Do you mean it varies only when you use it? 😀
A ClassD amp is one of the mostly varying load, it can go from small negative to Vdd/Rl*2 (2 channels). 36V/4*2=18A on secondary side! OK, only in peak, but you have to maintain peaks.
If I were you I redesigned the power section. And control section also... 😉
Distribution of primaries is good. Distribution of everithing else in primary loop is even better! Power switch. Input decoupling capacitor. Divide everithing by 4, and you can build it without any heat sink! Use polimer electrolite capacitors, and multilayer ceramic close to the mosfets!
No. I mean, when you only play the low frequency component (like bass bit), then current consumption will be higher. Obviously the class d amp is mostly varying load. But regulation here does not make any significant effect on sound. I think we may need an inductor at the 2ndary output to form an LC filter, else the feedback will not work properly. I am saying this from my previous experience.
Can you explain a bit " Divide everithing by 4, and you can build it without any heat sink!" I didnt understand it ..
I dont understand what you try to say. Middle freq doesnt need power? Why? And why is this "only"?
But does it make any difference? Load is varying. Thats all. Output voltage sags exactly when you need most. If you make it high enough under heavy load, then it will be too high, or at least unneccessarily high at idle.
Well. High leakage inductance can also be used for this purpose, but it's not easy to do it well this way, so yes, an inductor is advised.
4 segments (or more) of complete power stages distributed evenly over the toroid. You can use smaller components and proximity effect will be much lower. Also the heat is distributed on a higher surface.
TAS5630 is capable of outputting max of 300W. So, for 4 channel, the max secondary load current would be 300W/36V= 8.33A . So its not 18A !
Do you mean to use 4 toroid instead of 1 single toroid ?
1: output power is always lower than input.
2: there is current headroom.
3: 300 W is average value. Sine peak is roughly the twice (roughly, because of distortion).
If you look at the datasheet, you can see many values up to 600 W depending on many factors.
Actually calculation is simple based on Ohm's law. I assumed nominally 2 x 4 ohm stereo load, but it can go down to 3 ohms at certain frequencies.
I dont know why did you wrote 4 channels.
No. 1 toroid. 4 segments. I dont know how to explain for you without drawing.
I understand now.
I wrote 4 channels because TAS5630 is a 4 channel audio amplifier.
I understand now the winding things. You probably mean to distribute the 4 segment evenly on toroid of an winding.
I have rewinded the transformer with evenly distributed winding.
I used a gate drive resistor of 22R and added 1n4148 diode across the gate drive resistor so that turn on is slow and turn off is fast. Now performance is improved and the wavehspae is clean till 14.6V. After that the spike is again starting to appear.
Is there anything I can do with the snubber network ? How can I calculate the resistor and capacitor value of the network ? I am following AN11160 application note from NXP.
I used a gate drive resistor of 22R and added 1n4148 diode across the gate drive resistor so that turn on is slow and turn off is fast. Now performance is improved and the wavehspae is clean till 14.6V. After that the spike is again starting to appear.
Is there anything I can do with the snubber network ? How can I calculate the resistor and capacitor value of the network ? I am following AN11160 application note from NXP.
You can use it in single ended mode, but then it will be much more sensitive to supply voltage variation. And you have to use big coupling capacitors. I absolutely not recommend it unless you really know what you are doing. If you drive it with same phase signal, you will experience a very bad surpise. During negative half period it will pump energy back to supply rail. If puffer capacitor is not big enough, amplifier can kill itself (or switch off in better case). But even if nothing is damaged, quality will be far from expected. And what about the power? Peak voltage will be less than18 V. Its not much higher then with a bridged TPA3116 at 14 V PSU. Why is this big effort for this tiny result? I doubt you did the basic calculations.
Yes, but 1 segment consist of 2 primary windings in close coupling, 2 MOSFETs, and 1 decoupling capacitor. This all is then multiplied by 4.
Well, I was not perfectly right, because that coupling capacitors are split, this way ease the pumping back problem, but still SE mode is not the best way of usage, especially if one needs deep bass and high power. What is this going to drive? Satellite speakers of a 5.1 system? ;-) What ohm? For 4*75 W at 36V PSU you need 2 ohms. Not very usual.
Ya. There are 3 capacitors there, C67, C68 and C69. The amp will be used as SE because 4 audio channel will play different audio as well as same audio signals if required. The speakers are 4 ohms each.
I dont know what you are talking about. Which schematic? 3 capacitors? There must be 8 for 4 channels. And at 4 ohms you will get only less than 40 W/channels.
Are you talking about the two caps/channel reference design schematic on the datasheet for the SE ?
I am changing the turn ratio to 4:4:13 so that, at 14.5V the output power per channel will be 80W (Ref page 14 of the datasheet.)
Obviously. And 2x470 uF is enough only if 50 Hz LF cutoff frequency is OK for you.
Almost 80 W, at 1 kHz, with 10 % distortion, if you didn't have any drop on battery connection, dead time, transformer, and rectifier. But these are unavoidable. At lower frequency output coupler capacitors will also decrease output power significantly.
With smaller, bridged amplifiers you could reach the same or more power at only 25...30 V PSU, and you didn't need the 8 big capacitors, could reach as low freq as you want, and the demand for the PSU is halved. Actually it could be built with the same parts, you only have to omitt 2 diodes and 1 secondary coil. The remaining diodes could rectify the voltage of the primary. The problem of overshooting would be solved automatically also, without any snubber.
The reason why I don't really recommend it is this: however you don't have GND separation in your present design either, you still have the chance to do it later when you realise the problem with GND loop. If you used the voltage doubler I described above, GNDs are not separable anymore. GND loop can be broken also in the signal path, but I'm not convinced you could do it properly.
The reason why I don't really recommend it is this: however you don't have GND separation in your present design either, you still have the chance to do it later when you realise the problem with GND loop. If you used the voltage doubler I described above, GNDs are not separable anymore. GND loop can be broken also in the signal path, but I'm not convinced you could do it properly.
Last edited:
- Status
- Not open for further replies.