Ok, so towards the effort of reduction of noise in my circuit, through general reading I have come across the concept of - instrumentation amplifiers better than op amps. I need to understand what exactly is this instrumentation amplifier? Do I need to order expensive ICs, or can I just make one using my MCO6004 or LMV324?

 

Hi,

They are similar to op amps but with better specs.
You can make one with three or more op amps.

 

hi John,
This PDF has information regarding IA amplifiers.
E

Added a 2nd PDF.

 

Ok, so towards the effort of reduction of noise in my circuit, through general reading I have come across the concept of - instrumentation amplifiers better than op amps. I need to understand what exactly is this instrumentation amplifier? Do I need to order expensive ICs, or can I just make one using my MCO6004 or LMV324?

There are LOTS of Application Notes and websites that explain instrumentation amplifiers at all kinds of different levels. Have you even tried looking?

You can either buy an instrumentation amp IC or make your own from several opamps. The performance of an IA IC will almost certainly be sufficiently better so as to make that the way to do.

Why do you think they are expensive? DigiKey has them for as low as $0.63 in single quantities.

It really depends on what your performance requirements are. Have you already determined that the performance specs of the 6004 or a 324 are actually good enough to satisfy your needs? If so, then you don't need a high-performance IA and you are probably only talking a dollar or so.

 

An instrumentation amplifier is generally made from three op-amps. That means it will be 73% noisier than a single op-amp.
The right way to go and the components to use depend on the application.

 

An instrumentation amplifier is generally made from three op-amps. That means it will be 73% noisier than a single op-amp.
The right way to go and the components to use depend on the application.

I didn't get that math. How did you arrive at the 73% figure? Also I had heard that the instrumentation amplifier is less noisy. If it is more noisy, why do people even go for it?

 

I didn't get that math. How did you arrive at the 73% figure? Also I had heard that the instrumentation amplifier is less noisy. If it is more noisy, why do people even go for it?

More circuitry is always noisier than less circuitry.
There are three op-amps all contributing noise, assumed to be the same amount, so the total is √3 times a single one.
However, it might not be the most significant noise source in the circuit, so it might not even matter. It may allow you to reduce noise elsewhere, so the overall noise of the circuit may be reduced.
Advantages of instrumentation amplifier over op-amp:
An instrumentation amplifier can give you a differential amplifier with (almost) infinite input resistance. A differential amplifier made with a single op-amp will always have a low input resistance.
The instrumentation amplifier configuration will give better common mode rejection (especially if fabricated on a chip along with matched resistors) - so it will be less subject to interference than a single-ended or single op-amp configuration.

 

hi John,
An IA actual IC is less noisy overall than a 'homemade' 3 OPA version.
That's why they are commonly used in low noise applications.
E

 

For very low noise applications (microphone preamps etc.) the most common circuit is an instrumentation amplifier configuration where the two "input" op-amps are made from discrete transistors, and the "output" op-amp is an ordinary low-noise op-amp.
The two "input" op-amps generate most the noise of the overall circuit as that is where most of the gain takes place.

 

For a low noise op amp take a look at the TL07xx series from Texas Instruments. They aren’t strictly rail to rail and the recommended single supply voltage is 4.5V but I’ve been happy with using them for audio, even using a single supply of 3.3V.

You will get distortion if you push the output to near the rails so use a higher supply voltage if possible

These op amps have been around for quite a while so there may be better low noise rail to rail devices available now

 

Ok, so towards the effort of reduction of noise in my circuit,

Do you know how to calculate the noise in your circuit from the op-amp specs?

 

thanks, will look into these suggestion. On a different note, my circuit has AC coupling serving also as HPF. I just noticed that if instead of putting the resistance directly into GND I put it through a diode, the noise level drops. Is it a healthy thing to do? I also noticed the amplitude of the signal drops too. Anyways I was pushing the output close to the rail value, maybe that is causing my noise.

 

Do you know how to calculate the noise in your circuit from the op-amp specs?

No, I am fairly new to this stuff.

 

For a low noise op amp take a look at the TL07xx series from Texas Instruments. They aren’t strictly rail to rail and the recommended single supply voltage is 4.5V but I’ve been happy with using them for audio, even using a single supply of 3.3V.

You will get distortion if you push the output to near the rails so use a higher supply voltage if possible

These op amps have been around for quite a while so there may be better low noise rail to rail devices available now

Thanks, I think I'll try it with +- 12V supply voltage

 

No, I am fairly new to this stuff.

The overall voltage noise at the input from the op-amp will be √(vn^2+(R.in)^2).√f
where vn is the voltage noise in V/√Hz, and in is the current noise in A/√Hz.
Then there is the noise contributions from the resistances on the input which will be √(4kTfR)
where k=Boltzmann's constant, T=absolute temperature f=bandwidth and R is the parallel combination of all the resistance at each input.
Then multiply that figure by the gain of the circuit.
so Vn(out) = √(vn^2+(R.in)^2+4ktRp+4kTRn).√f
where Rp is the resistance at the non-inverting input and Rn is the resistance at the inverting input.
If your amplifier operates at frequencies below 1kHz you also need to consider 1/f noise.
The same equations apply regardless of whether your circuit is a standard op-amp gain stage or an instrumentation amplifier.
In many cases the current noise of the op-amp is what makes the biggest difference (and the one most often omitted from the datasheets)

 

When a manufacturer claims that an op-amp is "low noise" it generally means that it will be low noise for a low-impedance source. If you have a high impedance transducer, then it might not be low-noise at all!

 

How about posting your schematic and explaining what you are trying to do? For example, is this audio or higher frequency, and as just mentioned, what is the impedance of the input transducer?

 

How about posting your schematic and explaining what you are trying to do? For example, is this audio or higher frequency, and as just mentioned, what is the impedance of the input transducer?

Sure, will post a schematic once I settle on one. Thing is right now I don't have a schematic, I am continuously changing it on the breadboard. Don't even know what is the current configuration, will have to check it, but the noise is there no matter what configuration. Large caps help reduce it, but large amplification exasperates it. The transducer is a phototransistor.

 

Sure, will post a schematic once I settle on one. Thing is right now I don't have a schematic, I am continuously changing it on the breadboard. Don't even know what is the current configuration, will have to check it, but the noise is there no matter what configuration. Large caps help reduce it, but large amplification exasperates it. The transducer is a phototransistor.

Are you talking about interference, rather than noise?
If so, what does the interference signal look like? What is its frequency?

 

Sure, will post a schematic once I settle on one. Thing is right now I don't have a schematic, I am continuously changing it on the breadboard. Don't even know what is the current configuration, will have to check it, but the noise is there no matter what configuration. Large caps help reduce it, but large amplification exasperates it. The transducer is a phototransistor.

Are you talking about interference, rather than noise?
If so, what does the interference signal look like? What is its frequency?

Even a photo of an outline sketch would help. If we knew the phototransistor you are using, the frequency range and the output you are looking for from the op amp there are lots of eager contributors to this forum who will gladly point you in the right direction,