DIYHiFi.org

For the sake of audio
It is currently Thu Nov 23, 2017 4:33 pm

All times are UTC




Post new topic Reply to topic  [ 12 posts ] 
Author Message
PostPosted: Tue Aug 07, 2007 10:48 am 
Offline
Hen

Joined: Tue Aug 07, 2007 1:53 am
Posts: 2
Hello diyhifi,

As a non-electronics person (I work in IT as a Unix SA) .....yeah I know it's my own fault :banghead:

I was wondering what the good people of diyhifi had to say about "Class-I" ...or as the subject says
(Balanced Current/Opposing Current) amplifiers? I'm told the topology is the best solution for
switching amplifiers ?


Top
 Profile  
Reply with quote  
PostPosted: Tue Aug 07, 2007 1:43 pm 
Offline
Cow
User avatar

Joined: Sun Dec 19, 2004 1:18 pm
Posts: 265
Location: retired
http://www.crownaudio.com/amp_htm/amp_info/bca.htm

US Pat 5657219


Top
 Profile  
Reply with quote  
PostPosted: Tue Aug 07, 2007 8:03 pm 
Offline
Account disabled on owner's request

Joined: Wed Dec 29, 2004 8:31 pm
Posts: 899
Hi,

Just remember that despite them calling it "Class I", it's still a Class D amplifier.

It's one solution that works, the best? Likely not. It seems to overcomplicate things in order to address certain problem areas like say flywheel current flowing through the body diodes of the mosfets, when there are much cheaper, simpler and more exacting ways of minimizing it. Not saying it's bad though, but really, how balanced can it be when it depends on the matching of two coils, and what are all the other drawbacks of them?

Maybe this thread could be moved to the class d section ?

Cheers,
Chris


Top
 Profile  
Reply with quote  
PostPosted: Wed Aug 08, 2007 11:17 am 
Offline
Hen

Joined: Tue Aug 07, 2007 1:53 am
Posts: 2
TragicallyDistorted wrote:
.......how balanced can it be when it depends on the matching of two coils......?


Thanks Tragic, I think I get it.

I'm more then happy to move this to class-d.....how would I do that?


Top
 Profile  
Reply with quote  
PostPosted: Wed Aug 08, 2007 1:30 pm 
Offline
Sheep
User avatar

Joined: Sun Dec 19, 2004 2:56 pm
Posts: 63
Location: The Netherlands
Moved thread to Class D.


Top
 Profile  
Reply with quote  
PostPosted: Wed Aug 08, 2007 1:36 pm 
Offline
Account disabled on owner's request

Joined: Wed Dec 29, 2004 8:31 pm
Posts: 899
Thanks.


Top
 Profile  
Reply with quote  
PostPosted: Fri Sep 07, 2007 4:18 pm 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
I think even with unmatched inductors this solution is helpful to reduce the ripple in the output signal by great extend. It allows for a half bridge to get a similar ripple frequency doubling (and theoretically no speaker ripple at zero input) as you can normally reach with a double PWM full bridge.
The BCA vs a normal half bridge invests in:
-Two more diodes
-Second choke
-Slightly more complex PWM
But I can imagine that it pays back by allowing half switching frequency.
..especially if we consider that Crown came up with this idea 11 years back, when the MosFets were by far more troublesome than our types of the year 2007.

Furtheron if you consider to design a simple 2nd order output filter for a normal half bridge and demand less than 2db amplitude error over the full audio range for all load conditions between 4Ohms-16Ohms and at the same time want to have low ripple in the speaker, then you will find that you will need switching frequencies above 300kHz. Unfortunately this moves to higher frequencies if you increase the power, because the speaker ripple is directly related to your rail voltage.
This output filter issue is the reason why I am seriously thinking about a double PWM for
my class D project - even if I usually tend to trouble Chris by refusing any additional complexity... My project definitely will have a full bridge, so the frequency doubling would only cost me two more transistors (+surroundings..) for my discrete PWM solution.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Fri Sep 07, 2007 4:40 pm 
Offline
Account disabled on owner's request

Joined: Wed Dec 29, 2004 8:31 pm
Posts: 899
YYYyyeahhh, I argued against complexity such as toys like clipping detection, in place of simple buffers for discrete 2 transistor gate drivers on a 1000W full bridge .

