Converting small AC signals (46mV to 200mV) to DC

[kushal4269]

Is there any analog circuit which can convert small AC signals (46mV to 200mV) to DC, I know diode based half and full bridge won't works due to forward threshold voltage.

Need for this conversion is to use the out put of this DC volt as control voltage for Voltage controlled amplifier which amplifies prior mentioned AC signal.

Attached block diagram.

[Kleinstein]

For low frequencies (e.g.  < 10 kHz) one can use active rectifiers with OP-amps or similar. For higher frequencies there are square law power detector circuit - similar to one way rectifiers with a germanium or schottky diode and maybe prebiased. Another option is to first amplifiy the signal and than rectify.

[bdunham7]

One-chip solutions are readily available:

https://www.analog.com/media/en/technical-documentation/data-sheets/AD8436.pdf

[mawyatt]

Maybe consider a Log-Amp?

Best,

[Zero999]

Going from the block diagram, it appears as though you want an automatic gain control circuit, in which case, take the voltage from the output, not input signal and feed it into the control input.

Here's a microphone pre-amplifier circuit I designed awhile ago, for a project which was cancelled.

Q1 is a common emitter amplifier.

D1., D2, C3, C4 and R4 convert the AC part of the output to DC, which biases M1 on, once it exceeds its gate threshold voltage and it starts to short the microphone to 0V, thereby limiting the gain.



EDIT:
I forgot to post a link to the thread. The above circuit is designed to be two wire. The output terminal is connected to a power supply via a resistor and the output sits at a couple of VDC, with the amplified AC waveform superimposed on it.
https://www.eevblog.com/forum/projects/mic-pre-amp-circuit-discrete-or-ic/

[Benta]

Just use a NE570 from onsemi. Does exactly what your drawing shows.

[Sacodepatatas]

Is there any analog circuit which can convert small AC signals (46mV to 200mV) to DC, I know diode based half and full bridge won't works due to forward threshold voltage.

Need for this conversion is to use the out put of this DC volt as control voltage for Voltage controlled amplifier which amplifies prior mentioned AC signal.

Attached block diagram.

Yes, and indeed is pretty simple. It consists of just one bipolar transistor. Depending upon the polarity of the resulting DC, one can use either an NPN or a PNP. BUT, if you need accuracy down to 20mV or less, you need to use a BJT with a high reverse Hfe. Usually, medium power BJTs have an rHfe of few dozens, but there is a special case of the SS8050 having a Reverse Hfe of about 60, and its complementary, the SS8550, that has a great rHfe of 150 making this transistor very reversible (but you have to take into account that being PNP you will get a negative-going DC).

https://www.radiolocman.com/review/article.html?di=647409

[PCB.Wiz]

Yes, and indeed is pretty simple. It consists of just one bipolar transistor. Depending upon the polarity of the resulting DC, one can use either an NPN or a PNP. BUT, if you need accuracy down to 20mV or less, you need to use a BJT with a high reverse Hfe. Usually, medium power BJTs have an rHfe of few dozens, but there is a special case of the SS8050 having a Reverse Hfe of about 60, and its complementary, the SS8550, that has a great rHfe of 150 making this transistor very reversible (but you have to take into account that being PNP you will get a negative-going DC).

https://www.radiolocman.com/review/article.html?di=647409

Interesting circuit, but as might be expected it has many caveats and compromises around the full-wave signal rectification tag.

It does not use reverse HFE, the NPN simply saturates.  (Ic never changes sign, and actually never drops below 1.2mA at 12V, 2V Gen)
Once saturated, the inversion feature ceases and it attenuates the input, with no inversion.
The feed around RC adjusts relative gains of invert-non-invert.  I find  Rx = 12.07k exactly balances positive peaks for 2V drive, with 12V Vcc.

Source impedance must be low, and load impedance must be high.
The saturating nature of NPN means the load seen by source changes drastically and the drive current is asymmetric, which means some DC restoration will occur.
This also means the full wave is not 50%, but skews to 42%/58%, and it very dependent on signal, and actually only peak-balanced at one critical drive level.
At low Vin the rectify effect is very soft, and below about 180mV p-p is less than temperature variations, but a second copy and a transistor pair could help reduce that, along with a differential opamp.

