Use your stereo amplifier as an AC power supply

[drummerdimitri]

Hey guys,

A while ago I had this wild idea of using my stereo speaker's amplifier as an AC power supply for any low powered device that requires it.

I thought why not play a sine wave of 50 Hz such as the mains electricity frequency and adjust the volume until I reach the voltage specifications of the device I am trying to power.

Worked like a charm so long as I don't draw too much current and the amp shuts down its outputs for protection.

Too bad I don't own an amp powerful enough to generate a 220V 50 Hz sine wave. That would make for a very impractical and expensive AC filter 

Let me know what you guys think.

[Zero999]

It can be done but is a very expensive way to do it.

Unless you need a different waveform/frequency then a Variac should be fine.

No audio amplifier will be able to output mains voltages directly, unless you're talking about the public address type which can output around 100VAC. You'll need to connect a transformer to the output to get higher voltages.

I've done this myself when I needed 30VAC RMS with a low impedance over the audio range.

[Euphyllia]

I would worry about the load characteristics. Even good class AB audio amps tend not to like driving complex impedances with more than +/-60 degrees of phase shift. You'll see any good loudspeaker crossover try and keep the speaker's impedance in this range. Depending on what you have downstream the amp might not like it.

[NiHaoMike]

An audio amplifier is just a very specialized inverter.

[RJFreeman]

look for an Australian Monitor AM1600 on Ebay, these now go for under $500 and deliver 1100Wrms into 2 Ohms.

Come to that with the advent of modern light weight D class and Switch mode amplifiers there are now plenty of second hand 'Iron and copper'  Professional amplifiers on the market at low prices nowadays at low prices and if you get a decent amplifier (Crown, EV, HH, Yamaha, Australian Monitor, Perreaux etc) then they will usually happily drive pretty obtuse loads before you run into a VI limiter or otherwise have trouble (The Aus Monitor and Perreaux are Mosfet so don't have a VI limiter).

[tomizett]

I remember seeing, about 10 years ago now, some specialist piece of physics equipment that needed a nice clean mains supply being driven from a big amplifier (probably, now I think about it, through a transformer).

[diyaudio]

Hey guys,

A while ago I had this wild idea of using my stereo speaker's amplifier as an AC power supply for any low powered device that requires it.

I thought why not play a sine wave of 50 Hz such as the mains electricity frequency and adjust the volume until I reach the voltage specifications of the device I am trying to power.

Worked like a charm so long as I don't draw too much current and the amp shuts down its outputs for protection.

Too bad I don't own an amp powerful enough to generate a 220V 50 Hz sine wave. That would make for a very impractical and expensive AC filter 

Let me know what you guys think.

fun thing to do.

Class A, AB will just suffer from all kinds of inefficiencies. however with that said, a "well designed" Class-D amplifier will handle attached load with pleasure, with very little thermal problems, and yes, it is actually a switching converter with its front-end designed to modulate to an audio signal. 
 

[Pjotr]

Class A, AB will just suffer from all kinds of inefficiencies. however with that said, a "well designed" Class-D amplifier will handle attached load with pleasure, with very little thermal problems, and yes, it is actually a switching converter with its front-end designed to modulate to an audio signal.

Don’t think so. Keep in mind that audio amplifiers are thermally not designed to deliver continuous RMS power. They are designed to deliver temporarily maximum power in loud passages in music. Continuous average power is usually 1/10 of rated power.

Linear class AB are in general not suited, due their large inefficiency and difficulties with complex loads over 45 deg phase shift.  Better are Class-D amplifiers (switch mode amplifiers). I have had good results with a 400W Hypex class-D amplifier module and a reversed 230V/32V 500VA transformer. Those can easily handle complex loads up to 90 deg phase shift, if the PSU caps are big enough (>10.000 uF). But not for long time at full load. With sufficient heat sinking the Fets are not a problem but the on board powers supply bypass caps run pretty hot and the output choke as well (over 100 deg. C).

If you want to try, always put reversed bypass diodes from the output to the supply rails to handle leakage inductance of your step-up transformer. The feedback loop is usually not fast enough to track fast high voltage transients caused by that inductance and those transients can destroy your output transistors.

[KJDS]

Many years ago, when testing that equipment would meet everything from 88V to 264V and from 45Hz to 440Hz, we'd use a huge valve amp fed with an audio sig gen.

[Anks]

Crown macrotech VZ 5000's have been used in labs for years. you can even find the ones that have the special factory outputs on the front.

[linux-works]

this would be interesting; I read about a time-nut guy who created a local 60hz (exact!) power source just to power his ac-based digital clocks.  the clocks got their timing from the powerline 60hz and so if he created the 60hz from his rubidium divider (say) then the clocks would be automatically accurate short-term and long-term.

clocks don't take a lot of power, so this might be doable.  take a 10mhz standard or even a 1pps standard and scale it to 60hz for a sine, then amp it and power your time-based items from it.  lol

[fcb]

You don't specify the amount of power you need at 220VAC.

I've used various power amps in the past (driving small toiroidal transformers in 'reverse') to simulate mains.  Typically I put a few ohms in series with the secondary, and 10ohms across the amplifier terminals - in my case this helped with the stability.  As other's point out, you would need a huge PA to simulate mains if directly connected.

