"But these devices are low low reverse voltage, so they can´t be used in hi power amps."
IR makes some 200V 120A Schottky diodes.
The ON Semi 200V 20A Schottky diodes are only $1.60 each.
http://dkc3.digikey.com/PDF/T053/0822.pdf
IR makes some 200V 120A Schottky diodes.
The ON Semi 200V 20A Schottky diodes are only $1.60 each.
http://dkc3.digikey.com/PDF/T053/0822.pdf
I've had some experience with modifying a Proton DPD amp (which is essentially a close relative of NAD with slightlydifferent circuitry).
Originally, the amp has power rail switching common to both channels and uses 100V fast recovery diodes and bipolar transistors in the commutator. The rails were about 35 and 70V unloaded.
The amp was converted (at rather high expense in hours!) to a more optimal power supply, using the existing two power transformers - originally one provided the low voltage, and the otehr the high voltage rails, after conversion one provided both positive, the otehr both negative rails, for better power balance in the transofrmers. Fast recovery diodes were replaced by 100V schottky's. The driver portions had their own power supply rails of about +-80V, and the output stage used Sanken 16A 250V bipolars in a MT200 case, one pair per channel.
In the conversion, the switching netork was redesigned to use MOSFETs, and also, switching was split for the two channels.
The original switching network used two power transistors per HV rail, one for each channel but the drivers were common. Operation was by two methods: switching, and bootstrap from the output, which makes them class C.
Workhorse's diagram shows one common problem with these amps - insufficient bootstrap voltage (derived using zener diodes) would produce clipping artifacts in the middle of the waveform rising edge as the higher voltage rails would switch 'too late'.
Turning on MOSFETs (IRFP240 and 9140) by bootstrap plus small signal transistors was MUCH easyer, it is VERY easy to provide many mA of current to charge gate capacitances, and it is also very easy to control switch on and off characteristics to prevent overly large transients, especially in conjunction with schottky diodes. I have drawn the commutation waveforms in Workhorse's diagram in red, below...
I have also experimented with bootstrapping the MOSFETs, which is much more elegant since the driver circuit is enormusly simplified. This also results in better SOA for the bipolars at high currents, makes the MOSFETs work in class C and distributes heat amongst two easy to cool packages. Waveform is in blue, superimposed over Workhorse's original diagram.
Originally, the amp has power rail switching common to both channels and uses 100V fast recovery diodes and bipolar transistors in the commutator. The rails were about 35 and 70V unloaded.
The amp was converted (at rather high expense in hours!) to a more optimal power supply, using the existing two power transformers - originally one provided the low voltage, and the otehr the high voltage rails, after conversion one provided both positive, the otehr both negative rails, for better power balance in the transofrmers. Fast recovery diodes were replaced by 100V schottky's. The driver portions had their own power supply rails of about +-80V, and the output stage used Sanken 16A 250V bipolars in a MT200 case, one pair per channel.
In the conversion, the switching netork was redesigned to use MOSFETs, and also, switching was split for the two channels.
The original switching network used two power transistors per HV rail, one for each channel but the drivers were common. Operation was by two methods: switching, and bootstrap from the output, which makes them class C.
Workhorse's diagram shows one common problem with these amps - insufficient bootstrap voltage (derived using zener diodes) would produce clipping artifacts in the middle of the waveform rising edge as the higher voltage rails would switch 'too late'.
Turning on MOSFETs (IRFP240 and 9140) by bootstrap plus small signal transistors was MUCH easyer, it is VERY easy to provide many mA of current to charge gate capacitances, and it is also very easy to control switch on and off characteristics to prevent overly large transients, especially in conjunction with schottky diodes. I have drawn the commutation waveforms in Workhorse's diagram in red, below...
I have also experimented with bootstrapping the MOSFETs, which is much more elegant since the driver circuit is enormusly simplified. This also results in better SOA for the bipolars at high currents, makes the MOSFETs work in class C and distributes heat amongst two easy to cool packages. Waveform is in blue, superimposed over Workhorse's original diagram.
Attachments
Hi DJK & Anatech,
I have seen some QSC models which uses 3 Tier = 3 positive steps and 3 negative step supplies in Class - H...which increases the complexity of both the switching circuit and power supply...and the transformer utilization factor is quite low, as compared to Crown MA5000VZ amps......
What are your views regarding 3 Tier Class-H ?
regards,
K a n w a r
I have seen some QSC models which uses 3 Tier = 3 positive steps and 3 negative step supplies in Class - H...which increases the complexity of both the switching circuit and power supply...and the transformer utilization factor is quite low, as compared to Crown MA5000VZ amps......
What are your views regarding 3 Tier Class-H ?
regards,
K a n w a r
The transformer utilization really isn't any better on the Crown.
