Science Buddies: "Ask an Expert"

Hi,

Oh wow! I am so glad I asked you about that because I would have probably ended up spending hundreds of dollars. Thanks so much for the useful advice.
However, would it be possible for me to construct a band pass filter from those parts solely from the instructions online. Also, when I was researching it seemed that the band pass I would need would be very complex consisting of three different components in order for me to target my specific frequrncy. I just wanted to know whether or not that was the right kind of design.

Thanks again.

vikdha1 wrote:Hi,

Oh wow! I am so glad I asked you about that because I would have probably ended up spending hundreds of dollars. Thanks so much for the useful advice.
However, would it be possible for me to construct a band pass filter from those parts solely from the instructions online. Also, when I was researching it seemed that the band pass I would need would be very complex consisting of three different components in order for me to target my specific frequrncy. I just wanted to know whether or not that was the right kind of design.

Thanks again.

Hello, vikdha1!

Band-pass filter designs are rather simple. Yes, you might need to use three kinds of components - resistors, capacitors and inductors - but these are common. It's not unusual for a product to have dozens of such components.

The band pass filter is only one stage of your overall design. I see your project as requiring 4 major stages or parts:

1) Microphone and amplifier for picking up and boosting the signal

2) Band pass filter for blocking everything but the ring tone

3) Analog to digital to convert the level of your input signal to a 1 or 0 - on or off

4) A circuit to trigger a light, bell, or other output device.

Each of these stages requires you to research the design and then build the circuit. There are plenty of resources on the Internet to help you figure out how to design and build these four stages.

I hope this helps.

Hi,

Sorry to bother you again, but I have researched all the four components you have said I will require to build my project, but I cannot figure out how to combine all those components to make one central component. For example I do not know how to connect the microphone and amplifier to the band pass.

vikdha1 wrote:Hi,

Sorry to bother you again, but I have researched all the four components you have said I will require to build my project, but I cannot figure out how to combine all those components to make one central component. For example I do not know how to connect the microphone and amplifier to the band pass.

Hello, Vik!

The output of the first stage must be compatable with the input of the second stage. When doing audio signals, such as from the microphone to the audio amplifier, you must match what's called impedance. It's the same thing as stero speakers. Some are 8 ohms, some are 16 ohms, and so on. If your stero output impedance is 8 ohms, you should use 8-ohm speakers for best performance. (The reality is that 16 ohm speakers will work, but they won't be optimal.)

For other stages, such as the input to the stage that turns on a light, the output should be digital - perhaps swinging between 5 volts of "on" and 0 volts for "off".

You might want to do a little research on "Operational Amplifiers". Most Op Amps today are inexpensive integrated circuits. By themselves, they are extremely high gain, have a high input impedance, a low output impedance, and a very wide frequency response range from DC up to MHz which make them extremely useful in analog circuit design. With the addition of a DC power supply and a few resistors and capacitors, they can be tailored to a wide variety of applications. There is a whole class of analog circuit design devoted to working with Op Amps. Much of the expense of building filters can be eliminated by using Op Amps to eliminate the impedance interaction between L-R-C filters.

I agree that a high school sophomore can do the work, but you can only afford it by building most of the components from scratch.

I also agree that you should look at some of the inexpensive technology out there that has done a lot of the work already. In college we built a very sensitive vibration detector just by stripping apart a Radio Shack microphone and hooking it up to an amplifier and an oscilloscope.

The tech is a lot better now. A lot of what is challenging is just picking up the vibrations. A cheap modern microphone may do that for you. Then run the signal through an amplifier and a filter. The advice given already though is already well beyond my expertise. This is just to reinforce that you should not have to build everything yourself. Buy the off the shelf stuff that is cheap, then use components to amplify and filter the result.

Hi,

I'm sorry to keep bothering you. I'm having a little trouble. I just wanted to clarify the procedure. So first I obtain a normal microphone with a certain ohms and connect that to an op amp of the same ohm. Then I connect the op amp by an analog curcuit to a band pass filter that I have built myself. Then I finally connect that to another analog curcuit which is connected to a breadboard and then I connect a digital wire to that same breadboard which is connected to a light bulb.

Also I wondering, if I will be requiring a digital curcuit, how would create it and how would I convert it from an analog curcuit to a digital curcuit.

Would it make my project if I just purchased a receiver and tranmitter board (RLP and TLP 315).

Thank you,
Vikram

vikdha1 wrote:Hi,

I'm sorry to keep bothering you. I'm having a little trouble. I just wanted to clarify the procedure. So first I obtain a normal microphone with a certain ohms and connect that to an op amp of the same ohm. Then I connect the op amp by an analog curcuit to a band pass filter that I have built myself. Then I finally connect that to another analog curcuit which is connected to a breadboard and then I connect a digital wire to that same breadboard which is connected to a light bulb.

Also I wondering, if I will be requiring a digital curcuit, how would create it and how would I convert it from an analog curcuit to a digital curcuit.

Would it make my project if I just purchased a receiver and tranmitter board (RLP and TLP 315).

Thank you,
Vikram

Hi, Vikram!

Craig_Bridge made a good point about operational amplifiers with high impedance. My comments about matching impedance really apply to
output circuits. You will want to connect the output of your microphone to the input of an amplifier with high impedance. High input impedance is desirable so that your circuit does not put a heavy load on the input device - your microphone.

When I wrote that the output of your circuit might need to be digital, I ment that in the strictest sense of the word. It has to change from "off" to "on" when the ring tone is detected. If you want to drive an LED and you are powering the circuit with AA batteries, that might mean that your output swings from 0 volts to 3 or 5 volts.

