
Larry

On Jun 23, 2021, at 9:49 PM, jdow <jdow@...> wrote:

 That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED



Larry Dodd
 

Jdow
Yes agree. I think you have it described well. The Mark 1 eyeball you described isn’t highly accurate at averaging the bouncing line. A steady tone and step attenuator would be more accurate. Will do some more thinking on this. Thanks. 
Larry, K4LED 

On Jun 23, 2021, at 11:47 PM, jdow <jdow@...> wrote:

 Hot and cold noise sources are useful. Depending on the amplifier pure liquid nitrogen may be a good friend for accuracy. If you can pull in a precise calibrated step attenuator and nice hot noise source it might be interesting to see what you get as you add in attenuation. And ideally you'd work with a long term RMS average on the noise. (I think Simon does that on his spectrum.) Generally a Mark 1 human eyeball is not a very good tool to read spectrum analyzer noise levels.

An S-Meter is pretty poisonous to use with noise readings. But you might discover it can give you some useful readings when deltas are involved. If I had to do it I'd probably take some tools I already have and doctor them to do the precise job that's needed = set a precise IF bandwidth and measure and long term average the I^2 + Q^2 noise level coming out of the IF filter. I do not know that any tool already made out there caters to this. And I have spent over 3/4 century learning how to be lazy when faced with this sort of issue.

{^_-}

On 20210623 18:57:21, Larry Dodd wrote:
jdow,
Thanks. We use a calibrated on/off noise source to do a Y Factor Method for SDR noise figure measurements. I have found just using the spectrum dBm measurements easier than the dBm meter. It usually reads -140 dBm constantly with low noise floors. 
Thanks,
Larry

On Jun 23, 2021, at 9:49 PM, jdow <jdow@...> wrote:

 That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED



jdow
 

You are best sticking with all noise or all signal. And you need to calculate precisely what bandwidth is represented by each pixel on the spectrum analyzer screen AND how each pixel finds its way into the running power average. Noise compared to noise does not have this problem. Signal compared to signal faces typical SINAD issues.

TANSTAAFL

{o.o}

On 20210623 22:15:18, Larry Dodd wrote:
Jdow
Yes agree. I think you have it described well. The Mark 1 eyeball you described isn’t highly accurate at averaging the bouncing line. A steady tone and step attenuator would be more accurate. Will do some more thinking on this. Thanks. 
Larry, K4LED 

On Jun 23, 2021, at 11:47 PM, jdow <jdow@...> wrote:

 Hot and cold noise sources are useful. Depending on the amplifier pure liquid nitrogen may be a good friend for accuracy. If you can pull in a precise calibrated step attenuator and nice hot noise source it might be interesting to see what you get as you add in attenuation. And ideally you'd work with a long term RMS average on the noise. (I think Simon does that on his spectrum.) Generally a Mark 1 human eyeball is not a very good tool to read spectrum analyzer noise levels.

An S-Meter is pretty poisonous to use with noise readings. But you might discover it can give you some useful readings when deltas are involved. If I had to do it I'd probably take some tools I already have and doctor them to do the precise job that's needed = set a precise IF bandwidth and measure and long term average the I^2 + Q^2 noise level coming out of the IF filter. I do not know that any tool already made out there caters to this. And I have spent over 3/4 century learning how to be lazy when faced with this sort of issue.

{^_-}

On 20210623 18:57:21, Larry Dodd wrote:
jdow,
Thanks. We use a calibrated on/off noise source to do a Y Factor Method for SDR noise figure measurements. I have found just using the spectrum dBm measurements easier than the dBm meter. It usually reads -140 dBm constantly with low noise floors. 
Thanks,
Larry

On Jun 23, 2021, at 9:49 PM, jdow <jdow@...> wrote:

 That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED




Simon Brown
 

Larry,

 

Relative values will be perfect, or at least as perfect as the ADC / tuner chip in the SDR. No doubt there’s a professional methodology which could be used, I’m not aware of this so wait to be informed.

