EMRFD Message Archive 6863
Message Date From Subject 6863 2011-11-23 02:04:03 vlfun2 Sensitive AM Detector does anyone in the group have any further info they would share on this detector? EMRFD page 6.19, figure 6.47 D by Bill Carver. any info or recomendations welcome including comments or directions to somewhere else appreciated. I wish to thank everyone very much that replied to my first posting. from a comment I just saw online this may already be my last. thanks to all.
Dave PS if this site is taken down and you have info I can be reached at: email@example.com
6866 2011-11-23 09:38:27 kb1gmx Re: Sensitive AM Detector 6868 2011-11-23 19:33:26 vlfun2 Re: Sensitive AM Detector I've been playing with the circuit successfully today and like the detector very much. wondering if anyone is using it for AGC or possibly combining the detector output with some IF AGC to reign in the tremendous gain. just like hearing how others are doing with it if anybody is. I was lucky to get some notes Mr. Hayward sent me with testing he did some years back. if there's any more info available I'd like to add that as well to learn all I can about it and maybe not waste valuable time re-inventing the wheel. so if anybody has link info availble? thanks Allison.
6874 2011-11-24 08:23:23 Bill Carver Re: Sensitive AM Detector I'm in Colorado for Thanksgiving, Dave. Dont have a copy of EMRFD with me. I'll get back to you when I get home. I hope to be home by Sunday night, but Grandma always wants to stay longer.
W7AAZ (Bill Carver)
6876 2011-11-24 09:05:06 vlfun2 Re: Sensitive AM Detector good for Grandma. have a great time and enjoy the holiday. Allis 6877 2011-11-24 10:03:49 kb1gmx Re: Sensitive AM Detector 6878 2011-11-24 12:07:19 vlfun2 Re: Sensitive AM Detector If I followed your suggesti 6881 2011-11-24 15:32:35 Bill Carver Re: Sensitive AM Detector The AM detector does go to DC and generates AGC for an MC1350P. But there's some op amps and level shifting required to drive the gain control terminal on the MC1350P.
It IS sensitive: millivolts makes AGC and plenty of audio.
6882 2011-11-24 19:36:06 kb1gmx Re: Sensitive AM Detector Hi Bill,
We may have gotten out of sync. I was referring to the SA602 AM detector and using the OSC (pin6) input to throttle some of the
gain it has. That works when it's used as a mixer and I've even
used it a gain stage with AGC (about 17DB of gain and a 60+db agc range).
However the detector with adequate filter and level shifting
would work well as an AGC detector.
6884 2011-11-24 21:53:13 Bill Carver Re: Sensitive AM Detector As I said, I don't have EMRFD with me. I've used the NE602 as a detector in several different ways for slightly different purposes. All depended upon good DC balance and stablity of the NE602 and used a low powered op amp directly connected to the NE602. It isn't instantly obvious, but remember there are collector resistors inside the NE602, and those are well matched in value and the transistors are well matched. So the op amp operates as if it has differential inputs with approximately 1000 ohm matched source resistors. Only feedback resistors, and maybe zero-balance adjustment components, are required with this direct connection. My original op amps were "low power", drawing about 1.2 mA per op amp! A joke today: there are much, much lower power op amps with lower offset voltages and lower drift available now. One only needs to be careful to keep the inputs within the common mode voltage limits of the op amp.
I made a 3.5-7.5 MHz general-coverage receiver using an NE602 as an AM detector. The BFO terminal is driven to saturation by the AM signal and the same signal at normal levels drives the differential inputs. The differential output terminals of the NE602 are amplified by an directly connected op amp to make it a "syncronous" AM detector, essentiallyy a product detector whose BFO is a clipped version of the signal itself. I think this is conceptually akin to the AM detector in some Racal receivers. When driven so it still has enough drive even when carrier amplitude is down from QSB it is very resistant to distortion from selective fading of the carrier. When a conventional AM detector sees carrier lower than the sidebands it delivers a badly distorted output because, from its perspective, having sidebands larger than carrier constitutes overmodulation. As long as have "enough" of the pseudo-BFO this detector doesn't demand that the carrier exceed the sideband amplitudes.
There are other interesting variations on this theme. By driving both oscillator and differential inputs of the NE602 with the same signal, except hitting the oscillator harder so it is driven into saturation/cutoff each cycle, a DC shift is developed between the two output terminals proportional to amplitude of the signal. An op amp tied directly to the two outputs generates DC which feeds a hang AGC circuit for a dual gate MOSFET IF preamp and the MC1350 IF amplifier in my 40m mobile transceiver. That rig uses two NE602 each with op amps: one makes audio as a conventional product detector with BFO, the other generates the AGC voltage with CW signal driving both "oscillator" and signal ports of that NE602. Both audio and AGC are developed from millivolts of IF signal rather than the volts required to drive a conventional diode AGC detector.