Operational amplifiers (Op Amps) can be used to design a circuit that will drive the output in the manner described, above. In fact, Op Amps could be used to build every stage of your project circuit. (I wish I would have thought of the Op Amps at the beginning of this discussion!)

I don't know what the model numbers you listed in your last post mean. Are they devices from Radio Shack?

Mr. Castelli,

I'm really sorry to keep bothering you like this, but I still have some questions about my project. I've been looking up buliding band pass filters everywhere and I was just wondering if this site explained it well; http://72.14.253.104/search?q=cache:iOV ... cd=1&gl=us

Also I was wondering how I would use the breadboard you recommended to connect the various parts of my device. Would I be required to connect analog curcuits to the breadboard or something of that sort?

Thank you,
Vikram

vikdha1 wrote:Mr. Castelli,

I'm really sorry to keep bothering you like this, but I still have some questions about my project. I've been looking up buliding band pass filters everywhere and I was just wondering if this site explained it well; http://72.14.253.104/search?q=cache:iOV ... cd=1&gl=us

Also I was wondering how I would use the breadboard you recommended to connect the various parts of my device. Would I be required to connect analog curcuits to the breadboard or something of that sort?

Thank you,
Vikram

The web site provided is not the best I've seen. It's very theoretical. I think you'll have better results if you search on something like, "op amp bandpass". I used Google to run this and found some sites that I think would be more helpful.

I want to emphasize the point that Craig_Bridge made about the use of operational amplifiers (op amps). They will simplify the work you need to do, and you can find many cookbook-like tutorials on the web. In fact, op amp manufacturers often provide design tools on their web sites.

Breadboards are grids of holes for inserting IC, components and wires. Searching for "breadboards" will yield pictures and descriptions that will make this pretty obvious for you.

Hi Mr. Castelli,

I was searching the op amps and bandpass filters and I cam across some op amps that only cost about $2 and seemed to be pretty well developed. I just wanted to check if I was looking at the right sites and if the op amps were actually that cheap. Also, I found one that could go up to 50 MHz, but it said that it could be adjusted. Would that be suitable for my project or not? Lastly, I wanted to check that if I purchased an op amp with an adjustable frequency, would I still be required to construct my own bandpass filter?

Thank You,
Vikram

Integrated circuit Op Amps are very inexpensive as you have found. The expense comes in the test instruments, power supply, bread board, and components you put around them.

I wanted to check that if I purchased an op amp with an adjustable frequency, would I still be required to construct my own bandpass filter?

You adjust the frequency response of an Op Amp by adding bandpass components around it. There are several Op Amp "circuit cookbooks" around that describe different ways of doing this.

Do a little more research (reading) on Op Amp circuits and post back a link to a specific circuit if you want some specific help in understanding it.

Hi,

I have been working on my project ever since I posted my last question. Just to refresh everyone's memory, I was working on creating a receiver that would pick up the mosquito ring tone and convert the frequency into volts that would ultimately turn on a light bulb. I have the basic design of the receiver figured out , but I still I have one question. I was told that i forst must find out the number of volts produced by the 17.7 kHz of sound from the ring tone. I was told that this was neccessary because in order to buy the op amp, I would have to know how much to amplify the volts by. I was hoping that someone knew the volts in that frequency or a way that I could find out the number of volts in the 17.7 KHz. If there is a device that does it, where can I find one?

Thank you.

I have the basic design of the receiver figured out , but I still I have one question. I was told that i forst must find out the number of volts produced by the 17.7 kHz of sound from the ring tone. I was told that this was neccessary because in order to buy the op amp, I would have to know how much to amplify the volts by. I was hoping that someone knew the volts in that frequency or a way that I could find out the number of volts in the 17.7 KHz. If there is a device that does it, where can I find one?

You have chosen a fairly difficult problem to solve that needs to be broken down into stages. I don’t understand how you can have the basic design of a receiver figured out if you don’t have a better idea of the properties of the signal you are trying to detect. Personally, I wouldn’t classify the circuit you will be building as a receiver.

An electrical engineer with a background in signals and systems would probably break the problem down into various stages and think about the signal properties at the input and output of each stage.

What is the source of the signal you are trying to detect? In your case I think it is a “mosquito ring tone” sound wave from a cellular phone at a frequency of 17.7 KHz. This is only a partial characterization of the signal. You have a type of signal (sound wave) and a frequency (17.7 KHz) but you are missing an amplitude.

What is the desired output? In your case, turn on a light bulb. Again this isn’t a fully specified requirement; however, may I recommend a substitution of a LED (light emitting diode) to simplify the problem. A common 20mA variety is inexpensive and doesn’t take much power and is simple to work with, especially the ones with a built in current limiting resistors designed for 5 VDC operation.

First Stage:
The first stage in coming up with the rest of the circuitry is coming up with a suitable transducer to convert sound wave energy into an electrical signal (aka a microphone). Inexpensive microphones designed for telephone use typically have a limited frequency range of 250 HZ to 2500 HZ to pick up the human voice and limit the signal bandwidth needed for transmission and that more expensive studio recording microphones typically were designed for 20 HZ to 15 KHz.

Have you chosen a microphone? Do you have a data sheet for it? A typical data sheet for a microphone will tell you what its output impedance is (typically in ohms) and what its frequency response is. The frequency response information is typically a plot that will show the electrical output (typically in micro or millivolts) expected for some dbm of sound at all frequencies through its range. You might also think of the microphone as a receiver because it receives sound waves.

In this case, the appropriate question is how many volts of 17.7 KHz will the microphone produce. Even with a data sheet that has a 17.7 KHz point on its response curve, you can’t answer this question because you don’t know how “loud” or how much sound energy the cellular phone will generate and how much of the sound energy that the phone produces will propagate to the microphone. The best that you can come up with beforehand is an educated guess.

