AGC Design?
 

I'm looking for some advice/guidance on the design of AGC detection and
timing circuits, prompted by some level of frustration with a modification I
have been doing to a DX-394 SW radio. My questions, though, probably apply
to receiver design generally. I have a problem with stability - the receiver
gain oscillates at medium and fast release speeds.

Previously I had done a mod that pretty successfully provided 3 release
speeds for the DX-394 but fell short of what I thought was the ideal: an
attack time of ~1 millisecond, independent of the release time. That was
based on a survey of receivers from which I concluded that the attack should
be less than 13 ms and that 1 ms seemed to be the goal. Release speeds
should probably be on the order of 30ms, 300ms and 3 seconds, for fast,
medium and slow, respectively, although there seems to be lots of scope for
subjective preference. My mod required a rather large capacitor for Slow
release so my Slow was more like 1.2 seconds and the attack was slowed to
maybe 50-70 ms for the slow release..

The objectives of the enhanced mod are to:
a) improve the attack speed to better less than 13ms for all release speeds
b) extend the Slow release using smaller cap
c) reduce the loading of the AGC detector on the output of the 2nd IF amp
and also possible distortion due to the AGC and AM/Product detectors fed in
parallel

I used a JFET to buffer between the IF amp and the diode detector and an
emitter follower between the attack R-C circuit and the release R-C circuit,
dc coupled to the stock AGC amplifier. On the release side, about 1/10 the
capacitance vs the earlier mod is required for slow release and the attack
does seem to be similarly less affected by the release network.

However, at the fast and medium release settings, the receiver gain
literally oscillates at a rate that seems to be a function of attack and
release time constants, manual RF/IF gain setting, AGC gain setting and
signal strength. The depth of this gain modulation is affected by AGC and RF
gain. In order to get stability, it seems that I have to slow down the
attack (and/or release) time constant and carefully tweak the AGC gain
between the onset of oscillation and receiver peak distortion caused by not
enough gain reduction.

Have I completely misunderstood the meaning of attack/release speeds? My
'ideal' attack circuit has a R-C time constant of 1 ms, which means it will
even respond substantially to 1kHz modulation. That seems high. The R-C time
constant for my target fast release of 30 ms means that it will
substantially follow a 30Hz signal. I have had to pad these out to ~20ms
attack, 50ms release for stability or tolerably low gain oscillation depth
at medium and lower signal strengths. With this slower attack, stability is
much improved with the 500ms medium release speed.

The target attack/release of 1ms/30ms is not good for AM reception anyway as
it causes considerable distortion on heavy bass modulation - it is for data
services on steady carriers, e.g., PSK, FSK, DRM. But if the AGC causes
oscillation, then that's interference of another kind that would adversely
affect error rates. Several, including myself, have noted that DRM SNR is
improved by defeating AGC, on a wide variety of receivers.

Is this a typical problem for receiver design? Would 'hang' AGC stabilise
the AGC loop? Are my design objectives reasonable?

Comments from experienced radio designers/builders/experimenters much
appreciated.

Tom

This sounds like a classic negative feedback oscillation. You sense the
signal is too large, so you send a signal to kill the gain, and then
you sense the signal is too small, so you send a signal to increase the
gain.

Having different attack and release time means you have two different
time constants My guess is the quick attack leads to the instability,
since it is the lesser damped system. If this is true, then you should
concentrate on the attack time, i.e find how slow it has to be for the
sytem to be stable.

Of course this is really had to do without seeing the circuitry in
action.

I agree with you and slowing the attack is the only way that I have
been able to approach stable operation with a fast release. But 20ms or
longer attack runs counter to what I understand to be the objective -
an attack speed of less than 13 ms and ideally about 1 ms. So, unless I
have this wrong, how do other receivers accomplish similar speeds
without self-oscillation?