BTW what you discuss seems kind of like "full bridge emulation"... you already have a full bridge? Want a higher frequency, clock it faster? You're introducing a good source of nonlinear distortion this way. There are reasons for not making things more complex than they really need to be.I look forward to your results either way.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 3:02 am 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
Yes already in my babystep-thread I am using a full bridge, but with the most simple control.
There the second half bridge is acting exactly inverted to the first half.
If you invest in a separate comparator for each half, which are refering to the same triangle but get inverted music signals, then you will get the benefit of doubling the ripple frequency.
Doubling the clock is less fortunate, because double switching losses and -yes, as you say more nonlinear distorsions. I will definitely not step up my switching frequency from around 320kHz to more than 600kHz. But I am seriously considering stepping down to 250kHz and use the more intelligent PWM for doubling the resulting ripple frequency.
This doubling is resulting from the we use the differences between both halves as output. Besides double frequency this methode is adding one more output level. A simple halfbridge is delivering just positive or negative. The full bridge with double PWM is delivering as difference between both halves: Positive, negative and zero. Zero is always resulting when both half bridges are HIGH or when both half bridges are LOW.
To generate the double PWM you can use a comparator arrangement as in the crown paper, but use Sn for the first half bridge and Sp for the other half bridge (+ inverted signals for the high side driver).
Regarding the output filter you could theoretically use just one cap accross the speaker, but this is unfortunate vs EMI and distorsions. Why distorsions? Because of the missing ripple current in the half bridge itself. If you have the double PWM full bridge and zero music then both halves will always be HIGH or LOW at the same time. Difference zero. If the filter is connected to the difference only then we will have no ripple current in the half bridges.
BUT ripple current in the half bridge is our friend! As long as the ripple current is larger than the music current we derive almost neglectible low distorsions from the dead time.

A very nice behavior of double PWM is that we only add complexity in the logic, where it is easy to design&debug without connecting the power stage. Means without danger of smoke. And from performance view this logic does not increase the delay time, both comparators are acting time wise in parallel. So my two strongest reasons against complexity are no concern here. Only remaining the fact that every additional component has to be drawn, calculated, bought, placed and it's function to be tested in the circuit...

And yes, as you mention when you say full bridge emulation... From my perception the bca-patent is a method how to allow similar benefits for a half bridge as we know from full bridge designs with double PWM.

P.S:
I put all this here and not to my Gen2-thread, because up to now I have even in my Gen2 the most simple and straight forward single PWM. ...still struggling mentally with myself about adding the complexity or not... :rotfl:
Furtheron this essay about full bridges should be helpful to understand the intension of the crown-bca-paper.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 3:21 am 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
Add on: I am stating that it would be unfortunate to use just one filter cap across the speaker, because of EMI and dead time distorsion... I forgot to say that all is turning fine if you use two caps series and connect their center tapping to ground.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 7:40 pm 
Offline
Account disabled on owner's request

Joined: Wed Dec 29, 2004 8:31 pm
Posts: 899
Hi,

Increasing the switching frequency will lead to nonlinear distortion largely because you fail to take into account output stage delay in your clocked design, as I believe does what's known of BCA.

This is precisely where self oscillating designs have the advantage, and easily lead to a robust high fidelity amp. If you want that same level of performance in a clocked amp, it doesn't come as free, but you can't ignore it and expect it to perform as well, unless mid-fi /sub woofer use is your highest aspiration.

Keeping the switching frequency ~250khz or lower will not cure this.

What your idea for crown BCA will require are two bigger filter coils and then dealing with the problems that brings you, worse layout, worse EMI, and you're still suck with very rude delay issues between phases, but it's probably easier to make it "behave" (not blow apart via cross bridge conduction).

While the "logic" aspect of the timing of the second bridge wrt to the first is arguably minimal in variance between the two, the output stages are not operating equally between ouputs. They will suffer different delays, which in this case will directly reflect on the linearity of your modulation, since it is the difference between the two that's doing it.

Otherwise a lot of the advantages you mentioned are had with full bridge as well, so there's no reason for BCA in that alone.

There's also no need for an insane switching frequency. There is a serious need for good control over the output stage.

Cheers


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 16, 2007 2:25 pm 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
TragicallyDistorted wrote:
..., but it's probably easier to make it "behave" (not blow apart via cross bridge conduction).



That's probably a very good reason and real life advantage for mass production.
You still have to take care that the 'overlap' does not become large, otherwise it will integrate up to quite some undesired DC-bias current through the MosFets. But it should definitely be much more forgiving than a halfbridge.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 12 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group