Second Stage:
In order to keep the microphone happy and derive the most signal from it, you need to provide a load impedance that matches its output impedance. Typically this is done with a simple carbon composition or carbon film resistor equal to what the microphone data sheet calls for and a high impedance unity or low gain buffer amplifier. A simple Op Amp circuit will do here. Because the microphone will also be producing a signal for all sound frequencies, you probably want to provide some form of high pass filtering to this stage so that frequencies below 10 KHz will be attenuated.

You should look up “Bode plots” to get an appreciation for the frequency response of simple R-C filters.

Filter Stage:
Because the frequency response of any microphone falls off quickly at the high end of its frequency response range, you won’t need as much high frequency suppression as you will need lower frequency suppression. I suspect something like a two pole (again understand this from a Bode plot) high pass filter at 12 to 15 KHz and a single pole low pass filter at 25 KHz should be considered. This stage should also provide amplification to compensate for signal loss at 17.7 KHz introduced by the filtering. An Op Amp circuit can do this.

Detector Stage:
So far, all of the electronic stages so far have involved AC signals representing sound waves picked up by the microphone. To operate the LED, you need a DC signal that represents the presence of an AC signal at 17.7 KHz. The reality is that it would be extremely expensive to design and build a precise detector for just one frequency. If you are willing to accept false positives for all sound signals between 15 KHz and 20 KHz then the detector stage is fairly simple.

You first need to AC couple the signal with what is known as a DC blocking capacitor. To convert AC to DC, you need a small signal diode (half wave rectifier). To integrate and discriminate this half wave signal, you need an RC tank circuit. I would start by choosing the time constants such that tank fill time is approximately 10 cycles and the tank discharge time is approximately 30 cycles as a starting point.

Comparator and Output Stage:
You need to compare the output of the tank integration circuit with some threshold. The threshold can be established by a reference voltage derived from zener diode and a resistor divider network. I would pick a zener diode voltage of slightly above half of your power supply voltage. If you are using 5 VDC or a 6 volt battery, something like a 3.3 volt zener. An Op Amp makes a good comparator and will easily drive an LED.

Power supply, decoupling, noise, and trouble shooting
Electrical circuits operate in the presence of noise. There are many sources of electrical noise. Some are external and are picked up or induced into the circuit. Some come from components in the circuit itself like thermal noise or electron shot noise. If the DC power supply comes from AC source, no amount of filtering will reduce all of the conversion noise (again the Bode plot will explain this).

It is very important that the power supply leads of each Op Amp have a decoupling capacitor directly across them. Something like a monolithic 0.1 uFd.

To design, build, and trouble shoot a multi-stage circuit like this, you are going to need somebody local with electronics knowledge to help you along with some test equipment. There are just too many simple things that can go wrong to ever hope for a circuit this complex to work the first time even for a highly experienced electrical engineer with years of experience. Development of circuits like this are done a stage at a time and tested using an oscilloscope and signal generator. We can help with the concepts and design; however, we can’t spot construction issues, test equipment issues, or other simple problems remotely. Even with highly skilled technicians on site, remote trouble shooting is extremely difficult and communication error prone.

A good starting point would be to put together a setup with a cell phone with a mosquito ring, a phone you can use to call the cell phone, your microphone, a load resistor, and an oscilloscope in a quite room. Look at the change in signal when the cell phone rings. In other words, to answer your starting questions, you are going to have to measure it yourself!

WOW, thanks for your detailed and informative reply. I am very grateful that you took the time out to explain the mechanics of the device to me. You could have possibly saved me alot of time and money.

Over the last few months I have met with an engineer twice who had explained the calculations and the what he said was the basic design of my project. However after reading your detailed reply, I am afraid that he has misguided me.

His instructions seemed to simple to be true. He said that all I had to do was feed my ringtone into a microphone with an input of 20 kHz that was connected to an osilloscope to find the volts in the ring tone. Then I was supposed to find a suitable op amp after calculating the correct frequency pass. He said that I only need a high pass filter and that a low pass filter was not required. He told me that I just had to connect the mic to the op amp and connect the op amp to A RC circuit after doing the correct calculations.

My final step would be to hook the rc circuit to a 3 or 6 volt light bulb. He tld me that my whole device would be connected on a breadboard.

However, after reading your reply and researching the bode plot and various of the other components, I think the advice he gave me was incorrect. Was it?

Thanks

He said that all I had to do was feed my ringtone into a microphone with an input of 20 kHz that was connected to an osilloscope to find the volts in the ring tone.

This is similar to what I suggested as a starting point; however, I added looking up what the microphone's output impedance was and to put a matching resistor in the measurement circuit. If the microphone you were using had an output impedance of >10 Kohms, the difference probably won't matter; however, if it was 150 ohms it might make a considerable difference in being able to accurately measure it.

He said that I only need a high pass filter and that a low pass filter was not required.

He may very well be correct; however, many operational amplifiers, particularly those that have two or four on the same IC substrate can interact with each other because of parasitic coupling and can oscillate on their own at high frequencies (MHz range) especially if the gain is high and the signal to noise ratio is low. Throwing in 25 KHz low pass filter is just a simple way to help eliminate that possibility. My suggestion of using a double pole high pass filter may also be overkill and a single pole high pass filter may be all that is needed. Neither he nor I can predict without knowing what the signal level you are going to get when you run your first test.

My final step would be to hook the rc circuit to a 3 or 6 volt light bulb.

