Help Getting Band-Pass Filter Circuit to Work

[tehipwn]

Hello everyone. I am trying to implement a band-pass circuit with a narrow passband at 10kHz on breadboard and for some reason I can't get it to work. I am using the design procedure outlined in pages 6-7 in the following Texas Instruments pdf:
Filter Design in Thirty Seconds

This can be seen in the attachment as well.

I went through the simple design procedure and used C1=C2=0.01 microfarards (non-polarized). Then I went through and calculated the R values. Then finally for Cin and Cout I am using 1 microfarard (non-polarized).

I simulated the circuit and it worked out in PSPICE.

However, it does not work when I hook it up. I use a 741 and hook up the circuit as shown. For Cin and Cout I use non-polarized capacitors if that matters. At the input I hooked up a function generator and hooked up an oscilloscope at the output. I adjusted the function generator's output level to a 2.5V amplitude square wave. I then adjusted the frequency from around 100Hz to 11kHz and the output waveform on the oscilloscope didn't change much. The output just looks very noisy at all frequencies and the signal amplitude remains small for all frequencies. Some times I noticed that the signal keeps getting smaller as time goes by until it gets to zero.

To my understanding the signal should be about zero for any input frequency outside the filter bandwidth and should pass through a relatively undistorted squarewave signal at the passband.

I know this is not much description of the problem, but if anyone can think of any advice I'd appreciate it.

Thanks.

 

[ronv]

What do you have attached to +supply/2?

 

[tehipwn]

I thought that might be a problem. In the lab I was just hooking a 2.5V supply from the bench dual power supply. Do I need to use some other type of supply?

I've seen in another TI pdf the way to generate a Vsupply/2 with an opamp. If this is the method could you offer any pointers?

 

[MikeMl]

Also, the 741 will not work on a 5V single supply. Dump the 741 and get a modern low-voltage, rail-to-rail op-amp.

 

[tehipwn]

Also, the 741 will not work on a 5V single supply. Dump the 741 and get a modern low-voltage, rail-to-rail op-amp.

Oh, that's useful information. On the +Vsupply/2 do I need a more sophisticated supply voltage or will a voltage divider work?

I really hope that's what it is. I know a quick google would do the trick but do you recommend any single +5V supply op amp ICs?

 

[MikeMl]

I like the TLV246x family.

 

[tehipwn]

Thank you very much for your help. I didn't realize some op amps wouldn't work with a single supply. I am ordering a some today in the single op amp, dual, and quad just in case I need them.

Thanks again.

 

[tehipwn]

I have the correct opamps you specified on order. Just to try and make sure it goes smoothly when they arrive I would appreciate some clarification on some other aspects before I go to the lab and run into problems again.

1. Should I use a simple voltage divider from the supply voltage to give the Vsupply/2 or should I use another method?

2. Does it it matter if the input and output capacitors are polarized or not? If so, could you clarify which way the polarity needs to be in my circuit above? If not, are there any advantages to using polarized capacitors such as an electrolyctic?

3. If it functions properly, should I suspect a relatively undistorted squarewave signal at 10kHz at the output, while signals at frequencies falling outside the passband should be attenuated to a flat zero signal?

Thank you very much for answering even one of the questions!

 

[MikeMl]

On the voltage divider: The circuit you posted "loads" the voltage divider tap. You have two choices: make the divider out of power-burning low-value resistors chosen to make the divider have a much lower output impedance than the load, or use an opamp as a voltage follower to buffer a voltage divider made out of high value resistors.

Minor nit; a 741 will work on a single supply, but not one as low as 5V. It takes about 12V minimum. At that, the 741 is still a crummy opamp.

If the input and output of your circuit is referenced to 0V (Vss of the opamp), then the + should be on the right plate of the input capacitor, and on the left plate of the output capacitor.

 

[ronv]

1/2 vcc

Lucky for you you bought some duals. Use the other half of one to make your voltage.

 

[ronv]

Are you inputting square waves or sine waves?

 

[tehipwn]

Thanks ronv for the 1/2Vcc generator. I am inputing square waves from a 555 timer.

I'm building a full analog color sensor using a comparator, LEDs, and phototransistors as the sensor. It works right now to sense red, green, and blue but it's highly influenced by ambient light conditions. So I'm pulsing the emitter LEDs at 10kHz using the 555 timer and filtering at the gate of the phototransistors using this filter.

I'd think that would make it totally independent on ambient light...hope it works...

Can you guys think of a good non-microprocessor way to alternate the emitting LEDs to turn on and off in sequence so that they're not all lit all the time? If you know of an IC with 3 outputs or something that you can set up to output high in a sequence I'd could do the research.

 

[ronv]

Might be ok. Your output will look more like triangles cause the edges of the square wave are high frequency and the tops low frequency, but I suspect the photo transistor output is a little more like a sine wave anyway. Maybe just a counter driven by the 555 would put you in sequence. You would probably need to power up the counter with something like a 2N7002 to drive the leds.

 

[MrAl]

Hi,

If you need to sequence without a uC you can try a 74LS138 with a binary counter. You'll set the counter up to count 0,1,2, then back to 0, and that will provide you with three sequenced outputs.