Anyone interested in a digital amplifier project? - Page 5 - diyAudio
Go Back   Home > Forums > Amplifiers > Class D

Class D Switching Power Amplifiers and Power D/A conversion

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 20th December 2002, 12:00 PM   #41
diyAudio Member
 
Join Date: Nov 2002
Location: Essex, UK
Quote:
Originally posted by Brian Brown
"But the final 11dB represents digital gain and can, depending on the peak level represented by the incoming digital data, cause the amp to crash into high distortion. Believe it or not, TACT was obliged to introduce this extra gain because various dealers were unsettled by the fact that the volume control could be advanced to full without creating distorted sound. "
...
Brian. [/B]
Now the marketing droids can talk about an amp that has a volume control going all the way to 11!
  Reply With Quote
Old 20th December 2002, 01:05 PM   #42
diyAudio Member
 
Join Date: Jan 2001
Location: Scandinavia
Default This is all fine and dandy but ....

To get optimal performance we need a variable power supply. TacT does this with SMPS. How do you propose we do it?

Petter
  Reply With Quote
Old 20th December 2002, 02:26 PM   #43
MWP is offline MWP  Australia
diyAudio Member
 
Join Date: Oct 2002
Location: Adelaide, South Australia
Why mains->smooth->switchmode->smooth->class-d->smooth->speaker?

Why not just mains->smooth->class-d->smooth->speaker.

Yeh im thinking aloud, and yeh i know it probably cant be done... but wouldnt it be nice?

Anyone want to build up some high impedance drivers to do this?
  Reply With Quote
Old 20th December 2002, 03:16 PM   #44
diyAudio Member
 
Join Date: May 2002
Location: Switzerland
Quote:
Why not just mains->smooth->class-d->smooth->speaker.
The continuation of this would be:

> touch a wire >

Regards

Charles
  Reply With Quote
Old 20th December 2002, 03:40 PM   #45
diyAudio Member
 
Join Date: Jan 2002
Location: California
Default Re: Equibit Experiment

Quote:
Originally posted by Brian Brown
Most Class D amps require feedback for stability
most class D amps become highly unstable with the application of negative feedback. this is the main problem we face with class D in general. they are now trying to use DSP (hence the phrase digital amp / class T, ect...) in an effort to linearize the output without feedback. whether or not the "digital" amplifier accepts a digital signal or analog signlal at the input, the DSP employed is there to provide a "amplifier with feedback" like qualities without the application of real feedback.

there are currently some very interesting patents on class BD amplifier technology (which has nothing to do with traditional class B amplifiers). it employs a truely ingenious method of triple rail switching (-V / gnd / +V) at the output with only a single inductor. this allows application of some traditional negative feedback since the amplifier output is truely resting at a physical ground and there is a huge reduction in switching noise. damping factor is also increased significantly. all of this without DSP =). i believe high power class BD (and similar triple output switching systems) will eventually be the solution to the poor s/n, distortion performance, DF, and frequency responce of current class D and AD technology. combined with some DSP the possibilities are endless.
  Reply With Quote
Old 20th December 2002, 05:09 PM   #46
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Default Re: This is all fine and dandy but ....

Quote:
Originally posted by Petter
To get optimal performance we need a variable power supply. TacT does this with SMPS. How do you propose we do it?

Petter
This is another area that I'll be playing with a bunch of options once my first board is functioning.

A couple of points:

TacT uses a variable supply as a form of volume control. Once the volume is set, the supply voltage remains constant. The supply voltage isn't varying dynamically with the music signal like a Class H amp does. With the volume set, the available dynamic range is primarily determined by the digital word size. (The original Millennium used 18bits, the current version uses 24.)

The TAS5012 has a dynamic range of 102db with a fixed supply. The TAS5015 has a dynamic range of 112db with a fixed supply. There's probably very few linear amps that can approach these numbers in the real world. My thinking is that as long as the supply voltage of an Equibit amp correctly matches the amp's peak power to the speaker's capabilities, there probably isn't much to be gained from further using the supply voltage as a volume control. As the previously quoted article showed, TacT was willing to throw away 11db of their digital resolution to marketing. I have a hard time believing they would do that if they thought it would have an impact on sound quality.

Another point is that TacT doesn't make their amps for one specific speaker. They need a user variable power supply so that the amp can be matched to speakers with different efficiencies and power handling capabilities.

My intention is to match the power supply voltage to the speaker the amp is driving. So my power supply options will require the ability to set the voltage to a specific value, but they won't necessarily have to be continuously variable.

*******

My first digital amp board doesn't have the front-end supply included. It does have the main bus caps (separate for each channel) and a relay to aid in power up and power down sequencing. There are footprints for LT1083 voltage regulators for each channel's supply that I can either use or bypass. Per TI's guidelines, there are snubber circuit footprints at each node in the power input and output chains. The values for these snubbers will be empirically tweaked after the board is built and operating.

