1kW Class D

Status
Not open for further replies.
Jump to Latest
Netlist

Netlist

Moderator
Joined 2003
Paid Member
Let me ask it differently: What is the benefit for the diy community? Talking about companies, amounts of power and specs without the possibility to discuss the design and derive ideas to make people start building their own amplifier looks a bit strange. As long as you are not promoting your boss’s amplifiers, its fine. :)

/Hugo
 
F

fokker

Member
Joined 2006
Netlist said:
drzhuang,

What is the purpose of this thread if you can't disclose a schematic and can not talk freely about the design?

/Hugo
Click to expand...


for the same reason that Hypex UcD and NCD and other amps are discussed.

you cannot be selective in allowing commercial-like posts. Either you allow them all or you don't allow any of them.

As to what it is for DIY, I guess for inspiration and ideas for the DIY folks.
 
D

drzhuang

Member
Joined 2005
diaut.com
Netlist said:
Let me ask it differently: What is the benefit for the diy community? Talking about companies, amounts of power and specs without the possibility to discuss the design and derive ideas to make people start building their own amplifier looks a bit strange. As long as you are not promoting your boss’s amplifiers, its fine. :)

/Hugo
Click to expand...


kartino said:
A DIY range are from simple job like fixing a couple of bolt and nut until designing a complete kit. :D
Click to expand...


Do you have any orther questions except detail schematic document.
 
D

drzhuang

Member
Joined 2005
diaut.com
fokker said:



for the same reason that Hypex UcD and NCD and other amps are discussed.

you cannot be selective in allowing commercial-like posts. Either you allow them all or you don't allow any of them.

As to what it is for DIY, I guess for inspiration and ideas for the DIY folks.
Click to expand...


If you think it is commercial-like please report to admin to close it.
Else please stop useless discuss.
 
F

fokker

Member
Joined 2006
the prices are really attractive. Are those modules in production, or used by anyone in a commercial setting? How do the modules sound? any reviews on them?

the site actually has quite a lot of information and some of the daughter boards look like to use gate drivers. do you mind sharing al ittle info on that?

Thanks and this looks like a great product.
 
v-bro

v-bro

Member
Joined 2006
drzhuang said:



+/-60V can't transfer 1200W on to 8ohms unless THD up to 10%.
Click to expand...

Dunno, but shouldn't "+/ - 60Vdc and ground" be calculated with as an equal to 120V when compared to only "+60Vdc and ground"? In bridge mode the amp will get a double (pos. and neg.) 60V supply...

Can't find more info on bridged performance in the datasheet than:"up to 1000W into 4 ohms", but they say :"up to 2x300W into 4 ohms. The diagram "THD+N versus output power versus supply voltage" shows THD is only 0.2% at that level (at 54Vdc, this chipset can handle up to+/- 70Vdc)...
http://www.41hz.com/downloads/TK2350.pdf

But what is the purpose of 1kw amps anyhow, afterall you can create the same SPL levels with higher eff. speakers/more lower power amps+ more drivers.... :confused:
 
D

drzhuang

Member
Joined 2005
diaut.com
v-bro said:


Dunno, but shouldn't "+/ - 60Vdc and ground" be calculated with as an equal to 120V when compared to only "+60Vdc and ground"? In bridge mode the amp will get a double (pos. and neg.) 60V supply...

Can't find more info on bridged performance in the datasheet than:"up to 1000W into 4 ohms", but they say :"up to 2x300W into 4 ohms. The diagram "THD+N versus output power versus supply voltage" shows THD is only 0.2% at that level (at 54Vdc, this chipset can handle up to+/- 70Vdc)...
http://www.41hz.com/downloads/TK2350.pdf

But what is the purpose of 1kw amps anyhow, afterall you can create the same SPL levels with higher eff. speakers/more lower power amps+ more drivers.... :confused:
Click to expand...


Yes, it is right as your said and in dataheet. "up to 2x300W into 4 ohms" is 600W into 8ohms in BTL. 1200W/8ohms will be actualized up to +/- 70Vdc.

In professional field 1kW is not enough sometimes.
 
P

Pafi

Member
Joined 2004
snyaudio.hu/spafi/
fokker said:

how did you conclude that on an amp that you haven't tried? and how low is "lower impedance"?
Click to expand...

