EMRFD Message Archive 771

Message Date From Subject
771 2007-05-20 13:43:22 Kevin Purcell WN5Y Super Blue LED Mixer
This has got to be the oddest thing I've seen in a while.

I've been trying to determine if anyone has measured the IMD numbers
for a 1n914/1n4148 diode ring mixer to compare it to schottky diodes.
But I wasn't prepared for this ...

In the august journal if the Cinncinatti QRP Club "Bacon Bits" kd5kxf
has published an article from David White, WN5Y, on the "Super Blue


> A friend from Australia, Pascal Nguyen, sent me an email with the
> following information:
> “My suggestion is to replace the mixer with passive LED type.
> Please find more information from US patent 6,111,452. [1] It is a
> wideband gap SiC material; it means high mobility for electrons.
> Simply speaking it is a ring diode mixer with all diodes replaced
> by blue LEDs. I built this type of mixer in 1998 but only with red
> LED to convert my 500kHz to audio. ...It was designed for High IP3
> and jamming warfare by the US forces.”

I recommend you look at the PDF now (it has color pictures :-). He
skips over a few things (e.g. you still should match your LEDs it's
just easier than matching 48 diodes).


Unfortunately they make no IMD measurements. But there are some in
the patent.

> Abstract
> A wide dynamic range RF mixer is shown using wide bandgap
> semiconductors such as SiC, GaN, AlGaN, or Diamond instead of
> conventional narrow bandgap semiconductors. The use of wide bandgap
> semiconductors will permit RF mixers to operate in higher RF
> environments, to be less susceptible to out-of-band jamming and
> interference, and to be more effective in receiving weak RF signals
> in the presence of strong unwanted signals. RF receivers can be
> more closely collocated to transmitters and still receive weak
> signals without suffering intermodulation distortion products
In the patent they have a singly balanced two-diode mixer. Their
measurements are not specified too well -- but I think they leave
enough information of the SA output to derive some numbers.

For silicon schotkky diodes they get a conversion loss of 10dB with a
drive of (I presume) 7dBm. To get the same conversion loss with 3V
band gap SiC diodes they need 20dB more drive (27dBm). In a way this
is like connecting 5 Si PN diodes in series to get the same voltage
drop or 12 schotkky diodes.

They show the IMD products from two tones at -13.8dB separated by
about 2MHz around 106.10MHz for Si PN diodes. They see third order
products at -33dB (I estimate) from a single tone.

They show the IMD products for the same test with the SiC diodes
(with 20dB higher drive level) and they see no IMD products (more
than -50 dB down on a single tone). Doing that math that shows that
the intercept has gone up at least 6dB (which doesn't sound so
impressive given all that drive). They state that they couldn't
measure IMD products with their test setup though they don't define
the upper bound of their testing system.

They also make the interesting comment that the carrier lifetime in
SiC PN junctions is much shorter than the lifetime in silicon which
means these should make good diode switches.

4H-SiC p+non+ diodes with 6-kV PIV are made (I don't know how
available they are) but they might be interesting devices to try for
switching. Rather like 1N4007 on steroids (though the 1N4007 is cheap
and has a +50dBm intercept according to EMRFD).

> The PIN diode behaves like a current-controlled resistor to all
> signals higher in frequency than 10 times the cutoff frequency (fc)
> of the diode, given by:
> fc = 1/(2πτ), where τ s the minority carrier lifetime.

f > 10fc PIN diode

0.1fc < f < 10fc diodes behaviour is unpredictable alternates between
a current-controlled inductor and capacitor.

0.1fc < f PIN diode acts like a PN diode.

Bulk Si PIN diode τ= 300 ns
Epi Si PIN diode τ = 180 ns

> In the forward-biased PIN diode, harmonic and intermodulation
> distortion are created by the modulation of the I-layer charge
> density by RF currents. The distortion is influenced by frequency,
> stored charged, and junction resistance.3 The intercept point
> (IPn), a widely used figure of merit for switch linearity, is a
> fictitious point where the linear transfer function intersects with
> the power of the intermodulation product. Third-order
> intermodulation products, 2f1–f2 and 2f2–f1, are considered the
> most troublesome because they occur close to the desired signal.
> The third-order intercept point (IP3) of the PIN switch can be
> analyzed using the method of Caverly and Hiller (G. Hiller and R.
> Caverly, "Predict intercept points in PINdiode switches,"
> Microwaves & RF, Jan. 1986):
> IP3 = 24 + 15 log (f If τ / Rs)
> where f is in MHz, If is bias current in A, and τ is carrier
> lifetime in ns.

And see Infineon Application Note No. 034 shows reducing the carrier
lifetime reduces the forward resistance too. That's good for RF
switches. And it gives a method for measuring the carrier lifetime
(that doesn't involve a laser) so you could characterize those LEDs.


Just looking around the net (there are lot of different process for
making SiC) I find τ to be in the microsecond range (after annealing
-- it could be down to < 10nS before). But that's ten times larger
than bulk Si.

Of course it may be different in LEDs :-)

So this should work for SiC, GaN, AlGaN, or Diamond diodes.

SiC schottky diodes are available (Infinion) not just as LEDs. But 4
blue GaN or SiC LEDs do make a nice glow if you put half a watt into
them. And your LO will double as a QRPp transmitter.

Plenty of things for the RF experimenter to play with. If nothing
else they make a talking point.

Kevin Purcell
772 2007-05-20 13:43:47 Kevin Purcell Re: WN5Y Super Blue LED Mixer
I should have done the math ...