NO Problem ANYMORE with Full Wave rectifier using TL082CP. UPDATED on July 11th

[NHSA]

Hi,

Newbie here and having a problem with a full wave rectifier circuit using the Op Amp TL082CP.

Have set up the followong circuit using 10K and 5K resistors instead of the 6K8 and 3K4 resistors but guess this does not matter.

My problem is that the negative wave is rectified half of the Positive wave. Have tried during the last 3 days to get it even by changing the resitor values. Have read about how the inverting amplifier calculates gain but no result ...



Below a screenshot of the result measured in my oscilloscope. Before the full wave rectifier I have a non inverting amplifier.



Thanks for any tips and regards Rainer

[Andy Watson]

What power supply voltages are you using ?

[Zero999]

Are you using a single supply i.e. 0V for negative and +12V?

It looks like phase inversion, due to the common mode range being exceeded.

The schematic shows the LM358, so why are you using the TL082?

The TL082 won't work properly, if its inputs go within a couple of volts near its negative rail. The LM358 will work all the way down to 0V.

[NHSA]

What power supply voltages are you using ?

Hi,

I am using +-7,4 Volts

Rainer

[NHSA]

Are you using a single supply i.e. 0V for negative and +12V?

It looks like phase inversion, due to the common mode range being exceeded.

The schematic shows the LM358, so why are you using the TL082?

The TL082 won't work properly, if its inputs go within a couple of volts near its negative rail. The LM358 will work all the way down to 0V.

Hi,

Thanks. Will do the same test with some LM358.

I am using TL082 because I ahve plenty of them

Rainer

[magic]

Output should be same amplitude as input so where is the 40x gain coming from?

Check waveforms at Vin and P2.

[NHSA]

Output should be same amplitude as input so where is the 40x gain coming from?

Check waveforms at Vin and P2.

Hi,

The signal comes from a non inverting amplifier with its output delivering the 100X gain to the input.

I wrote in my original message " Before the full wave rectifier I have a non inverting amplifier. "

Will check the points you told me.

Rainer

[NHSA]

Output should be same amplitude as input so where is the 40x gain coming from?

Check waveforms at Vin and P2.

Hi Magic,

Looks like my preamplifier is doing something strange. Measuring the output of the pre-amplifier the curve is nice AC with same +- values.

There is a Voltage difference but I guess it comes from the values from R5 divided by R3. The values are R5 = ~10 and R3 = ~4K7. After writing that, I changed R3 for two 10K in parallel and now I have exactly bot curves equal on the rectifier. Thanks.

Rectifier with no input from my pre-amplifier at 100X is doing fine. See attached images

Now I need to find out why my output from the pre-amplifier when fed into the rectifier, the rectifier delivers half the voltage for the positive input and correct voltage for the for the negative output ¿? No idea where to start. The pre-amplifier is also a TL082 working as a non inverting amplifier.

Rainer




 

[NHSA]

Hi,

I guess I did not explain clearly my problem.

Non inverting amplifier, TL082, works OK creating a nice symmetric curve. I am feeding a Sine curve into the positive pin and amplifying 80X.

Violet line is input from AWG, 50 mV, and Yellow line is output at Pin 1, 4 V.



Full wave rectifier, TL082, is working OK. Feeding in a sine curve from my AWG it gets rectified correctly, e.g. negative portion is inverted correctly and has same voltage as positive portion.



Now when I connect the output of my non inverting amplifier, attach1, to the input of the full wave rectifier, attach2, then the negative portion is only inverted at half its value, e.g. 2 V instead of 4 V and that is my problem. I am clueless. What is different when feeding a sine curve from the output of my amplifier to a sine wave fed by my AWG ? I am using two TL082. One as amplifier and from there I go into the next one.

The Pins 5, 6 and 7 of the first TL082 are not connected to anywhere. Does that matter ? Both TL082 get power from the same source.



Here again the fullwave rectifier circuit



Thanks for any tips.

Rainer

[NHSA]

Hi,

I guess I did not explain clearly my problem.

Non inverting amplifier, TL082, works OK creating a nice symmetric curve. I am feeding a Sine curve into the positive pin and amplifying 80X.

Violet line is input from AWG, 50 mV, and Yellow line is output at Pin 1, 4 V.

