Home made audio amplifier powered by running PC PSU?

[buffallo]

Hi guys,

I made myself an audio amplifier and I have the high hopes of installing it in my PC's casing, so that I can have some sort of integrity.

The amplifier basically has 3 transistor stages - input buffer, driver and push-pull power output. I have 'matched' it at about 4 ohms output impedance, and viewing on oscilloscope the voltage output on a 4-ohm speaker I can tell that I get around 10 V peak to peak on it. Thus, I calculatethe maximum current, which will pass thru the load at that time wich is about 2 amps. The quescent current (without input signal is about 30mA ).

For now I am powering the device from the 12V terminal of a converted ATX power supply, made into a stand-alone bench power supply, but I really want to integrate the amplifier and the speaker in my PC without an additional power supply. However, my fear is that the amp may produce large voltage/current spikes which may interfere with the rest of what is powered from the PSU - the computer itself.

So... What do you guys think, should I give it a try on my running PC or search for a better option?

Any suggestions are welcome!

Thanks,
Dimitar

 

[betwixt]

It might work but the PSU may not take kindly to the varying load, especially if you have made it stereo and the peak load is doubled. However, I suspect you will have other problems, particularly noise on the supply and ground loops as the 12V line probably also powers the motors in hard disks/CD/DVD/BluRay drives. The only way to find out is to try it.

Brian.

 

[godfreyl]

IMHO, it should be OK. The current drawn by the amplifier is well within the limits of most PC PSUs, and I doubt the variation in current will cause much fluctuation of the supply voltage. As you have a scope, it would be easy to check the voltage fluctuations on your existing supply with the amp playing loud.

Just out of interest, could you show a schematic of the amp? It sounds a bit unusual.

 

[buffallo]

I was also thinking about the problem with the interference with other periferial devices, esp. hard drives (as they seem to be consuming large current from the 12V line). Maybe adding proper filters (i.e. capacitors) may fix the problem, but then the startup current could grow to be too large :? This should be investigated.

As far as the amp is concerned, I will post the schematic of the amp in ltspice as soon as I get home. What is the unusual of it?

 

[godfreyl]

As far as the amp is concerned, I will post the schematic of the amp in ltspice as soon as I get home.

Thanks

What is the unusual of it?

Just your description of the 3 stages. Most amps don't start with a buffer at the input, and if they do then there's normally more than 2 stages after that.

 

[Audioguru]

The output impedance of a modern audio amplifier DOES NOT match the load impedance. The amplifier output impedance is 0.04 ohms or less while your 4 ohms speaker is 100 times higher or more so that the amplifier can damp speaker resonances.
Old vacuum tube amplifiers matched the impedances with very little damping.

Your amplifier produces a whopping 3.1W into 4 ohms, a little more than a cheap clock radio.

Recently I purchased a couple of Name Brand clock radios for $1.75 each. They sound awful. I connected a half-decent speaker to one of them and it sounds pretty good.

 

[buffallo]

So, the title of thread clearly isn't "Should I match the impedance of an audio amp to a 4-ohm speaker", or "Please help me make the greates amp" or whatevs.

Please, don't stray off the topic. I don't have a life time experience in audio amps, nor am I an audio guru. I am satisfied with what I have managed to build.

Thanks

 

[Audioguru]

So, the title of thread clearly isn't "Should I match the impedance of an audio amp to a 4-ohm speaker", or "Please help me make the greates amp" or whatevs.

Please, don't stray off the topic.

In post #1 you said,"I have 'matched' it at about 4 ohms output impedance" which is wrong so I corrected you because I do not want people reading it to think it is correct.

 

[buffallo]

I don't understand why matching an output impedance to a load impedance is so 'wrong to you'?

Old vacuum tube amplifiers matched the impedances with very little damping.

I am not building a super modern hi-fi audio system! Dissapointingly, it is not even using tubes, just a couple of cheap transistors. But still it meets my humble audio requirements.

So, since you obviously know the right way of 'impedance mismatching', and you clearly want to shout it out loud, please, explain how a modern amplifier is 'matched' or 'mismatched' with a speaker. And why is it done so?

 

[Audioguru]

I don't understand why matching an output impedance to a load impedance is so 'wrong to you'?

Because if your amplifier has an output impedance the same as the speaker impedance then its output power will be half of what it is now, it will get hot and the speaker will resonate making bongo drum sounds.

I am not building a super modern hi-fi audio system! Dissapointingly, it is not even using tubes, just a couple of cheap transistors. But still it meets my humble audio requirements.

A simple audio amplifier circuit designed correctly is a super modern hi-fi audio system.

So, since you obviously know the right way of 'impedance mismatching', and you clearly want to shout it out loud, please, explain how a modern amplifier is 'matched' or 'mismatched' with a speaker. And why is it done so?

