Passband tuning in the Drake R4C
Richard
I drew a block diagram from a description and this is what I got.
SSB at 1st IF of 5645 Khz goes to V3 2nd Mixer. BFO at 50Khz
also goes to V3. BF0 is additively mixed with 1st IF and so a
signal of 5695 Khz goes to CW/SSB xtal filters in 2nd IF. BFO is
variable, so carrier can be made to sit at either side of filter
center frequency, or indeed dead center of a filter passband. So,
passband knob for SSB is either at left or right position,
depending on USB or LSB.
Okay, but so far this is just moving a SSB signal across the (at
5695 Khz) filter passband.
Resolving SSB requires BFO injection at correct frequency. I've
got the BFO also feeding a product detector. A 50 Khz signal is
fed to the P.D. as well as the BFO signal. When passband tuning
is at center the received carrier sits at 50 Khz in the 3nd IF
chain and the BFO is at 50 Khz. This is correct to resolve SSB.
When passband tuning is (say) on USB, the BFO will be on (say)
48.5 Khz. Then, received carrier (if there was one to receive,
it's suppressed) sits at 48.5 Khz in the "50 Khz", 3nd IF chain.
This is correct to resolve SSB, because the suppressed carrier
sits there now. In fact, with this system, changing the BFO
frequency has the effect of moving the signal across the filters
passband, with BFO injection automatically tracking with the the
sideband energy.
Like I say, that I think describes Drake's passband tuning.
Pretty neat. Of course, the frequency readout on your R4C
dial shows the correct receive frequency, as the VFO does
not change frequency. If I have it wrong, please let me know.
Thanks.
Floyd Davidson
>I think I understand it.
>
>I drew a block diagram from a description and this is what I got.
>SSB at 1st IF of 5645 Khz goes to V3 2nd Mixer. BFO at 50Khz
>also goes to V3. BF0 is additively mixed with 1st IF and so a
>signal of 5695 Khz goes to CW/SSB xtal filters in 2nd IF. BFO is
>variable, so carrier can be made to sit at either side of filter
>center frequency, or indeed dead center of a filter passband. So,
>passband knob for SSB is either at left or right position,
>depending on USB or LSB.
>
>Okay, but so far this is just moving a SSB signal across the (at
>5695 Khz) filter passband.
>
>Resolving SSB requires BFO injection at correct frequency. I've
You have it correct, but let me draw that out just to make it
more obvious for anyone else reading this. Here is a block
diagram for that portion of a Drake R4C:
5645 kHz 5695 kHz
+----+ +-----+ +-------+ +----+ +-----+ +-------+ +-----+
| rf | | 1st | |roofing| | IF | | 2nd | |AM, CW,| | 3rd |
|ampl|--|mixer|--|filter |--| AMP|--|mixer|--| & SSB |--|mixer|--+
| | | | | 8 kHz | | | | | |filters| | | |
+----+ +-----+ +-------+ +----+ +-----+ +-------+ +-----+ |
| | | |
+-----+ +----+ +-----+ | +--------------+ |
| pre-| |xtal| | 50 | | | 5645 or 5695 | |
|mixer|--|oscl| | kHz |----+ | kHz xtal osc | |
| | | | | BFO | | +--------------+ |
+-----+ +----+ +-----+ | |
| | 50 kHz |
+-----+ +-----+ +----+ |
| | |prod | | IF | |
| pto | Audio <---| det |--+--| AMP|-----------+
| | out | | | | |
+-----+ +-----+ | +----+
+-----+ |
| AM | |
Audio <---| det |--+
out | |
+-----+
There is a slight difference between how this works for AM and
for CW or SSB.
For AM, the BFO is turned off and the 2nd mixer becomes an IF
amplifier instead of a mixer.
The AM signal and filter are at 5645 kHz, and the injection
frequency for the 3rd Mixer is 5695 kHz, to produce an output
frequency of 50 kHz.
For SSB and CW the signal and filters are at 5695 kHz, and
the injection frequency for the 3rd Mixer is 5645 kHz to
produce a 50 kHz output. (Note that the filter frequency
and injection frequency are swapped with that used for AM.)
The location of the passband (in relation to the 1st IF) is
changed varying the BFO frequency either side of 50 kHz (by
about 3 kHz each side). This moves the 2nd IF and the 3rd IF,
but not the product detector output because it is also generated
by that same offset 50 kHz BFO.
Moving the 2nd IF moves a received signal to a different
location in the passband of the 2nd IF filter. The output from
the product detector, however, remains at the same frequency
because the 3rd IF moved exactly the same amount and of course
so did the BFO.
The BFO bandpass tuning knob is calibrated with among other
marks, a line indicating the "center" frequency. When a CW
signal is tuned to zero beat, the peak S-meter reading should
happen when the bandpass tuning is exactly on that center
frequency mark. That should also produce the lowest pitch
sounding "whoosh" when no signal is tuned in. The actual tone
that is produced is the difference between the 1st IF signal
and the 5645 kHz xtal oscillator injected into the 3rd Mixer.
Now lets think about what could be changed. This example would
be true for any two equal bandwidth filters, but I'm going to
use 4 kHz wide filters as an example because it makes the
numbers easy and obvious. In real life, 3 kHz filters might be
more appropriate, though even narrower filters would work.
1st IF 4 kHz Filter
_________|__________
/ | \
/ | \
/ | \
---------------------|----------------------
5643 5645 5647
+50 +50 +50
5693 5695 5697
Frequencies after translation to 2nd IF
2st IF 4 kHz Filter
passband set for maximum bandwidth
_________|__________
/ | \
/ | \
/ | \
---------------------|----------------------
5693 5695 5697
If those two filters were installed, and the BFO is tuned to
exactly 50 kHz, the two filters would be exactly overlayed, and
the total bandwidth would be a little less than 4 kHz.