(Attachment Link)

Full wave rectifier, TL082, is working OK. Feeding in a sine curve from my AWG it gets rectified correctly, e.g. negative portion is inverted correctly and has same voltage as positive portion.

(Attachment Link)

Now when I connect the output of my non inverting amplifier, attach1, to the input of the full wave rectifier, attach2, then the negative portion is only inverted at half its value, e.g. 2 V instead of 4 V and that is my problem. I am clueless. What is different when feeding a sine curve from the output of my amplifier to a sine wave fed by my AWG ? I am using two TL082. One as amplifier and from there I go into the next one.

The Pins 5, 6 and 7 of the first TL082 are not connected to anywhere. Does that matter ? Both TL082 get power from the same source.

(Attachment Link)

Here again the fullwave rectifier circuit

(Attachment Link)

Thanks for any tips.

Rainer

Hi,

Just to be sure I did set up the whole enchilada with two LM358 instead of the two TL082 and I get the same result except the positive is now half the voltage 

Have a nice weekend

Rainer

[Andy Watson]

I am using +-7,4 Volts

Pull-out the data sheets for the op-amps that you are using.
Op-amps require a certain voltage "headroom" in order to operate correctly. To operate correctly, the inputs to the op-amp need to be within a range of voltage - sometimes  called "common mode range" in the data sheets. Similarly, the output can only function within a defined range. However, extracting this information from the datasheets often requires interpretation (and guess-work) as to what the manufacturer is implying.

From memory, the TL08x series require approximately 2V headroom (on both power supplies) to operate to their full specification. You are trying to push a 5V signal with only 7.4V power supply - you are already sailing very close to the edge of the op-amp's operating range ... and we've not allowed for the 0.6-0.7V lost in the diodes in rectifier circuit.

I believe the LM358 claims to operate down to its negative supply, however, it has similar limitations with respect to the positive parts of the signal.

[magic]

Input common mode is at ground so that shouldn't be a problem. Output could be, but with 4V signal the outputs shouldn't swing more than ±5V and that seems OK on a ±7.4V supply. Nevertheless, the fact that TL082 and LM358 behave differently suggests that something may be amiss.

But I don't know what's wrong. I would try:
1. get more amplitude from the AWG and see if there is a particular level when things get weird
1b. if not possible, reduce gain of the preamp
2. simultaneously scope the output of the preamp and of the rectifier, see which of them misbehaves

[NHSA]

From memory, the TL08x series require approximately 2V headroom (on both power supplies) to operate to their full specification. You are trying to push a 5V signal with only 7.4V power supply - you are already sailing very close to the edge of the op-amp's operating range ... and we've not allowed for the 0.6-0.7V lost in the diodes in rectifier circuit.

Hi Andy,

Thanks. I just cranked the supply voltage up to +-15V and it remains the same. Amplification of voltage output max +- 5V. So there was headroom of -+ 10V.

Rainer

[Zero999]

The Pins 5, 6 and 7 of the first TL082 are not connected to anywhere. Does that matter ? Both TL082 get power from the same source.

That might be your problem. Op-amps can do funny things, if their inputs are not connected to anything, so it's important to correctly connect unused op-amps, otherwise they can affect other amplifiers, on the IC. Connect pins 6 & 7 together, to make a unity gain amplifier and connect pin 5, its input, to 0V.

[NHSA]

The Pins 5, 6 and 7 of the first TL082 are not connected to anywhere. Does that matter ? Both TL082 get power from the same source.

That might be your problem. Op-amps can do funny things, if their inputs are not connected to anything, so it's important to correctly connect unused op-amps, otherwise they can affect other amplifiers, on the IC. Connect pins 6 & 7 together, to make a unity gain amplifier and connect pin 5, its input, to 0V.

Hi Zero999,

Thanks for that and yes that is why I asked. From a former test many years ago, ~ 6 years, I somwhow knew it is not good to leave them unconnected. Sorry, I could have been less lazy and do a search   

Yesterday I wired the unused OpAmp as a voltage follower and still the same problem.

Today after reading your answer I wired the pins 5, 6 and 7 as you recommended and still the same.

I even changed both TL082 by two new ones and still the same assymetric rectified full wave.

