Class TD reversed engineering possible?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
A pro audio manufacturer (Labgruppen) have a very smart way to make efficiency with power supply. It's similiar to classH, where the voltage is always maintained +/-5V above and below the output signal, but this is done with PWM+buck regulator, not with analog dissipation like classic class H.

Anyone can help with making DIY version of this idea? My own idea is that I have to make 3 power amps. One is the main power amp for audio. The 2nd power amp is classD for making modulated +rail supply and the 3rd is classD for making modulated -rail supply.

How to make a classD power amp, that always give +5V more than the output signal for +rail, in the exact phase always to the output audio signal? And the other for making -5V for negative rail?

But a question arrises. Does the classD power amp can have exactly delay time compared to any power amp? If not the supply track and the audio track will be in the wrong phase.
 

Attachments

  • td.jpg
    td.jpg
    8.9 KB · Views: 995
This is what I want to achieve.
Trace B is output audio signal. Trace A is modulated +supply that is always +5V than the output signal (but needs to very sure that it always in phase with B). This A is made by classD. C is the same as A, made by another classD always -5V than B.

The supply rail is very high like +/-120VDC, but with this trick, the power amp B will be working lightly, only from +/-5V supply. The rest of the burden is taken by amp A and C which is classD, efficient.

One may ask, why not making a direct classD amp? Well, this idea is tobe implemented in an existing power amp, any power amp.
 

Attachments

  • hd.jpg
    hd.jpg
    47.5 KB · Views: 864
Hi,

Here is Carver Corporation patent describing requested technology. On the other hand, if you do your tracking converters well enough, you end up with a class D amplifier.

The patent is interesting because it uses active damping of LC filter resonance which can also be applied in any class D amplifier. I have not seen that mentioned anywhere before.

Best regards,

Jaka Racman
 
Jaka Racman said:
Hi, Here is Carver Corporation patent describing requested technology. On the other hand, if you do your tracking converters well enough, you end up with a class D amplifier.

The patent is interesting because it uses active damping of LC filter resonance which can also be applied in any class D amplifier. I have not seen that mentioned anywhere before.

Rob Williamson (the patent holder) was a friend of mine. I am sad to report that he died quite suddenly and unexpectedly several years ago much too soon (in his early fifties). He, I and another friend had many a late night kitchen table discussion at my house about feedback schemes. Active damping has been occationally used in switching power supplies before Rob's patent. Our discussions on such subjects were the spark that lighted the way to the formulation of my generalized leap frog design method (it can be looked at as active damping for an arbitrary number of filter elements). -- analogspiceman
 
Yves Smolders said:
Analogman,

So you must have know Mr. Carver himself as well then?
Not really, but I have worked with quite a few people who have worked with him. From all accounts he is more snake oil salesman and yarn spinner than knowledgeable engineer (when it comes to succeeding in the audio business, the former attributes are perhaps more important than the latter, unfortunately). I worked for a time at the Carver Corporation designing a class d amplifier, but it was after the board had kicked Carver out of his own company because he was so hard to get along with. He had the last laugh, though - without his marketing charisma the company went bankrupt and he eventually bought his name back for a song. -- analogspiceman
 
Hi analog,

thanks for the interesting (and sad) story. Now I understand your familiarity with active damping which seems to be one of the best keept secrets in SMPS and classD industry, especially when all that is required is ultimately only one feedback cap from the output filter cap to the input of the PWM comparator. I would say that even UcD profits from that.

Lumanauw,

I suggest you look at the UcD amplifier discussed in this forum. It has a single pole rolloff and 35kHz power bandwidth, so it could be compatible with linear amplifier possessing similar caracteristics. It has completely discrete modulator so you might get some ideas.

Best regards,

Jaka Racman
 
What would the advantages of such an amp be?

Bash uses this method, link providided, and has been discussed here already, link provided. You can see from the Bash website the switching noise on the rails appears on the output as well, the output will be less efficient, and still have whatever cross over distortion it would have had. You're using a more efficient amplifier to aide efficiency in a less efficient amplifier to gain a medium increase in efficiency and all the problems of both topologies combined, with alot more added complexity. I honestly don't see the point to it.

Still think it's worth it?

http://www.bashaudio.com/technologies.htm

http://www.diyaudio.com/forums/showthread.php?threadid=42276

Regards
 
IMO the better efficiency/performance compromise than a switching downconverter and class AB in series like BASH etc would be the paralleling of a high-quality AB amp with a class-d current source.
This would never ever achieve the same efficiency figures as a pure class-d (and would also be more complicated) but it would

- have almost no ripple at its output
- be possible to make a frightenly fast amp this way

Regards

Charles
 
phase_accurate said:
This could very well be the mediocre graphic resolution. At least theoretically BASH et al should have a very good suppression of switching residuals.
There is an amp by AudioPhysic that also uses a bridge amp that is fed by a tracking downconverter:

http://www.audiophysic.de/produkte/mono/index_e.html

It has received veeeeery good reviews !

Regards

Charles

Yeah, or simply they are smart enough to know that if they looked too perfect no one would believe them anyway :D

Bash strikes me as being unusually honest when it comes to what their products can do and how they do it. Very refreshing.

I personally tend not to bother reading reviews by anyone getting paid to write them, they've just no credibility as far as I can tell... for instance Shatki stones and their other various gimmick products got rave reviews from most Hi-Fi magazines I've heard of.

Non biased opinions from forum users hold alot more weight with me.

Anyway I'd love to see a project done here with this kind of topology.

Thanks
 
I'm wondering if anyone... Dan Frasier perhaps, has ever attained one of their evaluation boards?

Here's a nice find for anyone interested, the data sheet for the IC their evaluation board is based on:

http://www.bashaudio.com/images/sta575.pdf

Maybe if you can buy the chip for a reasonable price it would be a good way to get your feet wet with the topology:

These guys have them but no prices.

http://www.accordtech.com/parts7_6.php



Good luck.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.