This might work because a filament light bulb has some integration properties; however, if it doesn't or gives you too many false positives or negatives you will have to come up with something more descriminating. The concept of AC coupling into a small signal diode with a tuned tank RC circuit with separately tunable charge and discharge times followed by a threashold comparator is a much more sophisticated approach. Again, it maybe overkill; however, by changing the resistance values, you can easily do some tuning to improve the detection with this more sophisticated design.

He t[o]ld me that my whole device would be connected on a breadboard.

No disagreement on that point, if you read back to one of my earlier replies I mentioned that the cost of the breadboard would probably be a significant part of the cost of the project.

I think the advice he gave me was incorrect. Was it?

Definitely NOT! If he is going to be there helping you as you run into problems, then his approach to getting you started is similar to what I might have used. Circuits are developed a stage at a time. No need to get too far ahead of what the person you are mentoring can grasp and deal with at one time. You can always learn together. By being remote, I didn't have the luxury of down playing any of the stages. I had to take the steps that OneBrightGuy layed out in simplistic terms earlier in this thread and provide you a good recommendation on what I felt could definitely be made to work and to open your eyes all at once to everything that might be needed.

Your local mentor on this project may have plenty of practical experience with the particular microphone you have and know you aren't going to run into any problems making a VERY simple circuit work.

Oh, well thats a huge relief. I was getting worried that my mentor was not as qualified as I tought he was. Thanks for clearing that up.

However, he is not in the state for the next week and I was hoping that you could answer some questions that I had.

I'm still kind of confused regarding the microphone. I have researched various microphones and have found some that can pick up 20 kHz, but they are just normal recording microphones. Would those work? Also regarding the price of the microphones, the ones I have found range form prices of $29.99 to $329.99. I was curious whether purchasing the cheaper microphone would make a difference to my project. I am working on a very low, fixed budget and cannot afford such an expensive microphone.

Also, just to clarify, the application of op amps to my project would act as my band pass filters and I would not require any additional components, is that correct?

In the reply you just posted and a previous one, you have mentioned that the breadbroad will be a significant part of the cost of my project. However, I have been to a number of stores and I have found breadboards availible for as little as $6.49. Am I looking at the wrong kind of breadboards?

Thank you.

http://www.analog.com/en/prod/0,,759_78 ... %2C00.html

Hi it's me again,

I was just researching bode plots and op amps and all that good stuff when I came across this site that seemed to have a pretty wide selection of op amps.

I noticed that this one in site I have sent you seemed like a good choice for my high pass op amp as it has frequency pass of 20 kHz. However, I did not understand some of the details the site provided in the description of the product, such as Iq per Amplifier (max) and slew rate. Due to my lack of knowledge I was unable to tell whether this was a good choice for a high pass filter or not. is it?

I, unfortunately did not find an appropriate low pass op amp and that seemed to be the most varied site out of all the ones I went to. Do you know a place or site that has a wide selection of op amps and would have the ones that I require.

I assumed that part of your project would be to learn how to build some simple circuits around an inexpensive integrated circuit Op Amp such as the LM324 or LM348 which come in a 14 pin dual inline package and has 4 identical units in it.
http://www.ortodoxism.ro/datasheets/nat ... 009299.PDF
http://www.ortodoxism.ro/datasheets/SGS ... Xssuzr.pdf
Or some other widely available part.

The LM324 single voltage unit is harder to design with because it requies your signal to be given a non-zero DC bias than the LM348 which requires both a positive and negative supply but allows operation with a zero volt DC biased signal.

If your mosquito ring frequency is 17.7 KHz, then you want 17.7 KHz in the frequencies passed without much attenuation. This means you want the corner frequency of any high pass element (or pole) below 17.7 KHz (I recommended 12 to 15 KHz). You can make a simple two pole high pass "T" filter with two capacitors and a resistor placed between the output of a unity gain "buffer" amp and a unity gain "follower" amp. The capacitors form the top part of the "T" and the resistor forms the vertical to signal ground. It is just a matter of picking "reasonable" values and getting the correct RC time constant.

Once you understand that the "corner frequency" of these simple filter circuits is simply the frequency where the impedance (reactance of a capacitor) of the capacitor element equals the impedance (resistance of a resistor), you are almost freed from relying on pre-existing designs. There is a bit more to it in terms of understanding parasitic elements and having a feel for what range those values might be in any given circuit and doing a complete circuit analysis; however, that is where you can use test equipment to measure things and rely on help from your mentor.

OK, so what are resonable values? If the resistance is very high, the circuit will be suseptible to picking up noise. If the resistance is too low, the signal voltage will be very small and disappear into the noise floor. Somewhere between 1Kohms and 100Kohms is probably a good range for your circuit. See if you can figure out the values yourself and post back with what you come up with. Beware, common components come in standardized values so part of low cost design is coming up good pairings from standard low cost components.

A single pole low pass filter can be made by a "T" using two resistors as the top part of the "T" and a capacitor as the vertical. Since this is secondary feature and NOT the primary design consideration, you can add it to one of the amplifier stages where you already have a series input resistor whose value was already chosen to set the gain. Essentially throwing in a capacitor with a value chosen to provide the correct RC ratio to set the corner frequency between 25 and 35 KHz. Beware: The output impedance of the previous stage is also involved and may affect the time constant considerably.

The slew rate is a measure of how fast an amplifier's output voltage can change. I doubt that this property will be a limiting factor in your circuit given the 17.7 KHz frequency for any low cost commercially available IC Op Amp.

I'm still kind of confused regarding the microphone. I have researched various microphones and have found some that can pick up 20 kHz, but they are just normal recording microphones. Would those work?

yes.

Also regarding the price of the microphones, the ones I have found range form prices of $29.99 to $329.99. I was curious whether purchasing the cheaper microphone would make a difference to my project. I am working on a very low, fixed budget and cannot afford such an expensive microphone.

