Measuring Current Noise Density

[undergradphys]

I have a power source that is supposed to be low noise, and i want to verify that the data the company sent me is indeed accurate. In order to do this, i want to reproduce their graph showing current noise density (nA/root(Hz)) on the Y axis, and Hz on the X axis. I am an undergraduate physics student, and i am not familiar with making these types of measurements.

The method i have been using is connecting the source up to a dummy load (provided by the company for testing), to a pre-amp, to a oscilloscope. I took data off the oscilloscope but had issues in trying to understand how to get the correct values on the axis. An FFT seems needed, btu i am using igor for data processing and It only requires one set of data for an FFT and doesn't seem to require a time axis which perplexs me. It also doesn't help that i have not studied the math behind FFT's quite yet.

Does anyone have any suggestions? I have access to a Spec. Analyzer if that's useful, but it goes from 9KHz to 29.5Ghz, and the graph id like to reproduce goes from 0 to 10^5Hz.

 

[yungman]

Yes, you need a spectrum analyzer to do this.

You need to first look at the spec of the of your power source. It is a voltage source or a current source? If it is a voltage source ( voltage supply) then you don't worry about the current noise density as it depend on how much current you draw. You only worry about the voltage noise.

You need to give more info on the supply you want to test, the output characteristics, the noise spec.

 

[berkeman]

I have a power source that is supposed to be low noise, and i want to verify that the data the company sent me is indeed accurate. In order to do this, i want to reproduce their graph showing current noise density (nA/root(Hz)) on the Y axis, and Hz on the X axis. I am an undergraduate physics student, and i am not familiar with making these types of measurements.

The method i have been using is connecting the source up to a dummy load (provided by the company for testing), to a pre-amp, to a oscilloscope. I took data off the oscilloscope but had issues in trying to understand how to get the correct values on the axis. An FFT seems needed, btu i am using igor for data processing and It only requires one set of data for an FFT and doesn't seem to require a time axis which perplexs me. It also doesn't help that i have not studied the math behind FFT's quite yet.

Does anyone have any suggestions? I have access to a Spec. Analyzer if that's useful, but it goes from 9KHz to 29.5Ghz, and the graph id like to reproduce goes from 0 to 10^5Hz.

I agree that more info is needed. And why would the power supply company need to furnish the dummy load for testing? There must be some things that are very special about this power supply?

 

[sophiecentaur]

Does anyone have any suggestions? I have access to a Spec. Analyzer if that's useful, but it goes from 9KHz to 29.5Ghz, and the graph id like to reproduce goes from 0 to 10^5Hz.

A spectrum analyser for use up to 29GHz is likely to have a worse noise figure than one designed for signals less than 1MHz. This could be a limitation unless you can get hold of a suitable low-noise pre-amp. This can also apply when using an oscilloscope (time domain measurements).

Noise measurement can be difficult when very low levels are involved and there are some clever methods for getting the best accuracy. It would all depend upon what levels you are concerned with.

 

[f95toli]

A MW spec analyzer would also probably break if your connected it to a DC supply, they are usually rated for 0V DC (but can still have DC coupled inputs). If would STRONGLY advise against even trying. Also, most of the noise you are interested (the 1/f noise) in will be below 9 kHz.

What you need is a FFT analyzer. If your supply is very noisy and you have access to a good preamp you might be able to get away with using a DAQ and some software. However, a stand-alone benchtop instrument (often called a dynamic signal analyzer) will have much lower noise-level and is more suitable for "proper" noise measurements.

See e.g.
http://www.thinksrs.com/products/FFT.htm

Btw, measureing noise is far from trivial and there are about a million things that can go wrong (wrong cables, ground loops, interference etc etc), ideally you need someone to teach you how to do it.

 

[undergradphys]

What kind of info are you looking for? I am a layman when it come's to this topic. The only thing that is special about this power supply is that it was made to power our system specifically and was constructed to have low noise. The configuration of the power supply is it can be powered by batteries OR by a wall input. We usually run it on batteries, and the supply is said to reduce the inherent noise. I apologize now since i believe my vocabulary was incorrect in calling it a power supply, it is actually a driver.The output characteristics are it is an adjustable voltage supply which can range from 0-10 volts, and i have been testing it at around .5v DC. It is a current source however as well as it is used to power a laser.

I am re-reading the technical note the company sent, and it says that below 100kHz a FFT analyzer recorded the noise density performance and above 100kHz an RF spec analyzer was used. This is in line with what you have said f95toli. I am afraid i don't believe the lab i work at has a FFT analyzer, although i can check.

 

[undergradphys]

I have found a spec analyzer that is rated from 100Hz to 9MHz (might be more on the higher side, my memory is poor) Which is a suitable range for the measurements i am looking to do.

How would i measure the Current Noise Density using a spec analyzer? Is there a simple operation?

By the way, i am looking to confirm levels of around 10-1 nano amps per root hertz of noise. that is the level of noise i am working with.

EDIT: Yes i am using low loss short cables, to clear that up as well.