The Crown is a two-tier supply with a very clever rail switch.
Most Carver were three-tier:
the M400 was ±25V, ±50V, ±80V,
the M1.0 was ±25V, ±50V, ±100V,
the M1.5 was ±37V, ±75V, ±125V
The M2.0 had four tiers, the top being the same as the M1.5 (IIRC), it had a switching supply, the others had the triac supply.
I never heard a M2.0, the others all sounded good.
I never heard a QSC I liked, the ones I heard were all two-tier.
AB International makes a different breed of three-tier:
the 1100 was ±0V, ±50V, ±100V (I explained previously how the 0V tier worked).
All the Crest I have heard are two-tier, some are class G, and some are class H. I own eight of the CA series, which the larger ones are class H, the smaller are class AB. The class H models sound better than the Crown, the class AB models sound better than the class H models, the class AB from AB International sound better than the class AB Crest (probably the limiter stuff).
Pol Hsu
http://www.hsuresearch.com/
researched the M1.0 design when he was at MIT. The efficiency numbers he came up with for the three-tier M1.0 driving a reactive load were comparable to a PWM amplifier.
The Crown is a two-tier supply with a very clever rail switch.
Most Carver were three-tier:
the M400 was ±25V, ±50V, ±80V,
the M1.0 was ±25V, ±50V, ±100V,
the M1.5 was ±37V, ±75V, ±125V
The M2.0 had four tiers, the top being the same as the M1.5 (IIRC), it had a switching supply, the others had the triac supply.
I never heard a M2.0, the others all sounded good.
I never heard a QSC I liked, the ones I heard were all two-tier.
AB International makes a different breed of three-tier:
the 1100 was ±0V, ±50V, ±100V (I explained previously how the 0V tier worked).
All the Crest I have heard are two-tier, some are class G, and some are class H. I own eight of the CA series, which the larger ones are class H, the smaller are class AB. The class H models sound better than the Crown, the class AB models sound better than the class H models, the class AB from AB International sound better than the class AB Crest (probably the limiter stuff).
Pol Hsu
http://www.hsuresearch.com/
researched the M1.0 design when he was at MIT. The efficiency numbers he came up with for the three-tier M1.0 driving a reactive load were comparable to a PWM amplifier.
Hi Workhorse,
I haven't much else to add to what djk has said. The PM 2.0 sounds like the rest. The fans are louder
and go with the music more noticeably. Carver just ran them off their own winding, so the speed went up with current demand. A rack of them sounds hilarious when playing loud.
The PM 2.0 switched using a couple 50A bipolars around 2 KHz I think. Some of the rectifiers used to overheat and short, fans too. If you think the triac regulation gave some techs a hard time, you should see what they did with the switching supplies! Very reliable considering.
I think the best feature of these amps was the supply shut down. When an output went, generally the only damage was the output, possibly a commutator transistor as well. The speaker never got cooked unless it was not the amp that failed .
I enjoyed working on these products. I wouldn't mind at all doing it again.
-Chris
I haven't much else to add to what djk has said. The PM 2.0 sounds like the rest. The fans are louder
and go with the music more noticeably. Carver just ran them off their own winding, so the speed went up with current demand. A rack of them sounds hilarious when playing loud.The PM 2.0 switched using a couple 50A bipolars around 2 KHz I think. Some of the rectifiers used to overheat and short, fans too. If you think the triac regulation gave some techs a hard time, you should see what they did with the switching supplies! Very reliable considering.
I think the best feature of these amps was the supply shut down. When an output went, generally the only damage was the output, possibly a commutator transistor as well. The speaker never got cooked unless it was not the amp that failed
.I enjoyed working on these products. I wouldn't mind at all doing it again.
-Chris
HI everybody,
Here is the link to Studiomaster Class-H bipolar amplifiers Schematics...
http://mysite.wanadoo-members.co.uk/studiomaster/service/service.htm
regards,
K a n w a r
Here is the link to Studiomaster Class-H bipolar amplifiers Schematics...
http://mysite.wanadoo-members.co.uk/studiomaster/service/service.htm
regards,
K a n w a r
"I have recently checked Crest CA series Class-H amplifier....upon checking the waveform at the PS rails..lots of Spikes are seen when driving with HF content....isn't this adds coloration to the music ..."
Yes, you can hear it for HF use. They're fine for bass.
Crest does make some special amps for driving HF drivers that are not class H. I use some AB International amps for HF use that have about the same topology as the Leach amplifier, they sound better for HF use, but are bigger (3H vs 2H) even though they are less power.
Yes, you can hear it for HF use. They're fine for bass.
Crest does make some special amps for driving HF drivers that are not class H. I use some AB International amps for HF use that have about the same topology as the Leach amplifier, they sound better for HF use, but are bigger (3H vs 2H) even though they are less power.