If the BFO terminal is driven by an oscillator that is the same frequency as the incoming AM signal, there will be differential DC difference between the two output terminals of the NE602 dependent upon the instantaneous phase of the carrier and BFO. When they are the same frequency and have ninty degree (quadrature) relationshop that DC difference is zero. If deviates from zero, in either positive or nevative direction, if the phase deviates from quadrature. I have a little battery operated RX that uses the amplified DC difference to fine tune the LO so it phase locks to the AM signal. Unless the carrier is clipped, however, the DC shift is proportional to BOTH phase and amplitude. So it can't do linear AM detection and a really good job of "AFC" at the same time. Or, put another way, it can't be an ideal AM detector and ideal phase detector at the same time!
In that particular radio I don't have AGC. It has an RF gain pot from the antenna to a NE602 mixer, just like the first NE602 receiver that appeared in QST in the 1980s. I have a single transistor IF amplifier, then a second NE602 for AM detection and sorta-phase detection. The LO has diode AFC and it DOES lock to SWBC stations or WWV, but lock is sensitive to enough antenna signal to still produce enough tuning voltage to keep it phase locked to the incoming carrier. A rear-panel "zero adjust" was required to compensate for DC drift in the NE602 between usage sessions, which of course would be clumsy in a commercial product. But it's great for shortwave reception because selective fading doesn't cause the horrible distortion one gets with normal AM detectors. And it functions without a lot of power-consuming amplification to voltages required for ordinary amplitude and phase detectors.
In retrospect, I should have throw three milliamperes at that receiver and had another NE602 with clipped input for the phase detector. Or today I should throw a PIC inside to perform an auto-zero function on the NE602 each time it was turned on (just kidding). I eventually drifted away from super low power battery operated RX (especially after the bank of NiCAD cells leaked) and use of NE602s, in favor of more complicated circuitry that can handle bigger signals, and a Racal RA6790 for shortwave listening.
When I get home I'll refresh my memory of exactly what was printed in EMRFD and if there's a blunder in what I wrote above, I'll come back with errata. Barring snow in Wyoming I should be home Sunday evening.
73 - Bill W7AAZ
6886 2011-11-25 10:40:08 kb1gmx Re: Sensitive AM Detector Hi Bill,
I've been plying at the bench today with it as a general use RF
voltage detector. I"m not using a micropower opamp its' a OP27
since it was handy. Looks like it will serve well for the
millivolt range. With an input of 50 ohms created it should
serve as a useful microwatt power detector. I have to do a few
experiments to see what the dynamic range is. My initial guess
for the linear range is from about -50dbm to about -10dbm if
it does more so much the better.
Since the 602 is good to at least 200mhz that would be very useful.
6887 2011-11-25 13:35:41 AD7ZU Re: Sensitive AM Detector Maybe off topic but ...The following caught my attention:
If the BFO terminal is driven by an oscillator that is the same frequency as the incoming AM signal, there will be differential DC difference between the two output terminals of the NE602 dependent upon the instantaneous phase of the carrier and BFO. When they are the same frequency and have ninety degree (quadrature) relationship that DC difference is zero. If deviates from zero, in either positive or negative direction, if the phase deviates from quadrature. I have a little battery operated RX that uses the amplified DC difference to fine tune the LO so it phase locks to the AM signal. Unless the carrier is clipped, however, the DC shift is proportional to BOTH phase and amplitude. So it can't do linear AM detection and a really good job of "AFC" at the same time. Or, put another way, it can't be an ideal AM detector and ideal phase detector at the same time!
One of the problems with DC image cancelling receivers is the need to accurately balance the phase relationship between the in phase and quadrature lanes. Ideally this function would function much like an AGC loop in that if the I and Q signal phase difference could be measured and a DC error signal created to servo the phase relationship to near perfect quadrature.
So my questi
6889 2011-11-25 14:47:56 vlfun2 Re: Sensitive AM Detector thanks Bill. I will keep experimenting with your circuit. it is truly amazing. likely the best AM detector I've yet played with.
6890 2011-11-25 15:19:20 Bill Carver Re: Sensitive AM Detector -----Original message-----
6891 2011-11-25 15:34:21 vlfun2 Re: Sensitive AM Detector Allison I put turkey outta mind and re-read your pin 6/602 advice again. I did find using only the pot, pin 6 to gnd, would reduce gain slightly in Bill's sensitive AM detector before severe distorti 6892 2011-11-25 15:55:03 vlfun2 Re: Sensitive AM Detector wow. this is like having a front row seat at the best show in the world.
