Differences between revisions 47 and 48
Revision 47 as of 2021-04-21 15:43:14
Size: 24138
Editor: RudolfReuter
Comment: added other software
Revision 48 as of 2021-04-21 17:28:17
Size: 24169
Editor: RudolfReuter
Comment: added VMA Vitor
Deletions are marked like this. Additions are marked like this.
Line 227: Line 227:
The '''scren shot''' above shows a '''scan''' with the the '''tracking generator''' without calibration. I just wanted to see if it works. The '''scren shot''' above shows a '''scan''' with the the '''tracking generator''' (30 dB attenuator added) '''without calibration'''. I just wanted to see if it works.

/!\ Vorläufig / Preliminary, work in progress /!\

Translate (Übersetzte) with Google

/!\ If you mouse click on a picture it will be expanded. You can go back to the web page with the back arrow of your web browser.

Spectrum Analyzer LTDZ

  • Update 2021-04-11: Today there is a much better Spectrum Analyzer tinySA for about $50 on the market.
    OK, the bandwidt is only up to 960 MHz, but it comes in a housing, has an on board battery, and a 320x240 pixel color LCD, for portable use.

  • Below is a photo showing the spectrum of the FM Radio band.

  • The little antenna is part of the package, also 2 SMA plug coax cables.

  • Screen shots could be made with an USB connection to a PC.

  • You can work with it just out of the box.



A new NWT (NetWork Tester, scalar) clone shows up on https://www.banggood.com (search for ID: 1450129) for about 44 EUR, shipped from UK, within Europe.

You can have it also with housing, search for ID: 1494125.

I like to collect and document all useful informations about that Spectrum Analyzer.

I am using the program WinNWT4 version 4.11.09 under MacOS Mojave 10.14.5 with emulator wine.


  • Interface: micro USB (CH341 chip)
  • Power Supply: USB
    • Standby current: < 100 mA

    • Sweep current: < 350 mA

  • Spectrum Analyzer

    • Direct conversion with ADF4351 oscillator and mixer IAM81008, Bandwidth 120 (200) kHz
    • Sweep bandwidth: 35 - 4400 MHz
    • Sweep step: 33~68.75 MHz/125 Hz, 68.75~137.5 MHz/250 Hz, 137.5~275 MHz/500 Hz, 275~550 MHz/1 kHz, 550~1100 MHz/2 kHz, 1100~2200 MHz/4 kHz, 2200~4400 MHz/8 kHz
    • Sweep speed: ≥800 points / sec
    • Sweep dynamic log ratio: > 50 dB

    • Input detection: ≤10 dbm - on PCB: < +5 dBm

    • Background noise: ≈ -60 dBm
  • Tracking Generator with ADF4351

    • Point frequency or sweep output power: ≈ 0 dBm (measured: +2 dBm = 0.8 Vpp @ 40 MHz)
  • Software in archiv 20190421205201LTDZ35-4400M.rar, needs terminal program unrar version >= 5.5 to unpack

    • Computer operating system: winxp, win7, win10, MacOS with wine emulator
    • Provide software (English language) and instructions for use
    • Schematic
    • Program WinNWT4, version 4.09 (actual version 4.11.09)
  • Dimensions: 48 x 90 x 15 mm
  • Push button switch: To toggle Tracking Generator ON / OFF
  • LED's
    • Red: USB traffic
    • Blue: Tracking Generator active
    • Green: ADF4351, Lock Detect, Tracking Generator
    • Green: ADF4351, Lock Detect, Clock for Mixer


On the right side you see a picture of the board.

If you click on a picture it will be expanded. You can go back to the web page with the back arrow of your web browser.

The 2 green LED's on the right side are covered with some tape to dim the light intensity. This is useful if you want to measure at points nearby.

At the upper side are 4 holes for a pin header to flash the CPU STM32F103C8T6 with a new firmware (actual version 1.19). Pin 1 is marked with a square pad.

  • Pin 1 SWData CPU pin 34
  • Pin 2 SWClock CPU pin 37
  • Pin 3 GND
  • Pin 4 3V3

On the lower left edge of the PCB I added at the voltage regulator a capacitor 100uF 16 V to have less ripple on the USB 5 V supply.



In order to calibrate the spectrum analyzer, you need attenuators. Preferably good to frequencies up to 4 GHz. The best is to have the steps: 10 - 20 - 30 - 40 - 50 - 60 dB for linearity measurement.

