TinySA Is A $49 Spectrum Analyzer

The NanoVNA made network analyzers cheap enough for almost everyone. Now you can get a $49 spectrum analyzer to go with it. Is it worth it? Watch [IMSAI Guy]’s video after the break for his opinion. From the tinySA.org website:

  • Spectrum Analyzer with two inputs, high-quality MF/HF/VHF input for 0.1MHZ-350MHz, lesser quality UHF input for 240MHz-960MHz.
  • Switchable resolution bandpass filters for both ranges between 2.6kHz and 640kHz
  • Color display showing 290 scan points covering up to the full low or high-frequency range.
  • Input Step attenuator from 0dB to 31dB for the MF/HF/VHF input.
  • When not used as Spectrum Analyzer it can be used as Signal Generator, MF/HF/VHF sinus output between 0.1MHZ-350MHz, UHF square wave output between 240MHz-960MHz.
  • A built-in calibration signal generator that is used for automatic self-test and low input calibration.
  • Connected to a PC via USB it becomes a PC controlled Spectrum Analyzer
  • Rechargeable battery allowing a minimum of at least 2 hours portable use

A lot of cheap scopes and PC-based scopes can do spectrum analysis, too, of course, so this isn’t as exotic as a VNA. But at this price, having a dedicated instrument might be worth it to you, especially if you don’t care about frequency below 100 MHz.

There are some limitations, of course, but the price is right. [IMSAI Guy] shows a few oddities that he didn’t like, but overall, it seemed like a good value. If you have a modern scope it may already do this function, or you might be able to do a software solution. If you only need audio frequencies and you want novelty, try some ping pong balls

42 thoughts on “TinySA Is A $49 Spectrum Analyzer

    1. Well it doesn’t work as a spectrum analyzer, so if that’s what you need you’d be out of luck. You could probably modify the firmware to get some functionality, but I would guess the frontend would not be sensitive enough to be useful. Really you need an LNA.

      Probably something like the Siglent SVA1015X or Rigol RSA3015N would be a better choice if you need both a VNA and spectrum analyzer, although the price point is a lot higher. Anyone else know of good alternatives?

    2. Spectrum analyzer =/= VNA, They might both deal with frequency and even have similar technology. The presentation and what they are measuring is very different.

      Spectrum analyzer let you look at a signal and figure out its frequency components. It is passive measurement as doesn’t inject anything.

      VNA let you look at the transfer function of a circuit over frequency. It *injects* its own signals as part of the measurement process. Transfer fuction in layman’s term is output vs input and is usually in S domain.

    3. Dreaming of one with bigger screen. Badly.
      Those 3” displays are ridiculously limited wrt resolution and hard to read for us poor old greybeards. Is it asking too much to bump the screen size to 7”? The more space available could be turned into circuitry to obtain more functionalities (SDR, SA, function generator, multimeter, etc), not to mention bigger longer lasting batteries etc. I would definitely throw over €150 to an instrument like that.

    1. Like any spectrum analyzer, it is possible – easy, even – to overload the front end and generate harmonics and spurs. If you are aware of the tinySA’s limitations (which Erik has kindly posted on the wiki), it works surprisingly well. Keep signals below -35 dBm and most of thee spurs disappear.

  1. “But at this price, having a dedicated instrument might be worth it to you, especially if you don’t care about frequency below 100 MHz.”

    Don’t you mean 100 KHZ? That’s 0.1 Mhz…

    Watch your decimal places!

    1. “How does it compare to the RTL-SDR?”

      Significantly better bandwidth and a lower noise floor ~100dB for the frequency range spec. Works well below -30dB so consider using with attenuation or signals of interest. Yesterday there was an auto attenuate firmware upgrade. Small and stand alone operation too… no computer or portable pack required or ~-70db noise floor.

      With Spektrum or GRC, from what I’ve done, you can get a slow sweep so to have a wider bandwidth at the cost of sampling rate with the RTL-SDR’s.

      TinySA can also be used as a signal generator so with two tinySA’s you can make a tracking generator.