Buy if we tuned the BFO instead to 48 kHz, we get something
different though:
1st IF hi/center/lo frequencies
5643 5645 5647
+48 +48 +48
5691 5693 5695
Frequencies after translation to 2nd IF
2st IF 4 kHz Filter
passband set for 2 kHz bandwidth
| _______
| / \
| / \
|/ \
---------------------|----------------------
5691 5693 5695
Low end High end already
cutoff by cutoff by 1st Filter
2nd Filter
One consideration in this scheme is that many filters do not
have identical slopes on the high and low sides. This is very
common with SSB filters that are designed as a matched pair
using a common carrier frequency. If that type of filter is
used the 2nd IF should be chosen (above or below the 1st IF
frequency) to place the steepest skirts of each filter on
opposite sides of the bandpass.
Also note that in the first example, giving 4 kHz bandwidth, the
signal was located directly in the center of the bandpass...
which is not useful for CW or SSB. The crystal controlled 3rd
Mixer injection oscillator frequency must be made variable, by
replacing the trimmer cap with a front panel controlled variable
cap just large enough to move it over the same range as the BFO.
Sideband selection would be made by moving the crystal
oscillator frequency to one edge or the other of whatever
bandwidth has been set. This has the disadvantage of changing
the actual signal frequency to which the receiver is tuned (a
counter on the pre-mixer injection frequency no longer gives a
precise readout of the received signal's actual frequency offset
only by the injection frequency). Overcoming that would be
exceedingly difficult because it would involve mixing the 3rd
Mixer injection frequency with either the PTO or the band
selection oscillator... which would both be very difficult to do
considering that any feed through of the 3rd mixer signal into
the 1st Mixer would be a very serious problem.
>got the BFO also feeding a product detector. A 50 Khz signal is
>fed to the P.D. as well as the BFO signal. When passband tuning
>is at center the received carrier sits at 50 Khz in the 3nd IF
>chain and the BFO is at 50 Khz. This is correct to resolve SSB.
You've probably made a type there, and left out "not" before
"correct". That is correct to resolve AM or a zero beat, but
not SSB. Your following discussion is correct for SSB, which
requires the BFO, and thus the bandpass, to be offset to one
side of the received signal's carrier frequency.
>When passband tuning is (say) on USB, the BFO will be on (say)
>48.5 Khz. Then, received carrier (if there was one to receive,
>it's suppressed) sits at 48.5 Khz in the "50 Khz", 3nd IF chain.
>This is correct to resolve SSB, because the suppressed carrier
>sits there now. In fact, with this system, changing the BFO
>frequency has the effect of moving the signal across the filters
>passband, with BFO injection automatically tracking with the the
>sideband energy.
Exactly!
>Like I say, that I think describes Drake's passband tuning.
>Pretty neat. Of course, the frequency readout on your R4C
>dial shows the correct receive frequency, as the VFO does
>not change frequency. If I have it wrong, please let me know.
>Thanks.
You've got it. And you can see why a poor 1st IF filter causes
problems with blocking from adjacent strong signals. And how
that is still true even with a good filter if that strong adjacent
channel signal happens to be occupying the "other sideband" area
of the passband. One of the interesting modifications to the
Drake R4C is to provide switched filters at the 1st IF too, as
well as at the 2nd IF.
However, apparently a narrower 1st IF filter interferes with the
functioning of the noise blanker circuit. I'm not sure to what
degree, nor of exactly why. But there are some interesting
possibilities there too. The noise blanker's gate (as well as
all of it's amplification) comes after the filter. One very
interesting modification was done by Rob Frohne.
<http://www.wwc.edu/~frohro/R-4C/NoiseBlanker.htm>
This uses a Ten-Tec model 297 noise blanker modified to work in
a Drake R4C. The Ten-Tec unit does not include the noise gate,
and uses Motorola IC's for IF amplifiers that provide 120 dB of
AGC action. Rob's method was to put the noise gate in front of
the R4C's 1st IF filter, rather than behind it (and after the
1st IF amp and 3 transistor stages in the Noise Blanker itself,
which is the way the Drake 4NB is arranged).
I'm not sure if the noise gate prior to the filter is more or
less significant that the noise amplifier's need for transient
response. Hence, while Rob's design appears to work much better
than the old Drake design, I don't know if it is equally
affected by the bandwidth of the 1st IF filter or not. I am
definitely going to obtain a Ten-Tec noise blanker and install
it in my R4C, but testing it with various filters is just too
expensive, given that I don't need them otherwise. :-)
However, given your project... if your location is one where
strong adjacent channel signals are common, you definitely want
to go for the best filters possible early in the IF scheme. If
your location is prone to shotgun noise, you'll also want to
take a close look at the URL above (Rob does an *excellent* job
of describing it) as a possible basis for an added noise blanker
at a significantly lower cost that buying something like a Drake
4NB. (The Ten-Tec unit lists for $25 right now, plus it would
require at least that much more in parts for the interface.)
Another URL you might want to look at is www.sherweng.com, where
Rob Sherwood sells his mods for the Drake R4C and other
interesting things. It's good info, plus a couple of his kits
might actually be adaptable to your needs.
--
Floyd L. Davidson <http://www.ptialaska.net/~floyd>
Ukpeagvik (Barrow, Alaska) fl...@barrow.com