The way my circuit operates (I think) is as follows (I'd be happy to
send a schematic to anyone who is interested) :
a) assume an impulse of signal of duration very much longer than the
attack time
b) the rectified signal is filtered of RF by a series-parallel R-C
attack network whose adjustable output feeds an emitter follower
c) the emitter follower pumps current as a low resistance source into
the release R-C network so the attack is not greatly slowed - its
output feeds the AGC driver amp
d) at some point, equilibrium should be reached - the current flow
through the release resistor and AGC driver base should equal the flow
though the emitter follower - but maybe the emitter follower pinches
off and that could be a cause of instability?
e) the signal drops, the attack network discharges at attack speed and
shuts off the emitter follower, so the release capacitor discharges
through its parallel R at release speed, the voltage to the AGC driver
falls so the AGC bias rises at roughly release speed to increase RF/IF
gain.

Having written that out, I have an idea or two I will try.

Tom

In article ,
"Tom Holden" wrote:

I'm looking for some advice/guidance on the design of AGC detection and
timing circuits, prompted by some level of frustration with a modification I
have been doing to a DX-394 SW radio. My questions, though, probably apply
to receiver design generally. I have a problem with stability - the receiver
gain oscillates at medium and fast release speeds.

Snip

I don't know receiver design but I have a RX340 that uses the following
settings. Attack is in dB/mS, Hang in seconds, and Decay are in dB/S.

Attack Hang Decay
Fast .8 0 1600
Medium .8 0 100
Slow .8 0 25

Programable attack .01 to 1 dB/mS
Programable hang 0 to 99.9 seconds
Programable decay .01 to 99.9 dB/S

--
Telamon
Ventura, California

Telamon, those are interesting numbers, expressing the action of a more
sophisticated, programmable, digital AGC. Classic analog AGC speeds are
expressed as the length of time it takes to reach a certain percentage
or within a few dB of the desired gain setting, i.e., similar to and
based on RC time constants as that was the foundation of the classic
AGC control system. With an RC derived control, whether the gain change
is 10 dB or 100 dB, it takes the same time. With your digital control
in 'Fast' mode, attack would be 8ms for 10 dB and 80 ms for 100 dB;
release would be hang time plus 6ms or 60 ms respectively. It's
interesting how these compare with my target of 1-13 ms attack, 25-50
ms release.

I'm wondering how your RX340 behaves when you program to 0.01 dB/ms
attack, 0 hang, and 1600 dB/s decay (but I see that the programmable
decay is limited to 99.9? probably for good reason!). That would
correspond to my Fast target when you tune from no signal to S9+50.

Regards,

Tom

From: "Tom" on Wed,May 25 2005 10:30 am

I agree with you and slowing the attack is the only way that I have
been able to approach stable operation with a fast release. But 20ms or
longer attack runs counter to what I understand to be the objective -
an attack speed of less than 13 ms and ideally about 1 ms. So, unless I
have this wrong, how do other receivers accomplish similar speeds
without self-oscillation?

The way my circuit operates (I think) is as follows (I'd be happy to
send a schematic to anyone who is interested) :
a) assume an impulse of signal of duration very much longer than the
attack time
b) the rectified signal is filtered of RF by a series-parallel R-C
attack network whose adjustable output feeds an emitter follower
c) the emitter follower pumps current as a low resistance source into
the release R-C network so the attack is not greatly slowed - its
output feeds the AGC driver amp
d) at some point, equilibrium should be reached - the current flow
through the release resistor and AGC driver base should equal the flow
though the emitter follower - but maybe the emitter follower pinches
off and that could be a cause of instability?
e) the signal drops, the attack network discharges at attack speed and
shuts off the emitter follower, so the release capacitor discharges
through its parallel R at release speed, the voltage to the AGC driver
falls so the AGC bias rises at roughly release speed to increase RF/IF
gain.

Having written that out, I have an idea or two I will try.

