Audio Frequency Counter Recommendation

[ThomasHenry]

Would anyone care to recommend an audio frequency counter? I almost never work with circuits beyond 30KHz or so, so the fancy RF types are overkill for what I need. A five digit affair would be great, with 1Hz resolution if such a beast is available.<p>I don't mind building my own if anyone wants to recommend a kit or magazine schematic as an alternative to a ready-made unit.<p>Many thanks,<p>Thomas Henry

[dyarker]

How about a program that uses a PC's sound card. There must be be a freeware/shareware/demo on the web somewhere.<p>Cheers,

[ThomasHenry]

An interesting idea, but I was hoping for something a little handier (more portable) to use on the bench. My electronics shop is in the basement, while the computers are on the second story of the house, so that idea probably won't work well for me (although it does sound inexpensive). In any event, I would imagine the input would have to be conditioned somewhat.<p>So, let me rephrase the question: can someone recommend an audio frequency counter suitable for use on a workbench?<p>Thanks

[toejam]

There are several low cost vom.s available that have frequency counters in them for 50 or so bucks.I got one from radio shack.

[ThomasHenry]

Yes, I have one too. But aren't most of these 3-1/2 or 4-1/2 digits? I'm hoping for 5 digits, with a 1Hz resolution, if that's possible. In other words I would like 12,345Hz to show up that way, not as 12.3 with a 1K multiplier.<p>Any ideas?

[jwax]

http://www.qsl.net/zl1bpu/micro/CNTR/CNTR.htm
At a glance, should make for an interesting project for under $50.
Found this under a Google search under "audio frequency counters".
John

[Dimbulb]

Dimbulb is way out there today.<p>An interesting question and there are relativly few applications for people who really want to know precisely about audio frequency.<p>In studying bats a simple method of divider chip gave meaning to the ultrasonic by converting to audio. I like this because the variations are audible and some meaningfulness such as species was discovered.<p>Like studying bats the frequecys when brought together by converting into a standard unit gave differences that were interesting and the most interesting differences could be isolated and amplified these may be resonant harmonics.<p>The front end of a counter has always been interesting to me as phenomena and nature has many audible questions for a precision counter builder. <p>In audio there might be some usefulness to converting the ultra lows and the high range to mid tones in a harmonic function so that switching between three channels and counting while cross translating the ranges might have some interesting results.<p>[ April 04, 2004: Message edited by: dimbulb ]</p>

[Ron H]

Dimbulb, if someone asked you what time it is, would you respond with a history of timekeeping devices?

[Dean Huster]

"History of timekeeping devices"? Hah! I'd be more practical than that and teach him how to build a clock. Tell a man what time it is and he bugs you forever for the time; teach him to build a clock and ... he'll want NIST certification for the thing?<p>Hey, big Tom. You might scout out some of the Intersil counter chips. Good down one Hertz, simple, accurate (with the right timebase), and if you ever want to upgrade to a commercial unit, you can always reuse the parts for another project.<p>Otherwise, scout out ebay and look for a counter there. You might be able to get an accurate, commercial-grade box from there for peanuts. Here's the category for ebay test equipment if the link is still good: http://business.ebay.com/4676<p>Dean

[Dimbulb]

Thanks guys I'm having a good laugh at myself tonight. <p>Long ago someone said " YOU DIM BULB! "<p>So.

[ThomasHenry]

Thanks to everybody for the suggestions.<p>I ended up designing and building one from scratch, using the 80C32 microcontroller. It goes from 0Hz to 65kHz (I'm obviously using a 16 bit register to record the counts), with 1Hz resolution and a 5 digit LCD display. It counts on a fixed 1 second gate, which is okay for my purposes.<p>Believe it or not, I had everything I needed in my junk box, so my total outlay was $0! All it took was three nights of work, which included writing the firmware (easy for this chip), making a nice front panel (time consuming) and assembling the hardware.<p>I built the frequency counter in a handheld box, but it requires a wall-wart to power.<p>This really does the job for me. Like I said, I rarely design analog circuits which exceed the audio band, so the commercial gear was really overkill. 65kHz is really all that I need.<p>Thank Gauss for microcontrollers!

[xdissent]

thomas, just out of curiousity, how did you go about implementing this? adc into your mcu, using fft to find your strongest band? i played around for quite a while with something similiar using a mcu without an adc. it never worked well at all with anything other than a frequency generator. im just curious because im about to embark on a similar quest again in the near future. if you are using fft, are you windowing, etc.? are you measuring anything other than "clean" signals? just wondering what worked for you.

[ThomasHenry]

For the analog input I used nothing more than an NPN driving a buffer from a 4050. Some specifics: input jack to 2.2 non-polarized cap, to 22k to base of transistor. I put the cathode of a 1N4148 on the base to ground to dump negative excursions. On the collector is a 4.7K pullup, and then to the 4050.<p>Thus, the input is AC coupled and protected from overvoltage or negative voltages. (I only work with audio signals of 20Vpp max, in general).<p>The signal from the 4050 is applied to port P1.0 of the CPU. The negative-going transitions on this line are counted and stored in a register. An interrupt driven clock clears the register once a second, and counts the number of transitions which occur thereafter. This number (after some formatting) is sent to the LCD and the process starts over again.<p>It all works very nicely. I'm writing up the complete circuit for an appendix in a new VCO chip book I'm working on.<p>Hope this helps.

[ThomasHenry]

Also, the only signals I monitor on a regular basis are sines, triangles, ramps and squares (music synthesizer outputs). So, yes, I'm only working with clean signals.

[rshayes]

Since you already have a microprocessor in the system, you might want to make period measurements if the input frequency is low. At 20 Hz, a 1 second gate will give a 5% error due to the count ambiguity. A period measurement using a 100 KHz clock would have a .02% ambiguity at 20 Hz. The processor could automatically calculate the inverse, so the frequency readout would still be direct. For low frequencies, the period measurement is both faster and more accurate.