[AG6QR]

Here's a slightly related topic. 

http://www.leapsecond.com/pages/atomic-nixie/

A guy had a few old clocks that count the 60Hz AC cycles of mains power for timekeeping.  The power grid frequency varies significantly over the short term, but at least around here, the power companies periodically adjust it to keep its long term drift quite low.  For most practical household purposes, a clock driven by a synchronous motor can be set once, and needn't be adjusted unless the power fails.

But that wasn't good enough for this guy.  He wanted to improve the accuracy of his Nixie clock and a motor driven flip clock.  So he synthesized a 60.000000000000 Hz signal (and yes, those zeroes are all significant) using a cesium time standard, and then amplified that to around 100V in order to drive the clocks.  His clocks are much more stable and accurate than they were when driven by the everyday power grid.

Practical?  Perhaps not.  But definitely cool!

[Pjotr]

Indeed COOL

I did something the same with a 50 Hz pulse (mains clocks run at that in Europe) that was divided from a 0.1 ppm OCXO. But it didn't help, I still was often late, it didn't account for the traffic jams :lol:

B.t.w. here in Europe we have an interconnected power grid all over Europe and as a result a pretty accurate average 50Hz, locked to a cesium clock.

[diyaudio]

Class A, AB will just suffer from all kinds of inefficiencies. however with that said, a "well designed" Class-D amplifier will handle attached load with pleasure, with very little thermal problems, and yes, it is actually a switching converter with its front-end designed to modulate to an audio signal.

Don’t think so. Keep in mind that audio amplifiers are thermally not designed to deliver continuous RMS power. They are designed to deliver temporarily maximum power in loud passages in music. Continuous average power is usually 1/10 of rated power.

Linear class AB are in general not suited, due their large inefficiency and difficulties with complex loads over 45 deg phase shift.  Better are Class-D amplifiers (switch mode amplifiers). I have had good results with a 400W Hypex class-D amplifier module and a reversed 230V/32V 500VA transformer. Those can easily handle complex loads up to 90 deg phase shift, if the PSU caps are big enough (>10.000 uF). But not for long time at full load. With sufficient heat sinking the Fets are not a problem but the on board powers supply bypass caps run pretty hot and the output choke as well (over 100 deg. C).

If you want to try, always put reversed bypass diodes from the output to the supply rails to handle leakage inductance of your step-up transformer. The feedback loop is usually not fast enough to track fast high voltage transients caused by that inductance and those transients can destroy your output transistors.

I think you just re-answered me.

The class-d amplifier is far different from a traditional amplifier that uses a BJT, by design a well designed class-d output FET`s operates in their saturation region, Rds(on) dominates.
FET technology for class-d amplifiers has matured over the last 8 years dramatically. Have you seen the specification of some them? they are darn impressive.

Example

A irfb5620pbf exhibits a 72mOhm Rds(on), offers a VBRDSS of 200V and offers a gate channel capacitance of  Qg of only 25nC, with this data you surely understand that  if we designed a PA amplifier with a DC - bus +/-90VDC, and lets say for argument sake the supply dropped +/-89Vdc at full load, with a 2-ohm load connected, the continuous RMS rating would amount to V^2/R  = 3.9kW (rms) also note I assume the FET`s are paralleled to achieve this level of power. 

The  thermal efficiency is very good some well designed Class-D designs requires almost no heat sinking, I've built one (a kit) where I was able to play music at loud volume at 100W (for an hour or so) with "no heat sinking" its almost unbelievable!, note this is of course achievable because of  large programmed dead-time during switching times. to achieve efficiency. A commercial class-d PA amplifier using a no heat-sink approach like the Behringer iNuke amplifiers. (see attachment no heat-sink and this is rated at 2kw > @ 8ohms)

Why do you feel a Class-D configuration is incapable of delivering continuous RMS power into a load? In all reality its a switch mode converter with the same efficiency figures irrespective if it amplifiers audio, Its way more superior in efficiency than its A, AB brothers.

[woodchips]

Yes, just use a step up transformer, might have to experiment with a real load though.

But, have you seen these inverter generators? A usual generator has a 3000 rpm petrol engine driving a 2 pole alternater. The inverter generators use a 5000rpm engine driving a three phase alternater followed by what is in effect a variable frequency inverter used for induction motors. Must be a good base for experimentation?

[Pjotr]

Why do you feel a Class-D configuration is incapable of delivering continuous RMS power into a load? In all reality its a switch mode converter with the same efficiency figures irrespective if it amplifiers audio, Its way more superior in efficiency than its A, AB brothers.

I did not say a Class-D configuration is incapable of delivering continuous RMS power into a load. Sure they can, but not at full rated power given by the manufacturer. They can, if the are designed for. The topic title states "Use your stereo amplifier as an AC power supply". That means audio amplifiers for music reproduction, not lab amps or shaker table amps.

Also stereo class-D audio amplifiers are designed for their purpose: Music reproduction. To design them for continuous RMS sine load would be useless over-engineering.

In the example I gave, the output choke is properly designed for audio purposes, but not for continuous sine load at full power. Same for the PSU and current rating of the on-board bypass caps.  They will smoke in 10 minutes at full continuous sine load, but will last long for the intended music reproduction. The 400W modules I referred to, and these are very reliable, will spit out only 40 - 80 watts max. under continuous sine load without damage I guess.