开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育
  1. Tinysa
  2. Topics

Is the TinySA Ultra capable of measuring short radar pulses?

 

I recently purchased a TinySA Ultra in hopes of being able to measure the strength of a radar signal at various locations of my property. After weeks of comparing the (borrowed) Agilent FieldFox N9915A results to TinySA, I am coming to the realization that the TInySA is not capable. The radar signal is composed of 2.8752 GHz, +5dBm, 1us duration pulses every 1ms. The radar spins so I get a burst of about 20 of these pulses every 4.8 seconds.

This is the best I can do with TinySA. Unchecking the Spur Removal seemed to help but changing NSpeedUp, WSpeedUp and #points did not have any effect.?
Why are the power levels of the pulses so different using?the?TinySA while the FieldFox is very consistent (and much higher)?
image.png
image.png


Below is for comparison.
Thanks in advance.
~~Mike Flohr


Here is Fieldfox capture over 6 seconds:
image.png
And over 50ms
image.png
And over 50us
image.png
 

The TinySA sweep rate is not fast enough to capture short duration burst signals like this. ?

The Fieldfox will do the job but it costs 100 times more. ?Optimized circuitry for these types of signals. ?Here is a description of how they do it…
?
https://www.keysight.com/ca/en/assets/7018-05591/brochures/5992-2069.pdf
 

Thanks Roger.
<snip from keysight link> "With an FFT rate of almost 120,000 per second and FFT overlapping, each display update represents about 2,500 spectra of gap-free data capture."
OK I get that the?RTSA is very powerful and necessary to?differentiate complex?WiFi and BT signals.?I am wondering if all that power is necessary for this specific radar pulse.?
BTW, I can get the same plots with the HP8562E but I presume that is because it is analog and much easier to process.
image.png

Is there anything?we could do with the TinySA to get close? Is there a basic limitation such as the sampling frequency? Or the size of the FFT? Or buffers?
Is the challenge the narrow pulse width, or the dynamic range, or the pulse repetition?rate?
I see a previous topic about measuring doppler radar "PRT 120Hz-500hz (2ms-8ms), the pulse width is about 20us to 32us, and the center frequency is 9.33Ghz."
Is that easier since it is a wider pulse? Or greater PRT?



On Thu, Sep 26, 2024 at 11:07?PM Roger Need via <sailtamarack=[email protected]> wrote:
The TinySA sweep rate is not fast enough to capture short duration burst signals like this. ?

The Fieldfox will do the job but it costs 100 times more.? Optimized circuitry for these types of signals.? Here is a description of how they do it…
?
 

Mike,
Interesting experiment.
I took a signal generator at 10 MHz and -10 dBm and changed the signal burst length starting with 0.1 micro seconds and 10 ms burst repetition time
The tinySA was set to 10 Mhz zero span, 850 kHz RBW and to trigger in NORMAL mode at a level of -15 dBm
At 6 microsecond burst length the -10 dBm signal level was occasionally reached, longer bursts did increase the frequency of triggering.?
Reducing the burst length required lower level of triggering to see the burst.
This implies the minimum burst length for being able to occasionally measure the true signal level is around 6 micro seconds.
Due to the fixed DSP detector inside the tinySA that requires at least some samples for the single bucket DFT filter, it is impossible to reduce this time.
With reduced RBW the minimum time required for true level detection increases as more samples are needed for the resolution filter.
?
At -60 dBm trigger level the 10 microseconds bursts showed varying levels
This is caused by the roughly 40 microseconds interval between the samples taken where there is some dead time in the 40 microseconds sample interval.
Signal bursts of 50 micro seconds always where measured at the -10 dBm level as these are guaranteed to cover the whole 40 microseconds sampling interval.
Due to the design of the level detector in the tinySA it is impossible to reduce or eliminate this dead time in the signal sampling?
?
I will add this info the the tinySA specification.
--
Designer of the tinySA
For more info go to

 

One thing I need to add
The power of the short burst will only be measured correctly if their maximum power is at most 60 dB above the noise level as the AGC inside the tinySA can not react to such shorts bursts
Best is to set the attenuation/signal level? such that the signal level is expected to be no more than 60 dB above the noise floor of the tinySA.
With 50 microsecond bursts that are detected every sweep the AGC will correctly auto set the internal attenuation?
--
Designer of the tinySA
For more info go to

 

Thanks Erik,
I tried adding an attenuator on the off chance that the AGC might be clipping but it had no effect.
?
New thought ... Is there a way to tell the tinySA to plot the raw samples (in zero span) instead of trying to process them??
Can you create a "Sample Capture" buffer that would store raw samples when triggered and plot after it fills?
Can the buffer be large enough to capture 50us? Or better yet, 2ms? Selectable #points to trade resolution for time?
And can it be configured to capture samples 5 us prior to the trigger?
Extra credit: Could the sample capture buffer be made available for download?
?
 

I'm curious what radar type this is. You mention a power of +5dBm...is this an attenuated sample pulse on the radar transmitter's cabinet? If so, could it be an ASR-9?
?
73,
?
--
John AE5X
?
 

If you invoke "Max Hold", you will eventually catch a full radar pulse.?

Dave - W?LEV


On Sat, Oct 12, 2024 at 12:14?AM John AE5X via <ae5x=[email protected]> wrote:
I'm curious what radar type this is. You mention a power of +5dBm...is this an attenuated sample pulse on the radar transmitter's cabinet? If so, could it be an ASR-9?
?
73,
?
--
John AE5X
?



--
Dave - W?LEV
 

When waiting for a trigger, the tinySA captures the incoming samples at maximum speed in a raw circular buffer.
Once the trigger condition is satisfied the raw samples are converted to requested unit and displayed.
So to capture a short pulse in zero span mode, set a level trigger just above the noise floor.
--
Designer of the tinySA
For more info go to

1 - 9 of 9