The other thing is a SWR bridge (Standing Wave Ratio) for the frequency range up to 4 GHz.

As a connector SMA 50 Ohm is preferred.


In order to calibrate the spectrum analyzer, you need two attenuators.

  • Attenuator 10 dB
  • Attenuator 40 dB

A good offer is from company https://www.banggood.com ID: 1119478 (about 8 EUR), see the picture on the right. Two of them are needed.

  • Frequency range: DC - 4.0 GHz
  • Maximum power: 23 dBm (200 mW)
  • Standing wave ratio: ≤ 1.20
  • System impedance: 50 Ω
  • RF interface: SMA double female head (Outer screw inner hole)
  • Attenuation:

    • Direct: 0 dB (reference)
    • Level 1: 10 ± 0.8 dB
    • Level 2: 20 ± 1.1 dB
    • Level 3: 30 ± 1.5 dB


SWR bridge

For antenna measurement you need a SWR-bridge.

A good offer is from company https://www.banggood.com ID: 1497170 (about 9 EUR), see the picture on the right.

  • Working Frequency: 0.1 - 3000 MHz
  • Return Loss: more than 35 dB
  • Connectors: SMA
  • Size of the board: approx. 45 * 90 mm
  • With this RF SWR Bridge board, you can measure in thru-mode filters, amplifiers and in reflective mode antennas, impedances, inductance, capacitance and impedance matching etc.
  • IN: Tracking generator with 6 dB attenuator in between, for better performance.

  • DUT: Device Under Test.

  • REF: 50 Ohm

  • OUT: Spectrum analyzer with 6 dB attenuator in between, for better performance.


SMA cable

I am using 50 cm long flexible cable RG316 with SMA male connectors on both sides, for about 2 EUR (Ebay).

See the picture on the right side, which shows all mentioned coax cables.

For high frequencies I am using 50 cm long semi-rigid cable RG402 with SMA male connectors on both sides, for about 2.5 EUR (Ebay).

For the SMA plug attenuation I used this reference.

For dBm to Volt conversion see Links #1.


Attenuation (dB/100 m)

Velocity of propagation

Frequency [MHz]




























l = 0.7 m



















l = 0.5 m







semi rigid





semi rigid








l = 0.5 m







semi rigid




l = 0.5 m


Ecoflex 15


















30 cm













There are at least 3 different programs who support also the Banggood LTDZ spectrum analyzer.


Banggood supplies the software WinNWT4 version 4.09 from Andreas DL4JAL. Unfortunately it is an outdated version. The actual version 4.11.09 is no longer provided for download at the web page of Andreas DL4JAL, but you can find it in the internet.

  • MacOS: It work with wine under MacOS Mojave.

  • VirtualBox, Windows 10-64: OK

  • Windows 10-64: OK
  • Unfortunately it does not have any sweep function.

I use this Windows(R) software under MacOS Mojave 10.14.5 with the wine compatibility layer program. It is also available for MacOS 64 Bit, wine version 4.01.

There is a setup description for the serial and parallel ports.

In my case the setup for the WinNWT4 USB serial port was the following:

# Install the USB serial port driver for MacOS: 
# search in the internet for: CH34x_Install_V1.3 and install.

# Add a Windows(R) link in the registry.
# Start wine in the Terminal:
$ /Applications/Wine\ Stable.app/Contents/MacOS/wine
# A terminal window shows up

# Check for the serial port name:
$ ls /dev/cu.wch*

# Start the Windows(R) Registration Editor
$ wine regedit
# navigate to: HKEY_LOCAL_MACHINE\Software\Wine\Ports
# Edit->New->string
    COM1 /dev/cu.SLAB_USBtoUART      # for FA-VA5 analyzer
    COM2 /dev/cu.wchusbserial144140  # for LTDZ Spectrum analyzer 35-4400 MHz

# WinNWT4.exe location:
$ cd ~/.wine/drive_c/Program\ Files/AFU/WinNWT4/

# WinNWT4 data path:
$ cd  ~/.wine/drive_c/users/name/hfm9/

There is also a Linux version LinNWT available.




The program SNASHARP from David F4HTQ can also be used for the Banggood LTDZ spectrum analyzer.

The scren shot above shows a scan with the the tracking generator without calibration. I just wanted to see if it works.