6894 2011-11-26 06:48:44 kb1gmx Re: Sensitive AM Detector 6903 2011-11-26 17:21:57 ehydra Re: Sensitive AM Detector You had a controllable gain range of 60dB for the SA602 ? Is this right
understood by me? What was the frequency?
> Hi Bill,--
> We may have gotten out of sync. I was referring to the SA602 AM detector and using the OSC (pin6) input to throttle some of the
> gain it has. That works when it's used as a mixer and I've even used it a gain stage with AGC (about 17DB of gain and a 60+db agc range).
6923 2011-11-27 09:52:16 kb1gmx Re: Sensitive AM Detector 6924 2011-11-27 11:39:43 AD7ZU Re: Sensitive AM Detector Allison,
you should write a best selling book..
"Everything You Always Wanted to Know But Were Afraid To Ask About the NE602"
6925 2011-11-27 14:35:23 kb1gmx Re: Sensitive AM Detector Its been said by someone else that the 602 is the RF equivilent
of the 555 timer chip in that both have many more uses than
anticipated by their designers.
The DBM is a close second. It's a mixer, RF attenutaor, phase detector, and so on.
6927 2011-11-27 17:54:15 William Carver Re: Sensitive AM Detector On Tue, 2011-11-22 at 23:24 +0000, vlfun2 wrote:
>I'm home now.
> does anyone in the group have any further info they would share on
> this detector? EMRFD page 6.19, figure 6.47 D by Bill Carver. any info
> or recomendations welcome including comments or directions to
> somewhere else appreciated. I wish to thank everyone very much that
> replied to my first posting. from a comment I just saw online this may
> already be my last. thanks to all.
> Dave PS if this site is taken down and you have info I can be reached
> at: firstname.lastname@example.org
If signal on pins 2 and 6 are within the linear range of both the "LO"
and the differential stages of the NE602 then you get a double-frequency
sine (which in my case the 470 pF filter out) and a dc proportional to
the square of the input amplitude. It's a "square law" detector.
AGC is a feedback system, and getting good sounding AGC without
overshoot or undershoot over a wide range of signal amplitudes can be a
little tricky. In the case of MC1350P the gain reduction in dB is almost
a linear function of the current going into the gain control pin. Saying
"gain reduction is almost a logarithmic function of the AGC current" is
the same thing. In a feedback system when you have a controlled stage
with gain reduction a logarithmic function of AGC current, and a
detector that has a square law output voltage versus input signal, the
two mesh (almost) perfectly and your AGC dynamics (overshoot, attack
time, etc) will be (almost) independent of signal level.
If instead of tying 2 and 6 together you drive either one with a
squared-up version of the signal on the other pin, or you simply
overdrive either one with a much larger amplitude, then once one port
saturates the output becomes a linear detector. This still produces DC,
but it doesn't mesh with MC1350P gain control characteristics as well.
On the other hand, as an AM detector you want audio output to be a
linear function of the input. One of the two inputs to the Gilbert cell
multiplier needs to drive that input from cutoff to saturation. So in
the same transceiver the same MC1350P drove a second NE602 as a
conventional product detector with xtal and capacitors on pins 6 and 7
to form a Colpitts oscillator. The pin 4-5 outputs of this second MC602
also have an op amp connected directly, amplifying the audio to volts so
it can drive a volume control and IC audio amplifier without any other
As a general-purpose amplitude detector, say in some kind of wideband
instrument, it would be difficult to make the detector linear: the
squared-up signal would need to have the same phase relationship as the
linear signal on the other pin of the multiplier, and although the NE602
works into the VHF region the phase relationship between the two inputs
won't be maintained so the detector linearity would be a function of
frequency and pretty much uselss.
Tying the two inputs together, with 50 ohms to ground, would yield a
square law characteristic and I would expect it would not be frequency
dependent until bond wire inductance and capacitance of different parts
of the IC caused a phase shift. I have no idea at what frequency you
would find detector output starting to fall off. And "fall off" is what
would happen when the phase shift of the two signals inside the IC start
to differ: the DC output will go down from what you get for the same
amplitude signal at a lower frequency.
In truth I think the "cute and different" aspect of figure 6.47 (D) is
the direct coupling of the op amp to the NE602 outputs. It generates
volts of DC from millivolts of IF signal, eliminating the need for huge
IF amplifier gains to get big fractions of a volt for a diode.
I run the transceiver off a 12V gel cell, with regulated 5V for
oscillators and the NE602, and the DC output on pins 4 and 5 is pretty
well centered in the common mode range of almost any op amp, and the
output from the op amp can be scaled to be almost 12 volts p-p, if you
had some reason to need to wring the last bit of dynamic range out of