I need to design an amplifier for the project going on over here-http://www.diyaudio.com/forums/showthread.php?s=&threadid=85752
I'm designing speakers for my corporate america cube to play at around 65dB (estimated). Even the lowest sensitivity speakers aren't even going to need 1W, so this presents a special opportunity to design an extremely simple amplifier.
The amplifier design is the second design project. My goal is to design the most simple sub 1W amplifier possible. I have a power supply for it already, from a failed gainclone project. IIRC it puts out about ~30-35VDC. I hope to be able to come up with something that uses just a single transistor (FET, JFET, MOSFET, bi-polar) per channel. No need to ultra-filter the ripple, constant current source this and that, or make the distortion .000001 with massive feeback, have a high gain with multiple stages to get there, etc.
Source = PC. It's entirely possible that the PC's soundcard has enough gain, but I need a current buffer. Does that sound right? Again, more fuzzy logic from what I've soaked up on the net: the soundcard doesn't have the right output impedance to be hooked directly up to 8ohm speakers, so instead of hooking up a high gain amp, I can do a current buffer of sorts (with obviously the right output impedance), right?
Driver = Hi-Vi B3S.
Realistically, I need less than 0.25W, want just one transistor doing the amplifying, want it to be as simple as possible using just a handful of parts, don't care what it measures, etc. I can't think of anything else for now. Please ask questions on info I haven't given.
Should I be looking at the SEWA design on this forum?
I'm designing speakers for my corporate america cube to play at around 65dB (estimated). Even the lowest sensitivity speakers aren't even going to need 1W, so this presents a special opportunity to design an extremely simple amplifier.
The amplifier design is the second design project. My goal is to design the most simple sub 1W amplifier possible. I have a power supply for it already, from a failed gainclone project. IIRC it puts out about ~30-35VDC. I hope to be able to come up with something that uses just a single transistor (FET, JFET, MOSFET, bi-polar) per channel. No need to ultra-filter the ripple, constant current source this and that, or make the distortion .000001 with massive feeback, have a high gain with multiple stages to get there, etc.
Source = PC. It's entirely possible that the PC's soundcard has enough gain, but I need a current buffer. Does that sound right? Again, more fuzzy logic from what I've soaked up on the net: the soundcard doesn't have the right output impedance to be hooked directly up to 8ohm speakers, so instead of hooking up a high gain amp, I can do a current buffer of sorts (with obviously the right output impedance), right?
Driver = Hi-Vi B3S.
Realistically, I need less than 0.25W, want just one transistor doing the amplifying, want it to be as simple as possible using just a handful of parts, don't care what it measures, etc. I can't think of anything else for now. Please ask questions on info I haven't given.
Should I be looking at the SEWA design on this forum?
Hi Phorensic,
wouldn't Your one watt project, being supplied by a +-30V supply, actually become a ten watt project?
Power into the speaker is defined as P = U*U/R, which calculates as 1W = U*U/8Ohm, assuming eight ohms speaker impedance, which resolutes to 2,83V out of the amplifier. Current through the speaker is further given as I = U/R, giving us 0,35A.
According to Kirchhoff's law, this current flows through both the amplifier and the speaker - at the same time. This means the amplifier has to conduct 0,35A while dropping about 30V from the supply, resulting in a - wasted - power dissipation of 30V*0,35A = 10.5W inside a single transistor. This would be just crazy, even for an extreme class A attempt...
I think this project has to be reconsidered. Why did Your chipamp project fail? A chipamp would be perfect, as those chips are actually *intended* for that purpose, and there are low-power chips available, too!
The concepts You are looking for are e.g. "emitter follower" (for bipolars) or "source follower" (for fets), and they work with just a little more voltage than those 2,83V and only require some additional passive components. A search in electronics literature or on the web will yeld many results. An emitter or source follower can even be combined with an opamp and could actually make a quality amp. You can find such solutions in expensive headphone amplifiers for example. But as You describe it, even the most simple solution would work for You.
Hope this helps,
Sebastian.
PS: The SEWA is a good project, and for 1W, somy class A amps can sound incredibly good while still being cheap!
wouldn't Your one watt project, being supplied by a +-30V supply, actually become a ten watt project?
Power into the speaker is defined as P = U*U/R, which calculates as 1W = U*U/8Ohm, assuming eight ohms speaker impedance, which resolutes to 2,83V out of the amplifier. Current through the speaker is further given as I = U/R, giving us 0,35A.
According to Kirchhoff's law, this current flows through both the amplifier and the speaker - at the same time. This means the amplifier has to conduct 0,35A while dropping about 30V from the supply, resulting in a - wasted - power dissipation of 30V*0,35A = 10.5W inside a single transistor. This would be just crazy, even for an extreme class A attempt...
I think this project has to be reconsidered. Why did Your chipamp project fail? A chipamp would be perfect, as those chips are actually *intended* for that purpose, and there are low-power chips available, too!
The concepts You are looking for are e.g. "emitter follower" (for bipolars) or "source follower" (for fets), and they work with just a little more voltage than those 2,83V and only require some additional passive components. A search in electronics literature or on the web will yeld many results. An emitter or source follower can even be combined with an opamp and could actually make a quality amp. You can find such solutions in expensive headphone amplifiers for example. But as You describe it, even the most simple solution would work for You.
Hope this helps,
Sebastian.
PS: The SEWA is a good project, and for 1W, somy class A amps can sound incredibly good while still being cheap!
sek said:
"Headroom" I guess you could say, hehe
sek said:
The chimpamp failed because no matter how many times I COMPLETELY rebuilt the right channel, it would instantly fry a resistor and sound like AC was going right to the speakers. I tried EVERYTHING. Gave up because of the total waste of time involved on it and bought a Jolida 1701. The ironic thing? I had no idea the Jolida had the same exact chip inside of it until I opened her up, hahahaha. Tube preamp section in it, though, which is why I selected it in the first place.
sek said:
See below.
bigwill said:
From what I gather, making a LM386 based amp would be exactly like building a Chu Moy amp that I have already built--
http://www.headwize.com/projects/cmoy2_prj.htm . Correct? If so, I already have an amp! Just need a power supply for it. I was using it for headphones (like it is designed for) and it sounded excellent. Just need to mod it a bit for speaker use I guess.
This poses a new question. I already have a fantastic, overkill ~+/-30V power supply built as mentioned. Both the suggested LM386 chip and my already built Chu Moy require a max voltage of +/-18V. I'm just a newb solderer who doesn't actually know how the circuits are working, completely anyway. What would the schematic look like, or what components would be necessary to drop the 30V to 18V, efficiency again being no issue? I'm pinching pennies here as you can see. http://chipamp.com/docs/lm3886-manual.pdf page 4 is what my PSU looks like.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.