WHAT IS THE MATTER WITH YOU?? (I was not shouting before).
I simply corrected you. Why don't you understand about corrections? People reading this stuff do not know when something is wrong unless it is corrected.

Why don't you understand that a speaker resonates like a bongo drum and the extremely low output impedance of a modern amplifier damps the resonances?
Another expert also sees something wrong and wants to see your schematic.

 

[buffallo]

Because if your amplifier has an output impedance the same as the speaker impedance then its output power will be half of what it is now, it will get hot and the speaker will resonate making bongo drum sounds.

1)If a power source does not have any source impedance, the output power will only be determined by the load resistance. That is, the lower the ohms, the greater the watts. But only for sources without internal resistance. I honestly don't know how to make my final stange amplifier work as an ideal voltage source, but any ideas are welcome. I assume increasing quescent currents and adding some feedbacks might lower the output impedance, but there still be some. Increasing bias currents is something I don't want to do.

2)I can't remember what a bongo sound sounds like. Is it maybe a nasty effect, caused by a frequency that is being amplified more than all the rest, creating an unusal sounding. If so, I suggest a tank circuit would be needed to create such a filter effect. Yes, there is a decoupling capacitor and if we account for the inductance of the speaker we should be able to calculate the resonant frequency (for a series LC resonance).
So, we get a circuit like

amp (w/ resistance) -> LC circuit -> load

Isn't it that if the amp's resistance is larger than the load's resistance, the damping factor of the circuit would be greater? Correct me if I am wrong.

3)

WHAT IS THE MATTER WITH YOU?? (I was not shouting before).

The matter with me is that I don't want an overall assesment of my design. I am fine with it. I am receiving criticism (which is offtopic) and I feel kinda bad about it. I haven't said that my design is a role model, and I don't want to post it because it wouldn't help the situation any further.

4)

Why don't you understand that a speaker resonates like a bongo drum and the extremely low output impedance of a modern amplifier damps the resonances?

It seems I just can't. I maybe have never heard it.

So, don't be fooled, I don't want to get into fights with any of you, guys.
I can see that in general you are all very helpful and this is the way it should be.

People reading this stuff do not know when something is wrong unless it is corrected.

True! But then, people learn best from their mistakes.

Cheers

 

[Audioguru]

1)If a power source does not have any source impedance, the output power will only be determined by the load resistance. That is, the lower the ohms, the greater the watts. But only for sources without internal resistance. I honestly don't know how to make my final stange amplifier work as an ideal voltage source, but any ideas are welcome. I assume increasing quescent currents and adding some feedbacks might lower the output impedance, but there still be some. Increasing bias currents is something I don't want to do.

An audio amplifier has a fairly low output impedance and extremely high open-loop gain. When negative feedback is added to make the gain reasonably low then the distortion is reduced a lot, the bandwidth is increased and the output impedance is reduced a lot.

Look at the spec's for any modern audio amplifier or amplifier IC. The output power is almost doubled when the speaker impedance is halved.
My car radio uses 2 ohm speakers and a bridged amplifier circuit so that the 13V battery can allow the amplifier to produce the high power of 33W RMS at low distortion per channel.

2)I can't remember what a bongo sound sounds like. Is it maybe a nasty effect, caused by a frequency that is being amplified more than all the rest, creating an unusal sounding. If so, I suggest a tank circuit would be needed to create such a filter effect. Yes, there is a decoupling capacitor and if we account for the inductance of the speaker we should be able to calculate the resonant frequency (for a series LC resonance).

A speaker or bongo drum mechanically resonates (it is a weight on a spring) like a bell but at a low frequency. When a speaker is driven from a very low impedance then it is damped and it moves exactly like the signal without resonating. When a signal begins then the damped speaker produces sound immediately without delay. When the signal stops then the damped speaker immediately stops producing sound without "ringing". LC resonance does not resonate a speaker.

Isn't it that if the amp's resistance is larger than the load's resistance, the damping factor of the circuit would be greater? Correct me if I am wrong.

No, the opposite. A speaker is a motor. If you short a rotating motor or block a kid on a swing then it stops moving immediately.

The matter with me is that I don't want an overall assesment of my design. I am fine with it. I am receiving criticism (which is offtopic) and I feel kinda bad about it. I haven't said that my design is a role model, and I don't want to post it because it wouldn't help the situation any further.

We try to stop laughing at a nOOb's circuit. Sometimes we help a nOOb with his circuit.

 

[ftsolutions]

In any case, your amplifier should be using some form of a Zobel network on the amplifier's output to help with providing a proper load and reducing unwanted reactance with varying speaker loads. If you can show your design we can help if you want the input.

 

[buffallo]

Audioguru, I think we are generally discussing two different (yet actually similiar) phenomena - mechanical resonance and electrical resonance. What you are saying about damping the mechanical/magnetic energy, which builds up in the speaker - it is true. It is so, I couldn't disagree.

