http://forums.leaflabs.com/topic.php?id=10262 A place to share, learn, and grow... en-US Fri, 22 Jan 2016 00:07:46 +0000 http://bbpress.org/?v=1.0.2 q http://forums.leaflabs.com/search.php http://forums.leaflabs.com/topic.php?id=10262#post-22683 Wed, 27 Feb 2013 11:07:40 +0000 iainism 22683@http://forums.leaflabs.com/ <p>Final thought (after a brief re-read of my DSP notes), if you need to detect a number of tones (fixed frequencies) in the range then you could also use the Goertzel algorithm, this provides a lower computational load. There's a nice article over at embedded.com on the use of this along with a code listing: <a href="http://www.embedded.com/design/configurable-systems/4024443/The-Goertzel-Algorithm" rel="nofollow">http://www.embedded.com/design/configurable-systems/4024443/The-Goertzel-Algorithm</a> </p> http://forums.leaflabs.com/topic.php?id=10262#post-22679 Wed, 27 Feb 2013 08:56:46 +0000 iainism 22679@http://forums.leaflabs.com/ <p>stafil - additional information to gbulmer's question...</p> <p>If your signal is below the level at which you can set an interrupt using a GPIO pin (or if you don't know what level it will arrive at, or you are dealing with a 'complex' waveform) then you may need to use an ADC and this means you'll need to do some signal processing work.</p> <p>Based on what you've said then in a noiseless environment you would (theoretically) need to sample over a window of ~10 ms at a rate &gt; 1600 Hz to get a good result and then apply a DFT to get the information that you require. The DFT part is not too onerous, there are various code examples and libraries out there on the web. For example, <a href="http://sourceforge.net/projects/kissfft/" rel="nofollow">http://sourceforge.net/projects/kissfft/</a> (caveat: this was the first link I found on Google for 'embedded fft library', and may or may not be compatible with the Maple, I don't know, I've never tried it!)</p> <p>But of course we must (usually) deal with noise so I'd suggest sampling at 6400 Hz and a window of 10 ms to get a nice, 'power of two' number of samples within the window, pad this up with zeros to a total length of 128 samples and do any required filtering to minimise noise prior to the DFT. If you need help designing a filter then I can recommend this: <a href="http://www-users.cs.york.ac.uk/~fisher/mkfilter/" rel="nofollow">http://www-users.cs.york.ac.uk/~fisher/mkfilter/</a> (it's saved me a lot of work in the past!)</p> <p>If you have a lot of noise, a known signal frequency and just want to check for the presence of the signal then you could even use a software-PLL or the cross-correlation of the sampled data with 'specimen' data and make a judgement based on their output as to whether the signal is present. (But these are probably massive overkill!) </p> http://forums.leaflabs.com/topic.php?id=10262#post-22512 Tue, 26 Feb 2013 08:04:11 +0000 gbulmer 22512@http://forums.leaflabs.com/ <p>stafil - what approach have you tried? </p> <p>How accurate and precise do you need? Is <code>digitalRead</code>'ing the status of a pin and timing using <code>micros</code> good enough? </p> http://forums.leaflabs.com/topic.php?id=10262#post-22446 Mon, 25 Feb 2013 14:28:27 +0000 stafil 22446@http://forums.leaflabs.com/ <p>How I measure frequency on pin? Pls sample code. I have measure frequency 200Hz - 800Hz.</p> <p>Pls help me </p>