Help requested forimpedance matching of HMC414 from Hittite?
I'm trying to design and implement a 1 watt power amplifier for 2.45 GHz band. I 'm using the HMC 414 chip, from hittite.
In my first test, I designed a microstrip line for working at 2.45 GHz. The lines work well, but when I put all together, the chip, the input/output microstrip lines, the vias network and the block capacitor in the output just I can get only 26 dB of output signal. I have a sample board from hittite, and this sample yields a 29 dB at the output with 10 dB of input signal. And the first harmonic has more or less 20 dB lower than the fundamental. And in the sample board is 35 dB lower.
I low dawn the deembeded data from hittite and I made the smith chart graph. But at the frecuency of 2.45 the impedance for S11 is r=24.65ohm and X=-15.50 ohm and for S22 is r=110.348 and x= 2.917ohm in a 50 ohm system. Well I'm reading something about matching with smith chart, but I don't understand very well the procedure. I understand that is possible to match an amplifier only with the manufacturer data, i.e. the S11, and S22. But I read this fact from an application note from hittite.
What do you recommend me?.
Thank you for your help.
Thank's in advance.
The S parameters provided by Hittite are for low signal. You need S22 at saturated power.
Because the harmonics are high, and power is low, seems that you have a problem with the load line.
You have to match the PA output at saturated power. If we assume that the insertion loss of the transmission line on your board is ok, try to move the position of the output shunt cap (C1) longways of the output line, and check only for maximum saturated power. When you get a maximum, play with the value of the shunt cap, in small steps (0.1pF) and check for maximum saturated power.
Have to use High-Q caps in your design.
There may be so many thing that you limits to obtain high power. Because the opearating frequency quite high and power level is moderate.
- Output matching could be done at the "Optimum Impedance" for maximum delivered power for both port.
The material which was used on demo board might be different than yours..
- Different PCB plating could be implemented ( such as gold vs thin)
- Connectors should be chosen carefully ( but their influences are relatively low)
- Typical value for this amplifier already is 27dBm, so the should be some mismatching between samples
- Measurement errors might be occured ( calibration, losses by cables, etc )
So on , so on..
On Smith Chart, everything is possible but in reality nothing can occur. To find the maximum power use a slide tuner to find maximum power at which frequency is associated and then apply this impedance that will needded by amplifier.
Regards