 

Obviously there’s Windowing applied and with the Spectrum the optional smoothing.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Larry Dodd
Sent: 24 June 2021 01:53
To: main@SDR-Radio.groups.io
Subject: [SDR-Radio] SDR-Console Power dBm Accuracy

 

I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


jdow
 

A hot and cold noise source pair is probably his best path to accuracy. The math is fairly straight forward and useful. Bandwidth cancels out of the calculation. So as long as it is constant the Thot and Tcold can give you your device noise figure. This set of slides features some moderately heavy math, which breaks down to some fairly simple algebraic equations.

Agilent Technology - Hot Cold Nose Figure Measurement

Go down to the Y-Factor measurement that Larry has mentioned he's looking to using.

The only issue might be that signal outside the band pass may give you averaging effects that extend sensitivity. (The AN/USC-8  DSCS satellite modem used a simple SDR that actually needed an external signal applied when operating in loop back mode as there was not enough noise to wiggle the LSB. With the antenna and LNA gain that problem went away. It took me awhile to wrap my head around how it worked. I "think" I understand it a little, now.)

{^_^}

On 20210623 23:03:48, Simon Brown wrote:

Larry,

 

Relative values will be perfect, or at least as perfect as the ADC / tuner chip in the SDR. No doubt there’s a professional methodology which could be used, I’m not aware of this so wait to be informed.

 

Obviously there’s Windowing applied and with the Spectrum the optional smoothing.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Larry Dodd
Sent: 24 June 2021 01:53
To: main@SDR-Radio.groups.io
Subject: [SDR-Radio] SDR-Console Power dBm Accuracy

 

I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


jdow
 

Hm, of note the hot and cold noise source CAN become a problem during measurement. You must know both actual temperatures at the time of measurement. The assumed "290K" does not necessarily work. The old HP 340B had a noise head featuring a (gasp) actual 6AL5, empty state electronics, noise diode. It's filament power heated up the reference resistor above room temperature. So the meter COULD read out a negative noise figure. Boy howdy that weirded me out when it happened.

{^_-}

On 20210623 23:03:48, Simon Brown wrote:

Larry,

 

Relative values will be perfect, or at least as perfect as the ADC / tuner chip in the SDR. No doubt there’s a professional methodology which could be used, I’m not aware of this so wait to be informed.

 

Obviously there’s Windowing applied and with the Spectrum the optional smoothing.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Larry Dodd
Sent: 24 June 2021 01:53
To: main@SDR-Radio.groups.io
Subject: [SDR-Radio] SDR-Console Power dBm Accuracy

 

I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


Gisle Vanem
 

We are interested in measuring the noise floor of various SDR receivers.

FYI.

SDRangel recently added a plugin for such measurements: https://github.com/f4exb/sdrangel/tree/master/plugins/channelrx/noisefigure

I've not tried it. It seems to depend on an external noise-source controlled via this VISA library.


Bob Dengler
 

At 6/24/2021 12:19 AM, you wrote:
Hm, of note the hot and cold noise source CAN become a problem during measurement. You must know both actual temperatures at the time of measurement. The assumed "290K" does not necessarily work.
T(off) = 290K should be valid, provided there is a well matched attenuator greater than 20 dB between the actual noise diode/tube & the output of the noise source module that's always near (physical) room temperature.

Bob NO6B


Simon Brown
 

I’ve just looked at the SDRangel link below – looks good so this is the software to use for accurate NF measurements.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Gisle Vanem
Sent: 24 June 2021 08:25
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] SDR-Console Power dBm Accuracy

 

We are interested in measuring the noise floor of various SDR receivers.

FYI.

SDRangel recently added a plugin for such measurements: https://github.com/f4exb/sdrangel/tree/master/plugins/channelrx/noisefigure

I've not tried it. It seems to depend on an external noise-source controlled via this VISA library.


