fhx35lg

Dear all,
I have to measure DC behavior of a Fujitsu FHX35LG packaged low noise HEMT. In order to do that, I need to design a test fixture. I used a simple test fixture having 50ohm TRL at a fixed frequency connected to drain and gate respectively, two pads connected to the sources with a multihole metallization connecting the sources themselves to the ground plane of a microstrip. Performing measurements with a parameter analyzer, I get a quite strange Vds-Id curve, with a reduction of current in saturation region. I use two bias tee for terminating 50ohm lines. I tried with both pulsed and continuous measurements, without significant changes in results. I guess some instability takes places, probably due to a poor grounding. Can someone help me please in designing a good test fixture to get the correcti I-V curve?



Thanks in advance.

Ivan[img][/img]

The behaviour may be caused by self oscillations of the transistor in your fixture. Connect a a spectrum analyzer via a lightly coupled probe (or loop) for checkout. If the self oscillations are in the microwave region, try to put pieces of Eccosorb on the conductors to stop the oscillation.

Thanks a lot VSWR,
could you please explain to me how exactly ECCOSORB should work in self-oscillation suppression? Can I use it for surface wave suppression during RF operation as well? Moreover, I guess it can be useful for reducing mutual coupling between DC bias lines and RF circuitry in a final complex design. My question is: does it drastically change the electrical properties of substrate when used during RF operation in a uwave or mm-wave circuit? If it doesn't, as consequence, if I want to make a test fixture for measuring the S-matrix as well, it should come out very useful.But absorbing material should be sufficiently thin to be considered part of a complete RF circuit, I think...
Apart from absorbing materials, do you know any effective test fixture layout for optimizing HEMT measurement reliability?

Thanks again,

I.