For your project, the only technical consideration is will the microphone pick up audio around 17.7 KHz. I would find the least expensive one you can readily acquire that meets your only technical requirement and build the circuitry to match its other characteristics for your application.

Also, just to clarify, the application of op amps to my project would act as my band pass filters and I would not require any additional components, is that correct?

The resistors and capacitors you place around the op amps will act as your band pass filters. I mentioned using a twin "T" high pass filter (two capacitors and a resistor). If you consider the integrated circuit as the "op amp", then these three parts are the "additional components" required. I suggested a low pass "T" filter (two resistors and a capacitor) where the resistors were shared functionality with a gain control.

In the reply you just posted and a previous one, you have mentioned that the breadbroad will be a significant part of the cost of my project. However, I have been to a number of stores and I have found breadboards availible for as little as $6.49. Am I looking at the wrong kind of breadboards?

The breadboard kits have either come down in price or you will need a few more pieces than just the "white block". Typically, you will need something like some 5 way binding posts and something to mount them and the "white block" to in order to make your project more reliable. While you can directly solder your microphone and power supply or battery leads to some solid hookup wire appropriate to push into the white block holes, those connections won't be strain reliefed and will make it hard to move your project. I usually consider the "breadboard" as a combination of the "white block" and power supplies and interconnection posts and hookup wire. All these little things tend to add up if you have to buy them for this project. You can easily spend 20 to 30 dollars acquiring two 6 V lantern batteries, the hookup wire, the white block, Five 5-way binding posts, a piece of plastic and four rubber feet. If you spend $30 on the microphone and $30 on the quad op amp, resistors, capacitors, and LED or light bulb, I'd consider 1/4 to 1/3 for the "breadboard related components" to be a considerable part of the budget.

Okay,

Thanks for answering all those doubts. I think I have everything down nowexcpet for that whole "T" circuit you mentioned that should be placed around high and low pass filters. I don't understand that concept. I think I am having difficulty in understanding it because the mentor I met with before told me that I could just use one capacitor and one resistor placed after my op amp and that would be sufficient. However, if this "T" cicuit is more effective I will definitely use that.

My dead line is approaching and I think that it is time for me to start purchasing my equipment. I have already found a mic that fits the needs of my projects, but I am unsure of what op amp to purchase. I have looked at many op amps as I mentioned before and I looked at the ones you sent me as well, but I am confused. All those op amps I saw and you sent me had a bandwidth of 1 MHz or more, which is far above my needed frequency of 17.7 kHz. How would those be correct for my device.

Also regarding the capacitors and resisitors, how would I know which ones to purchase. I know how to do the calcualtions for the two, my mentor taught me that, but he didn't really tell me how to apply them. By this I mean, what do I use those calcualtions for and do I need to find anything out before I can purchase them. For example, will I need to know that volts int he ring tone before I purchase them or anything like that.

Thanks

I did a quick search for "Twin T Filter" and this one came up
http://www-k.ext.ti.com/SRVS/Data/ti/Kn ... aqs/tt.htm
If you use only at the "bottom half" of the circuit and connect the middle to signal ground, you have a single ended twin T high pass filter comprised of C1, C2, and R3. Earlier, I gave you a hint that R3 should be somewhere between 1K ohms and 100K ohms because of noise considerations. If R3 is lower than 1K ohms, your signal will likely be too small and too close to the noise floor. If R3 is higher than 100K ohms, the circuit will be suseptible to induced noise currents. You want to pick the capacitor reactance at the corner frequency to match the value of R3. Be sure to take into consideration the tolerance of the parts. If the resistors are 10% and the capacitors are 20%, then the corner frequency can be off considerably.
http://www-k.ext.ti.com/SRVS/Data/ti/Kn ... aqs/hp.htm Demonstrates the significant difference between a one and two pole filter design. For the cost of a second capacitor, a twin T high pass filter will allow you to detect your mosquito ring sound in the presence of background sounds more reliably.

All those op amps I saw and you sent me had a bandwidth of 1 MHz or more, which is far above my needed frequency of 17.7 kHz. How would those be correct for my device.

If the Op Amps are capable of accurately handling DC to >1 MHz, will it accurately handle 17.7 KHz? Yes. 17.7 KHz is definitely in their range. Is the cost of the Op Amps I suggested reasonble for your budget? If so, you have something you can work with.

I recommended that you add a simple single pole low pass filter around 25 to 30 KHz to one of the stages. What will this do to frequencies above 50 KHz? It is going to attenuate them 3 db (by half) for every decade above 50 KHz. In other words, the IC op amp that had 1 MHz or more bandwidth now has far less than 50 KHz.

Also regarding the capacitors and resisitors, how would I know which ones to purchase.

Experience is a big factor in these decisions. I recommend that you consider 1/4w 5% carbon film resistors for all of your filter, gain control, and impedance matching uses. These tend to be small, temperature stable, reliable, and relatively inexpensive for their tolerances. As far as capacitors, for your filter circuits, I recommend mica, monolytic, or tantalum depending on what precision is available for the capacitance value (mica for very small capacitance, monolythic for middle values, and tantalum for large values). You should choose ones with working voltages slightly higher than your supply voltages. If you are using two 6v lantern batteries as your supply, then you should use at least 15VDC working voltage. Higher ones are fine if they fit your budget.

If you have a signal generator and oscilloscope available, you can pick low tolerance capacitors and hand pick a resistor value to get the corner frequency desired and possibly save some money. This assumes that your mentor has access to a variety of resistor values and is willing to "trade" resistors in which case, you need to talk to your mentor so that you buy ones acceptable to him for trading.

