¿ªÔÆÌåÓý

  1. HP-Agilent-Keysight-Equipment
  2. Topics

Peak vs. Average Power Sensors

 

I'm trying to educate myself about the newer USB power sensors. When I look for example at the U2043XA (average) and U2044XA (peak and average) specs, I see that both can do at least 20K measurements/second. The 2044XA has a video bandwidth of 5 MHz, while that's not specified for the 2043XA.

My question is whether the averaging sensor can still provide a peak reading, if the modulation is slow enough.

If I can make 20K measurements per second, it seems I should be able to follow a modulation envelope of up to 10 kHz (Nyquist sampling) and use the computer to process that data stream to catch and report the peak level over some time period. In fact, in theory I could process all the samples to fully reconstruct the modulation.

Is that correct, or am I missing something?

Thanks,
John
 

¿ªÔÆÌåÓý

That¡¯s correct as far as amplitude modulations go, and also the sampling close to the Nyquist limit only works when it¡¯s synchronous with the input signal: the Nyquist limit isn¡¯t some general ¡°sampling works up to here¡± frequency. Without synchronous sampling, there will be large amplitude errors, getting lower in frequency as the signal approaches the limit. If you, say, modulate at 4.9kHz, and sample at 10kHz, the envelope itself will have a secondary envelope at 100Hz, and only its peaks will have the correct amplitude:

image1.png

Cheers, Kuba

30 okt. 2018 kl. 15:59 skrev John Ackermann N8UR <jra@...>:

I'm trying to educate myself about the newer USB power sensors. ?When I look for example at the U2043XA (average) and U2044XA (peak and average) specs, I see that both can do at least 20K measurements/second. ?The 2044XA has a video bandwidth of 5 MHz, while that's not specified for the 2043XA.

My question is whether the averaging sensor can still provide a peak reading, if the modulation is slow enough.

If I can make 20K measurements per second, it seems I should be able to follow a modulation envelope of up to 10 kHz (Nyquist sampling) and use the computer to process that data stream to catch and report the peak level over some time period. ?In fact, in theory I could process all the samples to fully reconstruct the modulation.

Is that correct, or am I missing something?

Thanks,
John
 

Thanks, Kuba. The idea about synchronous sampling being required near the limit is new to me, so I'm off to do some research.

I think you answered my question, and this limitation won't affect me in practice. My question was really whether I could use the 20k sampling rate to pick up peak power in an AM (actually, SSB) transmission that has about 3.5 kHz bandwidth. With a 20k sampling rate, that shouldn't be a problem.

Thanks,
John
----

On 10/30/18 4:27 PM, Kuba Ober wrote:
That¡¯s correct as far as amplitude modulations go, and also the sampling close to the Nyquist limit only works when it¡¯s synchronous with the input signal: the Nyquist limit isn¡¯t some general ¡°sampling works up to here¡± frequency. Without synchronous sampling, there will be large amplitude errors, getting lower in frequency as the signal approaches the limit. If you, say, modulate at 4.9kHz, and sample at 10kHz, the envelope itself will have a secondary envelope at 100Hz, and only its peaks will have the correct amplitude:
image1.png
Cheers, Kuba
30 okt. 2018 kl. 15:59 skrev John Ackermann N8UR <jra@... <mailto:jra@...>>:

I'm trying to educate myself about the newer USB power sensors. ?When I look for example at the U2043XA (average) and U2044XA (peak and average) specs, I see that both can do at least 20K measurements/second. ?The 2044XA has a video bandwidth of 5 MHz, while that's not specified for the 2043XA.

My question is whether the averaging sensor can still provide a peak reading, if the modulation is slow enough.

If I can make 20K measurements per second, it seems I should be able to follow a modulation envelope of up to 10 kHz (Nyquist sampling) and use the computer to process that data stream to catch and report the peak level over some time period. ?In fact, in theory I could process all the samples to fully reconstruct the modulation.

Is that correct, or am I missing something?

Thanks,
John
 

Not going to discuss the physics, rather the common use cases for peak reading sensors primarily used to look at either pulsed signals and/or look at the POWER waveform as opposed to the voltage waveform.

Peak reading sensors used frequently in WiFi development work because you can see the actual power of all parts of a wifi frame. Same goes for cell/bt development.

By changing measurement mode a peak sensor can do anything a average reading (square law) sensor can do


Content by Scott
Typos by Siri

1 - 4 of 4