This page was last changed on 21.04.2005 |
A couple of few years ago I had build a diode-sensor based RF powermeter using a 68HC11 microprocessor. Even before that I had build an analog RF powermeter with thermal power sensor. And I found datasheets of logarithmic amplifiers very interesting, that promised a measurement bandwidth of up to 2.5GHz. When I had familiarized myself with the Atmel AVR microprocessor family, it was just a logic step to build myself a nice powermeter for all three measurement principles.
So I did that and build a very small, battery powered powermeter with external power sensors. I called it UMD, the Universal Measurement Device. It was completely done with tiny surface mount components. Wanna see? Click here for a picture. Or here for another picture. Here the UMD is measuring a beacon transmitter with the diode sensor.
Unfortunately the device contained a few components that were very had to get (e.g. a mobile phone LCD display) and was so tiny that it was difficult to compare it with my documentation. I had changed the circuit a couple of times during the development, so I could not guarantee that my schematics were correct. When I was asked by some Belgian radio amateurs to write an article for their annual magazine, I decided to rebuild the circuit on a larger board with wired components and a standard 4x16 character LCD display. Lets call that version article-breadboard version.
The breadboard was laying around for a few month until I decided that I should rather use it to build myself another power meter. Since I already had the UMD with the external sensors, I have included the sensors now in the meter housing. I had almost never used the log-amp sensor, so I went for a design with only two sensors: a diode sensor and a thermal power sensor. Since it was kind of a third approach I have named this version DSPM03, the Dual Sensor Power Meter 03. If you want to copy one of my power meters, I strongly recommend to copy the DSPM03.
If you are interested in the older design, feel free to have a look at the unsupported files. For the hardware at hardware/ and for the firmware at firmware/.
For all that want to follow my advice to copy the DSPM 03, here is the information you need...
How to rebuild the DSPM 03
1.
Electronic hardware
The Powermeter consists
of two logic components. One is the digital part, consisting of the processor
(including the ADCs), the display and the batteries. The other component
is one of different sensors that transform the RF input power into
an analog DC level that can be read by the ADCs in the digital part. The
software knows the transfer function from RF power to DC level of the different
sensors and is therefore able to calculate the RF input power from the
read DC level. The result of the calculation is then written to the LCD.
In a special mode the software also displays the "raw ADC value", i.e.
the DC voltage level without transformation to RF power. Click
here to see a picture of the open case.
Power
Sensors
The theory on the
sensors, their advantages and disadvantages and which sensor suits which
application was worth an separate page, click
here to read it.
The sensors that you need for rebuilding the DSPM03 are:
Diode Sensor: How to build a diode sensor that converts an RF signal to a DC voltage is extensively discussed on a separate page, just click here to view it. The frequency response (i.e. the unwanted change in behavior over frequency) is really low, it is below 0.5dB up to more than 3.5GHz. So the accuracy of the complete system depends mainly on the quality of the lookup table in the software, which holds one pair of values (DC voltage vs. RF input power) for every dB of input power.
Between the passive diode sensor and the ADC inputs you need a little bit of OpAmp mumbo-jumbo. Here is the schematic of the circuit that you need.
Thermal
Power Sensor: Here you find the
schematic of the thermal power sensor that I have build for the DSPM03. Soon
I will add an extra page that gives a lot of background information on thermal
sensors and the do's and don'ts for building them.
If you have build the sensor, you will notice that the NTCs are never exactly of
the same value. In my circuit the lower NTC was of a bit lower resistance. To
compensate for that, I have included the 4.7kOhm potentiometer named "base
adjustment". You may have to put a poti in series to the other resistors if
that one is lower in your case. If you need help for doing the adjustment just
email me and I will be glad to assist you.
Digital part
There are two schematic
diagrams. First there is the DC supply schematic. The main circuitry
runs on +5Volts. However, some of the sensor run on 3.2 Volts and also
the reference voltage for the ADCs is 3.2 Volts. Therefore we need
an +5V regulator and one for 3.2 Volts. Since 3.2 Volt regulators are a
bit tricky to get, I have decided to build it myself with an opamp and
a NPN transistor. If you have a 3.2 Volt regulator available, please feel
free to use it. A +6V regulator is used to generate the input for a voltage
inverter (see main schematic).
As power source I have used 8 NiCd cells,
but any voltage from 8 to 14 Volts is ok. So if you want to use
a DC power supply feel free to do so (after all, its you circuit!). There
is a voltage divider from VBat to the input of ADC No. 3. It divides by
four, so the maximum supply voltage that can be measured with a 3.2V referenced
ADC is 12.8 Volts. Due to the high series resistor, the ADC will not be
damaged if you use up to 14 Volts, just the displayed battery voltage will
not go beyond 12.8 Volts.
The main schematic diagram consists of the microprocessor, the connection to the LCD, the internal connector to the power sensors and the negative voltage circuit that sets the LCD contrast. A 4x16 standard LCD needs a typical contrast voltage of -0.35 Volts for best readability. If you skip that and tie Vcontrast to ground, the LCD will still work well, just the contrast is a bit week. Since the amplifiers for the sensor signals require a negative voltage, we have to produce the -5Volts anyhow, so why not use it for a good LCD contrast as well.
The microprocessor in the schematic is a AT90S4433 which is discontinued, from Software Version 2.0 onwards the ATmega8 microprocessor can and must be used.
And here you can download the schematic of the main digital part.
2.
Software
The firmware is written
in C using the free WinAVR GNU C compiler, downloaded from the AVR-freaks homepage. If you
want to edit the code and recompile it, you can download the compiler together
with the simulator (AVR-Studio) at www.avrfreaks.net.
There you will also find links to numerous free and low-cost programmers
to load the firmware from your PC to the microprocessor.
The latest version for the AT90S4433 is 1.4 which additionally support to set a fixed pre-attenuation that is added to all measured results. This is very helpful if you have an attenuator or an directional coupler at the input of your meter to measure higher input powers. Also to compensate for the loss of your cable between signal source and sensor this is helpful. And since even negative pre-attenuation values are possible, you can also compensate for amplifiers that you use to measure very tiny signals below the sensitivity of the meter.
Also the new version 2.0 for the ATmega8 microprocessor is available in source and hex format now. The feature set is identical to version 1.4, which does not work on the ATmega8.
Version 1.4 for AT90S4433
Download the firmware here: dspm03.hex
Download the C source code here: dspm03.c
Version 2.0 for ATmega8
Download the firmware here: dspm03 (Version2.0).c
Download the C source code here: dspm03 (Version2.0).hex
Getting-Started Manual
After power-on you will see a greeting message for three seconds. After that, there are five different screens, which can be switched through by pressing button No. 1.
Button No. 2 lets you choose the power range (or switch to auto range when you use the diode sensor, which is default ).
With button No. 3 you can reset the relative power reading. The input power at that moment is used for later relative readings. Very useful if you want to know how changes in your setup affect the output power. E.g. longer or different cable, changed biasing values etc.
Button No. 4 is triggering the automatic zero adjust
for the thermo sensor. It is only functioning when thermal sensor data are
shown. A "Z" in the upper right corner of the LCD shows that automatic
zeroing is just on. Automatic zeroing will automatically switch itself off, as
soon as the zero point is reached. Please not that the internal zero point for
thermal sensors 1 and 2 is different, i.e. if you switch from on sensor to the
other, you have to trigger the zeroing again.