EMRFD Message Archive 1106

Message Date From Subject
1106 2007-10-18 12:57:40 Jim Miller Hybrid Cascode Noise Analysis?
I've got a diagram for a Driscoll oscillator from Chris Bartram. I'll post
it in the files area. It uses two FETs in cascode.

He suggests the NEC NE68833 for each.

Once consequence of that is a higher VCC required in order to bias both fets

I wonder if replacing the upper FET would have any negative effects on
perfomance? I'm guessing a bipolar with sufficient GBW would still provide
good isolation of the FET source from the load. Good guess?

I'm wondering about how the noise performance might be impacted by using a
bipolar as the upper device.

I'm considering the oscillator for use as a master oscillator for a DDS and
am concerned about phase noise, both close in (1/f) as well as wider spaced.

Any thoughts?


jim ab3cv
1108 2007-10-18 14:48:21 Wes Hayward Re: Hybrid Cascode Noise Analysis?
Hi Jim, and group,

I am afraid that I may have confused things with my description of
our hybrid-cascode circuit without posting a schematic. So I've
posted one as "hycas_imd_nf_ckt.jpg" in the FILES section. This is
one of the figures I used in the presentation in Dayton, FDIM07. I
hope this shows the basic nature of what it is we are doing.

The Driscoll oscillator as described in your posting looks
outstanding. The parts that he uses are NPN bipolars and not FETs.
Indeed, if you pull up the data sheet (follow the url trail from the
DigiKey catalog listing on line) you will see a wonderful part. Like
all of the NEC parts, it is well specified, yet it is reasonably
priced. Imagine: A 4 to 8 GHz F-t part for less than a buck!

Let's take a look at this circuit. The lower NPN is an inverting
amplifier with the crystal serving as an emitter bypass. Because
we are at VHF, the input impedance at the base is going to be fairly
low. The upper NPN is just the usual comm
1109 2007-10-18 15:38:52 Jim Miller Re: Hybrid Cascode Noise Analysis?

Thanks so much for taking the time to correct my brain seizure!!! No
other way to explain my lack of attention to an obvious detail.

I plan to BB this at 10Mhz with the "usual suspects" as training
wheels, then do it "for real" at 125Mhz with an overtone xtal, likely

FWIW, I've seen papers indicating that 10Mhz multiplied via multiple
schottky stages seems to outperform low noise overtone xtal
oscillators. I suspect I'll be look at ways to quantify the
performance of both.

Should be a fun building season!


jim ab3cv