Hi Sergio,
Ahh, you might have been thinking that the input pots to the IR2113 were 10k instead of 100k.
Ahh yes, even the 2SA970 input pair are rated for 120 volts. The 2SA1123 ones I am actually planning to use have a Vceo figure of 150 volts. They should see only a little more than the voltage from one supply rail too.
The delay using the series of gates should be OK for the sub amp. Doing it this way prevents the IR2113 from getting ambiguous signals.
I have experimented with hand made air core inductors in switching amps and like how they perform. I even have wound two smaller ones in series and adjusted their physical orientation 180 degrees in relation to each other. I just hadn't yet come up with a desgn for an amp I liked enough before now.
I usually use a drop down resistor and a 15V zener for the control power supply on the lower rail. Then another resistor and zener for the input of the 2113 and the 74C14. I usually run the 74C14 off of 10v
Thanks very much Charles. I like the 74C14 schmitt trigger very much. It is the same thing as the CD40106.
IVX, that is a good idea about the capcitance of the diodes. I have always gotten away with it though in actual circuits for some reason.
Ahh, you might have been thinking that the input pots to the IR2113 were 10k instead of 100k.
Ahh yes, even the 2SA970 input pair are rated for 120 volts. The 2SA1123 ones I am actually planning to use have a Vceo figure of 150 volts. They should see only a little more than the voltage from one supply rail too.
The delay using the series of gates should be OK for the sub amp. Doing it this way prevents the IR2113 from getting ambiguous signals.
I have experimented with hand made air core inductors in switching amps and like how they perform. I even have wound two smaller ones in series and adjusted their physical orientation 180 degrees in relation to each other. I just hadn't yet come up with a desgn for an amp I liked enough before now.
I usually use a drop down resistor and a 15V zener for the control power supply on the lower rail. Then another resistor and zener for the input of the 2113 and the 74C14. I usually run the 74C14 off of 10v
Thanks very much Charles.
I like the 74C14 schmitt trigger very much. It is the same thing as the CD40106.IVX, that is a good idea about the capcitance of the diodes. I have always gotten away with it though in actual circuits for some reason.
Yes, because although the logic signals to the opposite "corners" of the bridge are the same, one mosfet is high-side while the other is low-side, so the electrical levels are not the same.
You would need two IR2113 with both inputs (PWM and !PWM) connected to opposite inputs.
However, I wouldn't go that way. I think half-bridge is better. Or you can think this way: you can build two half-bridge amps and connect them in bridge mode as with any other amp, this way you have the same design and more flexibility.
Best regards.
You would need two IR2113 with both inputs (PWM and !PWM) connected to opposite inputs.
However, I wouldn't go that way. I think half-bridge is better. Or you can think this way: you can build two half-bridge amps and connect them in bridge mode as with any other amp, this way you have the same design and more flexibility.
Best regards.
Well, if I were you, subwo1, I would build a +/-40V version first in half-bridge and once you have it working well, I would think about using full-bridge or increase the rails or bridging, it is a matter of adding components only.
Remember to be very careful with IR2113 supply bypassing and layout.
By the way, what are your rise/fall times at the inputs of the IR2113 in the simulations?
Remember to be very careful with IR2113 supply bypassing and layout.
By the way, what are your rise/fall times at the inputs of the IR2113 in the simulations?
Yes, the 7414 can operate down to 5V but you should use the same supply as for the IR2113 to ensure correct thresholds. And you should feed the IR2113 with at least 12V to ensure proper enhancing of the mosfet gates. Conclusion: use 12V to 15V for both chips.
I don't understand your explanation about a diode a resistor in series or something like that. Could you explain it more, please?
Best regards.
I don't understand your explanation about a diode a resistor in series or something like that. Could you explain it more, please?
Best regards.
ssanmor said:
Actually my biggest concern is power supply pumping.
Good idea about bypssing the IR2113. I have a lot of tantalum capacitors.
The rise time is about 500nS with the pots at maximum. The fall times will be about 10nS for the first and 10nS for the other input of the IR2113 plus the delay of the gate.
PS pumping hppens when the inductive kick of the speaker feeds back through the output devices into the filter capacitors, increasing their charge above the level supplied by the power transformer and rectifier. The problem is often lessened by using large filter capacitors. But I was planning to use a switching power supply with small filter caps, so could have a problem
You are right. Good point!
About the negative suppy, that's what I did: I put a resistor from ground to the cathode of a 15V zener, with its anode to -50V. A big capacitor and a tantalum in parallel with the diode. This way you have 15V referenced to -50V. Just an idea.
You must pay attention to the dissipation of the zener and the resistor. I found that the average consumption of both my XOR chip and the IR2110 at work (with 10 Ohm gate resistors) was 50mA. With your 80V rail you will dissipate (80-15)*50mA=3.25W, so you need quite a big resistor with this approach.
Ahh, unless the resistor is connected to the diode which is connected to only half of the power transformer. Then, the resistor sees only half of the transformer voltage for only half of a cycle. If the peak voltage is 80v, the rms is 60v, but only for half the time, so the power dissipation is way down. 60-30=30, but for only half the time, so 15v. But at 100khz, only about 20mA is needed. 15v/.02 mA=750ohms. 15^2/1k=225/750=.3w. Then, that voltge is dropped down to 15v through another 750ohm resistor.
subwo1 said:
Indeed ? This must be dangerous for your components... why not design something to short the extra voltage to ground ? (something like a power-zener implemented with a Zener and a BJT) set to a few volts above your normal operating voltage ?
Also, about self-oscillating Class D, I personnally believe this is better. "Full Digital" amps or amps with a clock (like the Tripath) which output a bitstream aligned with a clock cannot avoid the same problems 1-bit DACs face. However, purely analog class D has no time quantization (it is continuous-time) so it avoids these problems entirely. Then, it is logical to go all the way to self-oscillating, and remove the triangular "clock"...
Has anybody got the Philips DSL UCD schematics ? I'm dying to see them...
A classic PWM modulator cannot be functionally compared to a 1-bit DAC, since the 1-bit DAC uses discrete level and discrete time, while the classic PWM modulator uses discrete-time but continuous value (the actual output voltage is a linear function of discrete voltage AND continuous duty-cycle).
And even a self-oscillating amp performs time quantisation, although at an unstable and much much lower rate than a delta-sigma amp like the Sharp and Tripath !
These criteria alone don't determine the sound quality of a final amp. As always it is not only a question of what principle you use, it is a question of how you implement which principle.
Regards
Charles
P.S.: D-S amps use overall feedback so they cannot exactly be compared to 1-bit DACs.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.