Bonkventures‎ > ‎

Thales 38GHz

This page exists to document my adventures with the TGTR38 38GHz microwave transceiver made by Thales. 
My main motivation for this is using the TX side as a high level LO for a f/2 harmonic mixer in a 76GHz transverter.
DB6NT mixers in particular look tempting due to their wr28 waveguide input for LO. 
The transmit chain produces about 300mW output power between 36GHz and 39GHz, so when fed with an Elcom synthesizer the setup should be fairly frequency agile and cover most of the 76GHz band. 
This is a valuable bit of future proofing as there exists a real danger that hams will be banished from the car radar part of the band. 

I did some measurements on the transceiver recently with these results. 
The LO input has reasonable return loss from 10.2GHz to 12.8GHz, a perfect fit for a Elcom synth. 
The transceiver has an internal bandpass filter in the IF path so measuring the return loss was an easy way to see the usable band. 
The same procedure was used for the TX IF path. 

The TX IF had dips in it's return loss on 2.4GHz and 4GHz. During testing we drove it at 3.6GHz without problems. 

The transceiver was powered up and transmit power measured. 
The measuring setup had a 30dB wr28 attenuator right at the TX output, it was then connected to a HP 8564E input with some flexible wr28 waveguide, a wr28 to 2.92mm 'K' transition and then with the required amount of adapters to fit that to the spectrum analyzer APC-2.4 input. 

A HP 86364B signal generator was used as the LO and Advantest R3763B VNA was set for 0Hz span and used as the TX IF generator.
LO port was driven with +8dBm as further increase in power did not result in more power output. 
TX IF port was driven with -10dBm at 3.6GHz, higher input did not result in more output power. 

+25dBm power output was measured from 36.4-39.6GHz. This should make this a nice LO for 76GHz experiments. 

Some pictures of the measurement process: 

The original wire colors for the transceiver DC power plug are:
Red for +8V TX control
White for -12V
Black for ground
Violet for +5V

Inside both the TX IF and RX IF converters are M/A-Com 2.5GHz to 6GHz mixers, two hairpin filters and LMX2325 controlled LO.
The AGC board was boring inside, just some MMIC amplifiers and some "diplex filter" inside. The AGC board might have done the baseband TX and RX IF separation in the ODU. 

One problem for 76GHz LO use is the requirement for two LO synthesizers. One for feeding the TX IF with a carrier and one for the transceiver LO.
The internal construction can be modified if wanted, but most of the SHF side is done with wire bonding so modifications are non-trivial.

However I recently read about modifying the TX IF mixer in a Thales transceiver in to a voltage varied attenuator.
One more or less ignores the original TX IF path and wires it via a resistor to the mixer IF port to apply DC bias to the mixer diodes.
Just DC biasing the mixer "on" turns the TX chain in to a tripler, allowing us to skip adding one 2-4GHz band synthesizer.
Biasing the TX chain "on" and varying LO drive with this voltage, without external variable 11GHz attenuators is nice.
Variable LO drive allows us to have different LO drive levels to the final mixer on transmit and receive.
This makes it possible to optimize for minimum receive noise figure and conversion loss and maximum output power, which usually don't happen with the same level of LO drive.