by Doug Jorgesen
In a recent tech note, we summarized 4 Ways to Implement a High Isolation Duplexer for a Transceiver in the specific case where the transmit and receive signals are separated in frequency, sometimes called frequency domain division (FDD). There are many electronic warfare and test and measurement applications that cannot separate the transmit and receive signals in frequency. In some cases the transmit and receive can be duplexed in time, with a switch. In the most challenging situations, the transmit and receive signals must share the same time and the same frequency, a situation known as simultaneous transmit and receive or STAR.
This is becoming an increasingly important problem, as recently recognized by the defense advanced research projects agency (DARPA) in their new wideband adaptive RF protection program (WARP). “Sometimes a system’s own transmitter is the biggest interferer to the receiver. To avoid this issue, transmitting and receiving at different frequencies has traditionally been commonplace, aided by the use of a frequency duplexer to keep the two bands separate. However, for defense systems there are a number of benefits to transmitting and receiving on the same frequency – such as doubling spectrum efficiency and increasing network throughput. This concept is referred to as same-frequency simultaneous transmit and receive (STAR),” said [DARPA program manager, Dr. Timothy] Hancock.
One of the main reasons this technique is not preferred is that STAR creates a major self jamming problem for the receiver. In a typical transceiver the received signal will be 80 dB or more below the transmitted signal. However, it is difficult to realize a STAR duplexer with 80 dB of isolation, meaning that the transmitted signal will be stronger in the receiver than the desired received signal.
Depending on the exact power ratios, the transmitted signal can be so high in power that it blinds the receiver completely during transmission by saturating the LNA or the mixer.
As mentioned in the previous note, one great way to mitigate this problem is by using Marki high isolation power combiners. These offer up to 35 dB of typical isolation from as low as 300 kHz to as high as 12 GHz. This can ease filtering requirements when signals are frequency duplexed, and it can reduce the self jamming effect in STAR systems. The drawback is the excess power loss and noise floor degradation caused by the insertion loss (6 dB nominally), but sometimes it must be accepted for STAR systems.
Another common way to alleviate this problem is using feedforward cancellation. In this scheme a small portion of the transmit signal is sampled off and fed into the receive path, but phase shifted by 180°. If the amplitude of the feedforward signal is identical to the amplitude of the transmitted signal but out of phase then vectorial signal cancellation can significantly reduce the transmitted signal seen by the receiver.
In a super heterodyne system this can be done at the IF frequency or at the RF frequency. While it is more challenging at the RF, this can prevent the LNA and mixer from being jammed by the transmit signal. In a direct sampling receiver performing cancellation at the RF frequency is the only option. Here is a block diagram view of how signal cancellation is performed:
Note that the transmit signal can be coupled off before or after the power amplifier and fed into the receive side before or after the low noise amplifier, depending on system requirements. Coupling off after the power amplifier will give a more accurate version of the transmitted signal (therefore better signal cancellation) at the expense of output power. Performing cancellation before the low noise amplifier will prevent jamming at the expense of degraded noise figure.
In either case the recommended device to tap the signal off is the recently released MC16-0222SM. This surface mount coupler will provide a 16 dB version of the coupled signal with the smallest loss possible in a surface mount form factor.
The next problem is the vector modulator. This device takes the transmitted signal as an input and performs a phase shift and amplitude modulation such that it matches the signal leaked into the receive path. A control loop is required to control the settings on the vector modulator to properly null the transmit signal.
For the device of the vector modulator itself there are two options:
- Use an IQ mixer such as the MMIQ-0218 as a vector modulator, as described in All About Mixers as Phase Modulators. The advantage of this technique is that it is easy to buy the IQ mixer as a single piece of hardware. The disadvantage is that the structure is optimized as a mixer, not a modulator. The double balanced mixer work as switches, but not as well as they could.
- Build a vector modulator using a surface mount quad hybrid, a surface mount power divider, and a switch. The quadrature hybrid is the really tough part, but Marki has already designed these with our MQH and MQS series of quadrature hybrids and splitters. In this application a splitter is totally sufficient. The switches can be variable attenuators of any type, digital or analog, or variable gain amplifiers.
Unless amplification is used the signal will be attenuated by the vector modulator. This is probably okay, since the loss of the vector modulator path (coupling loss on both sides plus the vector modulator) should be roughly equal to the isolation of the duplexer plus the path loss. If this is 50 dB then a 16 dB coupler on each side leave 8 dB of loss in the vector modulator.
As requirements become more stringent for electronic warfare and test and measurement applications, we expect to see more difficult STAR transceiver architectures in development. If you have any questions about this application, the most up to date components to use, or the possibility of custom component development for this application, please contact [email protected].