Half wave rectifier

Thread Starter

strathy

Joined Mar 26, 2008
10
I hate to jump in and jack this discussion, but I was wondering if someone could offer me some advice about which op amps to use.

I am trying to build a HALF wave precision rectifier circuit. I managed to test the circuit and it worked fine using a standard LM741, however, in my completed circuit I can't use this and must use one with a smaller supply voltage. Eventually I chose to use an LMC6041N.

However, since I'm not entirely sure how op-amps work I was wondering if someone else could help me with this problem.

The LM741 produced a perfect half wave rectified signal, replacing it with the LMC6041N however gives FULL wave rectified signal at the output.

Does anyone have any idea why this might happen?

The circuit I'm using is identical to this one: http://www.play-hookey.com/analog/half-wave_rectifier.html
 

beenthere

Joined Apr 20, 2004
15,819
Now you're no longer a hijacker. Please do start you own thread.

What value resistors are you using, and is the op amp in inverting or follower configuration? What are you doing with the other input pin?
 

Thread Starter

strathy

Joined Mar 26, 2008
10
I apologise for the picture I linked you to. This is what I've built:

http://www.ecircuitcenter.com/Circuits/op_HW_recitifier/Op_HW_Rectifier.htm

I'm using 10K resistors for R1 and R2.

I'm having to think back a few years now to when I was actually taught this stuff, but I think it's connected in inverting mode.

The bit I'm struggling with is what the difference is between the 2 op-amps I'm using. I can't understand why a 741 produces a perfect half wave then as soon as I change it I get an almost full wave.

strathy

Edit: THe non-inverting pin is grounded
 

beenthere

Joined Apr 20, 2004
15,819
That's a bit hard to understand. I've used that same setup plus another stage to make a full wave rectifier. A couple of different op amps, too.

You did say something about smaller supply voltages - what are you using now? Something has to make the difference.
 

Thread Starter

strathy

Joined Mar 26, 2008
10
Well previously I was using an LM741 so I supplied it from a protoboard at +/- 15V.

However, the final circuit that I will use it in doesn't have 15 so I got another amplifier to use. I'm now testing it with +12V and 0V as the V+ and V- (respectfully) before I put it on the PCB.

The only thing that I thought might be a problem would be the Op-Amp saturating when it's output tried to go negative but:

1. It shouldn't do that and
2. If it did it surely wouldn't magically turn into a full wave rectifier should it?

Maybe I should market this circuit I seem to be on to a good thing...
 

Audioguru

Joined Dec 20, 2007
11,248
The half-wave rectifier circuit needs a dual polarity supply.

Maybe you are exceeding the allowed input common-mode voltage range when the input signal goes positive (2.3V less than the positive supply voltage) which might cause the output of the opamp to suddenly invert and create the full-wave rectification.
 

Ron H

Joined Apr 14, 2005
7,063
As Audioguru says, you also need a negative supply. In order have a half wave rectifier, the virtual ground at the inverting input needs to be maintained. During the positive input half cycle, D2 can't conduct, because the output needs to go a diode drop (~0.6V) below ground. Therefore, the inverting input has no feedback, so the input shows up at the output through the two series resistors. If you make the signal amplitude larger than one diode drop, you will see alternate half cycles clipped to a diode drop.
 

Thread Starter

strathy

Joined Mar 26, 2008
10
Hmmm, I see.

Would a full wave precision rectifier work using a mono-polar supply?

If not does anyone know of any other options??
 

Thread Starter

strathy

Joined Mar 26, 2008
10
As Audioguru says, you also need a negative supply. In order have a half wave rectifier, the virtual ground at the inverting input needs to be maintained. During the positive input half cycle, D2 can't conduct, because the output needs to go a diode drop (~0.6V) below ground. Therefore, the inverting input has no feedback, so the input shows up at the output through the two series resistors. If you make the signal amplitude larger than one diode drop, you will see alternate half cycles clipped to a diode drop.
So the op-amp has to have a negative supply big enough to 'supply' the diode's voltage drop during the 'off' part of the half wave rectification?
 

Ron H

Joined Apr 14, 2005
7,063
Hmmm, I see.

Would a full wave precision rectifier work using a mono-polar supply?

If not does anyone know of any other options??
If your signal is a sine wave (or other 50% duty cycle signal), you can shift the signal to half the supply voltage with AC coupling, do the rectification, then shift it back to zero volts. See below. It ain't pretty, but it works, at least in principle.
 

Attachments

Thread Starter

strathy

Joined Mar 26, 2008
10
Yeah, there's something that I haven't mentioned though.

I have major space restrictions and also the only available supply rails I have are 0V and 12V.

