|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.123.24.133
I did this more for the fun and learning experience but was quite surprised at the results. Heres the parts I used.
1 - 500w pc power supply
1 - old 4 channel Kenwood car amplifier KAC646x
1 - pair of cheap radioshack 2-way bookshelf speakers Minimus7
1 - diy 8in subThe idea came up when my dad asked me to put together a super cheap desktop system for his PC so he could listen to music while hes working at the PC. He has a cheap 2.1 Logitech system and didn't like the sound at all. My mom wouldn't let him spend anymore money on audio/video equipment since he just bought a 42in lcd screen and surround system. So I looked around in the basement storage area and came across the stuff I mentioned above.
I hooked up the amp to the power supplies 12v rail and shorted the switch wires that are used to turn the PC on/off. This got the power supply to power up without being connected to a motherboard. I then connected the bookshelf speakers in stereo to two of the channels and let them run full range. The other 2-channels were bridges and connected to the 8in sub, the amps 80hz low pass was used. I connected my MP3 player to the inputs of the amp and powered everything on. To my surprise everything worked as I planned and even sounded very good. It sounds far better and plays louder and clearer then the 2.1 logitech system that my dad currently uses. I'm in the process of testing it now to make sure the power supply doesn't over heat. Its been on for a few hours playing loud bass heavy music and the power supply isn't even warm! If it gets through the testing phase Ill build a case for the amp and psu. Any ideas what I could do to improve this setup?? I was thinking of maybe some sort of filtering before the power goes to the amp to get rid of any of the switching power supply noise. And also was thinking of a storage capacitor or two. What do you guys think?
thanks
Follow Ups:
Rat Shack really cheaped out on the crossover caps on those speakers and by now they're really showing their age. Swap out the stock electrolytic caps for some modern poly caps of the same value and the "Mini-7's" will sound better than ever ;^) .
Check this link. It is showing a great power supply filter. Of course you would have to fit the design to your app, but that isn't really a big deal with this circuit. Using this technique you could nearly approach battery quietness and get some ultra fine regulation as well as long as you don't go too cheap.
Maybe im missing something, but if I use some type of regulator even a low dropout reg, I dont think ill get the 12v I need to power the amp?? What are some other ways of filtering the noise from a switching supply?? Right now I have a 10000/uf cap near the amp and also a small choke that I pulled from a car audio hook up kit that I had use years ago. Any other ideas?
Your amp will still work fine with slightly less than 12V. You could easily design an LDO to drop less than 100mV or so depending on the FET you bought. You probably would not be able to identify the amount of difference in the sound without special equipment assuming the drop was small enough.Another approach would be some type of PI filter, ie CLC. This requires really big capacitors on the secondary side of the filter in order to keep ripple low enough. However this approach could be even quieter than the LDO appraoch since the braodband silicon noises would get filtered. The main drawback with this type of filter will be the output impedance and the resulting ripple.
I think an LDO would give the most bang for the buck for what is important to me: Low noise, low ripple, and low output impedance.
Either way you decide to go, throwing more money at the problem gets you closer to the same ideal goal in the end but with different tradeoffs. You gaotta decide what is most important to you.
What size cap and inductor would you reccomend if i descide to go with a Pi filter?? I have alot of various capacitors laying around so it may be less expensive then a buildng some sort of LDO regulator circuit.
What size cap and inductor would you reccomend if i descide to go with a Pi filter?? I have alot of various capacitors laying around so it may be less expensive then a buildng some sort of LDO regulator circuit.
I am so far removed from my engineering math skills I won't even try to go there now. However IMO this is one design where it will be very easy to go overboard with good success. The ouput cap should be as big as bloody well possible to keep ripple as low as you can. Inductor size depends on power draw but I'm thinking 10mH should get you in the ballpark but bigger will always be better here. The input cap will likely be the power supply output filter cap which should be fine.Did I say that output cap size being big here will be important? because it will.
Hmm im think ive already done this actually if the the first cap is the output cap in the power supply. I have a choke inline with the 12v line and a 10000uf + .1uf from v+ to ground right before the amp. The only other thing i was thinking of adding is an rc network from +v to ground maybe a 1ohm in series with a .47uf cap from +v to ground to lower the indctance of the big cap a bit. I think building a voltage regulator may beyond my skill level. My only real expereince with regulators is the simple 78xx/79xx and 317/337 types.
10KuF isn't going to support much of an inductor at these power levels but then again you are presumably really only worried about fairly high freq stuff here so maybe a smaller inductor is in order. 100uH ish? In general the more inductance you add the worse the regulation will get so there is definitely a negative downside to this. However this exact negative effect can be overcome to an extent by bulking up on the amt bypass capacitance. It's a bit of a game to land everything right where you want it. It really is nice to have an oscilloscope to do this kind of development. Do you have access to one?Bypass the amps using normal good bypassing technique, ie short leads on all caps jammed right up against the power pins of the amp as close as is humanly possible and higher SRF (self resonant frequency) like a ceramic or polypropylene cap gets higher priority for the really tight to the power pin positions compared to lower SRF parts like the big aluminum electrolytic which can be farther away. BTW after the amp has been properly bypassed you may find no sonic benefit to adding the L in the pi or tyhe R of the RC. The better you can lay all the parts out in theory the more this is true.
In my opinion filtering is a bandaid and cascading filters would only make the downsides worse. But in the end you may want to cascade a pi with an RC to see how that sounds but again I'm betting that with the right pi or RC there is no need for rc after the pi or visa versa since in essence you are lowering the load/line regulation each filter stage you add.
BUT
You might want to try comparing the sonic benefits of an RC filter to the pi filter individually. The RC may actually end up the more reasonable design for this app!?!?! I've never designed for this exact problem so everything I've said so far is a big guess.
Hey thanks for all the help, I'm not sure you understood what I meant about the RC network. This would be a 1ohm resister in series with a .47uf capacitor, this is placed from V+ to ground after the large capacitor to help bring down the inductance, I got this idea from tnt audio Ill give you the link. I'm not sure if it will help with the type of power supply im using, but its worth a shot. I'm still getting a sort of high pitched buzzing type sound from the speakers when no music is playing and I put my ear close to the speaker. Its quite low and you don't hear it when the music is playing, but id like to get rid of it if I can. It doesn't get louder when I turn the volume up it always stays the same level???
I have a bunch of parts laying around not being used that i can experiment with. I have a few more of the 10000uf/63v BHX Aerovox capacitors that I used, I can try adding another for a total of 20000uf and see how it works out, I'm not to sure the PC power supply will like that on startup though, but its worth a shot. I also have a bag of capacitors with lots of smaller Panasonic FC, Cerafine and Oscons. I could try building a bank of smaller capacitors instead of one or two 10000uf. Ive found that a few smaller caps paralleled sounds better then a single large cap, so its worth a shot. Im really doing this mostly for the learning experience and fun, I wasn't even expecting it to sound as good as it did:) Thanks alot for your help.
Frank
This post is made possible by the generous support of people like you and our sponsors: