开云体育

On Wed, Sep 2, 2020 at 08:34 AM, hwalker wrote:

Another justification for two tinySA's ;)?

Well, just for fun I extended this page:?
so you have a justification for THREE tinySA
Two are used to create a two tone generator (with some limitations) and together with a third tinySA these are used to measure the IIP3 of a cheap eBay amp.
At least, that is what I think I did. Hope the measurement specialists in this group won't shoot to many holes in the line of thinking......

That explains why tinysa.org crashed this morning. Too much BW consumed

On Wed, Sep 2, 2020 at 07:46 AM, Erik Kaashoek wrote:

A new example is added to the wiki:?
This example uses
- Level sweep of the low output mode
- Measuring in zero span to record level change over time.
- Triggering the sweep at a specific level transition
- Delta marker to simplify calculation

? Another justification for two tinySA's ;)?

? ?The example is timely for me as I recently ordered one of the cheap 20dB gain, 5-6000MHz Chinese amplifiers on eBay and that's one of the measurements I plan on performing when it arrives.

? - Herb

A new example is added to the wiki:?
This example uses
- Level sweep of the low output mode
- Measuring in zero span to record level change over time.
- Triggering the sweep at a specific level transition
- Dela marker to simplify calculation

New FW released
Version V1.0-66

Changes:
- Improved auto attenuate algorithm

The new auto attenuate algorithm has more resolution (1dB i.s.o. 5dB) and updates only once per second to avoid quickly jumping noise floors when scanning in fast mode.

On Mon, Aug 31, 2020 at 06:32 AM, hwalker wrote:

Why does the level of the signals appear to change when switching from 0 to 10 dB attenuation?

Interesting question!
The clock ingress happens after the attenuator so it stays the same level when the attenuation increases. But the calculation assumes all signals are reduced by the attenuation so it increases everything with the set attenuation.
When you connect an antenna to the tinySA and you put that antenna close to the MCU (make sure the clock signals are substantial above -80dBm) the signals will NOT increase when you increase the attenuation unless you attenuate so the signal from the antenna is weaker than the direct ingress after the attenuator.
The noise floor is also "created" after the attenuator and also increases with the amount of attenuation

Erik,
? I have a question regarding the MCU clock related signals.? Why does the level of the signals appear to change when switching from 0 to 10 dB attenuation?

? ?MCU clk signals, 0dB attenuation.? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Same except 10 dB attenuation. Clk related signals ~10dB higher.
? ?? ??

? ?- Herb

The tinySA has some problems with measuring AM modulation so I added a measurement example:?

On Sun, Aug 30, 2020 at 07:22 AM, hwalker wrote:

Holy moly, a three tinySA set-up

You guys have those nice HP or even better signal generators, I have none, so I must use some cheap tinySA for my experiments

In this measurement example:?
you can see the internally generated harmonics disappears into the noise floor if the fundamental at the mixer? (after the attenuator) is below -25dBm?so -20dBm as target for auto attenuation was indeed not optimal for the novice user. Maybe -30dB is a bit on the safe side but I expect a lot of novice users.

New FW released
version v1.0-64

Change:
- Removed the 2.4kHz auto RWB setting

A youtube reviewer discovered the manual 3kHz RBW provided better narrow span results compared to the automatic 2.4kHz RBW so I removed the 2.4kHz RBW

On Sun, Aug 30, 2020 at 06:00 AM, Erik Kaashoek wrote:

New FW released
version v1.0-63

Change:
- The target maximum level for the auto attenuation is changed from -20dBm to -30dBm

In one of the user videos I noticed that, when not very familiar with the inner workings of a spectrum analyzer and relying on the auto attenuation functionality, it is important to ensure minimum internally generated harmonics.
This changed setting will raise the noise floor with 10dB with stronger signals but those that know what they are doing and in need of a low noise floor and not concerned about possible internal generation of harmonics can always switch? to manual attenuation.