Amplifier works OK alone and rectifier works OK alone e.g. not connecting the Amplifier output to the rectifier input.

I will keep investigating this. Next step is to set up the same enchilada using a LMC6484.

Have a nice weekend

Rainer

[Andy Watson]

Curious!
Can you confirm that you are using two power supplies and not one power supply with voltage splitter to derive the 0V?

[NHSA]

Hi Andy,

I am using a Korad KA3305P power supply for the voltage supply.

I am not using a voltage splitter from one battery.

I can test the result using two batteries and see what comes out then.

The negative output of channel 1 is connected to the positive output of channel 2 via a jumper lead. From here I take 0V or ground. Channel 1 delivers the + 7.4V and channel2 delivers -7.4V to my circuit.

Just measured it and I get exactly 7.4 volts from each output measured against the middle contact e.g. 0V or ground.

I have searched Internet and learned from there how to get the positive/negative supply.

Thanks Rainer

[NHSA]

Hi,

I found another curious thing.

The difference between negative and positive rectified wave is always ~ 1V.

Do I have in the full wave rectifier a problem of an amplifying the positive / negative wave and that gives me the difference ? Could that be due to the tolerance of the resistor values ? 

Yellow curve is amplifier output and violet curve is rectifier output.

Will check it changing the resistors in the circuit.

Rainer

[Pulsepowerguy]

You might find these videos helpful:

[magic]

Build the complete circuit with the amplifier and the full wave rectifier.

Put one probe between the amplifier and the rectifier.
Put the second probe on the output of the rectifier.

What happens?

[Andy Watson]

There is no reason you cannot operate this circuit at DC (... er, it is d.c. coupled throughout ... isn't it?) Make a resistive divider (or find another source) to supply the peak d.c. signal level to the input. Then follow the voltage levels through the circuit - you only need to test with +peak, 0V and -peak inputs and you should be able to determine where the offset is occurring with a voltmeter.

Could that be due to the tolerance of the resistor values ? 

Possible, but unlikely. Also, you swapped that op-amps and problem shifted to the opposite polarity - this would suggest it is not a resistor tolerance problem.

[viperidae]

Isn't the load seen at the input going to be different for each polarity? So you'd need a low impedance input to mitigate that.

[Andy Watson]

Isn't the load seen at the input going to be different for each polarity? So you'd need a low impedance input to mitigate that.

If I've followed this thread correctly, I believe there is an intervening "buffer" amplifier between the input and the rectifier.

[NHSA]

Build the complete circuit with the amplifier and the full wave rectifier.

Put one probe between the amplifier and the rectifier.
Put the second probe on the output of the rectifier.

What happens?

Hi magic,

The circuit is built complete and when all is connected and one probe is at the amplifier output = the wave is symmetric and so it goes into the rectifier. On the output of the rectifier the wave is assymetric by 1 V

Rainer

[NHSA]

Isn't the load seen at the input going to be different for each polarity? So you'd need a low impedance input to mitigate that.

The output of the amplifier is symmetric and that is what gets into the rectifier and out comes an assymetric rectified wave.

The output of the amplifier is -/+ 4.5 Volt and that goes into the rectifier.

The output of the rectifier is on the positive wave +5 V and the rectified negative wave is +4 V

Rainer

[NHSA]

Isn't the load seen at the input going to be different for each polarity? So you'd need a low impedance input to mitigate that.

If I've followed this thread correctly, I believe there is an intervening "buffer" amplifier between the input and the rectifier.

Hi Andy,

First the pins 5, 6 and 7 were floating. The first thing I did with that unused op amp was to make a voltage follower. Nothing changed at the rectifier output. I understand now. The recommendation for a low impedance input is created by the voltage follower but nothing changed as written before. So the problem is also not the input impedance, If I understood correctly. My amplifier as far as I ahve read also has a low impedance output.

Then I bridged pin 6 with pin 7 and grounded pin 5 as recommended by Zero999 and same assymetric output.

 

[NHSA]

You might find these videos helpful:

Hi Pulsepowerguy,

Thanks for the videos. Interesting to watch them and it is never a time waste to llok at them a few times. 

I knew them already before you gave me the link but my problem is not really frequency nor lag like explained in this video. My rectifier works nice, I tested up to 50 kHz and have nearly no distortion at the zero point between the waves.

