Looking for a hi Q voltage controlled bandpass filter....

[Learner]

voltage controlled bandpass filter

Hi guys,
I am looking to build a audio bandwidth spectrum analyzer and I am looking for a most compact/simple circuit that can do this eg. using a monolithic IC voltage controlled bandpass filter.

I have done a search on all the major semiconductor manufacturers but without any luck, all I have found are either clcok controlled or resistor controlled filters. Am I missing something here or that is really the only thing available?

I guess what I can do is to use digital potentiometer in replacement or the resistors or I could use a variable clock source to make the filter sweep controlling the Fc, either way I would need a variable clock source.

Which has lead me to the chip Max038 funtion generator, I am planning to use 2 of these chips in order to be able to generate a sweeping clock pulse. The first will have to generate a ramp wave in order to trigger the second chip which generates the clock pulse, however this is where I am stuck.....

I can not find any where in the data sheet that tells me how to gererate a ramp wave, can anyone help me out?

What are the other approaches that has been done since the birth of spectrum analyzer? Am I on the right track?

Any thoughts will be greatly appreciated!!

Datasheet removed, use this link instead:
hxxp://pdfserv.maxim-ic.com/en/ds/MAX038.pdf
/Cl

 

[flatulent]

bandpass filter ic

One option you have is switched capacitor filters. Their frequencies are set by a clock. Linear Technology IC companies make these.

Another option is to use a direct conversion type of superheterodyne receiver with a DC IF.

 

[Learner]

high q bandpass filter

One option you have is switched capacitor filters. Their frequencies are set by a clock. Linear Technology IC companies make these.

Another option is to use a direct conversion type of superheterodyne receiver with a DC IF.

Thanks flatulent!!!! :sm9:

Currently I am just trying to explore my options and see what would be the most simplest and effective method, am not quite farmiliar with the superheterodyne method would you be able to direct me to a source?

Thanks again! :spoko:

 

[flatulent]

learn q filter

The superheterodyne method is as follows. You multiply the input signal by a sine wave. Then low pass the output with the bandwidth you want (low pass BW is half of band pass BW) You then multiply the low pass filter output by the same local sine wave and get the output at the same frequency as the input.

That is the simplified description. Add to that the fact that the multiplies are quadrature. That is, there are two multipliers, they are fed from the local sine wave and itself shifted by 90 degrees. There are two IF strips and then two quadrature multipliers at the output.

This is much more expensive than the switched capacitor filter but the SC filter produces clock noise in the output.

If your system already has an ADC and DAC you can do the superheterodyne method in math.

 

[strabush]

voltage controlled low pass filter ic

Another way is to discard the analog approach, sample the signal and compute its spectrum using FFT. Numerous free software packages exist in the web doing just that, using the PC's sound card as input.

 

[echo47]

voltage controlled bandpass

Back in the early 1970's, HP made the model 3580A spectrum analyzer. It covered 5Hz to 50kHz using traditional swept LO, mixer, 100kHz IF, and some really nice 100kHz bandpass filters (down to 1Hz resolution) using a fancy multi-stage synchronously tuned crystal filter. Oh boy, you don't want to re-invent that beast! I recommend strabush's PC approach, or perhaps build a small board with an A/D converter and a DSP running the FFT.

 

[Learner]

voltage controlled nitch filter

The superheterodyne method is as follows. You multiply the input signal by a sine wave. Then low pass the output with the bandwidth you want (low pass BW is half of band pass BW) You then multiply the low pass filter output by the same local sine wave and get the output at the same frequency as the input.

Hi flatulent,
Thanks for taking the time out to explain the concept, I have tried to read a book on this topic I think it was called Construction of Spetrum Analyzer(forgot the name of the book..) but as an EE newbie I am stilling trying to grasp the concept eventhough I been a student of EE of 2 years.......

In the explanation above, it sounds like amplitude modulation the input signal but I don't understand the purpose of LP filtering the output and AM it again will achieve?

Sorry about the newbieness...... :?

That is the simplified description. Add to that the fact that the multiplies are quadrature. That is, there are two multipliers, they are fed from the local sine wave and itself shifted by 90 degrees. There are two IF strips and then two quadrature multipliers at the output.

So I guess I would need 2 phase splitter chip or a TI dvr143 chip for the sine wave and the input signal to drive the + and the - terminal of the multiplier?

This is much more expensive than the switched capacitor filter but the SC filter produces clock noise in the output.

If your system already has an ADC and DAC you can do the superheterodyne method in math.

I am aiming to build a digital controlled analog circuit because I am not quite competent yet at programming a DSP chip to do the processing not to mention understanding the math involed in FFT....

My analysis and design project is to build a spectrum analyzer to display the frequencies of music notes up to 8 octaves, I am plannig to use a ucontroller to generate 7 clock pulses(7 notes of the highest octave) and then use divided by 2 counters to create the clock rates of the lower 7 octaves to control the Fc of the bandpass filter.

I am thinking of using commutational switches to switch the output and the clock rates controlling Fc of the bandpass filter in a synchronized manner, then use a **broken link removed** with a buffer at each Q out to control each switch. Will use the highest clock frequency possible controlling the D type shift registers but I am worried about the switch noise might distort the output of the bandpass filter?

Any thoughts?

 

[echo47]

ebay fft spectrum analyser

I think you would learn a lot from this famous application note:
AN-150 - Spectrum Analysis Basics
**broken link removed**

 

[Learner]

voltage controlled bandpass filter maxim

Back in the early 1970's, HP made the model 3580A spectrum analyzer. It covered 5Hz to 50kHz using traditional swept LO, mixer, 100kHz IF, and some really nice 100kHz bandpass filters (down to 1Hz resolution) using a fancy multi-stage synchronously tuned crystal filter. Oh boy, you don't want to re-invent that beast! I recommend strabush's PC approach, or perhaps build a small board with an A/D converter and a DSP running the FFT.

I have just recently bought a HP3581 off ebay to use it with my CRO, since it is the cheapest way to have a spectrum analyzer


**broken link removed**

Am looking forward to get it and hook it up.

I am trying to apply my limited experience in analog and digital application aiming to gain some practical experience and looking to learn something on the way by building a project, and may be use this as a reference for my future employment....

 

[echo47]

synchronously tuned crystal filter

I have just recently bought a HP3581 off ebay to use it with my CRO, since it is the cheapest way to have a spectrum analyzer

Good for you! Buy cheap instruments from eBay, fix them up, and then start playing ... oops ... learning.
I have a bench full of great eBay stuff. For me, the most educational has been the HP 8505A network analyzer.

Yes, the HP 3580 and 3581 are cousins. One has a CRT, the other a meter. Nice analog instrument. You will want to get the Operating and Service Manual.

 

[flatulent]

band pass high q

The low pass filter controls the bandwidth of the tunable filter. Only signals that are inside it get put back out again at their original frequency.