[time-nuts] Quad Driven Mixer 5 to 10 MHz Doubler

[Charles Steinmetz]

For those who may be curious, I dug out the schematic for my version 
of the quadrature-driven DBM frequency doubler and posted it to 
ko4bb.com.  It uses a quadrature hybrid coupler to generate +45 and 
-45 degree signals at 5MHz, buffers them with emitter followers, and 
applies them to the RF and LO inputs of a Level 7 DBM.  The file name 
is "Frequency doubler quadrature DBM" (it should be searchable once 
it is moved out of the recent upload directory).

Due to the excellent balance of the quadrature hybrid coupler and the 
DBM, the spurious products in the mixer output are all odd-order 
harmonics of 10MHz, the strongest being 30MHz at about -35dBc.  The 
5MHz feedthrough is about -50dBc.  Note that these depend on the 
mixer balance and circuit layout, as well as the quadrature coupler 
design and construction.

Also shown is an active filter/amplifier that brings the 10MHz output 
back up near 1Vrms and lowers the 30MHz harmonic to ~ -60dBc and the 
5MHz feedthrough to ~ -65dBc.  If a simple series LC is used to feed 
the load, H3 and 5MHz feedthrough can be reduced further to ~ -80dBc.

Bandpass filters at the output frequency can create phase modulation 
with temperature changes.  Accordingly, they are presumptively 
disfavored.  However, I built the whole circuit into a smallish cast 
aluminum box, which integrates any external temperature changes with 
a long time constant (tens of minutes), and have never had any 
problem in this regard.

For many applications there is no need for any filtering (other than 
a simple LP noise filter) after the mixer.  Because the even-order 
output products are very well suppressed, the output waveform is 
nearly perfectly symmetrical.  This means that it has a 50% duty 
cycle and maintains its 50% duty cycle when it is AC coupled or 
DC-restored, so triggering can always be stable at the center point 
of the waveform.

This inherent freedom from spurious outputs and suppression of 
even-order output harmonics makes the circuit attractive wherever low 
spurious output is desired, compared to diode doublers or active 
push-push circuits.  For any required suppression of spurs, the DBM 
doubler needs less aggressive filtering because of its inherent 
balance.  It also avoids the flicker noise that is characteristic of 
diode, BJT, and FET doublers.

Of course, any design is an exercise in compromise and meeting 
required performance goals, and there are always many means to reach 
the end, so different designers will choose different paths.  This is 
one I happen to like.  If you try it, you may find that you do, too.

Note:  The circuit works without the emitter follower buffers, but 
the spurious outputs are somewhat higher due to the nonlinear loading 
on the hybrid coupler.  This may still be useful if an all-passive 
circuit is required.

Best regards,

Charles