  • MacOS: It does not work with wine under MacOS Mojave.

  • VirtualBox, Windows 10-64: It works, see the screen shot above.

  • Windows 10-64: OK
  • Unfortunately it does not have any sweep function.

VMA Vitor



The program VMA Simple Spectrum Analyser from Vitor CT2JSA can also be used for the Banggood LTDZ spectrum analyzer.

The scren shot above shows a scan with the the tracking generator (30 dB attenuator added) without calibration. I just wanted to see if it works.

  • MacOS: It does not work with wine under MacOS Mojave.

  • VirtualBox, Windows 10-64: Program starts, but no connection to the COM port.

  • Windows 10-64: OK
  • Fortunately it does have sweep function, which can be used for sweeping the tinySA.

Below you see which VMA parameters must be set for a sweep on the tinySA. Because the sweep is asynchrones, the Step width and the Pause time must be set at least double the scan time of the tinySA.

Because the VMA Pause maximum is 1000 ms, in case of a tinySA scan time of more than 500 ms, that can be compensated with lowering the VMA Step size accordingly.




On the right side you see a screen shot of a tinySA scan from 35 MHz (LTDZ minimum) to 350 MHz (tinySA LOW maximum).

The LTDZ tracking generator has an output level of about 0 dBm, and is connected to the tinySA LOW input with an 30 dB attenuator.

In order to get an envelop curve, you need to set on the tinySA the parameter: DISPLAY->CALC->MAX HOLD.

The automatic RBW (Raw Band Width) is this case where 621 kHz.






Hardware Improvements

Bad soldering

I have read at the banggood customer review of the LTDZ board that mrardj reflowed some soldering points around the AD8307 log amplifier to make it work.

I also had a contact problem at the Tracking Generator ADF4351, no output signal. With a lot of resin flux at the 32 pins of the ADF4351 and reflowing the solder points with a fine Hako 900M-T-1C soldering tip I could make it work.

See the picture on the right.

100 uF at 5 V

In order to stabilize the +5 V supply from the USB connection like on the D6 board, I added a SMD 100 uF 16 V capacitor at the voltage regulator AMS1117_3V3 Input to GND.


120 (200) kHz filter

The 120 (200) kHz low pass filter between mixer IAM81008 and log amplifier AD8307 is not matched to the higher input impedance of 1100 Ohm of the log amplifier. David F4HTQ has redesigned it to match better.

The following changes need to be done, with the part numbers of the LTDZ:

  • L7: 470 uH -> 100 uH

  • C59: 1.5 nF -> 10 nF

  • C53: 1 nF -> remove

  • R18: 820R -> remove

  • C60: 680 pF -> 2.2 nF

  • C56: 1 nF -> 10 nF

The result can be seen in the diagram on the right. The output level is now more than factor 2 higher than the input level.


Input resistance

David F4HTQ describes on his web page some hardware improvements for the D6 Spectrum Analyzer, which is very similar to to the LTDZ Spectrum Analyzer when you compare the schematics.

A simple thing is the input impedance of the mixer IAM81008 (HP, SMD marking M810). It is already matched from the chip manufacturer to 50 Ohm, as written on page 13 of the IAM-8 Series Active Mixers Application note S013, "The RF input port of an IAM-8 mixer is matched on chip to 50 Ω, eliminating the need for any external matching circuitry.".

So, an additional 50 Ohm resistor R15 to GND is probably not necessary. In the schematic of the D6 SA the same resistor R16 has a value of 510 Ohm.

As a test I removed the resistor R15 (50 Ohm) and measured the mixer input impedance with the FA-VA5 see Links #4, direct connected (no cable used). See the diagram on the right.

I could measure just up to 600 MHz. The input capacity C46 and C46 (5 nF) have at 35 MHz an impedance of about 1 Ohm, see Links #3. That does not explain, why the input |Z| is so much increasing below 35 MHz.


GND Test Point

Unfortunately under the 4 screws the ground area is covered with silk screen (also the lower side), so there is no GND contact. In order to get a GND Test Point for the scope probe, I removed the silk screen with a knife below one screw and soldered a soldering eyelet in place, see the picture on the right.

The picture also shows the modified 120 kHz Low Pass Filter in detail.






After calibration with 40 dB and 0 dB attenuation, the linearity was checked with attenuations of 0 dB, 10 dB, 20 dB, 30 dB, 40 dB and 50 dB, see the diagram on the right.

