-> Projects ->ATF54143 Maximum RF input power

ATF54143 Maximum RF input power


During (contest) operation we used for the 70cm band some pre-amplifiers based on the design of oz1pif (OZ1PIF, 2009). As an active element this design uses a Agilent ATF-54143 Low Noise Enhancement Mode Pseudomorphic HEMT. Pre-amplifiers were mounted near the antenna and switched by means of some RF coaxrelais. Although performance is very good a number of HEMTís where destroyed during transceive operation. After ruling out any mall configuration in receive-transmit sequencing the only cause of destruction could be the RF power at the input due to the finite isolation of the coaxrelais. This paper investigates the permissible RF power level at the input of the pre-amplifier and as such provides some metrics to the required isolation level.

Investigation is done by simulation using Aplac a Circuit Simulation and design Tool (Aplac, 2005).


Simulation setup

The original oz1pif design was captured into the circuit simulator, slightly simplified by omitting all DC related components. For the active element a linear 2-port model was used (ATF54143, .s2p, 3V, 60mA). Extracted are the source, drain and gate (peak) voltages as well as the gate current when the amplifier is excited with 100 uW. Frequency swept between 200 and 500 Mhz.

The used Aplac simulation is given here together with Aplac text input file

Simulation results

In the figure below the peak gate current is given as well as the maximum allowed gate current for the HEMT according the manufactures specification. As can be seen the device is at its limit with an input excitation of 100 uW.

Figure 1: Peak gate current

Another interesting parameter are the peak gate-source and gate-drain voltages. In the figure below those are given together with the permissible levels. As can bee seen, with an excitation of 100uW those maximum levels are not reached.


Figure 2: Peak gate-source and gate-drain voltages

To conclude also the gain of the total amplifier is simulated. Simulation result is very close the actual measurement at the bench.

Fig 3. Power gain

3 Conclusion

For this design, RF power levels at the input in access of 100uW (-10dBm) will lead to HEMT destruction and therefore should be prohibited at all means. To put this permissible level into perspective, the TX coaxrelais should have an isolation over 70 dB when applying TX power levels of 1000W (+60dBm). A measurement on regularly used types such as CX520 reveals that those are marginally adequate and should not be used or in conjunction with a second (small, but low loss) coaxrelais to improve isolation.

Further improvements on the simulation results could be done by using not the small signal model of the HEMT but incorporating the large signal model in conjunction with e.g. harmonic balance circuit simulation.