There are some basic calculation rules. When you want to reverse the polarity of a forward biased diode, you have to pump in a given amount of charge. This charge depends on previous forward curent, temperature, and di/dt. This is specified on datasheet as 4,3 uC at If=24A, 25C, 100 A/us. Actual Qrr is less because of less forward current (11 A effective), but more because of higher di/dt, and higher temperature. In the most optimistic way we could assume 3 uC. Since this charge comes from the opposit rail, energy loss is VDD*Qrr=2u*300=900 uJ. At 360 kHz this is 324W power loss in best case. Plus conduction loss, plus traditional switching loss. This is somewhere in the upper limit of dissipation capability of output devices.

and how low is "lower impedance"?
Click to expand...

There is no point in calculating exact impedance. It depends on cooling, type of audio signal, etc...

drzhuang said:


The Q isn't too high for high power MOSFET.
Click to expand...

You are right: Qrr in not high for MOSFET. It is high for audio amp!

It is current not Q limit the load impedance.
Click to expand...

Load impedance is limited by power dissipation, and power loss is affected strongly by Qrr (reverse recovery charge of body diode). I don't know what kind of Q do you think of.

Please measure efficiency!
 
D

drzhuang

Member
Joined 2005
diaut.com
Pafi said:


There are some basic calculation rules. When you want to reverse the polarity of a forward biased diode, you have to pump in a given amount of charge. This charge depends on previous forward curent, temperature, and di/dt. This is specified on datasheet as 4,3 uC at If=24A, 25C, 100 A/us. Actual Qrr is less because of less forward current (11 A effective), but more because of higher di/dt, and higher temperature. In the most optimistic way we could assume 3 uC. Since this charge comes from the opposit rail, energy loss is VDD*Qrr=2u*300=900 uJ. At 360 kHz this is 324W power loss in best case. Plus conduction loss, plus traditional switching loss. This is somewhere in the upper limit of dissipation capability of output devices.



There is no point in calculating exact impedance. It depends on cooling, type of audio signal, etc...



You are right: Qrr in not high for MOSFET. It is high for audio amp!



Load impedance is limited by power dissipation, and power loss is affected strongly by Qrr (reverse recovery charge of body diode). I don't know what kind of Q do you think of.

Please measure efficiency!
Click to expand...


Is there additional ultra fast recovery diode in your class d?
 
Eva

Eva

Member
Joined 2003
Pafi said:


There are some basic calculation rules. When you want to reverse the polarity of a forward biased diode, you have to pump in a given amount of charge. This charge depends on previous forward curent, temperature, and di/dt. This is specified on datasheet as 4,3 uC at If=24A, 25C, 100 A/us. Actual Qrr is less because of less forward current (11 A effective), but more because of higher di/dt, and higher temperature. In the most optimistic way we could assume 3 uC. Since this charge comes from the opposit rail, energy loss is VDD*Qrr=2u*300=900 uJ. At 360 kHz this is 324W power loss in best case. Plus conduction loss, plus traditional switching loss. This is somewhere in the upper limit of dissipation capability of output devices.



There is no point in calculating exact impedance. It depends on cooling, type of audio signal, etc...



You are right: Qrr in not high for MOSFET. It is high for audio amp!



Load impedance is limited by power dissipation, and power loss is affected strongly by Qrr (reverse recovery charge of body diode). I don't know what kind of Q do you think of.

Please measure efficiency!
Click to expand...

In practice it's far simpler than that. Conventional MOSFETs just latch in an on state and explode (without even a chance for overheating :D ) when a certain di/dt limit is reached during reverse recovery, or when the maximum allowed dV/dt slope is exceeded when Vds rises after diode recovery. That slope is less than 5V/ns for most devices, thus requiring long crossover times for reliable operation (and low efficiencies).

Also, the power ratings mentioned in previous posts are just false. With +-60V rails and 95% maximum duty cycle only 200W are achievable on 8 ohms, and 400W on 4 ohms (multiply by 4 for bridged operation). That comes from some basic math that everybody should know before attemptiong to discuss audio amplifiers seriously. Some understanding of body diodes, Trr and Qrr is also required when it comes to discuss class D seriously.
 
P

Pafi

Member
Joined 2004
snyaudio.hu/spafi/
Mosfet body diode isn't fast enough.
Click to expand...

Yes, there are MOSFETs with slow body diode, but not the one I use.

IGBT is < 200kHz.
Click to expand...

There are different IGBTs. Some of them are able to switch at 300 kHz, but I found that 200 kHz is enough. An ultrafast IGBT is still better then a slow FET with very slow body diode.

Schottky is fast but voltage is low. Do you have any models up to 300V?
Click to expand...

SiC: SiliciumCarbide. Infineon SDP20S30, SDT12S60. These are extremely expensive, but in my latest development I had no other choice.
 
Status
Not open for further replies.
Top Bottom