      Here’s a link with the newer videos noting how to use for better performance: https://www.youtube.com/c/IMSAIGuy/search?query=TinySA

      Here’s the TinySA video regarding: https://www.youtube.com/watch?v=aGy1p3_2qAY

      Ham Radio Crash Course has a video also regarding: https://www.youtube.com/watch?v=o4CnecfAzng

      There is even a branch that is open source (last I knew) now called the simpleSA that is based off the original tinySA design.

      Really neat what could be done with the cheap SI4432 modules, the variety of cheap MCU’s and now to see a marketed product I even own. Really neat experience for RF design studies and learning too. The HBTE and tinySA groups.io have a lot of great information regarding.

    2. Most obviously, the RTL-SDR does not have an integral battery, display screen and tracking generator output. For only about an additional $10 US, you get a stand-alone instrument with the TinySA. On the other hand, the RTL-SDR gives you real time I and Q samples at 2 megasample/sec or more over USB, making it a very flexible receiver when used with a computer. You can build almost arbitrarily complex virtual receivers the GNU Radio software and an RTL-SDR. RTL-SDR also has a jumperable bias tee that allows power injection to power an antenna mounted pre-amplifier. One of the bigger problems with the RTL-SDR is that there are no pre-selector filters, making it susceptible to overload and intermodulation of out-of-band signals with the signal you want. The RTL-SDR is OK but not great for 1.0 MHz – 30 MHz. The 820T2 chip belies its original intended use as a DVB-T UHF TV tuner.

    3. It’s a whole lot better than the RTL-SDR for lower frequencies – The RTL-SDL (with the front end bypass for low frequencies) has a fixed internal sample rate 28.8 MSPS at 8 bits/sample, then decimated to a lower rate (1-2 MSPS) for the host. That 28.8 MHz means there’s plenty of aliasing for low band stuff, and the 8 bits limits the dynamic range. The TinySA also has a more open design (the RTL-SDR is mostly implemented in a proprietary chip from Realtek) and has various useful SA features (attenuators, etc.)

      The TinySA doesn’t have a front end tracking filter, so strong out of band signals can cause problems. The attenuator helps. Nobody expects a low NF for a spectrum analyzer, so that’s not a big deal.

      Overall, for $50, it’s a pretty handy device. Between it and the NanoVNA (which can also work as a signal generator) you’ve basically got $100-120 to get the RF (up to L-band) equivalent of a DMM – very, very handy stuff.

    4. Apples to oranges. The tinySA is a spectrum analyzer. The RTL-SDR is a receiver; there are ways to turn it into a spectrum analyzer, albeit with limited dynamic range and slow frequency scanning.

      That said, the RTL-SDR can be a useful supplement to the tinySA for spectrum analysis. The minimum RBW of the tinySA is 2.6 KHz, so it’s not much use for looking at the spectrum of a transmitter or oscillator near its main frequency, or the response of a narrowband filter. The RTL-SDR, along with appropriate external attenuators (and a noise source for the case of filter analysis), serves those purposes well. A higher performance SDR such as the ones made by Airspy and SDRPlay are even better for that but are more costly.

  2. What differentiates TinySA from Nano VNA is the TinySA includes switchable input attenuators. To get the best measurement performance, it is important to pre-scale the input signal to the level where the detector has its best trade-off of linearity and sensitivity. There are also other Spectrum Analyzer factors such as resolution bandwidth, IF bandwidth and type of detector (peak/average/log/minimum) that are not as important considerations to a vector network analyzer that is generating its own controlled stimulus signal.

    I do have a Nano VNA, the model with a few improvements over the original bare circuit board. It is a surprisingly good instrument for less than $100 US. Keep in mind that the Nano VNA is not a lab grade instrument, it is fairly fragile, has minimal shielding, has minimal input overload protection, and other compromises. The Nano VNA is a great teaching tool which I have used in my Communication Systems lectures, and it can also be used effectively by radio Amateurs for antenna tuning, etc. On the other hand, there is a reason an Anritsu Site Master costs thousands of Dollars.

    I will probably buy a TinySA. At fifty bucks it is hard to say, “no.” If anything, it would be nice to have a, “disposable,” piece of test gear for situations where a lab grade instrument with NIST traceable calibration is not required. You could probably buy a briefcase full of TinySAs for what it would cost to repair a blown 50 Ohm input on my Keysight analyzer.

