|
There has been some interest in the past about building a tracking generator for the tinySA.? Here is an excerpt from an article in the current issue of QEX magazine from thje ARRL.
?
?
?
?
|
Here is a readable version of the first page as a zip file
?
?
|
What's the point if for the price of this generator you can buy a Nano/Tiny VNA that will give you much more data than just a frequency response?
|
On Thu, Jan 16, 2025 at 10:09 AM, MaxPi wrote:
What's the point if for the price of this generator you can buy a Nano/Tiny VNA that will give you much more data than just a frequency response?
On the first page of the article the author points out that the tinySA with a tracking generator has more dynamic range in dB for measurements of filters than S21 mode of a typical NanoVNA.
?
Roger
|
I can hardly imagine the need to get another ten dB of dynamic range beyond the 80 dB available from the litevna. But if I needed it, I would simply add a mixer to this pair of devices.
|
On Thu, Jan 16, 2025 at 11:43 AM, MaxPi wrote:
I can hardly imagine the need to get another ten dB of dynamic range beyond the 80 dB available from the litevna.
? I just checked my litevna and above 3.5Gz you don't get anywhere near 80 dB of dynamic range.? I haven't read the full article but for tuning duplexers a tracking generator could be a handy accessory.
?
Herb
|
To adjust diplexers,?duplexers filters?etc better off using a VNA. A spectrum analyzer / tracking generator only provides amplitude, a VNA measures S parameters i.e. amplitude?and?phase etc. Always terminate all unconnected ports of a DUT.? ?
On Thu, Jan 16, 2025 at 11:43 AM, MaxPi wrote:
I can hardly imagine the need to get another ten dB of dynamic range beyond the 80 dB available from the litevna.
? I just checked my litevna and above 3.5Gz you don't get anywhere near 80 dB of dynamic range.? I haven't read the full article but for tuning duplexers a tracking generator could be a handy accessory.
?
Herb
|
On Thu, Jan 16, 2025 at 09:42 PM, VK3DPM Dave wrote:
To adjust diplexers,?duplexers filters?etc better off using a VNA.
Ideally, yes.? But finding a VNA that costs less than $500 that has the dynamic range and sweep speed to do so real-time does not appear to be currently available.? My main point was that there are still some scalar measurements where having a tracking generator might allow you to make S21 measurements that the current crop of NanoVNA and LiteVNA's cannot perform as well.
?
Again, I have not read the full article, so I don't know what dynamic range the author's build is capable of achieving.
?
Herb
?
?
|
Il Nanovna è uno strumento a basso costo, molto versatile e facile da usare. Ha una gamma dinamica mediocre ed è molto utile per visualizzare la curva di risposta di un filtro. Tuttavia, se utilizzato per calibrare i filtri, è lento nella visualizzazione della curva e questa lacuna ne impedisce l'utilizzo nella calibrazione dei filtri, diventa stressante se ci sono molti componenti da regolare perché tra una regolazione e l'altra possono passare un paio di secondi. Inoltre non è in grado di eseguire misure sui miscelatori.
|
The N2PK VNA can achieve measurements as low as -120 dB, but its frequency range is limited to 60 kHz to 60 MHz.
|
For those that do not have a translator here is the text of Valentino's post:
?
The Nanovna is a low-cost, very versatile and easy-to-use instrument. It has a mediocre dynamic range and is very useful for displaying the response curve of a filter. However, if used to calibrate filters, it is slow in displaying the curve and this shortcoming prevents its use in filter calibration, it becomes stressful if there are many components to adjust because between one adjustment and another a couple of seconds can pass. It is also not able to perform measurements on mixers.
|
Also please note: This is a tracking generator for the TinySA not the TinySA ultra..
?
To build a tracking generator for the tinySA is relative easy even with the need of high isolation. To build a tracking generator for the ultra is not quite straight forward.
?
?
John
|
There are some problems with building a tracking generator for the Ultra
LO output is possible through the CAL port using CONFIG/MORE/LO OUTPUT but the level is somewhat lower then with the tinySA basic. Not a big problem.
