How do you get the bandwidth from a frequency sweep?

[yankeebush]

I had to design and construct a real-time DAQ system to measure the characteristics of a simple audio power amplifier. This involved building a control panel to meet system requirements, performing spectrum analysis and determining the distortion figure of merit, and characterizing the frequency response of the amplifier and determining the half power bandwidth.

Could someone suggest how I do this next part? Do I do a frequency sweep? Do I record the info using a write measurements file in my block diagram or is there a VI I could use to add to my block diagram that would calculate the bandwidth, Could someone give me a suggestion?

     Obtain frequency spectra of the amplifier with Vi=0.02x10^-3 sin wt from 50Hz and up.

o       From the spectra, obtain the 3dB or half power bandwidth(when the output power is reduced by 0.5 of the input power or by -3dB; or when the amplitude ratio is reduced by 70.7%. The difference between the 2 frequencies is called the 3dB bandwidth.

[johnsold]

Your problem has several parts, which can be separated: Excitation of the DUT (Device Under Test), measurement of the response, calculation of the bandwidth, and presentation (and saving) of the results.

Which parts are you having trouble with? What equipment (signal sources and measurement devices) do you have? What version of LV are you using? What are the requirements for accuracy, repeatability, and time of test? What formats do you need to display or save data? Is the system to be highly automated or does it have significant user interaction (selecting tests, connecting DUT, naming files, etc.)?

You will not find a single VI which does everything for you, but LV make it easy to build a customized test system from many subVIs.

Lynn

[yankeebush]

The project is posted on my site in the pdf section under Real Time DAQ Project. Could you download it, it explains the project and shows what I have so far. Here's my site where you can download the pdf file: http://yankeebushsoftware.net/index.cfm?id=113&fuseaction=browse&pageid=32

[johnsold]

Your circuit looks like a good distortion generator. What does the SPICE analysis show? If you are not sure what is wrong with the circuit, tell me what you have found wiht your analysis and I will help you try to figure it out.

Is this a school project? If so, it does not mean that you will not get help here. However, we will not do your project for you.

The block diagram is pretty fuzzy in the .pdf file. It looks like you are on the right track. One way to sweep your frequency is to set the frequency on each iteration of the loop and add or multiply the current frequency with another term to create the next frequency using a shift register.

Lynn

[yankeebush]

I was going to put a manual loop around our continuous loop, change my continuous loop to fix at 1000 samples. When it runs we get 1000 samples and the Vi waits for use to change the freq and attains the max peak db power and saves it to a shift register., we hit a button and we get another sample block. Once we are done another button allows a frame to evaluate all of our collected data finding peak DB per Freq step, we then have one peak for each frequency step. Not a hard change and very simple to control. Do you think this will work?

[johnsold]

Sounds reasonable.

Lynn