Sometime ago I wrote the original blog entry by this name. ([
CLICK HERE] to read that).
This topic was introduced based on equipment configurations in use in many of my different contest operations from Concord, Ca. Since then, a number of equipment reconfigurations have been made.
Because WQ6X operations use cascading filters, back in 2014 I asked the QST Dr. about the ramifications of a given order of devices in each audio chain. Joel's comments about device
linearity and overload were very prescient. ([CLICK HERE] to read that).
Recent audio filter re-cabling @W7AYT brought his comments back to life.
Several questions come to mind
- How many devices can be practically cascaded on each audio line?
(Left ear and Right ear) - What order should those devices be?
- Can we intermingle Analog and Digital circuits?
Because noise (QRN) is always a problem, it was discovered that
overall there are 5 types of noise limiting/processing circuits:
- Peak Limiters
- Trough Limiters
- Noise Silencers
- Noise Blankers
- DSP (surgical) Noise limiting
Realize that the above categories are approximate. Different manufacturers use different names
for their NR circuitry, in the same way that the terms "IF-Shift" and "Passband Tuning" (PBT)
are often interchanged. The classic MFJ-752c ("Signal Enhancer II") utilizes both Peak and
Trough noise filtering.
PEAK LIMITERS
Peak limiters are essentially 2 (or more) diodes across the audio line to clip the +/- noise peaks.
An adjustable potentiometer (between the diodes and ground) allows adjustment of the depth of clipping. Early ANL (Automatic Noise Limiter) circuits were tube diode circuits; often using the
stubby 6H6 (and later 6AL5) dual-diode tubes to accomplish the clipping. Not surprising,
the results were not that spectacular. In my mind, most <NL> switches on all the classic
shortwave receivers were nothing more than a JOKE; add a DEPTH potentiometer and
they become useful.
TROUGH LIMITERS
Trough limiters are essentially 2 or more (reversed) diodes "biased" to NoT conduct unless the signal components exceeds a (pot-adjusted) threshold. The advantage of this approach is that low-level noise is not allowed to pass while signals above that level come right through.
NOISE SILENCERS
Noise silencers are essentially all based/influenced by the 1936 work by James Lamb and are often called Lamb silencers. The idea of this approach is to remove the excess noise pulses from the IF signal path before they have a chance to key the AGC line. Collins used this approach by including
a separate 40-mc receive circuit; capturing JUST the "noise", the output can be allowed to influence the AGC line.
NOISE BLANKERS
As we moved into the 70's various adaptations of the Collins Radio adaptation of the Lamb circuit began to appear. In the Collins approach the blanking signal was usually made outside of the amateur band being received. The more economical/common approach takes the blanking signal right at the IF frequency. The downside to this approach is that with the blanker circuit inline and no noise present, strong signals on nearby frequencies can trigger the blanking circuit, creating a form of static/distortion; another name for this is artifact.
DIGITAL SIGNAL PROCESSING With Digital Signal Processing, the idea is to digitize an analog signal in order that it be
"dissected" on a Bit-by-Bit and Byte-by-Byte basis. At least theoretically, the extreme PEAKS
can be completely removed; as well as signals below a minimum threshold. Once all this "JUNK"
is removed, the digital is transformed back into analog, as if nothing had happened. Unfortunately most DSP-algorithms are not flawless and can [theoretically] contribute their own artifact.
I use the above DSP units for the Right Channel (NIR-12) and Left Channel (MFJ-784) audio coming from the Yaesu FT-1000mp. Because the Sub-RX lacks any kind of notch facility, the NIR-12 gives the equivalent of the Main-RX eDSP. They both can auto-notch carriers and can shape the audio passband. The NIR-12 also offers a DYN PEAK (essentially a trough limiter). For practical use,
the signal threshold of this circuit is set WaY too high; however when it works, the effect is AWEsome. LooK for me to devise a circuit modification introducing a threshold adjustment PoT.
Because Noise Reduction (NR) is needed MORE for the Main-RX, the MFJ-784 does an AWEsome job. While the Yaesu's eDSP NR clips the noise-heads off rather nicely, the MFJ-784 actually seems to "root out" the noise from underneath - what a great combination. The FILTERs selection knob produces effects quite similar to the eDSP Contour control; the pair used in cascade produces
some unusually interesting listening effects.
Earlier on I raised the question of whether or NoT Analog and Digital can be intermingled. Eventually it ALL becomes analog. It seems to me that as long as we don't introduce undesirable artifact into the audio we are processing, we can/should give consideration to all manner of audio improvement.
Remember: IF You Can't Hear'Em, You Can't Work'Em.
What Do YOU use to accomplish signal-copy improvement?
Write me and let me know.
