measure centre frequency and bandwidth of VCO

[haohaodk46]

Hi all!
I want feedback of VCO frequency about 2.1Ghz.What do the method measure this frquency,anyones know ?
Thanks !

 

[haohaodk46]

Help me! It is very important with me.
Thanks!

 

[betwixt]

2.1GHz is well within the range of many frequency counters. If you have a divider circuit atached to it, as maybe the case if it's PLL controlled, you can measure the divided frequency if it's easier then use a calculation to multiply it back to the original.

Brian.

 

[haohaodk46]

Thanks ! But Fequency variable continuos interval delta t.Eg Freq variable :2105-2170Mhz(bandwidth = 70 Mhz). So to measure this centre frequency and bandwidth is difficult .How can I measure this frequency ?

- - - Updated - - -

Same attach pciture,how to measure centrer freq and bandwidth?

 

[betwixt]

You didn't say it was a modulated signal in the original post!

I think you answered your own question though, look on the spectrum analyzer, you can read the center frequency and use the scale (stop frequency - start frequency)/10 to find the span per division then count them across the waveform to get the bandwidth. To find the 'dB' down point of the signal edges use the amplitude level marker.

Brian.

 

[haohaodk46]

Perhap i tranfer content unfully,I want measure frequency feedback about microcontroller.It mean that i want use a microcontroller + Graphic LCD to read frequency replace a spectrum analyer.

 

[betwixt]

This is outside my field of expertise but my first thoughts are to simply sweep across the frequency spectrum with a controlled VCO and mix the signals in a zero-IF receiver then measure the resulting amplitude -or- to down convert the frequency to something much lower and filter it like in a superhet receiver. The graph on the LCD would be frequency against amplitude in both cases. If mixed to a lower frequency it might be possible to sample it and perform an FFT on it but that depends on processor speed and the nature of the signal.

Brian.

 

[haohaodk46]

I intend use divider frequency IC and put into ADC pin of microcontroller ,After FFT algorithm to measure frequency but i dont understand this algorithm and caculate frequency from FFT function.Who done ,can help me ?Thanks!

 

[haohaodk46]

Who did done about measure frequency use FFT alorithm can help me ?

 

[RFDave]

So you're attempting to measure a signal in the range of 2100 to 2200 MHz. Do you want to have information on

1) The frequency of the signal

or

2) the spectral plot in the 2100 to 2200 MHz range?

1) is pretty straightforward, and can be done with a divider circuit feeding into a timer input on your ADC.

2) is somewhat difficult, as you need to determine information on the amplitude vs frequency over that 100 MHz range. An FFT could do these calculations. Why don't you do some calculations on the FFT resolution you need in terms of sample frequency/Sample accuracy vs the frequency resolution you need. This would help you with some design approaches.

Googling for DIY spectrum analyzer could also be useful.

 

[haohaodk46]

Thanks you for comments.I want measure to dispay feedback frequency from output of VCO.But I am difficult about measure frequency when it alway variable in a range.How to applied FFT resolution and what is the formula bettwen fft output and frequency?

 

[betwixt]

RFDave is better qualified to help you then I am but I can tell you that you will have difficulty trying to find a single frequency when the energy is dispersed over a range. Probably the best you will achieve is to find the edges of the spectrum and assume 'THE' frequency is the mid point between them. Most modulation schemes are symetrical so the original carrier frequency will be half way between the edges.

Brian.

 

[RFDave]

OK, I just looked at your plot. You are showing a spectrum analyzer display of a modulated signal. If you want to display that information, then you need to design a spectrum analyzer. It's not as simple as running an RF signal into a microcontroller and performing an FFT. For example, you will need to sample this (Assuming a semi-digital approach) at greater than 200 MHz to get the spectral information you want in a 100 MHz window. No Microcontroller had an ADC that works that fast.

 

[haohaodk46]

I intend use divider frequency IC (eg MC12079) to reduce frequency.after use FFT to caculate this frequency.I intend do this medthod.So can i do it ?

 

[RFDave]

When you go through the divider, you'll lose all of the amplitude vs frequency information you'll need to do an FFT. But it's your project, so do it like you intend to.

 

[haohaodk46]

After measure use a software multiplier to back origin frequency.

 

[betwixt]

What RFDave is telling you is that the dividing process changes the wave shape. Frequency division is done digitally so on a single frequency it works well and produces (theoretically) a square wave at f/n where 'n' is the division ratio. The frequency of a square wave is easy to measure but it's amplitude will be that of the divider, not the original input. It gets more complicated when you feed a wide bandwidth into the divider because it will give an output which may jump from one component frequncy to another. To perform FFT you need both frequency and amplitude input to be as the original.

A better solution in your case is to sweep the signal with a narrow band filter and measure the energy within the filter passband. This can be a relatively slow process which most microcontrollers can manage easily because the sweep rate can be decided by the speed of the ADC. You take a measurement then move the filter to the next section of the signal, take another measurement and so on. It is usually easier to do this by keeping the filter on a single frequency 'IF' and moving the incoming signal by mixing it with a VCO.

Brian.