Ultimately, I'm hoping to come up with a good off-line switching power supply (possibly with power-factor correction) with feedback to maintain 1.0V to 1.5V across linear post-regulators. This architecture should have the best combination of efficiency, compactness, and dynamic performance. If desired, it would also easily be adapted to continuously variable voltage using digital control.

The linear post-regulators should help make the supply very stiff in the audio band. This should allow me to use some fairly heavy LC filtering at the switcher output to get rid of any high frequency trash. Without the linear post-regulators, heavy LC filters would limit the switcher's response bandwidth to the point that it would get down into the audio range. (A side benefit of the LC filters is that they help limit the precharge inrush current of the bus caps.)

As previously discussed, Equibit amps need a very stiff power supply with low noise in the audio band because it's open loop. The power supply is analogous to a voltage reference in a DAC. There is zero power supply rejection ratio in the audio band. Ultrasonic power supply noise is a completely different matter. By its very nature, an Equibit amp produces 384KHz switching trash. Most switchers operate in the 100KHz to 500KHz range (I'd probably want to use at least 250KHz). The literature isn't very clear if there's a concern with the 384KHz intermodulating with 250KHz to produce products down into the audio band. If this isn't an issue, then the heavy LC filters and linear post-regulators could be eliminated. This would allow the response bandwidth of the switcher to be increased enough to make the audio band very stiff.

Another approach would be to make the switcher run synchronous to the amp's 384KHz carrier. This would ensure that there would be no intermodulation products.

*******

There's no reason that an Equibit amp won't work with a more traditional supply consisting of a line transformer / rectifier / capacitor arrangement. It will work just as well as it would for a linear amp. Of course it will be larger and more inefficient, but I suspect that isn't much of a concern to most of the DIY community.

Just as with a linear amp, the challenge is to make this supply as stiff as possible with a super low impedance at audio frequencies.
In fact, the super high damping factor of an Equibit amp will reveal the quality of the power supply even more. The Equibit output stage directly shunts any back EMF from the speaker straight into the supply. This doesn't mean that an Equibit amp will have higher distortion with a given supply. Rather it means that the elimination of distortion components resulting from the back EMF influencing a linear amp's output drivers will make it easier to hear what's happening in the supply.

High frequency power supply trash is a big concern with linear amps. I believe it will be less so with an Equibit amp.

A basic transformer / rectifier / capacitor supply is easy to design, and there's low risk of problems. This makes it a safer choice for many designs.

A super low impedance, tightly regulated conventional supply is a completely different matter. IMO these are very difficult to do well. I think that the quality of a conventional high-end linear amp is more likely to hinge on the quality of its power supply than on the quality of its output stage.

A high quality switching power supply is a tricky proposition, no doubt. But I think that for an Equibit amp it will be easier to achieve the best results from a switching supply than from a conventional one.

Once again, I'm planning to try both.

Just to be clear, I don't consider myself a switching power supply guru, especially when it comes to magnetics. But I have been involved with them for a number of years and have spent quite a bit of time troubleshooting system noise and performance issues.

More importantly, I've got some friends who are transcendent gurus with these things that I can turn to if I run into issues.

******

Another approach that I'm looking at is to use a stack of gell-cell batteries in series (one for each channel). With bus caps in parallel, this should be a very low impedance stiff supply. I'll probably try them with and without linear regulators. (Using batteries direct, the voltage can be controlled by the number of cells in series.)

This approach wouldn't fly for most commercial applications, but DIY is a completely different thing.

The high efficiency of Equibit amps makes batteries much more practical to consider for a power amplifier.

******

To initially get my amp up and running, I'm planning on using a commercial switching power brick.

Regards,
Brian.
  Reply With Quote
Old 20th December 2002, 07:20 PM   #47
Rookie is offline Rookie  Serbia
diyAudio Member
 
Rookie's Avatar
 
Join Date: Nov 2002
Location: Serbia
Would be a variable power supply with OPA549 (page 13 of OPA549 datasheet) usable? How would it react to a capacitive load? Would it cause the OP-AMP oscillate?

You mentioned that an Equibit amp has super high damping factor, but in the TAS5100EVM Data Report (sleu011a document) it is stated that the damping factor is only 15 (1kHz, 8 ohm). That is far away from super.

Regards,
Dejan
  Reply With Quote
Old 20th December 2002, 10:23 PM   #48
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Default Damping Factor

Quote:
You mentioned that an Equibit amp has super high damping factor, but in the TAS5100EVM Data Report (sleu011a document) it is stated that the damping factor is only 15 (1kHz, 8 ohm). That is far away from super.
I probably shouldn't have used the specific phrase 'damping factor', but the concept being considered is the same.

Normally a damping factor of 15 would be a downright lousy figure. But this is an issue of the term's definition and the way it is traditionally measured.

Damping factor is a figure of merit to indicate how well an amplifier can control a load. In terms of how well an Equibit amp can do this in a real-life situation driving an actual speaker with lots of electrically and mechanically induced back-EMF, it is superior to most traditional linear amp designs.