Having good isolation between the almost 1 GHz LO in the tracking generator and the IF of the tinySA Ultra will be VERY difficult. At 1 Ghz almost everything becomes transparent which will reduce the dynamic range
But the biggest complicating factor, in my view, is the tuning mechanism used in the Ultra. The frequency steps of the main LO are in the order of kHz which is not good enough when using a 200 Hz RBW. This is solved by tuning both the LO (for the coarse steps) and the IF frequency (for the fine steps). Furthermore the IF filters have some manufacturing tolerance and for optimal performance the IF frequency is at startup time optimized for IF filter. And then there are some situations where the IF is shifted to solve various problems in the architecture.
As there is currently no way to know for sure the frequency of the IF at every point of the sweep you can only use the widest RBW with a tracking generator which also causes a loss of dynamic range.
--
Designer of the tinySA
For more info go to
|
May I suggest yet another cheap solution for a tracking generator for use with the TINYSAs?? This has been used extensively with many SDRs as well.? A good flat output noise generator makes an excellent substitute for a tracking generator.? Build or buy one and use it.? If there is not enough output, use one or several MMICs to increase the amplitude.? You might even want to envoke the MAX HOLD or AVERAGINGoptions for archiving smoother records of the plots.? I've used this technique many times with my SDRs.? Of course, I have a number of SAs, vector network analyzers, and signal generators, but just using the broadband noise source is just "fun" if you didn't have all these instruments.? Personally, I see no need with this option to build a tracking generator for the TINYSAs.
Dave - W?LEV
On Fri, Jan 17, 2025 at 4:28?PM valentino barbi via <barbivalentino= [email protected]> wrote: Il giorno ven 17 gen 2025 alle ore 17:21 Erik Kaashoek via <erik= [email protected]> ha scritto: There are some problems with building a tracking generator for the Ultra
LO output is possible through the CAL port using CONFIG/MORE/LO OUTPUT but the level is somewhat lower then with the tinySA basic. Not a big problem.
Having good isolation between the almost 1 GHz LO in the tracking generator and the IF of the tinySA Ultra will be VERY difficult. At 1 Ghz almost everything becomes transparent which will reduce the dynamic range
But the biggest complicating factor, in my view, is the tuning mechanism used in the Ultra. The frequency steps of the main LO are in the order of kHz which is not good enough when using a 200 Hz RBW. This is solved by tuning both the LO (for the coarse steps) and the IF frequency (for the fine steps). Furthermore the IF filters have some manufacturing tolerance and for optimal performance the IF frequency is at startup time optimized for IF filter. And then there are some situations where the IF is shifted to solve various problems in the architecture.
As there is currently no way to know for sure the frequency of the IF at every point of the sweep you can only use the widest RBW with a tracking generator which also causes a loss of dynamic range.
--
Designer of the tinySA
For more info go to
|
BTW:? Zener diodes make wonderful BB noise generators when biasef just breaking into the Zener knee!? Follow that with a few MMICs to increase amplitude and finally a 6 dB resistive attenuator for impedance stabilization.? You're ready to roll.
Dave - W?LEV
May I suggest yet another cheap solution for a tracking generator for use with the TINYSAs?? This has been used extensively with many SDRs as well.? A good flat output noise generator makes an excellent substitute for a tracking generator.? Build or buy one and use it.? If there is not enough output, use one or several MMICs to increase the amplitude.? You might even want to envoke the MAX HOLD or AVERAGINGoptions for archiving smoother records of the plots.? I've used this technique many times with my SDRs.? Of course, I have a number of SAs, vector network analyzers, and signal generators, but just using the broadband noise source is just "fun" if you didn't have all these instruments.? Personally, I see no need with this option to build a tracking generator for the TINYSAs.
Dave - W?LEV
On Fri, Jan 17, 2025 at 4:28?PM valentino barbi via <barbivalentino= [email protected]> wrote: Il giorno ven 17 gen 2025 alle ore 17:21 Erik Kaashoek via <erik= [email protected]> ha scritto: There are some problems with building a tracking generator for the Ultra
LO output is possible through the CAL port using CONFIG/MORE/LO OUTPUT but the level is somewhat lower then with the tinySA basic. Not a big problem.
Having good isolation between the almost 1 GHz LO in the tracking generator and the IF of the tinySA Ultra will be VERY difficult. At 1 Ghz almost everything becomes transparent which will reduce the dynamic range
But the biggest complicating factor, in my view, is the tuning mechanism used in the Ultra. The frequency steps of the main LO are in the order of kHz which is not good enough when using a 200 Hz RBW. This is solved by tuning both the LO (for the coarse steps) and the IF frequency (for the fine steps). Furthermore the IF filters have some manufacturing tolerance and for optimal performance the IF frequency is at startup time optimized for IF filter. And then there are some situations where the IF is shifted to solve various problems in the architecture.