? ?I seem to recall thinking in one of the YouTube videos that what was being called internally generated tinySA spurs were actually harmonics of the HP signal generator.? The so called "tinySA harmonics" moved when the signal generator frequency was changed which should have been like a big sign popping up saying, "signal generator related".

? The demonstrator obviously didn't spend a lot of time with the tinySA or read the wiki before making the videos.? In his first video the measured levels were off because he had not performed a level calibration (red reference level characters).? I thought all tinySA's were shipped with in-factory level calibration so I'm not sure how that happened. Subsequent videos showed he eventually figured it out and measured levels were within 0.5dB of the HP generator setting.

? You would expect a beginner to need extra hand holding when setting optimum levels but the YouTube demonstrator had a room full of equipment and should have understood that over driving a spectrum analyzer input will generate spurious signals - as it also did on his HP analyzer in his second video.?

?More experienced users manually change the tinySA settings for best dynamic range, with some attenuation up front for better input impedance.? Auto reflevel, attenuation, and rbw are all set to manual as soon as I power on my tinySA.

- Herb

On Sun, Aug 30, 2020 at 01:25 AM, Erik Kaashoek wrote:

The real reason why you would used the tinySA with a tracking generator is to measure frequency translation devices such as a mixer as this can not be done using a nanoVNA.
I added an example to the wiki

Holy moly, a three tinySA set-up. Who'd of thought :)

- Herb

On Sat, Aug 29, 2020 at 11:13 PM, Erik Kaashoek wrote:

Yes,
One of the early testers used it to find leakage paths through an EMC shielding.
I made a crude NF probe and used it to check for radiated EMC.
Is the sensitivity is not enough in low mode you can add an LNA as shown here:?

? 1. Ensure you maintain probe positioning between circuit changes for accurate comparison.? Also, if you are narrowed in on a specific problem frequency ensure you widen your scan after fixing that frequency. Fixing one frequency can sometimes have unintended consequences at another frequency.

? 2. The hand probe detects emissions in the near field and, for compliance testing, you may sometimes wind up chasing ghosts that are not seen in the far field with an antenna.

- Herb

New FW released
version v1.0-63

Change:
- The target maximum level for the auto attenuation is changed from -20dBm to -30dBm

In one of the user videos I noticed that, when not very familiar with the inner workings of a spectrum analyzer and relying on the auto attenuation functionality, it is important to ensure minimum internally generated harmonics.
This changed setting will raise the noise floor with 10dB with stronger signals but those that know what they are doing and in need of a low noise floor and not concerned about possible internal generation of harmonics can always switch? to manual attenuation.

In one of the user youtube videos I noticed confusion on what to expect from spur removal so I extended the wiki with a page showing the impact and (im)possibilities of the various countermeasures.

The real reason why you would used the tinySA with a tracking generator is to measure frequency translation devices such as a mixer as this can not be done using a nanoVNA.
I added an example to the wiki

A fascinating comparison.

However, ?there are apparent differences in the results, in that the sa/tracking generator does not show the endstop attenuation at -40db, despite the scale (presumably 10db per div) going to -70db.

It brings back memories of when I first played with a home-made spectrum analyser in the early 90’s (0-100mhz, 60db dynamic range, display on a scope). I made a bridge and noise generator to align filters etc.

interestingly, I got it out recently, as I had originally modded it to have an extra rbw filter of 1khz but it was useless because the first swept oscillator ( ramp generator into a variac diode) had too much jitter, but it was because the original designer had made a circuit error. I would like to get it out again and compare it with the Tinysa.

Anyway, I have several nanovna devices, as well as an old Deepace KC901s.

Steve L. G7PSZ

Yes,
One of the early testers used it to find leakage paths through an EMC shielding.
I made a crude NF probe and used it to check for radiated EMC.
Is the sensitivity is not enough in low mode you can add an LNA as shown here:?