(Note that the following discussion is presuming a very stiff, wide-bandwidth, high-quality power supply).

A linear amp with bipolar output devices is usually highly susceptible to back-EMF, which will alter the conduction of the output devices.

A linear amp with MOSFET output devices is much better in this regard, because the MOSFETs don't react to the back-EMF, and it doesn't affect the actual output of the amp the way it does with bipolar devices. However, the MOSFETs are only partially turned on, so there is considerable impedance between the back-EMF and the power supply.

A switching amplifier, because its output drivers are either fully turned on or off, will directly shunt the back-EMF into the power supply. There is only the Rds-on of the MOSFETs and the impedance of the low-pass output reconstruction filter to overcome (this is one reason that the output inductors should have as low of series resistance as possible, preferable below 50mOhms.)

Strictly speaking, damping factor is defined as the ratio of the load impedance to the amp's output impedance. It is given as a single number with a given test frequency at a certain load resistance. It is also measured with a dummy test load resistor!

Let's talk about negative feedback. Most of you are probably familiar with the way that negative feedback can be used to 'force' the output to have very low THD measurements when driving a resistive load, but then causes all sorts of stability issues and distortion with an actual speaker. This is a similar issue with damping factor measurements.

Negative feedback lowers the effective output impedance of an amplifier, especially into a load resistor. Back-EMF, by its very nature, contains lots of high frequency trash that most negative feedback loops don't have the response bandwidth to keep under control. This can mean that the low effective output impedance into a static load is lost in a real life situation. Even worse, the back-EMF can get into the feedback loop and further screw up the amplifier's output. This is the exact opposite of what a high damping factor is supposed to represent!

This is why an Equibit amplifier, running open-loop with zero negative feedback, scores poorly when testing damping factor with a dummy load.

In real life, an Equibit amplifier has no feedback at all to be screwed up by the back-EMF, and it has a very direct low-impedance path between the power supply and speaker that is more effective at shunting out back-EMF.

So in terms of the 'Spirit of the meaning' of damping factor, an Equibit amplifier is superior to most others.

I'm surprised that TI included this spec in the report without any explanation.

Regards,
Brian.
  Reply With Quote
Old 20th December 2002, 10:24 PM   #49
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Default Re: OPA549

Quote:
Originally posted by Rookie
Would be a variable power supply with OPA549 (page 13 of OPA549 datasheet) usable? How would it react to a capacitive load? Would it cause the OP-AMP oscillate?
This is an interesting idea. I especially like that the OPA549 can operate up to 60V.

I wouldn't want to use any type of linear regulator as the sole means of voltage adjustment. The voltage drop would be considerable, and it'd take a pretty hefty heat sink to dissipate the waste heat. That's why I was proposing a switching preregulator to keep the voltage drop across the linear regulator to a minimum.

The stability into a capacitive load is certainly a concern. I think this could be worked out, but it'd be a little tricky.

By comparison, an LT1083 is more efficient, probably lower noise in this application, and is stable into unlimited capacitive loads. It does have a 30V input limitation, however.
  Reply With Quote
Old 20th December 2002, 10:44 PM   #50
diyAudio Member
 
Join Date: Aug 2002
Location: Wisconsin
Default Re: Class BD

Quote:
Originally posted by R. McAnally


there are currently some very interesting patents on class BD amplifier technology (which has nothing to do with traditional class B amplifiers). it employs a truely ingenious method of triple rail switching (-V / gnd / +V) at the output with only a single inductor. this allows application of some traditional negative feedback since the amplifier output is truely resting at a physical ground and there is a huge reduction in switching noise. damping factor is also increased significantly. all of this without DSP =). i believe high power class BD (and similar triple output switching systems) will eventually be the solution to the poor s/n, distortion performance, DF, and frequency responce of current class D and AD technology. combined with some DSP the possibilities are endless.
We're only beginning to see what I'm sure will be a flood of new switching amplifier techniques. While most efforts seem to be focusing on low cost, small size, and high efficiency, there's lots of stuff coming to raise the bar on sound quality. I'm very excited about the prospects.

I just hope that we, as individuals, get access to this new stuff.

It seems like so many of the new technologies coming out are arriving with licensing agreements that only large companies can afford. This isn't an issue only to hobbiests, but also to small audio manufacturers.

Mitsubishi is now producing a digital amp chip (M65817AFP) that works with either IIS PCM data or DSD data from an SACD. Apparently they're making it for Sony. I sure wish I could get my hands on some of these to play with!

I can't say enough about how happy I am that I can get the TI chips!

Brian.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Noise measurement amplifier - interested? jackinnj Solid State 15 27th March 2008 09:36 PM
New digital XO project borges Multi-Way 2 8th January 2008 06:49 PM


New To Site? Need Help?

All times are GMT. The time now is 02:50 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2