DL5NEG - A high-quality 2.3 GHz Amateur Radio Transceiver

A high quality 2.3GHz FM receiver

1. What does "high quality" mean?

Low phase noise (typ. 80dBc/Hz at 1kHz from the carrier)
High receiver sensitivity (1.5dB Overall Noise figure)
Very accurate, high dynamic S-Meter (>65dB dynamic, dBm-reading in LCD Display, see chart)
Small step-width for fine tuning if the other stations frequency does not fit (2.5kHz steps)
Wide tuning range (2.3 to 2.43 GHz, wider range possible with software change)
Menu user-interface like in a modern cellular telephone (using soft-keys)
Separate RX and Tx frequency with several coupling methodes
Large 2x16 characters LCD display with software controlled backlight

2. How does it look like?

As the saying goes: A picture says more then a thousand words...

3. The receiver concept - the block diagram

The receiver is a tripple superheterodyne receiver with two PLL-guided local oscillators. The first LO (2.1 GHz) can be set in steps of 500kHz and sets the RX signal down to 188 - 188.5 MHz. It has a high PLL loop bandwidth of around 20kHz, therefore gives low phase noise of below -80dBc/Hz at 1kHz from the carrier. (=> very high S/N-ratio on the audio output, providing the received signal is strong enough.

The second LO can be tuned from 177.3 to 177.9 MHz and sets the signal further down to 10.7 MHz. The VCO is very carefully design and exhibits -87dBc/Hz at 1kHz phase noise. Due to the low noise of the VCO itself a small loop bandwitdh of 70Hz can be used and therefore a step width of 2.5 kHz is possible.

FM-demodulation is performed at 455kHz using a NE614 baseband IC that has a very good RSSI output to drive the S-Meter (up to 80dB dynamics, giving 50mV per dB). High quality Murata IF filters are used for channel filtering. A LM386 audio amplifier is used to drive the loudspeaker.

The first mixer is a image-reject mixer (UAA2077), negleting the need for a narrow input-filter to attenuate the image frequency. This makes it possible to tune over a wide frequency range with constant sensitivity. While the UAA2077 is made for frequencys up to 2.0 GHz (the american PCS cellphone system), it still give 28dB image rejection at 2.3 GHz if a carefull layout and good grounding is used.

In front of the first mixer is a very low noise amplifier with a noise figure of 1.5dB and a gain of 22dB. It includes a mild band-pass-filter to make sure that the maximum gain is achived at 2.3GHz. Frequencys below a few hundred MHz are strongly attenuated to avoid overloading the next stages with strong VHF broadcast signal, cellphone signals, man made noise etc. The image frequency (1.9GHz) is attenuated about 3dB compared to the wanted frequency at 2.3 GHz. Together with the image reject mixer this gives an overall image rejection of 31dB.

Have a look at the block diagram...

4. The schematic diagrams and pictures of the inside

Unfortunately the development of all the different parts of the transceiver uses up most of my spare time right now. Therefore I really do not have the time to draw nice schematic diagrams with the computer. I took the freedom to draw them manually and scaned them in. The files are in GIF format and are therefore resonable small. The quality seems to be good enough to read everything.

Lets have look at the RF-modules first.

The low noise amplifier, Picture
The image-rejecting first mixer, Picture
The 2.1 GHz VCO and PLL system, Picture
The 176 MHz low phase noise VCO and PLL system, Picture, c/u VCO, c/u PLL-IC

Below you can find the schematics and pictures from the input of the 188 MHz 1. IF stage all the way to the loudspeaker. That means if you adjust the filter to 144 MHz and add an preamplifier this circuit would make a very fine 144 MHz receiver. The mixer works with low-side injection in my application (LO frequency is lower than RF frequency). Since the VCO works well down to 150 MHz you don't even have to change it at all, simply use high-side injection (LO = RF + IF, e.g. LO = 144.200 + 10.7 = 154.9 MHz).

5. The software

The firmware for the 68HC11 microcontroller is written in assembler and is very specific for the configuration that I used. It reads calibration data from an external I2C-EEPROM and cannot be used without all the mumbo-jumbo in the digital part. However it is quite easy to control the PLL circuits with a PC via the LPT1 port. I am using the PC program in my lab whenever I would like to try something without the digital part. This is the source code in C.
The program will work under DOS, Win95 and Win98. Since it directly programs the LPT1 it will not run under Windows NT.