开云体育

Have I just been unlucky and got a bad unit, or does everyone suffer spurious signals on their TinySA?

I see specific signals at 144.20MHz, 192.15MHz, and 240.19MHz.

They are unstable, and wander over a few hundred kHz, during a period of about an hour. Each are 12-15dB above noise floor in 10kHz RBW.

They are definitely internally generated. I have applied 50Ohm loads to the SMA connectors, and even operated the device inside a Faraday cage, but still the spurious signals remain visible.

Any feedback welcome.

John

These are 48MHz harmonics generated by the MCU. For now all tinySA will show them at 48, 96, 144, 192 and 240 MHz

On Mon, Aug 17, 2020 at 03:16 AM, Erik Kaashoek wrote:

These are 48MHz harmonics generated by the MCU. For now all tinySA will show them at 48, 96, 144, 192 and 240 MHz

Erik,
? ?On my tinySA there also appears to be a sub-harmonic of the 48MHz clk at 4.8MHz.? Possibly a divide by 10 output?

LOW and HIGH INPUTS terminated in 50ohm loads.


Your explanation and an illustrative screen shot deserve a Wiki entry, possibly under 'Limitations' or as an addition to the FAQ topic, "Question:?Even without anything connected I still see some signals ".Perspective buyers should know what to expect if the internally generated MCU signals will be an issue to them.- Herb

On Mon, Aug 17, 2020 at 06:00 AM, hwalker wrote:

Your explanation and an illustrative screen shot deserve a Wiki entry, possibly under 'Limitations' or as an addition to the FAQ topic, "Question:?Even without anything connected I still see some signals ".

Done,
The 4.8MHz is indeed present.

Yes the unit certainly has limitations, though for the price is not to bad really.

Here's a -4dBm carrier from an ADF4351 at 50MHz.

The TinySA shows a wide flat top and noise floor of only 45dB down at 12kHz away from center, where as in actual fact the carrier's noise floor is 80dB down by the time your only 1kHz away from center.

A useful unit all the same.

Now the 64,000 dollar question.

What could be done to suppress them?

Either hide them with software tricks, such as the spur-removal option, but that might leave “holes” in the spectrum, or additional filtering/ decoupling/ screening on the board, which is out of the question for most current owners.

There is of course the human software modification - learn to live with it and adjust practice accordingly.

Out of interest, I was playing with my nanoV2 to improve screening. I found that things seemed to be worse if I applied conductive (absorption) tape inside the case and even laying the device (in FR4 unclad case) on a flat metal sheet also made things worse, especially above 2.5ghz.

Steve L. G7PSZ

Regarding ADF4351, would the Tninsa perform better with these rather than adf4350?

Steve L ? G7PSZ

The TinySA doesn't use the ADF4350, it uses a couple of Si4432's.

Above 1GHz RF is like water. It flows? through every hole
Hiding with the spur removal does not work as the spur removal removes spurs that do not actually exist but this 48MHz harmonic leakage is an actual signal.
In a next HW revision (not planned) I will pay some extra attention to them.

Which adf4350?

On Mon, Aug 17, 2020 at 08:31 AM, OneOfEleven wrote:

The TinySA shows a wide flat top and noise floor of only 45dB down at 12kHz away from center, where as in actual fact the carrier's noise floor is 80dB down by the time your only 1kHz away from center.

Have a look here:?

On Mon, Aug 17, 2020 at 09:31 AM, Erik Kaashoek wrote:

Above 1GHz RF is like water. It flows? through every hole

? ?[Our lab called it the Dutch boy effect.? You plug one hole and RF finds another path of least resistance.

Hiding with the spur removal does not work as the spur removal removes spurs that do not actually exist but this 48MHz harmonic leakage is an actual signal.

? [There are innovative DSP solutions but I always view them as treating the symptom instead of the problem. I favor the old fashioned way of actually curing the problem by rounding off sharp clock edges so that the harmonics are severely attenuated.? some engineers favor using spread spectrum clocks that attenuate the emissions by spreading the signal power.

? The tinySA meets my needs and expectations as is.? User's with more advanced applications in mind will most likely try mitigating the MCU related emissions using their own bag of tricks.? The tinySA surface mount components damper my enthusiasm to do so.]

In a next HW revision (not planned) I will pay some extra attention to them.

[Hopefully a fellow hardware type will submit an inexpensive solution to the group in the meantime.? Source suppression seems the best route.]

- Herb