Having encountered a similar problem many years ago, I'll offer
this as a suggestion: Analyze the behavior of the total signal
amplification chain at LOW frequencies, not at the RF or IF
carrier. Know the control characteristics of the AGC voltage
input to the amplifier versus the total amount of gain of the
receiver chain. Approach the whole receiver AGC action as a
low-frequency servo loop (which is what the AGC actually does).
Think servo control systems theory.

Control systems theory is a rather abstract thing and there
probably will be no sudden bright light of understanding
switched on, but here's a bit of that: The AGC loop action
works by BOTH magnitude and phase at low frequencies.
"Nyquist" and "Bode" plots are helpful there, even though
both of those subjects are also rather abstract. In general,
if the AGC control action results in instability or even motor-
boating, the overall receiver gain - related to the control
voltage range - is too high. Adding a voltage divider at the
low-pass R-C filter of the AGC voltage input will demonstrate
that. Also, the low-frequency phase shifting in the AGC
voltage "decoupling" can upset the phase versus magnitude of
the control voltage. Note: Vacuum tube or FET RF/IF
controlled amplifiers probably use such R-C decoupling, working
only on AGC voltage; other amplifier types might have some
other form of R-C filtering at low frequencies. That low-
frequency magnitude AND phase relationship is important for
total loop stability.

What has to be considered in the AGC loop is the response
through all the decoupling newtorks between the ACG control
source and the controlled device(s). For a "non-linear"
loop (separate attack and decay times) that analysis will be
difficult. It is much easier to analyze with a Spice
simulation that has the capability to model a controlled-gain
amplifier. The whole loop at low frequencies can be modelled
that way. In starting that, forget the RF and IF components
and consider only the amplifications at low frequencies; the
source of the AGC control (detector output) may have to be
modelled slightly differently in that the detector is, in
effect, similar to a power supply rectifier. If that model
is tweaked to be stable with sudden transitions on its input,
then it will be stable at RF and IF.

Let me add one more general note about AGC design. The BFO frequency is
very close to the IF, and it typically puts out volts of signal while
the AGC circuit is trying to operate with millivolts. Unless you're very
careful with layout, shielding, and balance, a lot of BFO signal can get
into the AGC circuit and cause disturbances and malfunctions of various
kinds.

The last AGC circuit I did was very conventional, and it's the sweetest
operating one I've ever used. But I went to great pains to keep the BFO
out of it, and feel that was one of the essential ingredients in getting
it to operate so well.

Roy Lewallen, W7EL

From: Roy Lewallen on May 26, 5:39 pm

Let me add one more general note about AGC design. The BFO frequency is
very close to the IF, and it typically puts out volts of signal while
the AGC circuit is trying to operate with millivolts. Unless you're very
careful with layout, shielding, and balance, a lot of BFO signal can get
into the AGC circuit and cause disturbances and malfunctions of various
kinds.

I agree on the need for isolation of various circuits but fail
to see the relevance. A "BFO" is on for OOK CW reception and
normally a manual RF/IF amplification control is used to set a
comfortable listening level. Yes, AGC could be used on OOK CW
but it would be a mistake to derive the AGC control from an AM
detector getting "BFO" input...that would be the same as
introducing a DC bias into the AGC control loop...which would
change the AGC servo-action control...perhaps severely so.

Note: A "BFO" source is steady-state. The detector mixes the
incoming signal (usually at the IF) with the "BFO" to derive
the audio. If the AGC control line is picked off this same
detector, the DC component is akin to having a nearly fixed
DC bias inserted. To use AGC on an OOK CW signal, the audio
tone would have to be used...and that necessaitates a different
sort of AGC control source derivation. A peak-riding, perhaps
selective audio circuit could do that, but the complexity of
that part of the receiving chain is growing. It might be
easier all-around to just pick off the IF to a separate AM
detector as the AGC control line source. The "original"
detector could remain as the OOK CW output with isolated BFO
feeding it.

For SSB voice reception, a "BFO" is still present but a single
diode detector all-purpose sort of detector is far from
optimum as a combined audio source and AGC control line source.
It WILL work, but it is non-linear for both audio and AGC
purposes and that alone could be the source of AGC instability.
It depends on the IF signal level at the detector diode (or
"product detector" which is really a mixer stage).