As there is currently no way to know for sure the frequency of the IF at every point of the sweep you can only use the widest RBW with a tracking generator which also causes a loss of dynamic range.
--
Designer of the tinySA
For more info go to
--
|
Dave, can you explain how you can reach a high dynamic range, such as 100 dB, using a noise source?
This because this thread seems to focus on high dynamic range measurements.
--
Designer of the tinySA
For more info go to
|
For filters, not active devices (!!), one can amplify the BB noise to the point where a good SA is able to display (maybe barely) 100 dB of dynamic range.? All my "professional" SAs are capable of? 100 dB of on-screen range.? I don't use the TINYSAs for aligning cavities.? That's a job for the HP8753C or equivalent.?
Lot's of gain from the Zener, if that is what's used, is required.?
Here is a link to the Elecraft BB (500 MHz) noise source (a number of others are available on Amazon and epay):
? ? ?
These are specified at 35 dB of excell noise from 100 kHz through 500 MHz, depending on the characteristics of the noise diode.??
I just checked my calibrated noise source (25 dB of excess noise) through the HP 8447D BB amplifier into the 8568 SA.? The source is attenuated down to 20 dB excess noise for other purposes.? With that combination, I observed 20 dB of excess noise.? Therefore, an additional gain of roughly 80 dB would be required to hit the top, 100 dB above the bottom of the display, to yield 100 dB of dynamic range.? Agreed, this is a lot of gain, but easily obtained with MMICs or (Amazon or epay) BB amplifiers.
Note, my direct conversion receiver I use for analyzing ionospheric scattering of the WWV transmitters includes a BB 75 dB gain block (and filters) home brewed from MiniCircuits MMICs. ? Only three MMICs are used. ? I'm only 30 or so miles south of Wellington so I receive both scattered and GW signals from Wellington.? Gain is easy so long as the builder is aware of best engineering practices addressing isolation between stages.?
Dave - W?LEV?
Dave, can you explain how you can reach a high dynamic range, such as 100 dB, using a noise source?
This because this thread seems to focus on high dynamic range measurements.
--
Designer of the tinySA
For more info go to
|
The problem with using wideband noise sources is that the total power output can easily exceed the safe maximum power input limit of the Spectrum Analyzer if one is not careful.?
?
If you keep the total power output below the safe limit you don't get much dynamic range for the measurement.
?
Sure you can run a high power noise source and hope the filter output will reduce the total output power to a safe limit but just one mistake when you hook things up and bye bye front end of your SA.
?
Roger
|
OK.? I have never used the BB noise technique for high dynamic range measurements.? I have other test equipment for those purposes.? But, for amateur application (not for tuning cavities), it works quite well.? For 100+ dB applications, still, a good professional VNA is the correct tool.?
Dave - W?LEV
On Fri, Jan 17, 2025 at 7:49?PM Roger Need via <sailtamarack= [email protected]> wrote: The problem with using wideband noise sources is that the total power output can easily exceed the safe maximum power input limit of the Spectrum Analyzer if one is not careful.?
?
If you keep the total power output below the safe limit you don't get much dynamic range for the measurement.
?
Sure you can run a high power noise source and hope the filter output will reduce the total output power to a safe limit but just one mistake when you hook things up and bye bye front end of your SA.
?
Roger
|
On Fri, Jan 17, 2025 at 01:15 PM, W0LEV wrote:
A good flat output noise generator makes an excellent substitute for a tracking generator.
Be careful of how much broadband power your noise generator puts out lest it overload the SA front end.? You definitely won't be able to use the LNA!
?
I just measured and my 3-MMIC "NOISE SOURCE BG7TBL 2016-03-06" puts almost exactly +7dBm of broadband noise into my 8GHz power sensor.? ?That's a big signal.
?
It might not damage the SA, but during normalization it could compress the front end, distorting amplitude of your measurements when lower power levels are present and the gain is less compressed. ?30dB more or less of attenuation after the noise source should keep the SA happy. ?
Definitely don't try to measure an amplifier using a noise source unless you really know what you are doing.
73, Don N2VGU
?
|
Roger Need
MaxPi
Juan M0WWA
John Cunliffe W7ZQ
Donald S Brant Jr