The Black Hole......

Jump to Latest
#4,729
Actually auto transformers do run really cool! Keep in mind some folks like the added distortion. They are regularly used with 70 volt speaker level circuits to provide zone volume controls from a single amplifier.

Of course even normal folks who use potentiometers do have some increase in distortion. A good resistive control may add distortion at a level of -140 dB or better referenced to full scale signal. It does drop as the signal level falls. A switched attenuator with the right resistors can do -160 dB or better at audio frequencies. The layout of the parts is important to avoid excess capacitance between the resistor bodies!

I doubt you can get better than -110 dB from a transformer and even less at very low frequencies.

There are definitely some folks who prefer increased low frequency distortion.
 
Last edited:
#4,732
I have some experience with digital vs analog volume control in my home system. Many years ago I initiated a major upgrade, converting my 5.1 system to full active operation using a pair of miniDSP nanoAVR boxes. These units work on HDMI audio, and were intended to go between the music source and an audio-video receiver with HDMI inputs. They provide the usual various DSP functions like crossover, para EQ, time delay, etc.

For some time I was running an HDMI splitter into 2 of these nanoAVR boxes, then into a pair of Yamaha multichannel receivers, using the Yamahas mainly for DAC and (analog) level control. Eventually miniDSP introduced the nanoAVR-HDA with self-contained 8-channel DAC outs, including digital level control via IR remote. I upgraded my boxes to these models, and eliminated the Yamahas, running the nanoAVR analog outs directly into a pair of multichannel power amps. Here are some observations about the first setup vs the second (excerpted from this post over on the miniDSP forum):

----------------

...The other thing I miss from Gonzo v1 is analog volume control. I understand the misconceptions that many seem to have about digital volume, and I'm not hearing any degradation in the actual music - it's as lovely as ever. But as near as I can calculate, the system noise floor has come up about 10 dB. Even this is not as bad as it sounds, since I still can't hear any noise at the listening position - I have to have my ear about a foot away from a tweeter to hear anything at idle. But that distance used to be more like 2 or 3 inches.

The thing that concerns me more (and I don't see this discussed as much) is the safety cushion provided by analog, post-conversion level control. Previously, using the Yamahas (which use multichannel analog ICs for master volume), at normal listening levels the volume indicators would read in the -10 to -12 dB attenuation range. With the nanos, I'm at roughly -14 to -16 when I check levels in the plugins. This lines up, since I measured roughly 4 dB more level out of the nano HDA compared to the Yamaha pre outs, for a given input. But here's the thing: A few weeks ago, I managed to temporarily "crash" one of my nanoAVR boxes, and it seemed to go into some sort of a test mode, producing a 1 KHz sine wave at what sounded like 0 dBFS at each output. For a conventional system with passive speakers, this would not necessarily be cause for panic. But with an active setup, I had a fairly loud 1 KHz connected directly to my dome tweeters! Fortunately this occurred with the old setup, so there was that -14 dB or so of analog attenuation between the nano full-scale output and my speakers, and no damage occurred. But with digital level control, that cushion is gone - full-scale is full-scale, and if anything like this happens again, I doubt I'll be so lucky.

----------------

Just tossing that into the salad for what it's worth.

I still intend to return to analog level control of some sort, or even fixed passive attenuators between the DACs & power amps. I've definitely got a bit too much gain here.
 
#4,733
Jim the Oldbie said:
The thing that concerns me more (and I don't see this discussed as much) is the safety cushion provided by analog, post-conversion level control. Previously, using the Yamahas (which use multichannel analog ICs for master volume), at normal listening levels the volume indicators would read in the -10 to -12 dB attenuation range. With the nanos, I'm at roughly -14 to -16 when I check levels in the plugins. This lines up, since I measured roughly 4 dB more level out of the nano HDA compared to the Yamaha pre outs, for a given input. But here's the thing: A few weeks ago, I managed to temporarily "crash" one of my nanoAVR boxes, and it seemed to go into some sort of a test mode, producing a 1 KHz sine wave at what sounded like 0 dBFS at each output. For a conventional system with passive speakers, this would not necessarily be cause for panic. But with an active setup, I had a fairly loud 1 KHz connected directly to my dome tweeters! Fortunately this occurred with the old setup, so there was that -14 dB or so of analog attenuation between the nano full-scale output and my speakers, and no damage occurred. But with digital level control, that cushion is gone - full-scale is full-scale, and if anything like this happens again, I doubt I'll be so lucky.

----------------

Just tossing that into the salad for what it's worth.

I still intend to return to analog level control of some sort, or even fixed passive attenuators between the DACs & power amps. I've definitely got a bit too much gain here.
Click to expand...
I can imagine active speakers being designed this way, so that each amplifier's power output is no more than what the driver can take. The "other" consideration is what the listener can take. I mention this here:

Drive level for headphone?

The DAC outputs should theoretically be able to drive the amp output into clipping, but if that's more power than you want into the driver (whether it's to prevent burning out the driver or limiting the SPL to a safe level), then an attenuator has to go somewhere (alternatively, the power amp's voltage rails could be reduced to reduce max power). There's a good argument for it going between the amplifier and driver, as it will also reduce the residual noise of the amplifier as well as the rest of the system.

Of course this generally doesn't apply to the woofer range, which usually needs as much power as it can get.

A challenge is letting through loud musical transients such as kick and snare drums while still protecting from a continuous tone at an unacceptable SPL, especially for the ear's more sensitive ranges above, say, 500Hz (which might include some of the woofer range). This would be yet another DSP function, checking output level vs. time. Again this could have two functions, keeping high, sustained SPL from damaging ears as well as overheating voice coils and other driver damage.

Surely some company has done this for some higher cost design of active speaker. The technology has been available for as long as DSPs used for crossovers.
 
Last edited:
#4,735
simon7000 said:
Of course even normal folks who use potentiometers do have some increase in distortion. A good resistive control may add distortion at a level of -140 dB or better referenced to full scale signal. It does drop as the signal level falls. A switched attenuator with the right resistors can do -160 dB or better at audio frequencies. The layout of the parts is important to avoid excess capacitance between the resistor bodies!
I doubt you can get better than -110 dB from a transformer and even less at very low frequencies.
There are definitely some folks who prefer increased low frequency distortion.
Click to expand...
The transformer has a very high efficiency compared to resistors. Energy, including musical energy, in the transformer is not wasted. In any case, the speakers themselves distort at these low frequencies by an order of magnitude more than transformers - attenuators.


benb said:
I can imagine active speakers being designed this way, so that each amplifier's power output is no more than what the driver can take.
Click to expand...
The power amplifier does not specifically limit the power supplied to the speaker, otherwise the speaker will burn out even faster. For example, in JBLM2 monitors, the signal from the very powerful Crown iTech5000HD digital amplifier is fed to the tweeter.
 
Last edited:
#4,736
Jim the Oldbie said:
I understand the misconceptions that many seem to have about digital volume, and I'm not hearing any degradation in the actual music - it's as lovely as ever.

I still intend to return to analog level control of some sort, or even fixed passive attenuators between the DACs & power amps. I've definitely got a bit too much gain here.
Click to expand...

I wonder if the music can remain "lovely as ever" going through digital volume and then to a I2S input amplifier? Where it wouldnt be possible to put, say, a tube stage in the signal path.

I know...what fun is that?
 
You must log in or register to reply here.
Top Bottom