EMRFD Message Archive 257

Message Date From Subject
257 2006-12-23 15:51:53 Ted Bruce KX4OM MMIC PC Board mounting
Picking up on the thread from W4ZCB's MMIC application, I'm currently
building a transmitter driver-amplifier stage for HF based on Fig
11.14 in EMRFD. I've decided to separate the PA and driver stages,
particularly the MMICs and the PA due to layout considerations. The
PA built by WA3EBY that was published in QST emphasized the critical
nature of layout for his use of the MOSFET push-pull PA.

My other concern is the MMICs. I have that stage designed for
double-sided .032" glass-epoxy PCB material, with a considerable
isolation zone around the MMICs. I've tried to duplicate the ground
pads for the MMICs in my CAD program (this design is going to be used
in at least two transceivers). In The application notes regarding
test fixtures for MARs and MAVs from Mini-Circuits, they show the top
side of the board with only foil traces, and many holes drilled
through the extended ground pads for the MMICs to the other side of
the board, presumably ground plane. Other MMIC board photos I've
collected show some with top-side foil traces only, and some show both
sides ground plane. My current design is for the latter, as I
mentioned, with many holes drilled through the ground plane, using "Z"
wires to connect them, and of course many more holes around the ground
pads for the MMICs.

When I initially CADded the entire transmit RF stage, I had a lot of
spaghetti traces leading into the PA transistor mounting locations,
and from looking at the board photos in EMRFD, and taking to heart the
experience of other builders as well, I've decided to separate and
simplify driver and PA physical layouts. For a fairly straightforward
schematic, this stage of the rig has been the most difficult...I was
up to Rev. H before I decided to make the split!

I would appreciate any comments regarding the layout considerations,
and the use of double and single-sided board.

259 2006-12-24 05:11:09 Harold E. Johnson MMIC PC Board mounting
GM Ted, and Merry Christmas to you and the group.
Your comments (about my comments hi) and MMIC's and board construction. Agilent has a beautiful little piece of free software called AppCad where you can do all sorts of manipulation of electronic design. They call the practice of a single land over a ground plane "microstrip" and the use of a groundplane on both sides of a land as well as complete coverage of the opposite side with another ground plane "coplanar waveguide". My experience doesn't extend to the microstrip approach, but I've had nothing but success with the coplaanar approach when done properly. I can't imagine that you'll find any problem with the layout you're proposing, I have several utility amplifiers in the shop using cascaded 0685's and even a couple of 60 dB gain amps using 02186's that are completely tame as a pussycat. LOTS of decoupling, and lots of vias and you're practically guaranteed success.
265 2006-12-24 15:44:48 Ted Bruce KX4OM MMIC PC Board mounting
Thanks for the input, Harold. It occurred to me after I made the
comments that the Mini-Circuits test boards are most likely for well
above the HF range, where striplines are used. The ground pads I made
for the MMICs are pretty substantial, to be able to drill many holes
around them through to the bottom ground plane. As I said, they're
well-isolated from the rest of the top foil ground plane. The planes
will be well-knitted together, and while there will be distributed
capacitance as Rick mentioned with respect to your vias, I'm obviously
going to be well below the range where I'd be considering the coplanar
sandwich a waveguide. Assuming I don't have oscillati