Dual RF Head sensor - Analog
1 MHz to 10 GHz power sensor/SWR

I needed something to measure low RF levels, primarily from directional couplers etc. This created a need for this "Dual RF Head" sensor (and some other products, base on the same frontend, see below).

The two RF inputs can measure from around -50 to +10 dBm with great linearity.

Connected directly to a power source, -50 dBm is the same as 10 Nanowatts! Connected to a directional coupler, you can measure several Kilowatts with the Dual RF Heads, all with a dynamic range of 60 dB (1:1000000)!

If you need a directional coupler for the Dual RF Heads, take a look here.

Check also my REPAM (Remote PA Monitor) product

Currently I have three different "Dual RF Head" devices:

Dual RF Head - Analog - Gives out analog voltages dependent on the RF input power
Dual RF Head - USB - Connects via USB-C to PC (virtual serial port), Windows application w. source available
Dual RF Head - M5 - Attaches to a Tab5/CoreS3 touchscreen, used as a standalone power meter (or built in)

I am able to deliver finished Dual RF Head devices for €135,- each plus shipping. Send me an email if interested, contact info on the CV/Contact page.

Devices delivered will be calibrated at a (by you defined) specific band/frequency (I can only do up to up to 3 GHz).
An example of a calibration sheet is shown below. The calibration sheet (an Excel file) contains some calculators that will allow you to experiment with measured values etc.

Background

I based the design on a dual logarithmic sensor chip, the ADL5519 from Analog Devices. The chip is not exactly cheap, but it has some very nice features, primarily:

Wide bandwidth: 1 MHz to 8 GHz (useable to 10 GHz)
Dual-channel and channel difference output ports
Integrated accurate scaled temperature sensor
62 dB dynamic range (±3 dB)
>50 dB with ±1 dB up to 8 GHz

Datasheet for ADL5519

I added a 10 dB attenuator on both inputs, this ensures a better return loss on the two inputs. I also added two opamps to give a little bit of gain to the output signal. The ADL5519 delivers approx 1.7V with no RF input, the opamps scales this to around 4.25V (multiplies the output from ADL5519 with 2.5).

As the sensor will be connected to my "REPAM" device that has 0..5V analog inputs, the addition of the opamps made sense.

Block schematic of the interconnections between the various components in a typical PA.

The overcurrent switch can be found here.

The REPAM device can be found here.

Installation of two Dual RF Head - Analog modules, REPAM device, Overcurrent module, directional coupler, High power 90 degree hybrid, low power 90 degree hybrid, 150W 23cm driver PA and 1500W attenuator in my dual 23cm PA for my 4.8 meter dish project:

Norbert, OE3NFC, installation in his water cooled dual module 144MHz power amplifier using a Dual RF Head - Analog module, REPAM device, Overcurrent module and a directional coupler:

The design

The PCB board is made on double sided FR4, 0.8 mm thick. I did not strive for anything spectacular with regards to RF precision, the sensor will be calibrated for whatever frequency range it will be used at anyway.

The sensor is supplied with 8 to 14V DC (nominal 12V, absolute maximum 15V) and consumes around 60 mA.

There are 4 output signals, OUT A/B, these are the logarithmic outputs from the two channels, A and B. The output is close to -55 mV/dB. No RF in will give approx 4.25V output, +10 dBm will give approx 1.6V.

The OUT P/N are outputs directly from the ADL5519, these are the sum and difference outputs. You can use one of these for detecting high SWR (or you can calculate the SWR from the forward and reflected power as I do in the REPAM device).

The REPAM box has settings for calibration factors etc. for the sensor, it will also monitor forward, reflected and calculate SWR, and enable alarms, setting of relay outputs etc. when thresholds are exceeded.

Box design

I designed a small aluminum housing for the board, this makes the board somewhat RF "tight" (not completely true because of the plugin screw connector).

I used JLCPCB's 3D milling service to make the box (total cost €18,- for a single one, if ordering 5 pcs, price each is only €8.80 !). The surface is conductive anodizing. All screws are M2.5, the box also holds two SMA female bulkhead connectors.

Frontplate

Many times I design frontplates in PCB material. This has a number of obvious benefits. For the Dual RF Head sensor I made the frontplate in 0.8mm PCB material, black with white silkscreen.

First (small) batch

I made a small batch (6 devices). The price (at JLCPCB) for the Anodized alu box was €8.80 when buying 5 pcs at once! Devices will be tested and calibrated, I expect the results to be in line with the prototype.

Returnloss.

First the returnloss is measured for both port A and port B on the prototype. As my VNA (R&S ZNLE3) only covers up to 3 GHz, I did not measure at higher frequencies than that

RL for Channel A:

RL for Channel B:

Smith-chart for channel A:

Smith-chart for channel B:

Uncalibrated output voltage versus frequency

Below is a graph showing the mV reading on IN A for both -10 dBm and -40 dBm input levels. Measurements are taken at:

1, 10, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500, 2600, 2700, 2800, 2900 and 3000 MHz.

The error seen here will be removed once you do a calibration of the device for the frequency (band) you will be using the Dual RF Head on.

Frequency sweep, 100 MHz to 3 GHz

Below is a sweep taken from 100 MHz to 3 GHz for 10 different power levels (from -40 to +5 dBm). This is done in increments of 100 MHz.

The X axis is frequency in MHz, Y axis is voltage in mV (approx 55 mV/dB) and the different traces are for the different power levels from -40 to +5 dBm in increments of 5 dB.

Below the graph are the raw data.

Frequency sweep, 1 MHz to 96 MHz

Below is a sweep taken from 1 MHz to 96 MHz for 10 different power levels (from -40 to +5 dBm). This is done in increments of 5 MHz.