  3. There is an update to the review by IMSA Guy that explains that the tests he performed were overloading the mixer front end. When a 20dBm antennuator or more is added so that the reference is below -30dBm, it performed favorably in all but RBW. In fairness, I would think that you should add the followup link.

  4. The conclusion that IMSA Guy reaches about the spurious indications was incorrect. He says that the fact that the signals other than the main one moving faster than the main one indicates these being mixer artifacts. That’s not the case. If a signal has harmonics, they WILL move faster than the fundamental. Twice as fast for the 2nd harmonic, three times as fast for the third, and so on.

    The tell-tale indications for internally generated spurs is that they move faster or slower than the expected rates for harmonics, or even moving in the opposite direction, or moving at the wrong rate as the center frequency is changed, i.e., if you change the CF by 1 MHz, any signal that does not move 1 MHz is a spur.

  5. Hi, is there a similar device, but with a proper BAS output ?
    I’d like to have a very basic spectrum analyer, but with a real screen.
    LCDs are cute, but don’t offer a lot of comfort. I’d rather use an old b/w TV set or green monitor.
    Their picture can be seen much better from the shelve in the shack.

    vy73s! Joshua

  6. The base specifications on this are remarkably close to the NanoVNA which makes me think that they’re roughly the same thing under the hood. The menus/UI are remarkably close to the NanoVNA as well which makes me think that there’s a lot of shared code under the hood.

    One of the best parts of the NanoVNA is that it’s released under GPL v3. This has allowed it to grow much beyond its initial limitations over time. This project looks to be an attempt to close off and monetize a bunch of work carried out by others.

    I’ll be voting with my wallet and waiting for access to the source code before I buy one of these. Please feel free to correct me if I’m wrong.

      1. I stand corrected. Thank you for the link. I should have looked harder prior to posting my previous comment. I love my NanoVNA and will be adding one of these to my collection.

        Thank you!

  7. I may be a bit biased, since I was a beta tester for the tinySA, but I have to say the tinySA is a bargain. Sure, it’s got its shortcomings (which have been plainly spelled out on the wiki), but for the price it’s outstanding. The firmware has features that require spending big bucks to add to my expensive spectrum analyzer. I love that the tinySA is portable and runs on batteries – I can take it anywhere with me! Great for tracking down power line noise and other interference. (I do need to wear my glasses to use it, but that’s true of the NanoVNA, too. And you can use it with the tinySA-Saver software and display it as big as you like.) The built in signal generator function is handy, too, and the output level is high enough that you can use it as the LO for common double balanced mixers. As a beta tester, I received two tinySAs, versions 0.2 and 0.3, and as soon as they were available for purchase, I bought one. One is on loan to the president of the local ham club, and he’s already planning to buy his own. For 50 bucks, this thing is sweet!

  8. I see the tinySA has the ability to store a sweep and use the stored sweep to “normalize” (invert and add) subsequent sweeps, this is needed If the tinySA is to be useful with a tracking generator as a Scalar Network Analyzer (SNA). However I am not seeing a sweep output ramp and/or a trigger out that would be nice to sync the tracking generator. Otherwise, using the built-in sweep trigger and span width adjust might suffice as a messy work-around.

  9. Too bad the tinySA’s frequency range stops at 960MHz, that’s right around where the output of a consumer satellite low-noise block downconverter (LNB) output starts. If tinySA covered 950 MHz through 1525 MHz, anyone on Earth wanting to point a free-to-air (FTA) satellite TV dish would seriously be thinking about buying a tinySA.

  10. This thing is obviously a NanoVNA hardware, the labels are painted a bit differently and it’s running an entire new software. Is it possible to combine the two softwares into one on a single device?

    1. While there are similarities, the tinySA is NOT NanoVNA hardware. The tinySA hardware has components that the NanoVNA doesn’t have, and the NanoVNA has components that the tinySA doesn’t have. As far as combining the two softwares, the short answer is no, as there is currently no hardware that could provide both the spectrum analyzer and VNA functions.

    1. Andreas Spiess did a review of that SA. It seems interesting, although huge scanning steps and other aspects make it less interesting than a “real” instrument. The price however is so low that there’s no reason to avoid it since it still can be really useful.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.