I think I mentioned the need for 0.1 uFd power supply decoupling capacitors. If you are using a DC power supply that plugs into the electrical grid, you will need a bulk filter capacitor, something like a 50 to 250 ufd tantalum for each voltage as well.

Matched resistors. In some situations, you want two or more resistors to be very close in value. One of the most cost effective ways is to use a single inline or dual inline resistor package. Resistors change values with temperature, so having them on the same substrate causes them to track. This is often a far more cost effective solution than buying individual high precision resistors. While the precise resistance value of the resistors in these SIP and DIP packages are not controlled, the manufacturing process almost gaurantees that all the resistors in the same package are very very close to the same value. Note: This technique can also get you precise ratios of 1:1, 1:2, 1:3 as well.

Even with years of experience and working with a team, I've never been able to build a multi-stage circuit like the one you will be experimenting with and be able to get everything right the first time. Expect to have to need to go back for different resistors to adjust the gain or corner frequencies and budget accordingly. Engineers who do this full time for a living ususally have access to a wide range of resistance values on their bench or within 50 feet in a well organized common stock pile.

Hello Craig,

As I thought I had finally understood the make up of the device and how to construct, I set off to complete the first task, finding the number of volts in the ring tone by feeding the ring tone into a microphone that was connected to a oscilloscope. As I did not have an oscilloscope and neither did my school, I had to find a company that had one and arranged an appointment with one of the engineers there.

I found that the volts in the frequency was 200 mV.

However, he was very interested in my project and wanted to know more about, so I told how I was planning to construct it and he gave me an alternate design. I wanted to check with you which would be more efficient and one that I would be able to do with more ease and on my own.

Up until the op amp, his instructions were the same. However, afterwatrds he said that for the low and high pass filters I did not need resisotrs, only capacitors. He then went to say that after the sound was filtered out I would need a detector circuit. He suggested I use a trigger with a latch circuit to catch the sound waves. The trigger would be connected to a 555 timer which would send out a specific number of volts. he then further suggested that I hook this up to a TIP 150 to further amplify the sound waves. This would finally be connected to a light bulb with would be turned on by the volts output by the TIP 150. He also said that the 555 timer would tell how long the light would stay on for.

I was just wondering which design you think I should go through with as both sound very qualified, but I wanted to build it in a way that I would actually be able to do it myself, but it would have a chance of working. (If that makes any sense.)

Thank you

However, afterwatrds he said that for the low and high pass filters I did not need resisotrs, only capacitors.

One can design OP AMP filter circuits using the intrinsic resistance/impedance elements that are internal to the integrated circuits themselves; however, one has to determine/select/tune the external capacitance values experimentally. The output impedance of an OP AMP typically is a low resistance value (less than 100 ohms) and a very low shunt capacitance value (less than 300 pfd). The input impedance of an OP Amp is typically > 1 Mohms and the "white block" breadboard adds a shunt capacitance of 20-50 pfd. By adding your own series resistor of at least 5000 ohms and series (high pass) or shunt (low pass) capacitance values of > 500 pfd, the added resistor and capacitor predominates which means you can calculate the RC time constant to pick the corner frequency of the filter element ignoring the OP Amp and breadboard impedances. For a science fair project where you have to explain how it works, it will be much easier to describe physically obvious components than burried parasitic components.

He then went to say that after the sound was filtered out I would need a detector circuit.

We agree on this point. Earlier I proposed a series capacitor, diode and an RC tank circuit as a detector. The origins of this is an AM Radio detector. Something that is fairly easy to explain the rational for in terms of AC coupling, "half-wave" rectification, and "integration" (summation).

Your latest advisor went with a "digital" style detection scheme which makes your circuit a hybrid between analog and digital techniques. I've designed lots of very sophisticated combinations of these over the years; however, they require extreme care in cuircuit layout (construction) techniques with extreme attention to decoupling of power supply currents to the components to prevent digital switching transients from coupling into the analog circuitry. The ubiquitous 555 timer is a particularly troublesome component to add to the mix and I personally will not use one when there is any low level analog circuitry around. I've had to redesign several 555 based designs developed by others over the years to improve reliabilty and eliminate safety issues. It is a fun part to experiment with and can do a lot of neat things with a few resistors and capacitors; however, it has a crude, obnoxious and boisterous persona that isn't tolerated well by some other circuits.

200 mv is a great signal strength and isn't what I consider low level; however, you need to realize that sound propogation is usually a square law relationship (doubling the distance of your cell phone from the microphone will drop this level by one fourth) 10 times the distance will divide it by 400 and you are now down in what I consider a low level signal range.

I guided you toward designing with a quad OP Amp because it is one IC part to learn about that has a lot of different uses and you can show case a few of them in your application.

he then further suggested that I hook this up to a TIP 150 to further amplify the sound waves.

I'm not sure what a TIP 150 is. If it is a Darlington Power Transistor to operate a DC lamp, that is certainly an alternative; however, it is yet another thing to learn about. If a TIP 150 is a thyristor to operate a standard 120 VAC 40 W light bulb, then I have a lot of safety concerns. Construction of custom safe devices that plug into the power grid is a beyond what should be done for a science fair.

My recommendation for a visual display element was to use a low cost low power LED and simply drive it from a section of a quad Op Amp package used as a comparitor because you already needed other amplifier elements for other purposes in previous stages. This would be keeping with the "KISS" principle (Keep It Simple Stupid) of not making something any more difficult than necessary. In my interpretation, learn how to use Op Amps and make multiple use of that knowledge. The detection of the mosquito ring tone was the main focus of the project and the display element was just something you had to do to demonstrate your device worked.

