Selfoscillating class D amplifier

[81bas]

Originally Posted by Eva

Comparators exhibit very high gain and slew rates at the expense of the inability to make them work in a NFB loop. 5mV of overdrive are enough to change the state of the outputs completely within a few dozen nanoseconds.

Put 5mv overdrive to the hi-speed opamp inputs, and it will change it's state very fast too
The only differences between op-amps and comparators I see is the special cares to avoid a deep saturation of BJT devices inside of a comparator and avoidance of any phase corrections (making them unable to work with NFB).
And famous LT1016 does not have a very high gain, actually it has even less gain than any other precise op-amp...

Originally Posted by Workhorse

CD4070 XOR is used as PHASE SPLITTER OF PUSLES [for driving the gate driver chip]not a comparator, opamp is already acting as integrator and comparator simuntaneously

I know very good, what the CD4070 is intended in this schematic for But if you say, that the CMOS element does not have any gain, then please explain, how it is possible to build crystal oscillators using it...

[Eva]

Can any of your high speed op-amps produce full 50Mhz square wave output with 5mV of overdrive on the input?

High speed comparators can. They can't be made stable with NFB because they keep the high gain up to very high frequencies, unlike op-amps. Gain at DC may not be as high as in op-amps, but there are no poles on it until very high freqencies.

[Workhorse]

Originally Posted by 81bas

I know very good, what the CD4070 is intended in this schematic for But if you say, that the CMOS element does not have any gain, then please explain, how it is possible to build crystal oscillators using it...

I never said that cmos is not having any gain, all i told you was that the application of cd4070 in that schematic was for phase splitter rather than a comparator.

[Workhorse]

Originally Posted by 81bas

Put 5mv overdrive to the hi-speed opamp inputs, and it will change it's state very fast too
The only differences between op-amps and comparators I see is the special cares to avoid a deep saturation of BJT devices inside of a comparator and avoidance of any phase corrections (making them unable to work with NFB).
And famous LT1016 does not have a very high gain, actually it has even less gain than any other precise op-amp...

Can your hi speed opamp could switch in less than 10nS????

Comparators are mostly known by RESPONSE TIME which is of importance and in opamps the response time doensnt matters that much.

[Eva]

Given that the TL081 op amp used is not "fast" at all, the first XOR gate will act as a comparator, something that a CMOS gate is not intended to do, since the analog input level, always close to 2.5V, may result in high dissipation in the gate due to prolonged cross conduction.

I use CMOS gates as 1/2 Vcc comparators sometimes, but only when the input voltage is going to be far from the commutation threshold most of the time. The circuit can't rely on the threshold being exactly at the center either.

[AP2]

Hi All
tl081..very good.
see pic.

[81bas]

Originally Posted by Workhorse

Can your hi speed opamp could switch in less than 10nS????

would this hi speed op-amp have no any phase correction networks inside and would this op-amp have some shottky diodes between base and collector of BJT devices to avoid their saturation, then it would be able to do this... But in general, they are the same... And please notice, that the CMOS gates are not having any saturation effect

Also, CMOS elements of CD4000 series are not having low Rdson, so it is not so dangerous to feed their inputs with Vcc/2 at all. But the HC74 series may indeed overheat and die in such case.

[AP2]

Hi,
I recommend fast hfe4070 very good, change the input stage comparator ex lm319d in common mode (60ns to conversion), small change bjet and polarizzation for jumping, adding an op-amp input (also for balanced) and phase correction. result is good sound.

[AP2]

hi,
I'd be curious to know what parameters of the TL081, have decided to use this circuit?

[vogor]

Greetings!

I have returned

It is glad that the circuit will work. There was one more question: what power(W) can be received with a +/-100V supply if to include such two amplifiers the bridge?

I have chosen tl081 for experiment because of its low cost if the circuit will earn that I am going to put something better ad817, ad8001 for example

Originally Posted by 81bas

UcD will give more stable switching frequency and better performance results... Simply replace DA1 with a comparator IC (T1+XOR will be not enough here ), replace your feedback with a UcD style feedback and it will be a UcD amp!

It will be better agree that ucd but here is how it to make here while precisely I do not know, can share as better to transfer the given circuit in ucd It is possible sent on my personal email: vogor@yandex.ru

Thanks for answerrs 2 all.