In order to not overdrive the mixer always an additional 10 dB attenuation was inserted.

It is just a 2 point calibration.

If you want to have more accuracy, it is better to just calibrate that frequency span you are interested in.

The noise floor is at about -55 dB.

The absolute Input dBm values are:

  • 0 dB = -15.3 dBm = 0.1 Vpp
  • -50 dB = -66.0 dBm = 0.3 mVpp
  • -55 dB = -71 dBm = 0.18 mVpp








The next diagram (Figure 9) on the right shows the linear curve Vout versus Input Level (dBm) at various voltages at the Input INT of the Log Amp AD8307.

Use the lower curve, because Input INT is not connected.

If you calculate with a slope of 25 mV/dB the output value does not fit. Because of R17 & R16 (48 KOhm) the slope is changed (see chapter SLOPE AND INTERCEPT ADJUSTMENTS of the data sheet) to about 21 mV/dB. Then it fits, that an input voltage of 2.0 Vpp = 10.0 dBm gives an output voltage of 1.94 V, which gives a Slope value of 1940 mV / (82.9 + 10)dB = 20.9 mV / dB.

With an open Mixer Input the output voltage of the AD8307 is 0.80 V = -82.9 + 38.3 (800 / 20.9) = -76.6 dBm.

There is a gap of 76.6 + 71 = 5.6 dB, which is caused by the attenuators (2 x +/- 1.5 dB), cables (3 x 0.8 = 2.4 dB) and DSO tolerance.



Spectrum Example

As an example, I used a 40 cm long wire soldered to a SMA plug as an antenna, in order to receive some FM radio stations.

  • Span 3 MHz
  • Step width 1.5 kHz

I looked at the whole FM band for strong stations, and selected the strongest ones in a 3 MHz span.

The strongest signal was just about 20 dB above noise floor.











The lower diagram has the input from an 10 MHz Square Wave oscillator, for details of the oscillator see RFCalibratorSineSquareWave. You see that the duty cycle is not 50%, so you have also even harmonics. The Tracking Generator was switched off while the measurement.



It surprised me, to see some spectrum lines, which do not fit into the harmonics raster. So, I documented the frequencies and level, in the hope someone will find an explanation for it.

My explanation up to now is, that the even harmonics are modulated. The main side bands are symmetric 3.3 MHz off the 40 MHz harmonic. The lower side bands are symmetric 2.0 MHz away. Where does it come from?

It is eye-catching, that 3.3 MHz is 1/3 of 10 MHz, and 2 MHz is 1/5 of 10 MHz.

In the calculated spectrum of my Hantek DSO5072 I could not see those 3.3 MHz side bands, see SquareWave_Oscillator.2C_harmonics.

WinNWT4 does have 5 cursors only, so I stopped here.



In the mixer IAM-81008 data sheet I have seen for the RF Input Power an Absolute Maximum of +14 dBm, but the Test Conditions tell about -20 dBm. So I reduced the Mixer Input Power to -20 dBm with an 10 dB Attenuator. But I can see only a 10 dB lower spectrum, so I assume -10 dBm is not an overdrive of the mixer input.



When you look at the 60 MHz harmonic, you see also the two 3.3 MHz side bands, cursor 1 and 2.


  1. Online dBm - Volt converter

  2. Comment from Vitor CT2JSA to this Spectrum Analyzer

  3. Geekcreit® Spectrum Analyzer USB LTDZ 35-4400M is now supported!, vitor

  4. Capacitor Impedance Calculator

  5. VA5 Antenna Analyzer

  6. Comment from Vitor CT2JSA to the Spectrum Analyzer D6

  7. Some considerations about the limitations of the ADF4350/ADF4351/MAX2870, vitor

  8. Firmware source for the Spectrum Analyzer D6 from joseluu

  9. LTDZ_35-4400M application information from DJ7OO

  10. A good video from Andreas Spiess explaining Spectrum Analysis and the LTDZ

  11. dBm, Volt, Watt table

up to top

Contact Email, please enter your Email address

/!\ The entered Email address will not be published, or given away.

Comment this page
Are you human?

List of pages in this category:

-- RudolfReuter 2019-06-30 16:29:28

Go back to CategoryAmateurRadio or FrontPage

SpectrumAnalyzer_LTDZ (last edited 2021-04-21 17:28:17 by RudolfReuter)