Full solder and working well EU1KY V3 antenna analyzer frontend board:

The EU1KY antenna analyzer V3 is an open source project to build your own, reasonably cheap but very functional antenna analyzer that is a handful tool for tuning coax-fed shortwave ham radio HF/VHF, and CB antennas. Big color TFT LCD, full control using capacitive touchscreen, many features outperform many expensive antenna analyzers present on the market. Moreover, you have fun building this tool on your own and save some money.

Features

The device is actually a 1-port VNA, it allows measuring the parameters of any passive load connected to it in the range of 500 kHz ... 150 MHz (and even up to 450 MHz, depending on your hardware, see below), and displaying the results in graphic form. Additionally, it functions as TDR (Time Domain Reflectometer).

Open-Short-Load calibration is supported to compensate transmission line influence on measured parameters. SeeVersion 2 pagesto learn more.

All the settings, as well as calibration files, screenshots and Touchstone S1P files, are stored on SD card which can be accessed via the on-board USB HS port (it works as card reader).

ST-Link 2.1 debbugger built into the dev board provides virtual COM port that, when the device is connected to PC with Mini-USB cable, can be used to control device remotely usingRigExpert AntScope softwarewhen the device shows main menu window. The device emulates some features of the RigExpert AA-1000 antenna analyzer.

Si5351 revision B specs dated 2015 declare that the chip can output maximum frequency of200 MHzinstead of 160 MHz as they stated before. I am not sure that Si5351 chips manufactured before 2015 can output this frequency reliably in the entire operational temperature range, but, I hope, most of them can. Actually, the IC is designed so that 225 MHz is its maximum output, but the manufacturer limits it in the spec, as you can see. There must be a reason for this limitation.

Anyway, you can try using device up to 200 MHz on the first harmonic by enabling hidden parameter in the menu. But only if your Si5351a is able to output these frequencies reliably.

And, moreover, the device is able to workup to 450 MHz(using 3rd harmonics of both signal and LO above 150 MHz, or above 200 MHz if confgured as above). Although the IF signal is 20 dB weaker, this is not a big issue since there are no strong interfering signals induced in the UHF antennes, unlike on HF. But the performance of many 602/612 type mixers degrde significantly at the frequencies above 200..300 MHz. Probably, increasing supply voltage of the mixers (only) to 8 Volts could help, but I did not try. In order to extend the working frequencies up to 450 MHz, enable hidden params in configuration menu, and choose the upper bound. Note that above 150 (or 200 if configured so) MHz the measurements will appear noisy (the signal is 20..40 dB weaker). I recommend to minimize another hidden param (LIN ATTENUATION), set it to 0 in this case. But still some mixers, or your RF choke may limit the maximum usable frequency to a significantly lower bound. E.g. in my case the signal level drops rapidly at frequencies above 320 MHz. Another possible reason of this drop is the half-wave resonance of the coax shield length on the RF choke, make it not longer than 1/4 wavelengths at the highest frequency. Also note that full recalibration of the device (includng HW cal) is needed if you increase the maximum frequency, and the calibration runs much longer.

Some screenshots and photos

Moved to separate page:see theGallery

Hardware

Device is implemented usingSTM32F746G-DISCO Development Board, which is cheap but full of features needed for this project:

• fast but low power ARM Cortex M7 CPU core running at 216 MHz, with floating point hardware accelerator that boosts digital signal processing
• big and fast 480x272 4.3\" TFT LCD display, driven by controller built into the CPU and mapped directly to the on-board RAM
• capacitive touchscreen on the display - no buttons are needed
• excellent on-board audio codec with stereo line input, a great addition for signal processing
• on-board ST-Link V2.1 in-circuit debugging interface
• Arduino Uno compatible shield socket
• a lot of peripheral interfaces
• open-source code of the drivers for all peripherals and on-board devices, provided by ST Microelectronics. Tons of bugs in it, but it was very challenging to find and fix them.
• DSP math libraries provided by ARM, with FFT code highly optimized for speed

And, finally, there is no more zoo of various poor quality boards and displays that Chinese guys supply and that were used in version 2. This development board is designed and supported by ST Microelectronics.