On the other side what I am going to feed into my amplifier is low frequency, not more then maybe 100 Hz.

Thanks again. Rainer

[magic]

Any chance that your PSU runs into current limiting during signal peaks when both circuits are active at the same time?

It's starting to look like some paranormal activitiy

Yes, it's a good idea to just test it with DC input and probe every point in the circuit until you find out what's wrong

[Kleinstein]

The asymmetric signal looks like there is an offset voltage, no different gain for the positive an negative voltage. With a x 80 gain of the amplifier and up to some 5 mV at the amplifier input this could be 400 mV offset for the amplfier ouput. This would be about the error observed.

[NHSA]

Any chance that your PSU runs into current limiting during signal peaks when both circuits are active at the same time?

It's starting to look like some paranormal activitiy

Yes, it's a good idea to just test it with DC input and probe every point in the circuit until you find out what's wrong

Hi magic,

The PSU is reporting a current consumption of 5 mA

Thanks Rainer

[NHSA]

The asymmetric signal looks like there is an offset voltage, no different gain for the positive an negative voltage. With a x 80 gain of the amplifier and up to some 5 mV at the amplifier input this could be 400 mV offset for the amplfier ouput. This would be about the error observed.

Hallo Kleinstein,

Could be, but look at the yellow curve which is the output of the amplifier and there I have a difference of about 160 mV, why then this difference turns into ~ 1 V at the output of the rectifier ? When I feed a +/- ~4 V sinus curve directly into the rectifier there is no difference at the rectifier output ?

Thanks. Will check that by playing with my inputs.

The fact here is that my sensor will deliver very low voltage, in the mV area.

Is there an alternative for amplifying this better then using an op amp ?

At the moment I have been only feeding my amplifier as well as the rectifier with my AWG.

Will start to feed using my sensors and see what happens then. I will also analyze the amount of noise I have and try to minimize that.

Grüße Rainer

[NHSA]

Hi All,

Thanks a lot for all the comments which helps me to learn more about all this and analyze my problems.

Rainer

[Damianos]

Using AC coupling at the oscilloscope input, any offset voltage is "hidden"!
To eliminate the offsets from the signal use AC coupling in the circuit not on the instrument.

[Kleinstein]

1 V shift between the positive an negative half way would be 500 mV offset at the rectifiers input. One half would go down by the offset and the other half would go up by the offset. The TL082 OPs can have a relatively large offset. So for the amplifier it may be worth using an OP with a smaller offset. An alternative may be to adjust the offset once, by adding some constant voltatage.
If one is not interested in the DC part at all, one could have AC coupling for the signal.

It also depends on what to do with the result - the full- wave rectification may not be the best solution either.

[magic]

The asymmetric signal looks like there is an offset voltage, no different gain for the positive an negative voltage. With a x 80 gain of the amplifier and up to some 5 mV at the amplifier input this could be 400 mV offset for the amplfier ouput. This would be about the error observed.

[NHSA]

Hi Damianos and Kleinstein,

Thank Guys, for the AC coupling recommendation. Did not know what that is and searched in the internet. Now the output of the amplifier is connected to the input of the rectifier via a capacitor and the difference is far less and what is left now I attribute it to a possible Gain due to the resistors not being exactly 2:1 in ratio.

Will experiment with a trim pot instead of the fixed resistors.

Great learning here. 

Rainer

[NHSA]

Hi,

Well have been testing the circuit with a 3-axis sensor and I like te result which is exactly what I was looking for. It is very sensitive. Tested it on a base of wall mounted loudspeaker.

Also tested it on a tripod while stomping slightly with my feet on the floor. Very sensitive. It does even detect just plain walking.

Below some images of the result. The oscilloscope image is while the sensor is attached at the wall mounted loudspeaker base.

It will be powered from 2 9V batteries in order to get positive / negative voltage for the TL082.

Now working out the PCB in KiCad and then test it on the tripod of a friends tripod for an astronomical mount with telescope. My piers are not adequate for testing it as they are 6 metres high ( ~ 20 feet) and each has a weight of about 7 tons of concrete ( ~ 15.500 pounds) .

Thanks again for all your help

regards Rainer