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


jdow
 

There was a chance he was using the well honed Amateur Radio approach, find some surplus gear and use it. As noted in a different note, HP304B can surprise you.
{^_-}

On 20210624 00:29:47, Bob Dengler wrote:
At 6/24/2021 12:19 AM, you wrote:
Hm, of note the hot and cold noise source CAN become a problem during measurement. You must know both actual temperatures at the time of measurement. The assumed "290K" does not necessarily work. 
T(off) = 290K should be valid, provided there is a well matched attenuator greater than 20 dB between the actual noise diode/tube & the output of the noise source module that's always near (physical) room temperature.

Bob NO6B 







Nicholas Shaxted
 

With the proviso that the amateur user assumes the calibration sticker is still valid of course.

 

Nick – g4ogi

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of jdow via groups.io
Sent: 24 June 2021 08:28
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] SDR-Console Power dBm Accuracy

 

There was a chance he was using the well honed Amateur Radio approach, find some surplus gear and use it. As noted in a different note, HP304B can surprise you.
{^_-}

On 20210624 00:29:47, Bob Dengler wrote:

At 6/24/2021 12:19 AM, you wrote:
Hm, of note the hot and cold noise source CAN become a problem during measurement. You must know both actual temperatures at the time of measurement. The assumed "290K" does not necessarily work. 
T(off) = 290K should be valid, provided there is a well matched attenuator greater than 20 dB between the actual noise diode/tube & the output of the noise source module that's always near (physical) room temperature.
 
Bob NO6B 
 
 
 
 
 

 


hb9ajg
 

Larry:

the accuracy depends on 3 things:

-  ADC accuracy of the RX under test

-  Accuracy of the software used

-  Accuracy of your calibrated test generator.

I have done accuracy tests with my Hermes board (16bit ADC) using SDR Console and a calibrated (within +-1dB) HP8657 signal generator.

For levels between -60 and -120dBm from the generator the dBm readings of SDR Console were within +-1dB.

73, Walter HB9AJG

 


Larry Dodd
 

Bob
Thanks. The brand new BBGen noise source is mounted in an rf tight metal box that is peltier temperature controlled at 72 degrees F. Have a brand new JFW 110 dB step attenuator mounted in that same box. No surplus stuff. 
Larry, K4LED 

On Jun 24, 2021, at 4:28 AM, jdow <jdow@...> wrote:

 There was a chance he was using the well honed Amateur Radio approach, find some surplus gear and use it. As noted in a different note, HP304B can surprise you.
{^_-}

On 20210624 00:29:47, Bob Dengler wrote:
At 6/24/2021 12:19 AM, you wrote:
Hm, of note the hot and cold noise source CAN become a problem during measurement. You must know both actual temperatures at the time of measurement. The assumed "290K" does not necessarily work. 
T(off) = 290K should be valid, provided there is a well matched attenuator greater than 20 dB between the actual noise diode/tube & the output of the noise source module that's always near (physical) room temperature.

Bob NO6B 







Larry Dodd
 

Simon,
Thanks!  Will download it. 
Larry, K4LED 

On Jun 24, 2021, at 3:34 AM, Simon Brown <simon@...> wrote:



I’ve just looked at the SDRangel link below – looks good so this is the software to use for accurate NF measurements.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Gisle Vanem
Sent: 24 June 2021 08:25
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] SDR-Console Power dBm Accuracy

 

We are interested in measuring the noise floor of various SDR receivers.

FYI.

SDRangel recently added a plugin for such measurements: https://github.com/f4exb/sdrangel/tree/master/plugins/channelrx/noisefigure

I've not tried it. It seems to depend on an external noise-source controlled via this VISA library.


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


Larry Dodd
 

Thanks to all. Very valuable information. Appreciate it all!
Larry, K4LED 

On Jun 24, 2021, at 3:14 AM, jdow <jdow@...> wrote:

 A hot and cold noise source pair is probably his best path to accuracy. The math is fairly straight forward and useful. Bandwidth cancels out of the calculation. So as long as it is constant the Thot and Tcold can give you your device noise figure. This set of slides features some moderately heavy math, which breaks down to some fairly simple algebraic equations.

Agilent Technology - Hot Cold Nose Figure Measurement

Go down to the Y-Factor measurement that Larry has mentioned he's looking to using.