Basically I only JUST have space for 1 op-amp chip (DIP 8) and a couple of resistors and diodes for the rectification. All I need is a DC signal. I went for half wave because that circuit had the least components (cheaper and a smaller footprint).

It doesn't have to be particularly pretty, it just has to be compact and work with a +12/0V supply.

Edit: Also my signal is a sine wave. I'm passing a mains rated live wire through a current transformer. So it's a 50Hz sine wave. And input voltages to the op-amp will be a maximum of ~2.5V.
 

Ron H

Joined Apr 14, 2005
7,063
Yeah, there's something that I haven't mentioned though.

I have major space restrictions and also the only available supply rails I have are 0V and 12V.

Basically I only JUST have space for 1 op-amp chip (DIP 8) and a couple of resistors and diodes for the rectification. All I need is a DC signal. I went for half wave because that circuit had the least components (cheaper and a smaller footprint).

It doesn't have to be particularly pretty, it just has to be compact and work with a +12/0V supply.

Edit: Also my signal is a sine wave. I'm passing a mains rated live wire through a current transformer. So it's a 50Hz sine wave. And input voltages to the op-amp will be a maximum of ~2.5V.
What's the minimum voltage you want to be able to handle? And what are you going to do with the output?
 

Pich

Joined Mar 11, 2008
119
Try this, If you have room for 3 more components you could use 2 5 volt zener diodes and a resistor to limit the current. The resistor and the diodes are connected in series so that both diode will drop 5 volts, the GND is connected to the center point of the 2 diodes then you will get +- 5 volts at the other ends of the diodes. The 12v volts used is very marginal for running the 2 zeners but it may work, 15v would be much better also the resistor will determine how much current is available to the circuit and how much power it will draw from the power supply. You probably will not need more than 20ma for the circuit. If you are using 12v try a 68 ohm resistor to start.
 

Thread Starter

strathy

Joined Mar 26, 2008
10
What's the minimum voltage you want to be able to handle? And what are you going to do with the output?
Well theoretically I would like to rectify as small a voltage as possible, but the smallest voltage I would like to rectify accurately would be about 0.5V.

The voltage represents a measurement of current drawn from a home appliance (using a CT). The rectified output will be going to a microcontroller (micaZ sensor mote) to give me a read out of power used by the appliance.

It's all part of a project on demand side management that I'm doing at university.
 

Thread Starter

strathy

Joined Mar 26, 2008
10
Try this, If you have room for 3 more components you could use 2 5 volt zener diodes and a resistor to limit the current. The resistor and the diodes are connected in series so that both diode will drop 5 volts, the GND is connected to the center point of the 2 diodes then you will get +- 5 volts at the other ends of the diodes. The 12v volts used is very marginal for running the 2 zeners but it may work, 15v would be much better also the resistor will determine how much current is available to the circuit and how much power it will draw from the power supply. You probably will not need more than 20ma for the circuit. If you are using 12v try a 68 ohm resistor to start.
Well the circuit so far is as follows:

I have mains voltage connected to a transformer that gives me out 15Vac. This is then put through a bridge rectifier chip that gives me out about 17 or 18V. Then I use a zener diode to give me a 12V supply.

So I could use the pre - zener rails to supply the new zeners. Although it would make my PCB into even more of a rat's nest than it already is!
 

Ron H

Joined Apr 14, 2005
7,063
Are you trying to rectify in order to increase your resolution? Otherwise, you could just bias one end of the CT at Vcc/2, and A/D convert the 50Hz signal directly, assuming you have control of the conversion process.
 

beenthere

Joined Apr 20, 2004
15,819
The use of an active rectifier may be unnecessary. A single diode and capacitor may give plenty of precision to track the current.
 

Ron H

Joined Apr 14, 2005
7,063
Beenthere, I don't think he would get accurate results (or any results) with the requirements he posted:
Well theoretically I would like to rectify as small a voltage as possible, but the smallest voltage I would like to rectify accurately would be about 0.5V.
 

Thread Starter

strathy

Joined Mar 26, 2008
10
Thank you all for your input and suggestions.

However, I've found a way around the problem without having to rectify the signal.

Now that that panic is over it's onto the next sumbling block in this project :)
 

Ron H

Joined Apr 14, 2005
7,063
Thank you all for your input and suggestions.

However, I've found a way around the problem without having to rectify the signal.

Now that that panic is over it's onto the next sumbling block in this project :)
I'm glad you found a solution, but damn! :D I just finished a design. I'll post it anyway.:cool:

Edit: BTW, I breadboarded this, and it works well.
 

Attachments

Top