We try to stop laughing at a nOOb's circuit. Sometimes we help a nOOb with his circuit.

And we are back to the all time curse of forum conversations. There was a very nice joke about forum members in our country. It is as follows:

"You ask a question in an American forum and you get answers. You ask a question in a Russian forum - you get even more questions. You ask a question in a forum in our country - everyone starts saying how stupid you are (and no one actually comments on your question).

So what is going on here? I have made something. It is a crappy amplifier, doesn't sound good in your opinion, should be made in a differen way, it is not standard for what typical amplifiers look like, etc. And I am asking if it lould interfere in some way with a computer if powered from its PSU. I thought giving detail about the amp would give a much better understanding of my real problem. And I want to stress on this - my problem is not how the sound from the speak sounds. Thank you for your suggestions, but this is not what I need.

I feel like you're trying to demonstrate overpower in knowledge and experience.

 

[Audioguru]

A computer power supply has a lot of power (at least 60W) and the amplifier you made has a pretty low output power (3.1W) so the power supply will be fine.
You did not and will not post the schematic of your amplifier so we do not know if it is good or bad, it might be excellent.
I simply commented and corrected you when you wrongly said its output impedance matches the speaker.

 

[godfreyl]

So, the title of thread clearly isn't....

Thanks for the chuckle! Add "How much did you pay for your clock radio?" to the list of questions not asked. :lol:

Another expert also sees something wrong and wants to see your schematic.

Hang on, don't drag me into this mess - I never said there's anything wrong with his amp, I was just interested to have a look.

- - - Updated - - -

Anyway, back on topic....

I was also thinking about the problem with the interference with other periferial devices, esp. hard drives (as they seem to be consuming large current from the 12V line). Maybe adding proper filters (i.e. capacitors) may fix the problem, but then the startup current could grow to be too large :? This should be investigated.

These are good thoughts. I think one of the most important steps is to separate the high current wiring (especially ground) of the amp from that of hard drives etc as far as possible. PC power supplies normally have plenty of leads with Molex connectors, so it shouldn't be a problem to dedicate one to the amp, and use the others for hard drives etc.

A supply decoupling capacitor at the amp itself is always a good idea anyway, and may help with the interference as well. A modest value shouldn't cause any problems with start-up. My gut feel says 100uF should be fine but 10000uF may be a problem. However I've never tried it, so some experimenting would probably be a good idea.

 

[audiomik]

Re: "My gut feel says 100uF should be fine but 10000uF may be a problem. However I've never tried it, so some experimenting would probably be a good idea."

Suggest also parallel connecting a 10 to 100nF Disc Ceramic onto the electrolytics as there are some high frequency switching 'spikes' on PC power lines.
Otherwise try electrolytics specified for switch mode PSU outputs, with a high frequency ripple current rating.

Did try this and it improved amplifier output noise, but signal grounding was a far bigger issue and not easy to solve properly to remove 'digital noise' breakthrough.
Eventually had to use a separate PSU for the Amplifier circuit, but a different motherboard/PSU may well perform differently.

Mik

 

[Audioguru]

Some amplifier designs have an input bias resistor connected directly to the raw supply voltage then the amplifier amplifies all the noise on the supply.
A better way to bias would be an RC filter from the raw supply that feeds a smoothed voltage to the input bias resistor.

 

[ALERTLINKS]

https://ludens.cl/Electron/audioamps/ta8215.html
I made this amplifier some years back. I did not bothered about power requirements and hooked into my PC. The only problem was the extra pickup of noise due to ground loops. There was a lot of noise from CPU and hard disk. It was minimised by finding a ground point of supply for the amp from motherboard which was from near sound card, and also by not connecting ground of input shield wire to audio output socket.
It did not bothered the power supply even when speaker leads were accidently shorted and IC blew. I had not put any fuse with amplifier.

 

[buffallo]

https://ludens.cl/Electron/audioamps/ta8215.html
I made this amplifier some years back. I did not bothered about power requirements and hooked into my PC. The only problem was the extra pickup of noise due to ground loops. There was a lot of noise from CPU and hard disk. It was minimised by finding a ground point of supply for the amp from motherboard which was from near sound card, and also by not connecting ground of input shield wire to audio output socket.
It did not bothered the power supply even when speaker leads were accidently shorted and IC blew. I had not put any fuse with amplifier.

Great work! This is basically what I want to do, except for lower power amp.
So, if I connect the Audio signal output GND of my sound card to the molex GND wire which feeds to my amp, I may be able to minimize the noise from ground loops?
Furthermore, what if a differential amp (or any with high CMRR ) is used as initial input stage? Could that proove to also be effective?

P.S. I haven't yet tested it on my PSU. I will update the thread with my results as soon as I can.

Thanks to everybody participating in the discussion

Cheers