The only issue might be that signal outside the band pass may give you averaging effects that extend sensitivity. (The AN/USC-8  DSCS satellite modem used a simple SDR that actually needed an external signal applied when operating in loop back mode as there was not enough noise to wiggle the LSB. With the antenna and LNA gain that problem went away. It took me awhile to wrap my head around how it worked. I "think" I understand it a little, now.)

{^_^}

On 20210623 23:03:48, Simon Brown wrote:

Larry,

 

Relative values will be perfect, or at least as perfect as the ADC / tuner chip in the SDR. No doubt there’s a professional methodology which could be used, I’m not aware of this so wait to be informed.

 

Obviously there’s Windowing applied and with the Spectrum the optional smoothing.

 

Simon Brown, G4ELI

https://www.sdr-radio.com

 

From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> On Behalf Of Larry Dodd
Sent: 24 June 2021 01:53
To: main@SDR-Radio.groups.io
Subject: [SDR-Radio] SDR-Console Power dBm Accuracy

 

I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED


--
- + - + -
Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.


Larry Dodd
 

Yes. I use 500 hz CW as a standard. Use calibrated noise or signal generator sine wave for tests. 

Using actual live signals has some practical aspects for sure but complicates accurate noise figure measurements. 
Larry, K4LED 

On Jun 24, 2021, at 1:51 AM, jdow <jdow@...> wrote:

 You are best sticking with all noise or all signal. And you need to calculate precisely what bandwidth is represented by each pixel on the spectrum analyzer screen AND how each pixel finds its way into the running power average. Noise compared to noise does not have this problem. Signal compared to signal faces typical SINAD issues.

TANSTAAFL

{o.o}

On 20210623 22:15:18, Larry Dodd wrote:
Jdow
Yes agree. I think you have it described well. The Mark 1 eyeball you described isn’t highly accurate at averaging the bouncing line. A steady tone and step attenuator would be more accurate. Will do some more thinking on this. Thanks. 
Larry, K4LED 

On Jun 23, 2021, at 11:47 PM, jdow <jdow@...> wrote:

 Hot and cold noise sources are useful. Depending on the amplifier pure liquid nitrogen may be a good friend for accuracy. If you can pull in a precise calibrated step attenuator and nice hot noise source it might be interesting to see what you get as you add in attenuation. And ideally you'd work with a long term RMS average on the noise. (I think Simon does that on his spectrum.) Generally a Mark 1 human eyeball is not a very good tool to read spectrum analyzer noise levels.

An S-Meter is pretty poisonous to use with noise readings. But you might discover it can give you some useful readings when deltas are involved. If I had to do it I'd probably take some tools I already have and doctor them to do the precise job that's needed = set a precise IF bandwidth and measure and long term average the I^2 + Q^2 noise level coming out of the IF filter. I do not know that any tool already made out there caters to this. And I have spent over 3/4 century learning how to be lazy when faced with this sort of issue.

{^_-}

On 20210623 18:57:21, Larry Dodd wrote:
jdow,
Thanks. We use a calibrated on/off noise source to do a Y Factor Method for SDR noise figure measurements. I have found just using the spectrum dBm measurements easier than the dBm meter. It usually reads -140 dBm constantly with low noise floors. 
Thanks,
Larry

On Jun 23, 2021, at 9:49 PM, jdow <jdow@...> wrote:

 That's a big subject.

To a first approximation the spectrum is very accurate for differences in reading. It's theoretically perfect if there is a perfect A/D converter involved. The A/D converter's step to step linearity can mess up this theoretically perfect relative calibration.

In any absolute sense any calibration present is rather loosy goosy unless you provide a precision source and take care getting your readings. Otherwise "loosy goosy" can mean +/- 20 dB or worse.

I refuse to address Simon's S-Meter accuracy until I check to see if I can still spoof it.

{^_^}

On 20210623 17:53:16, Larry Dodd wrote:
I have been asked the question, "How accurate are the SDR-Console dBm spectrograph and dBm meter readings?"  Can anyone answer that question? We are interested in measuring the noise floor of various SDR receivers.
Thanks,
Larry, K4LED