Analog Devices' Matt Duff describes how to convert RMS noise into Peak-to-Peak noise. Distributed by Tubemogul.
Read the full transcript »
Calculating RMS Noise to Peak-to-Peak Noise Hi, my name is Matt Duff. I am an Applications Engineer covering Precision Amplifiers specifically Instrumentation Amplifiers. And today, we’re going to go over a real world example of how to calculate noise with an instrumentation amplifier circuit. This is something that comes up often with customers that I talk to about how to calculate noise, and so I wanted to show you a real life example. So what I’ve drawn up here is a very standard instrumentation amplifier circuit with a bridge configuration. Now, the bridge resistances that I’ve used are quite a bit larger than you’ll typically see in an all bridge application but I’ve made them bigger so that we can illustrate some concepts. But the actual idea of using instrumentation amplifier with the bridge to measure pressure or 0:48 or something like that is very, very common. So you’ll noticed I’ve started with five Volts and ground on the bridge but when we do a noise analysis, what we can do is we can get rid of, I can turn any independent voltage sources into a ground. So what I’m going to do is I’m going to turn that from 5 Volts and I’m going to make that ground. Now you’ll notice now that I have these two resistors are in parallel, so I’ve connected ground, ground, and it’s going to this input. And I have these two resistors in parallel so what these resist… 250 K’s in parallel equals 25 kilo-Ohms. So what we can do is we can go from this type of circuit to this circuit over here. I’m drawing my grounds, now we can do the noise analysis on this circuit here. So when you do a noise analysis for an instrumentation amplifier, there are three things that you need to calculate, and this is not just for an instrumentation amplifier but for any type of amplifier. You need to calculate the ‘Resistor Noise’. You need to calculate the ‘Voltage Noise’ of the amplifier, and you need to calculate the ‘Current Noise’ of the amplifier. So that’s what we’re going to do. So first were going to cover ‘Resistor Noise’. Now remember a one kilo-ohm resistor has four nano-Volts per Hertz of noise. We can use this to figure out what the noise of these resistors are. So if a one kilo-Ohm resistor has four nano-Volts, we can take the square root of 25, we get five and multiply that by four to figure out what the noise of these resistors are. So noise of these resistors are 20 nano-Volts per Hertz, 20. So we go through, we calculate with 20, and we get the total noise from our resistors as 28 nano-Volts per Hertz. The next step that we want to calculate is the ‘Current Noise’ of the instrumentation amplifier. So we go to the data sheet of AD8226. And we find out that the ‘Current Noise’ going into the inputs of the AD 8226 is 100 zepto-Amps per Hertz. 100 zepto-Amps per Hertz. So this 100 zepto-Amps multiplies by the 25 kilo-Ohms so I’ve got here, I’ve shown this, I’ve put 0.1 because we’re in, I want to convert Zepto to Nano. So 0.1 times this 25 and you do the math and you end up getting 3.5 nano-Volts per Hertz contributions from the ‘Current Noise’ multiplied by the resistances. So now we’ve got the ‘Current Noise’. The next step is to calculate the ‘Voltage Noise’ of the instrumentation amplifier. And instrumentation amplifier is a little bit strange compared to say, a standard 3:48 in that they typically have two noise specs. So I’ve got a noise spec if the output, so if we were to go to data sheet, it would say that our output noise or our ENO is 120 nano-Volts per Hertz and our input noise over here is 22 nano-Volts per Hertz. Now the funny thing about instrumentation amplifiers is this noise changes with the game so, or I guess when we’re talking about input referred, this noise tasty constant and this noise goes down with the game. So what I’ve said in the circuit is we’re going to use again a 10 for this instrumentation amplifier, so what we’re going to do is we’re going to, this 120 nano-Volts per Hertz, we’re going to divide by 10. So now