You\'ll only need to build and add a small RF Frontend in the form of Arduino Uno similar shield, and make some additional interconnections.

You\'ll need also a Micro-SD card (4 GB is more than enough) where the board will store it\'s configuration files and screenshots, and a 5 V power supply (any cheap power bank used to autonomously charge smartphones can be used).

Development board should be modified a little to control the RF frontend via I2C: solder bridges SB5 and SB3 should be removed. Solder bridges SB1 and SB4 should be soldered in instead. All these bridges are near the Arduino Uno shield connector CN5 at the corner of the board, near the SD card slot, they are clearly marked. See the picture below.

RF frontend schematic

Yes, it is very simple and contains no parts that are difficult to find. Though it works significantly better than the RF frontend of Version 2.

Please read the\"Principle of operation\" articlefor version 2 of the device to understand how it works. Only the differences are described below.

The frequency synthesizer is built on a well proven Si5351a chip, like in version 2 of the device.

There are two channels in the device: Voltage channel (V) and current channel (I). The device actually measures magnitude and phase difference between these channels to calculate impedance of the load connected to the input.

There is a noticeable novelty in the measurement bridge schematic. It avoids mixer imbalance influence on magnitude measurements, i.e. the mixer no longer \"hangs\" with high level RF signal applied to both inputs. No matter what you do, the actual in-phase signal suppression is 25-30 dB in this case, this significantly limits the device precision. This new design introduces low Ohm measurement resistors connected to the ground. The inputs of the mixers are not exposed to high level in-phase signal when the load impedance is high in this case. The RF choke at input has a very high impedance and does not shunt low-Ohm resistor to significant extent. Any phase and magnitude change it brings is easily compensated by calibration. I did not see this idea used before in similar devices.

The RF choke is 3-4 turns of 4mm diameter 50-Ohm coax (I used RG-303) on a ferrite bead taken from the signal cable (D-SUB) of an old CRT monitor. The impedance of it is several orders higher than that of the current measurement resistor (5.1 Ohms used). But the length of the cable wound on the core, from the bridge to the grounding point, must not exceed 1/4 wavwlength for planned device\'s maximum operating frequency in order to avoid shunting the measurement resistor due to resonance. But this can aggravate performance on the lowest operational frequency if you plan to use device up to 450 MHz. It is up to you to find the best core to satisfy these requirements. Please note that the shield of the coaxmust be grounded to PCBright behind the choke, see the schematic. Also take into account that since the RF choke is included into the bridge, it should be properly mechanically fixed, otherwise the device calibration may be severely disturbed if the choke is moved, especially on VHF frequencies.

Two mixers (cheap and widely available 612 type) convert the I and V signals to the intermediate frequency of 10031 Hz which is then passed to Low pass filters.Note that these mixers don\'t divide the LO frequency by two like in version 2.And the low-pass filters are now used instead of bandpass, this improves calibration stability in wide range of ambient temperatures because phase vs frequency dependence is much steeper in BPF than in LPF.

OUT_VI is connected to theleft channelof linear audio input (blue 3.5 mm connector), OUT_VV to theright channel(at the same blue connector) of the Discovery board. The rest is handled mathematically by the firmware.

Parts:

• All SMD resistors: 0805 1%
• All SMD capacitors: 0805, >=16V
• Electrolytic capacitors: V >= 16V
• Beads: any, 0805
• SA612 (can be replaced with SA602, NE612, NE602): in SO-8 package
• MC4558: in SO-8 package. Can be replaced with any opamp in the same package with the same pin assignment, that works from 4.5V unipolar voltage, and with GBW product > 3 MHz.
• Si5351a: 10-MSOP package
• Crystal: any 27000 kHz fundamental frequency quartz

RF frontend PCB

The prototype PCB design in Sprint Layout 6.0 format can be downloaded in theDownloadssection, the assembly drawing is also there. There is also an edited version by YL2GL.

The PCB is double-sided. The other side (not shown in the file) is not etched and is used as ground plane. Some vias should be connected to it, but those vias that should not be grounded (e.g., unused pins of the shield connector) must be countersunk at the ground plane to prevent short circuits.

---