Modulating amp supply rails with 8-bit DAC

[Circlotron]

In the old days when a master tape was transferred to disc, to facilitate the variable groove spacing necessary for long playing time, on the tape deck an auxiliary replay head was placed a short distance ahead of the main one so any upcoming transients or loud passages in the programme material would be read by it before they actually appeared on the main output, thereby giving the cutter head notice that it had to begin spacing the grooves wider to fit in the bigger wiggles about to appear. The wasteful alternative was to have constant groove spacing wide enough to fit in the loudest anticipated sound.

A similar thing exists with power amplifiers. The voltage rails are set for the highest expected sound level and most of the time they are not needed. Especially is this so with class A amplifiers which most of the time waste power unnecessarily and as a result need great big heatsinks, power supplies etc. Here is my idea. Suppose you have a standalone DAC for your CD player. Upstream of this, place a simple 8 bit DAC and then a digital delay of say 1 second. The 8 bit DAC controls the amplifer power supply rails 1 second ahead of the main DAC output. That's it!

I know this kind of thing has been done before with Carver amplifiers etc, but only in an analogue way. Now that diy CD DAC's are a reality, I think it would only be a small step to include this facility in them. Especially would it be useful in a "big iron" class A amp where the bias current could be reduced in proportion to the rail voltage as well.

GP.

 

[ThingyNess]

I hope the amp never gets used for home theater or A/V use...

1 second delay would destroy any synchronization the audio + video had.

 

[deandob]

What about varying the supply voltage relative to volume setting? Normally your volume setting stays pretty much constant so you dont need to worry about tracking the input signal or transient response of the power supply etc. So when you dont want much volume you dont need to waste power. The TACT amplifier uses this technique.

Use a variac as your volume control, which adjusts mains voltage input to your mains transformer (note a variac does not isolate the mains so you still need a transformer). I have not tried this myself, but it should work, although watch out for the amplifier becoming unstable due to undervoltage conditions.

Alternatively look to build a variable voltage switch mode power supply, but this is more difficult than a variac.

Someone more informed than me may have a better suggestion or reasons why a variable voltage power supply may be a problem. I want to build one of these for an all digital amplifier to avoid doing a volume control in the digital domain or by converting to analog.

Regards,
Dean

 

[Brett]

Cute idea, but it won't work for analogue sources, unless you want to digitise them, or use a methodology similar to the Carver and NAD amps.

If efficiency is really a concern for you, get a Tripath.

 

[deandob]

If only the Tripath chips were all digital (why convert from a digital source like CD to analog, connect analog to a Tripath amp to be converted back to digital again)...

Some would argue that a tripath amp does not compare to a good class A amp. A big problem with class A is the huge amount of heat dissipated, so using a mechanism like lowering supply voltage with volume or tracking supply rails with the input seems reasonable if it can be done without too much complexity or side effects.

 

[phase_accurate]

Hi Circlotron

As far as I remember Meyersound uses a topology like a class-G amplifier to be able for running on high bias AND being capable of delivering high peak power, while maintaining reasonable efficiency. I assume they can do it with very high signal quality since they use it on their active studio monitor (the X10).

In fact it is a variant of the circuits Carver and NAD use (as Brett suggested).

Regards

Charles

 

[Brett]

If only the Tripath chips were all digital (why convert from a digital source like CD to analog, connect analog to a Tripath amp to be converted back to digital again)...

For me, most of my music is on vinyl or reel.

Some would argue that a tripath amp does not compare to a good class A amp. A big problem with class A is the huge amount of heat dissipated, so using a mechanism like lowering supply voltage with volume or tracking supply rails with the input seems reasonable if it can be done without too much complexity or side effects.

Have you heard a Tripath properly implemented to comment? I have a friend who's ears I trust, speak very highly of some of them. Not as good as the best tubes, but far from ear bleeding too.
Now if heat isn't a big issue, and if you have speakers with decent efficiency, and therefore only need a modest amp it shouldn't be, then your last 12 works are the key aren't they? Especially, the side effects.

Cheers

 

[deandob]

I have not heard the tripath (yet) but I have been doing a lot of research, the concensus summary from what I have read is that they make great amps, efficient with excellent bass resolution, but some distortion (IMD?) at high frequencies.

I plan on building a tripath amp once they release a version that supports digital input - previous threads on this topic indicated sometime this fall. Has anyone heard lately when tripath are releasing their new chips/modules?

Regarding class G - can someone post links to examples of class G schematics?