Webster defined "engineers" as "schemers". As you are finding out, talk to several engineers, and they will all come up with a wide variety of different schemes to do the same thing and will happily argue with each other about which is best. In most cases, these differences of approaches come from particular background and experience (skill level) and comfort level with different approaches. When combined as a team with cost goals, this friendly competition often drives inovation and combinations that no one individual would have come up with.

My advice to you for this project is to pick a local mentor and stick with them. Your project will be a learning experience for both you and your mentor. Hopefully your mentor will be learning how to be a better mentor and not a better circuit designer and you will be learning design, construction, and measurement techniques.

Thanks for the helpful suggestion. I have researched the method you suggested and the method the mentor I met with over the past week suggested and I think your method is more suitable for me as i am only a high school student with very limited knowledge of electrical engineering and the methood that the engineer I met with suggested was very advanced and more suitable for someone who was versed in the field.

I am still trying to figure out what amplifiers I need. I was looking online as I wanted a reference point for my calculations when I stumbled across this site that actually let did the claculations for you. All I would have to do would be to plug in my values. Now I know that for the frequency for my low pass filter I want a 16 kHz band pass and for my high pass filter I want a 20 kHz band pass and that DC offset should be in mV, somwhere between 100 and 200 mV, but when I clicked on the site that allowed you to design your op amp I was really confused by all the stuff it required me to enter in order to complete the calculations. I have attached the site. I was hoping maybe you could explain to me the other stuff as I have never even heard of most of those applications in my life.

http://designtools.analog.com/dtAPETWeb ... #RunAnchor

Also, concerning the detector circuit, I am a little confused. I know you have talked about this before but all I can seem to find on the subject is that you reccommend that I use an LED which i have researched but do not fully understand. I mean I understand what it is used for but I ma not quite sure how I can implemt it into my device. Also would that be all I required for my detector circuit. In the sense, would this pick up the volts of the that were filtered out and relay that to the light bulb. I remember you talkin to me about the detector circuit but I was unable to find anything except for the LED circuit suggestion. I would really appreciate it if you could explain that to me, btu I wil continue looking through the pages for more information that you might have told me earlier.

Thank you

Sorry,

I was reading over my previous post and I think I might have been unclear as to what i was asking about the detector circuit. I did not mean to suggest that you have only mentioned the LED, but the LED and the RC tank and also the capacitors, but I meant to ask how would i combine all those to form a dtector circuit. By this I mean would I need to attach a capacitor on each side of the LED and then attach a RC tank to the back of the second capacitor. Also, I was wondering would I need to do any calculation do find out the different applications of the capaciotrs and the RC tank circuit. If there are I was hoping that you could explain them to me or maybe, if you knew a site that does them for you, if you could send me that link.

Thank you,
and sorry for the confusion.

The web page you referenced is an inverting amplifier. Its calculations are far more detailed than you need. For this kind of inverting configuration, the gain of the amplifier is -Rfb / (Rg+Rsrc). In critical applications, the value of Rbias should be selected to be roughly equal to (Rg+Rsc) in parallel with Rfb if you want to minimize any bias caused by the leakage current. In your case, the exact value isn't critical if you stay under 100 Kohms for noise reasons I tried to explain earlier. If you use a single power supply and attach -Vs to the common (signal ground), you will need a bias voltage Vref of approximately 1/2 of +Vs. Earlier I recommended using two 6v lantern batteries as your power supply so you -Vs is -6v and +Vs is +6v. In the dual supply case, you can simply attach the + input to the common (signal ground). If Rseries is 100 ohms or greater then it determines the output impedance from this stage (if it is smaller, then the IC output properties will no longer be swamped by the external resistor's impedance). The Cl and Rl are the impedance of the load (input impedance of the next stage) and can be ignored for your frequences and the Op Amps I mentioned (or ones with 1 MHz bandwidth or greater).

You can make a simple comparator out of an Op Amp by leaving out Rfb so that the gain approaches infinity (the output will be either very close to +Vs or -Vs as a result of whether the voltage on the + input is higher or lower than that on the - input.

You can make a simple unity gain stage by connecting the - input directly to the output and feeding your signal into the + input. Look up unity gain or "follower" OP Amp designs. I mention this as a possible input buffer amplifier stage.

The detector circuit I recommended consists of a series DC blocking capacitor (say .01 to .1 ufd) attached to the Op Amp output pin of the previous stage, a series small signal diode, feeding something similar to Rseries and a tank RC shown as Rl and Cl in the schematic on the referenced web page. Lets rename Rl and Cl to Rtank and Ctank. From my earlier recommendations, I recommend you start with the Rseries Ctank time constant to be about 5 cycles of 17.7 KHz long and you want the Rtank Ctank time constant to be > 20 cycles of 17.7 KHz long.

The blocking capacitor prevents any DC offset voltage from the previous stage from affecting the detector as it will only pass AC. The small signal diode acts like a rectifier and clips off either the negative or positive half cycle of your audio signal.

The Rseries Ctank acts as an integrator that will increase (or decrease if the diode is reversed) the voltage on Ctank with each half cycle present. The Rtank acts to bleed off the voltage on Ctank. I recommend starting with time constants chosen so that the charging will occur 4 times faster than the discharging.

Look up half wave rectifiers and power supply ripple voltage for a simple filter capacitor circuit to get a feel for what the waveforms might look like.

You can connect the RC tank circuit to one input of a comparator Op Amp circuit and attach the other input to a couple of resistors used as a voltage divider. Again look up voltage divider resistor networks for a description of these kinds of circuits.

