Greetings, A person was kind enough to show me these forums and it got me started on the idea of modding my comp (most of my time before now was spent playing with circuts that were rather destructive). I have a basic knowledge of electronics and am currently starting on my first mod for my computer (can be used with anything else that outputs audio if altered a bit). I just have an idea, but I have every intention of completing as soon as I can work out how to get it all to work. My idea thus far is to build a device similar to the graphical part of the equalizer on a home stereo that just shows the siganls coming off of the source and is displayed in bars which are divided into several bands of the frequency. I want to make a grid of led's that is 3 in by 3 in and 1/2 in in depth (hopefully the entire unit will be 3"x4"x1/2" with all components). I have gotten some help on how to achieve the grid of led but I could use some advice on how to get it to interface with my sound card (Sound Blaster Audigy 2 ZS Platinum) through the use of the speaker jacks and still allow the signal to go to the speakers undistorted. I would greatly appreciate any advice on how this can be done. And once I complete this project I will upload the schematics and pics to the guides forum. Any advice/suggestions would be greatly appreciated.
Proper Spectrum Analyser circuits are a bit thin on the net but 3-channel sound-to-light are more common, and the front end is about the same. Take & boost your signal with an opamp, the soundcard won't even know its gone, then fan out, through an opamp active filter for each channel, to your bar drivers. Sanyo do some cheap 5-led vumeter chips that Rapid Electronics sell (LB1443N).
i would have to disagree slightly, there are quite a few spectrum analisers, just most end up been horrifically complicated, if complex numbers + calculas don't do it for u. http://8ball.ofb.net/~heath/eeg/laco-98.pdf - is a prime example of a fantastic design. i mean realy good for the time, bit dated now 6 years on.
They've added some more, that's my excuse & I'm sticking to it. If you think PICs are for wimps, http://www.web-ee.com/Schematics/ThreeChanSpecAnalyzer/ThreeChanSpecAnalyzer.htm is a cheap, simple 3-channel, and you could use an LM324 & the Sanyo bargraph ics.
If you think pure logic is for pre-pubestent training-bra-wearing girlies who have such tiny breasts jacko would do them. http://www.library.cornell.edu/nr/bookcpdf/c12-0.pdf a bit about FFT more at http://www.library.cornell.edu/nr/bookcpdf.html
I am probably going to use a LM3914 IC it drives 10 LED and can cascade to more if needed but I am just going to use one IC per frequency band and I am going to try to divide it into 10 bands. I thank all of you for all of your help any more advice/suggestions would be greatly appreciated.
Sorry to ressurect something from the dead but I finally got back to working on this idea. I have most of it worked out but it will be a pain to do since most of it will be point to point wiring. But I still need to figure out how to rip a signal from my sound card split it into the frequencys, run it through this circiut and then send it off to the speakers without any distortion. This is something I threw together earlier (hope it makes some sense). Sorry about the image quality as this was done in an hr. or so on paint. The resistor values have yet to be determined. But any help on how I can finish this circiut would be very helpful as I am still pretty new to electronics, though I have done some desigining of my bays (just the hum of a dremel and the soldering of the new cages)
You don't send it through the circuit then on to the speakers, you take the signal from line_out in parallel with the signal going to the speakers. Op-amp U1 has a very high input impedance, won't affect sound quality if you use screened cable to get to it. Then split off into your filters.
Thanks for that piece of advice, but how would I split the signal into 10 frequency ranges? All I can remember on doing that was that my Electronics teacher briefly mentioned it and said something about caps and inductors. If you would like i could post a link to a .bmp of my file so that you could modify the pic.
I would think you would utilize a multi-step crossover but I'm not sure. Someone would have to elucidate about it.
The circuit I linked to shows the method. The first buffer op-amp fans out to 3 active filters built around other op-amps, each handling a section of the audio spectrum. You need to design 10 active filters, or google for a set with component values already calculated.
After much searching I finally came across schematics for a filter, but I am a bit confused by the end components or whatever they are, is anyone willing to clarify for me? I have no idea how to hook up the filters to each other so if anyone is willing to just show me how to wire 2 of those filters together I could probably figure out how to cascade the rest.
The symbols for Vin and Vout are switched mono audio jacks. In this case though, the output of the last op-amp would go to the input of the LM3914 rather than to another audio jack. Problem though: that design is dual-rail (i.e. you have a positive supply voltage, ground, and a negative supply voltage). While you could still use it, you'd need to tap into what'd most likely end up being -12v on the ATX connector (depending on the audio output, it might peak at 2v or 10v, and supply should higher than that, so 12v it is). However, if you can find a good single-rail op-amp meant for audio duties, you can still use the schematic.
The circuit you've found looks like a graphic equaliser, so after adjusting the volumes of the individual frequency bands inside the dashed box, the final op-amp outside the box mixes them back together before they go into the power amplifier. In your case you need to keep them seperate and send each filter output to its 3915. Put an op-amp on the input end as a low-gain buffer. It's a bad example to take, as you don't need the 10 volume pots, apart from it using a dual-rail supply. You can replace the 10k-100k pot-10k string with a pair of fixed 62k resistors, and it's worth finding out what would happen if you used LM324 op-amps on a single 12V rail. I've a suspicion you'd end up with half-wave rectified audio signals that would sound terrible but would drive the bargraph chips no problem. Time to learn circuit simulation.
Well I finally found some time to pick up the project again and I figured what the heck. I made the schematic this far, why not do it all in paint. As for the simulation I found some program that was supposed to be able to simulate it and even allow me to design a PCB and then send it off for someone else to develope. I tried working in it, I gave up after 4 hrs. I have most of the filter done, and in my searches I found that most of the schematics for existing sound to light circuits just connected the v- to ground on the opamps, so I think that I should be ok just doing the same. As you will see in the giant pic below, I have begun to label the values of the components. I know there is a mess with the op amps right now but it is a minor discrepancy and will be corrected later. I would greatly appreciate it if you could give me a hand in finishing up the circuit. After it is done I will create a guide on the forums with all the appropriate files uploaded.
Sorry for bringing this thing back from the dead again but I really want to get this done since I have gotten this far. Upon some further research, the op-amp that was used in those 3 channel spectrum analysers was an LM3900 quad op amp. And in the examples I came across it was used in such a manner that all it needed was a + 12 volt supply, the v- pin on the op amp was just connected to ground. That is about all I have been able to figure out, I read the data sheet on the op amp and was lost, so I would be very grateful for any help on this circuit from those who could help me. Many thanks to those who have helped me get this far, Sea Shadow
Well, not a ton of help, but here is the schematic for the Velleman K4300 10 band spectrum analyzer. It's been discontinued since one of the ICs is no longer made, but you may be able to use the filters from it. http://www.animeguru.com/images/velleman/k4300.html Good luck.
Great, thanks, I recall coming across those once on the forums before, but I couldn't remember where it was or what it was called when I began this endeavor. I think I will swap out the filter section and use the one from the vellman design. Thanks! Edit: Now that I look closely at that filter schematic you provided, I wonder if there is a way to use the existing filter setup I have right now, I want to keep this as cheap as possible and with as few components as possible w/out having to deal with multiplexing, I hope to keep the cost of building this under $30 or so.
