Class D half-bridge amplifiers as we known to have some disadvantages. The main one is the requirement for a power supply. As a consequence of the pamping effect Hi-quality bipolar SMPS almost not available on the market, especially if you need to get a small power.
We have designed a circuit to adapt the Texas Instruments bridge power stages for use in a self-oscillating amplifier circuit. In such a case, a single monopolar voltage is sufficient.
TAS53xx, TAS51xx type chips are well suited for self-oscillating operation: they provide linearity in the of 0.2% without any special means. Propagation delay of 50-60ns contributes to the improvement of dynamic characteristics.
Even for a UcD modulator, this makes it possible to get better than 0.001% THD+noise at first watts without loss of quality at high frequencies.
Considering further of the technology, modern modulators like ncore/purifi can be easily applied to this platform. In this case the THD+noise will be below 0.001% in the whole frequency and power range.
We have made a variety of prototypes with different variations of the bi-polar supply for the modulators.
As is often the case, the most primitive one turned out to be the most effective as well.
As a result, a virtual ground source and a level shifter added.
The prototype is built on TAS5342 in PBTL mode.The block diagram of the amplifier is shown below:
Due to the peculiarities of the output stage, the amplifier cannot clipping, it simply disconnects the half-wave of the signal. but this is not a bug, but a feature
![20231230_203747[1].jpg](https://www.diyaudio.com/community/data/attachments/1172/1172776-32c28ebbbc6088b37a63e2bd523762e6.jpg "20231230_203747[1].jpg")
Tipical performans ( modern modulator like purifi et400)
15W 4ohm real speaker, we can't measure better because our DAC/ADC is of insufficient resolution
We tried our best and fit everything into a two-layer PCB. PP capacitors panasonic are used on the output. 8x470µf panasonic FR are installed in the power main, 2x220µf panasonic FM in the supply of the analog part. OA OPA1602+ high-speed comparator TL3016 are used in the modulator, preamplifier is assembled on OPA1612.
Subjectively it turned out to be quite pleasant sounding.
The module supports 2ohm load.
The signal spectrum in the picture is obtained with the power supply for 9 dollars, 24V 6A of the lowest quality.
We have designed a circuit to adapt the Texas Instruments bridge power stages for use in a self-oscillating amplifier circuit. In such a case, a single monopolar voltage is sufficient.
TAS53xx, TAS51xx type chips are well suited for self-oscillating operation: they provide linearity in the of 0.2% without any special means. Propagation delay of 50-60ns contributes to the improvement of dynamic characteristics.
Even for a UcD modulator, this makes it possible to get better than 0.001% THD+noise at first watts without loss of quality at high frequencies.
Considering further of the technology, modern modulators like ncore/purifi can be easily applied to this platform. In this case the THD+noise will be below 0.001% in the whole frequency and power range.
We have made a variety of prototypes with different variations of the bi-polar supply for the modulators.
As is often the case, the most primitive one turned out to be the most effective as well.
As a result, a virtual ground source and a level shifter added.
The prototype is built on TAS5342 in PBTL mode.The block diagram of the amplifier is shown below:
Due to the peculiarities of the output stage, the amplifier cannot clipping, it simply disconnects the half-wave of the signal. but this is not a bug, but a feature
Tipical performans ( modern modulator like purifi et400)
15W 4ohm real speaker, we can't measure better because our DAC/ADC is of insufficient resolution
We tried our best and fit everything into a two-layer PCB. PP capacitors panasonic are used on the output. 8x470µf panasonic FR are installed in the power main, 2x220µf panasonic FM in the supply of the analog part. OA OPA1602+ high-speed comparator TL3016 are used in the modulator, preamplifier is assembled on OPA1612.
Subjectively it turned out to be quite pleasant sounding.
The module supports 2ohm load.
The signal spectrum in the picture is obtained with the power supply for 9 dollars, 24V 6A of the lowest quality.
Attachments
![20231230_204231[1].jpg](/community/data/attachments/1172/1172777-a375a4acf93023c4ec1063e406b4e3b5.jpg)
Last edited:
Thanks, these nice build boards deserve good salesHi Eric!
I remember your request. Now the first batch is over, instantly
~ PM you for the detailed orders.
Thanks for the tip.
TAS5261 can also be used in this mode, but it will be necessary to form a separate inverted PWM signal, as the chip does not know how to work in 1N mode. There may be some difficulties with this because of the small offset time of the inverted PWM signal. But most likely it will work.
If I can buy one, I'll try it
TAS5261 can also be used in this mode, but it will be necessary to form a separate inverted PWM signal, as the chip does not know how to work in 1N mode. There may be some difficulties with this because of the small offset time of the inverted PWM signal. But most likely it will work.
If I can buy one, I'll try it
@theAnonymous1, thanks
We're targeting small batch production. The chip is very convenient, but I am afraid that when we make the board we will not be able to buy it. Most likely we will use TAS5162 for more powerful versions.
Wondering how do the spectral line noise skirts compare to purifi et400?
Measurement technique discussed at: https://www.diyaudio.com/community/threads/phase-noise-in-ds-dacs.387862/post-7063038
Measurement technique discussed at: https://www.diyaudio.com/community/threads/phase-noise-in-ds-dacs.387862/post-7063038
