DIYHiFi.org

Hey thats not BCA, in BCA the value of inductors range from 100uH to 250uH

Whereas in that approach, the inductors are of very low value around 100nH which add an offset to bodydiode VF and thus enables the FRED to conduct more current.....with efficient recovery also

_________________
The Power Fanatic

Awesome, quick, re-patent it lol

I feel better if I reinvent the wheel rather than just copying and using it.
For sure from commercial side that's not most efficient, but here is my hobby playground...
...ahem, well... in fact I could even imagine that my di/dt limiter is also nothing new under the sun, and may be even already patented... you never know...
But for me it is new & nice.

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

Hi Choco
did you use antiparallel fast recovery diodes also??

_________________
The Power Fanatic

You mean two antiparallel diodes in series with drain?
No I did not try, but I would not expect to much advantage, because you would still make use of the body diode and for reverse recovery you would still have to remove the related charge. And for this path you would use one forward direction of the antiparallel diodes.
...but that's just grey theory... Did you try it? If yes, please post comparative screen shots.
Is there a reason for fast recovery high voltage diodes? If antiparallel you should be able to use the lowest voltage schottkys...

Anyway it would be again 3 ugly power semiconductors instead of just one.
Yes, yes... I know, I am driving you all crazy with my anti-component attitude.

Of course I am also fully aware that using the body diodes is causing by far more losses during hard switching compared to keeping the body diode without current. For continuous high power applications the combination of a series schottky plus a fast soft recovery freewheling diode is for sure the way to go. I do not doubt that (just compare i.e. MUR3020 vs a MosFet body diode.... :Tongue: ) . But music program does not show continuous high power. If you settle a filter with resonable higher inductive current ripple (means small choke), then you will not often have hard switching conditions. May be 10% of the switching events will be hard even if you play music close into slight clipping. And by this I do not expect issues with the losses during musical use. The power stage of my amp is not intended for continuous DC-welding applications. Such use would for sure overheat it....

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

Markus,
I always design my stuff to be reliable under extreme worst conditions, therefore you can compare an PWM amplifier [if it says made by "Workhorse Technologies"] with a welding set also.......
I was just curious about the effect of anti-parallel diodes in your application......
When an amplifier is subjected to 22dBu LF input signal overdrive along with 2 Ohms load, then its almost working in a hard switching mode, there comes the reliability which matters most in pro-audio....
Meanwhile i have ordered some of these just have a look
http://ixdev.ixys.com/DataSheet/16fd3bd ... 100898.pdf

_________________
The Power Fanatic

One instance i would like to quote is of a commercial Class-D using IR2110 and APT Mosfets [Trr400nS]was launched by Crest audio, the "LT series" which used similar arrangement of series coils like yours and patented that scheme also, that turned out to be a big disaster for crest, just because the usual DJ folks ran that amp with severe clipping and that thing simply cant handle that much abuse and they had to discontinue the series,
Now they have relaunched another Class-D "CD series" which uses series schottky and anti-parallel Fred......now they are big hit with mobile rigs........

_________________
The Power Fanatic

I cannot tell you with which effect these guys killed the Crests,
but I curious for the schematics of that amp and the patent abstract, or a link or a white paper....

22db overdrive should be easy to handle for my halfbridge.
With such heavy clipping you will have the half bridge operating mostly in LF.
Switching losses are neglectible then.

Even if we construct the worst artificial signal:
A rectangular shape that is high enough to ensure high output currents, but does not clip. Then each MosFet will do 50% of its switching in hard switching. The resulting losses just heat up the heat sink until the overtemp shut down is activated.
So far the theory.....

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

http://www.pat2pdf.org/patents/pat4724396.pdf
Look at figure 5

http://www.crestaudio.com/media/pdf/lt_schematic.pdf

_________________
The Power Fanatic

Hi Kanwar,
Crest is not at all using my di/dt limiter, they are using the circuit which have have posted.
Here is their schematic.
http://www.crestaudio.com/media/pdf/lt_schematic.pdf
And if their MosFet Trr is really 400ns, then they will 4 sure have a hard fight....

My MosFets have a Trr of 89ns at 100A/us, when pushing with 500A/us then Trr goes down to 50ns (but charge amount is of course higher than with 100A/us).
The reverse recovery peak in my application is around 25Apeak. It takes 50ns with about 500A/us and makes around 600...650nC.
With 352kHz and 110V rails this can cause 24W losses during 100% hard switching. That's no headache, but of course to much for continuous operation with my small heat sink of just 10cm x 4cm, but easy to fetch by a bimetal. Up to I see no flaw..., let's see what my ongoing examinations will bring.

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

*lol* time overlap...
But again their schematic is not using my circuit but the circuit which you have posted and where I said that it would be patented. ...OK, now we are sure that it is patented.
But it has noting in common with my approach.

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

I said they were using coils similar to your but not common approach just like you.....

In your case your body diode is quite fast therefore not a problem........I was talking about the situation with large die High Voltage mosfets with Trr>400nS and Qrr>1uC........

Thats why i have choosen to use this
http://ixdev.ixys.com/DataSheet/16fd3bd ... 100898.pdf

_________________
The Power Fanatic

This IXYS device is looking like somebody understood the topic and integrated the proper arrangement. Hey that's nice! I have to dig if they have something similar for lower voltage.
On the other hand, this devices would be fine to make a 2kW/8 Ohms half bridge instead of
my current approach of 2x 500W/2 Ohms or full bridge 1x 1kW/4 Ohms.

...very very nice device... doing the right thing integrated... ...low component count without compromising the performance...

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

I am planning to go for 2 parallel pairs of these devices for 5KW/2Ohms music and 3KW/4Ohms .......

_________________
The Power Fanatic

U are a crazy horse. I like nerds like U..

But take care for safety with this voltages. Even if compliance to safety standards is not mandatory for all products in your country, your local laws do make the manufacturer liable for the product. If something happens, then somebody might sue you and you have to proove that your amp is safe. Prooving this is easy if you fullfill Indian safety standards or IEC safety standards. If you just make it more or less safe according your feeling, such a proof might become difficult.

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