An LED will emit light when current flows through it. Because it is a diode, it will only allow current to flow in one direction. Some LED packages come with internal current limiting resistors designed for about 5v operation, others do not and you have to supply your own current limiting resistor. Common inexpensive LEDs come in 2 mA and 20 mA varieties. If you use 6v lantern batteries as your supply and have a 20 mA LED, then you need about a 270 ohm current limiting resistor (2.7 Kohms for the 2 mA). (6v / .02a = 300 ohms; however, the diode will have a forward voltage drop of .5 to .7 v so the next smaller standard 10% resistor value is appropriate). Note: The 6v lantern battery recommendation is close enough to 5v that an LED with an internal current limiting resistor will work without any additional resistor. Simply attach the output pin of the Op Amp comparator to the series current limiting resistor and LED indicator to signal ground.

http://www.tpub.com/neets/book2/3d.htm describes time constants.
http://en.wikipedia.org/wiki/Rectifier describes half and full wave rectifiers; however, it doesn't deal with the filtering capacitor aspect.
http://www.play-hookey.com/ac_theory/ps_filters.htmlShows a full wave rectifier with a single filter capacitor. If you redraw the waveform so that every other blue half wave is missing and let the red line drop until it meets the following blue wave, that is what you can expect the result to look like.
http://www.stmicroelectronic.com/stonli ... s/1707.pdffigure 5 shows a half wave rectifier filter waveform (better picture, more complicated verbage).
http://ourworld.compuserve.com/homepage ... /opamp.htmshows inverting, non-inverting, and unity gain voltage follower designs. I don't recommend you use the 2nd order (two pole) filter designs for your circuit. This is not because they don't work well, it is simply because they are far more difficult to comprehend than the twin "T" circuits I referenced earlier. Their complexity comes because they are placed in a positive feed back loop of the Op Amp which makes them difficult for electrical engineering students until they have had two courses in circuit analysis and a signals and systems course dealing with feedback loops.

Hope these references help.

Thanks so much for all those sites.

They really did help me alot understand the time constants and half wave and full wave rectifiers.

I don't want to sound stupid or anything as I think it is not a very intellectual question, but I am confused as to where the 6V battery comes into play and where it will be plaved and what the use of its placement is.

Also just to clarify, the intial RC circuit will be connected to the op amp which will filter the volts and then the series of capacitors you mentioned I should use for the detector would be connected to that RC circuit and then the RC tank would be connected to the capacitors and then the anode of the LED would be connected to the RC tank?

Thanks

I am confused as to where the 6V battery comes into play and where it will be plaved and what the use of its placement is.

I'm not surprised by this confusion. Electrical engineers and technicians typically use some shorthand designations for power supply and other common connections and may even omit them entirely from the main graphic part of the drawing and provide a table or other "typical or representative of..." diagram.

All OP Amp integrated circuit packages have two power supply pins. On http://designtools.analog.com/dtAPETWeb ... 0#RunAnchor they are labeled +Vs and -Vs.
On http://www.ortodoxism.ro/datasheets/SGS ... Xssuzr.pdf they are labeled +Vcc and -Vcc. Some part diagrams might label them V+ and V-.
On parts like the LM324 that are designed to operate from a single voltage source http://www.ortodoxism.ro/datasheets/nat ... 009299.PDF, they are typically labed V+ (Vcc, or Vs) and GND. Note: The LM324 can be used on dual supplies by "relabeling" pin 11 as V-.

All amplifiers require an external power source in order to operate. For the LM348, the maximum Vcc voltage is listed as +/- 22 volts under Maximum Ratings. There is a Power Supply Current vs Power Supply Voltage set of parametric curves that shows how much power the part will draw assuming it isn't asked to supply current through its output pins. The curves assume a balanced supply -Vcc = - (+Vcc) starting at +/- 2 volts and going up through +/- 22 volts. The Electrical Characteristics values are listed for +/- 15 volts; however, coming up with two sets of batteries each supplying 15 volts would be expensive. By choosing an LED as an output indicator, all your circuit requirements are low power requirements so you can choose something like +/- 6 volts that can be obtained using two 6 volt lantern batteries.

The negative terminal of one 6 volt battery gets labeled -Vcc (or -Vs if you prefer).
The positive terminal of the other 6 volt battery gets labeled +Vcc (or +Vs).
The other two battery terminals are connected together and are your common connnection which is sometimes refered to as signal ground often shown as a symbol with three parallel horizontal lines (top longest, bottom shortest). See http://ourworld.compuserve.com/homepage ... /opamp.htm under "Operational Amplifier (Op-Amp) Basics" for a typical usage. Bill Bowden is using two 9 volt batteries in his diagram and labeling them +/- 9V Battery. Note: The use of the "ground" symbol (or any other labeled symbol) means that all circuit points with the same designation are connected together. In the diagram in this section, he does not show the power supply pins of the Op Amps; however, for the Op Amp sections to work, they must be powered by these "understood" connections. Only in diagrams like "Low Power Op-Amp..." where one or more of the power supply pins is connected to something other than directly to the power supply is the connection always shown.

Other than making sure that the power supply voltages are balanced with respect to common (e.g. they are the same but opposite polarities) and the correct polarities are connected, the actual power supply (battery if you follow my recommendations) voltage doesn't matter to most of your circuitry.

It will matter to the current limiting resistor for the LED being driven by a Op Amp section used as a comparator as the output pin will be close to either +Vcc or -Vcc except for very brief switching transients.

I tried to describe the detector and comparator in words earlier and I failed to make it clear. I don't know of a way to post a circuit diagram of the detector and comparator circuit into this forum. A diagram would probably be worth a thousand words in this case. I'll supply one to the science buddies staff and see if they can post it or forward it to you.