Commentable Thoughts
HI CHARLES , Where u been around these days, seems quite busy.
I simply dont want to use transformers for level shifting.
a question arises in my mind that can i use SCHOTTKY Diodes in the place of reverse Recovery diodes or not at the output.
Secondly what must be the safest voltage rating for the bootstrap capacitor for the gate drivers , with H-Bridge working at +-60 Volts, PLZ suggest.
COMPLIMENTS
AMPMAN
phase_accurate said:
HI CHARLES , Where u been around these days, seems quite busy.
I simply dont want to use transformers for level shifting.
a question arises in my mind that can i use SCHOTTKY Diodes in the place of reverse Recovery diodes or not at the output.
Secondly what must be the safest voltage rating for the bootstrap capacitor for the gate drivers , with H-Bridge working at +-60 Volts, PLZ suggest.
COMPLIMENTS
AMPMAN
Commentable Thoughts
Recently I had a discussion with IRF for the type of Comparator scheme , they insisted me to design amp with single comparator which will not pose problems with threshold.
Now I am going to Construct the PCB in SMD components format, and then I will fabricate then test it-->Then i will post my results
AmPmAn
Recently I had a discussion with IRF for the type of Comparator scheme , they insisted me to design amp with single comparator which will not pose problems with threshold.
Now I am going to Construct the PCB in SMD components format, and then I will fabricate then test it-->Then i will post my results
AmPmAn
Re: Commentable Thoughts
amp_man, I would definitely use the 4th order filter if the switching frequency is going to be that low. Alternatively, I would suggest increasing the switching frequency to something about 400khz or greater. There's a Leach article out there that suggests the switching frequency be 400 khz, and I believe the IR tutorial has a switching frequency around that. I'm sure you could also save a decent amount of money not needing another high powered inductor and cap on each output.
You also mentioned using rails of +/- 60V. Assuming you're using a full bridge output, I wouldn't worry about making a bipolar supply. Again, its a cost factor, and also if you're trying to make a large amount of power output, your choices become fairly small for the right transformer.
Just my 2 cents once again
amp_man_1 said:
amp_man, I would definitely use the 4th order filter if the switching frequency is going to be that low. Alternatively, I would suggest increasing the switching frequency to something about 400khz or greater. There's a Leach article out there that suggests the switching frequency be 400 khz, and I believe the IR tutorial has a switching frequency around that. I'm sure you could also save a decent amount of money not needing another high powered inductor and cap on each output.
You also mentioned using rails of +/- 60V. Assuming you're using a full bridge output, I wouldn't worry about making a bipolar supply. Again, its a cost factor, and also if you're trying to make a large amount of power output, your choices become fairly small for the right transformer.
Just my 2 cents once again
Re: Re: Commentable Thoughts
Acc. to me 400KHz will decrease the thermal efficiency as compared to 250KHz.
Secondly , We are in High Power Phenomena [1 KW above] and would only use Full bridge topology[eliminates power supply pumping effect].
As far as the filters are concerned , we are using Class-BD modulation technique which uses small size filters due the cancellation effect of in phase Switching frequency from both half Bridges , as a result the ripple is small .
bknauss said:
amp_man, I would definitely use the 4th order filter if the switching frequency is going to be that low. Alternatively, I would suggest increasing the switching frequency to something about 400khz or greater. There's a Leach article out there that suggests the switching frequency be 400 khz, and I believe the IR tutorial has a switching frequency around that. I'm sure you could also save a decent amount of money not needing another high powered inductor and cap on each output.
You also mentioned using rails of +/- 60V. Assuming you're using a full bridge output, I wouldn't worry about making a bipolar supply. Again, its a cost factor, and also if you're trying to make a large amount of power output, your choices become fairly small for the right transformer.
Just my 2 cents once again
Acc. to me 400KHz will decrease the thermal efficiency as compared to 250KHz.
Secondly , We are in High Power Phenomena [1 KW above] and would only use Full bridge topology[eliminates power supply pumping effect].
As far as the filters are concerned , we are using Class-BD modulation technique which uses small size filters due the cancellation effect of in phase Switching frequency from both half Bridges , as a result the ripple is small .
Amp_Man,
Is your website also in function? http://www.workhorseaudio.com/
I am very curious to your products.
Regards,
Jan-Peter
www.hypex.nl
Is your website also in function? http://www.workhorseaudio.com/
I am very curious to your products.
Regards,
Jan-Peter
www.hypex.nl
Hello Amp man, I am new to this site but I do have some comments regarding you criteria. I have worked with switching amplifiers for several years having some good experience with this type of amplifier.
I would agree that stray inductance/capacitance is an unwanted evil, inductance being the biggest parasite contributing to excessive ringing in the output stage: it is good to apply SMT with short traces where possible.
Equal track lengths for carrier signals, i.e PWM High, PWM Low.
Dead time must be very small, I have a deadtime of no greater than 20ns, also be carefull when using rc delay type deadtime production it is non-linear.
A better option for deadtime is a resistive divider into your window comparator giving a difference in 'change' voltage at each comparator. This gives more accurate results when your at the edge of your modulation window, with an rc type you may find that there is no pulse at times when approaching the limits.
If you need to use large snubber's then you have too much ringing and a change of layout is required.
When I designed a clock driven class-d I found that you would need to buffer this clock from a central point and feed it through out the board, also they should all come of the same clock or you may experience a beating sensation which is audible through the speakers, but not traceable.
Also the damping factor of an open loop class-d amplifier is dictated by the impedance of the output transistors and filter inductors. these will be seen as a series resistance with your speaker. If the loop is closed, say, after the output filter your damping factor will become very large: - )
I would agree that stray inductance/capacitance is an unwanted evil, inductance being the biggest parasite contributing to excessive ringing in the output stage: it is good to apply SMT with short traces where possible.
Equal track lengths for carrier signals, i.e PWM High, PWM Low.
Dead time must be very small, I have a deadtime of no greater than 20ns, also be carefull when using rc delay type deadtime production it is non-linear.
A better option for deadtime is a resistive divider into your window comparator giving a difference in 'change' voltage at each comparator. This gives more accurate results when your at the edge of your modulation window, with an rc type you may find that there is no pulse at times when approaching the limits.
If you need to use large snubber's then you have too much ringing and a change of layout is required.
When I designed a clock driven class-d I found that you would need to buffer this clock from a central point and feed it through out the board, also they should all come of the same clock or you may experience a beating sensation which is audible through the speakers, but not traceable.
Also the damping factor of an open loop class-d amplifier is dictated by the impedance of the output transistors and filter inductors. these will be seen as a series resistance with your speaker. If the loop is closed, say, after the output filter your damping factor will become very large: - )
Commentable thoughts
hi mr. dibley surprised to hear a comment on Class-D, again Thanks SIR.
Yeah ur absolutely rite man The Clock must be buffered and the delay mechanism which u illustrated is far good than rc network.
But for now i am using differential comparator LM360 for the out of phase signals rather than window comparators due to threshold problems.
With Regards,
AmPmAn
hi mr. dibley surprised to hear a comment on Class-D, again Thanks SIR.
Yeah ur absolutely rite man The Clock must be buffered and the delay mechanism which u illustrated is far good than rc network.
But for now i am using differential comparator LM360 for the out of phase signals rather than window comparators due to threshold problems.
With Regards,
AmPmAn
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.