Oscillator ALC Rectifier

For a SWITCH. In the oscillator it's used as a peak detector, pulsing current into an integrator, when the "reference", also thru a diode (for compensation), is exceeded.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

Hello Jim, Hello Roger,

The fwd drop issue could be tackled nicely with the inverted transistor, as suggested in EDN Magazine Nov 9, 2000, page 172ff:

Regards, Joerg

[...]

Peak detection is generally noiser than 1/2 wave averaging. The noise will increases the near carrier noise in the oscillator output. If you are going with an ALC oscillator to minimize the near carrier noise, it is best not to peak detect in the rectifier.

--
--
kensmith@rahul.net   forging knowledge

The HP 8654 RF generator has an ALC detector which passes current through a pair of diodes even when no signal is present. This seems to make the detector follow the waveform envelope without holding the peaks too long.

The Linear Technology App Note : AN43-32 "Bridge Circuits" shows Wien Bridge oscillators with two diodes in the ALC.

ALC rectifier attack time and decay time combine with loop filter response to make a circuit which I find hard to tweak experimentally.

Please share your insights, experience, references.

thanks, Roger

It is so easy to get a sawtooth amplitude jitter when using peak detection. The peak detector grabs even short noise pulses and then the amplitude ramps back down again. Winding up the CRO gain and looking at the envelope can show this. This level of jitter is obscene for low noise RF oscillators, but bench lash ups often show this.

- Jim tells us that the second diode is for temp compensation.

- Ken tells us half wave rectification is less noisy.

Looking at equipment manuals, I see both of these features. A couple of basic principles for me.

thanks Roger

[you can ignore me, call me a troll, whatever, but...]

I was looking through info on Wein bridge oscillators, including AN43, a year ago when I wanted a stable amplitude, fixed-frequency oscillator for the exitation voltage for an LVDT for position measurement, as I've seen LVDT schematics that used such an oscillator. I didn't like all the complications and tweakings of those types of oscillators, so from parts I had laying around, I got a 1.6MHz crystal ocillator driving a CMOS 4024 7-stage ripple counter for frequency divide by 128 (resulting in 12.5kHz - I considered 1MHz / 64, but

15.625kHz rings a bell, it's the US TV horizontal sweep rate, and if there's a TV on in the house it's sure to cause interference), then put that square wave output through three or four poles of low-pass filtering [thanks once again to Don's "Active Filter Cookbook") to get a reasonably good sine wave. With the whole thing powered by voltage regulators, it's as steady (both amplitude and frequency) as I need it to be.

While needing the complication of a microcontroller to control it or otherwise set up the registers, a DDS chip would be a VERY stable alternative that needs no tweaking.

-----

long.

There's a name for diode-detector-following distortion, but I cannot recall it. But I agree with Jim, the circuit you see is probably to remove tempcos in diode forward voltages.

response

Poor choice of time constants results in a phenomonon known as "squegging". The control voltage going through undesired and sometimes chaotic oscillations does this. I know at least one oldish (60's?

70's?) RF textbook descusses this in terms of poles and zeroes. Is this the problem you're having?

Tim.

and

Almost certainly I'm thinking of Clarke and Hess. (Not all that old, either...) They discuss AM detectors other than diode detectors IIRC. Tim.

I succeeded in modelling a level loop on (LT) spice. The build up of oscillation takes many cycles so simulating can take minutes

In message , Ben Bradley writes

--
dd

Mere minutes! My DC-DC converter circuit took all weekend to simulate up to steady state. Lucky for me, the sim matched what I expected (well close enough) so I only had to run it twice as the full up test.

I was brave enough to test it at 2 voltages and assume it didn't do anything surprising between them.

--
--
kensmith@rahul.net   forging knowledge

I read in sci.electronics.design that Tim Shoppa wrote (in ) about 'Oscillator ALC Rectifier', on Tue, 25 Jan 2005:

Diagonal clipping.

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

The reason for the distortion under negative modulation is the long discharge time constant, which is needed with conventional peak detectors (diode, comparator, etc.) For example, the typical ratio between the charge and discharge time constants may be 10,000 or more.

A different method offers greatly improved performance, with reduced noise and thermal drift, and better tracking over wide signal amplitudes.

A charge/discharge ratio of 10:1 is easily obtained, which greatly reduces the distortion on negative modulation. It will easily extract the envelope of a double sideband suppressed carrier signal, where a conventional diode or comparator detector would give unusable results.

For more info, see US4603299 "Constant duty cycle peak detector", at

Mike Monett