RF Utility Knife UHF & microwave

[profdc9]

I have been working on a PCB which contains five useful circuits for RF work.  The five circuits are:

1.  Active RF Probe
2.  RF Noise Source
3.  Active Differential RF Probe
4.  Single Balanced RF Mixer
5.  RF Amplifier

The current version is on github

http://www.github.com/profdc9/RFUtilityKnife

It is a project that can be made using JLCPCB's PCB and SMT assembly service.

So far, I have the ActiveRFProbe and RF Amplifier working, the differential probe is working single-ended but I don't really have a good way of testing it with two sources yet.  The RF Noise Source has a nonuniform spectrum and so I am trying to use different noise diodes that might produce a more even spectrum, so I have added a boost converter so I can use a high voltage Zener diode, which is then amplified by about 10 dB to have sufficient excess noise.  The mixer is new and I simulated it, but it is not tested yet.  This uses the UHF transistor 2SC3356 which is a 7 GHz transition frequency, so the maximum usable frequency is about 1 GHz.

I would like to eventually try to build this using a much higher frequency transistor, the BFP740 which is a SiGe BJT with a 44 GHz transition frequency.  To do this, I am trying to use my 1 GHz design as a starting point.  Basically, the plan is to shrink everything down to 0402 size, use 5 volts maximum, confine the whole circuit to the top layer of the PCB to try to complete all the RF signal paths on top, keep all of the grounds on one side of the circuit so there's as much uninterrupted copper between them as possible (at least in each amplification stage), and then where the ground must be broken, sink as many vias as possible as close to the ground pads as I can to minimize inductance.

The results of this crazy plan are the PCB I laid out this weekend.  The schematic is also attached.  All of the RF signals are on top.  The bottom plane is largely an uninterrupted ground plane with very small breaks for delivery of the 5 volt signal where necessary.  Where there are breaks in the ground plane, there is stitching of the top and bottom layers together with vias to try to minimize the inductance of the break.

Is this approach totally crazy?  It is flyspeck engineering I know, but it might work.  I attached a PNG of the PCB as well as a PDF of the schematic.

[9aplus]

Nice development.
Please keep us posted about progress.
73
Djani

[sab]

Nice idea.
Sanjay

[Gribo]

I would make the layout so I can cut individual boards out, preferably with a V-Cut.
Also, since you have board space, add more stitching vias.

[profdc9]

Ok, I changed the lines so that they may be v-cut through, but I can not add v-cuts at this time.

I also added a lot of vias as well to stitch the ground planes together.