[fourfathom]

Google "opamp active rectifier" and lots of circuits will show up.  This is a typical one (a half-wave rectifier):

You can set the gain with the R1 and R2 resistors.  That D2 diode is just there to keep the opamp out of saturation.  Putting D1 inside the opamp feedback look effectively cancels the diode forward drop.

You didn't say if you needed any particular precision, RMS vs peak, frequency, etc, but the circuit above is a good starting point.  Or, as mentioned, there are probably ICs that will do all this for you.

[Sacodepatatas]

It does not use reverse HFE, the NPN simply saturates.

The first time I met this circuit, and read about the working principle refering to the reverse Hfe, I thought the same as you do, because when saturated, there is no current gain. For working in reverse mode, besides forward biasing the BC diode, there should be higher voltage at the emitter than the collector.

But, when I tried a circuit like this (without fb) in a breadboard, I found that a BJT such as the SS8050 (Hfe=250, rHfe=57) behaved much better at the corner case than a 2N2222 (rHfe=25) or 2N3904 (rHfe=7), although every transistor did a pretty decent job rectifying over 100mV peak. SS8050 did a sharp transition from inverting to non-inverting even below 20mVpeak, while 2N3904 "rounded" the bottom corner. I guess the rHfe has something to do here eventhough this BJT isn't working in reverse.

Source impedance must be low, and load impedance must be high.

The saturating nature of NPN means the load seen by source changes drastically and the drive current is asymmetric, which means some DC restoration will occur.

That's when using the capacitive coupling (It seems that the schematic exposed at the link is a bit unfortunate), but you get an acceptable symmetric rectifying if you DC bias the source properly and set the inverting gain to just 1.

This circuit is rather old. If i can recall correctly, It was used in the Moog synthetizer for converting sawtooth waves into triangles (that is, performing an ABS function).

I have succesfully made and tested a circuit that converts a ramp into a parabole. It can be of use for obtaining an RMS value.

[boB]

Maybe consider a Log-Amp?

Best,

I was thinking that too. Maybe this one but it also contains a VCA.

https://www.thatcorp.com/datashts/THAT_4305_Datasheet.pdf

[PCB.Wiz]

But, when I tried a circuit like this (without fb) in a breadboard, I found that a BJT such as the SS8050 (Hfe=250, rHfe=57) behaved much better at the corner case than a 2N2222 (rHfe=25) or 2N3904 (rHfe=7), although every transistor did a pretty decent job rectifying over 100mV peak. SS8050 did a sharp transition from inverting to non-inverting even below 20mVpeak, while 2N3904 "rounded" the bottom corner. I guess the rHfe has something to do here eventhough this BJT isn't working in reverse.

I'd expect that more an artifact of low VCEsat, than reverse HFE operation.
Muting Transistors tend to have very high reverse HFE, and they are spec'd sub ~1ohm  RBC on
See
https://www.rohm.com/site-search/-/rohmsearch/scope/site?searchKeyword=Muting%20Transistor

[PCB.Wiz]

Is there any analog circuit which can convert small AC signals (46mV to 200mV) to DC, I know diode based half and full bridge won't works due to forward threshold voltage.

Need for this conversion is to use the out put of this DC volt as control voltage for Voltage controlled amplifier which amplifies prior mentioned AC signal.

For such AGC system you may rather look at a peak detector. (as in #4) and also see ADI's notes here
https://www.analog.com/en/resources/technical-articles/ltc6244-high-speed-peak-detector.html
For Audio you can relax the Opamp specs.

[Zero999]

To answer the original question.

What gain?
Full wave or half-wave?

Here's a full-wave precision rectifier with a gain of 10.
R1 = R2
A_V = 1+R3/R4

U1 and U2 need to be a single supply op-amp i.e. must work with the inputs at the negative rail. The LM358 will do for most applications,.

[Terry Bites]

Barrie Gilbert lives on.......