A single diode with large time-constant on its voltage output
is a peak-riding source for the AGC control line. Whether or
not it follows fast "attack" conditions depends on the source
impedance capabilities of the final IF stage. If that is too
high then the "attack" time is slowed from the necessity to
build up a charge on the diode's load capacitance; that can
be seen on examining an ordinary AC rectifier circuit in
response to a step transient of AC input through various values
of AC source resistors. The peak-riding capability is usually
distorted on the leading edge...which then reflects on the AGC
control characteristics (when loop is closed) in trying to hold
the received signal constant at the detector.

Thought of as a servo-control loop, the AGC subsystem can get
rather involved and complex, affected by a number of different
factors, ALL of which are important insofar as AGC instability
is concerned. "BFO" level is just one item and I will disagree
that it is a very important. It is no more important than
anything else in that loop in my experience.

As a suggestion to anyone else, I would recommend first either
measuring or calculating the AGC control line versus both the
antenna input level and the IF level at the AGC detector input.
That yields a DC baseline datum on the controllable level of
the receiving chain. From that, one can "back-track" calculate
how well the closed-loop AGC action behaves; i.e., the antenna
input RF level versus the peak audio output with AGC on. If
that is using old-style "variable-mu" pentode tubes, then the
control characteristics will show whatever non-linearity it
has steady-state. That can be used as a special controlled-
gain model baseline for a Spice analysis of the AGC loop.
Differing time-constants IN the AGC control feedback can be
set to observe closed-loop response with transient signal
input to the antenna.

The last AGC circuit I did was very conventional, and it's the sweetest
operating one I've ever used. But I went to great pains to keep the BFO
out of it, and feel that was one of the essential ingredients in getting
it to operate so well.

Having had a National NC-57 receiver since 1948, I decided to
"play" with it in 1959 and "improve" its performance, such as
increasing IF gain. The first IF stage as well as the RF stage
were AGC-controlled. Not knowing enough about Control Theory
then, nor considering the low-frequency characteristics of the
AGC control voltage line R-C decoupling, that modification
became a disaster for anything but manual RF gain control. The
motorboating (very low-frequency oscillation) extended to having
the VR-150 screen supply regulator (gaseous shunt regulator to
those of solid-state era times) going on and off. It was
restored to its original components and not played with for over
a decade. Much later, on having had to get into Control Theory
and servo control loops, I could analyze how bad it was and see
what I SHOULD have done. The control was too "tight" in trying
to hold the audio output too constant over a wide signal input
range. There was low-frequency phase shift in the AGC voltage
control decoupling that was responsible for most of the motor-
boating; the VR-150 shunt regulator control range was a bit
too narrow so naturally it had dropped out of regulation and
added the final insult to the original "mod." [forty somethings
and younger may not be familiar with such relaxation oscillator
circuits :-) ]

National Radio Company had made an acceptible product in the
NC-57 but it was a low-end item in their product line. It
worked well enough as a single-conversion HF receiver but it
wasn't optimum in design and no doubt stock logistics at the
factory probably accounted for some of the parts values.
Several passive components seemed to be rather arbitrary in
value choices. I had learned (or should say re-learned) that
NO product is an example is "what something should be" as a
design example.

There just aren't any "easy" answers for some things in
electronics. But, they can be WONDERFUL, challenging
"cross-word puzzle" kinds of thing to solve! :-)

"Roy Lewallen" wrote in message
...
Instead of solving the fundamental
problems, increasingly complex circuits are developed until one
accidentally works correctly, then the improvement is credited to the
complex circuit rather than its accidental relative immunity to the
results of poor fundamental design.

I prefer the solution of, "Hmm... there's already a CPU in this radio
anyway... and we've got an ADC around... hey, let's make it the software guy's
problem!"

:-) :-)