What happens when you cross a Tringlotron with a Regulator-Chip Amplifier?

[Elvee]

Answer:

The result is a Chtringlunator

The Tringlotron
The Regulator-Chip Amp

And their progeny: the Chtringlunator is a headphone amplifier, capable of delivering 500mW into 32ohm at low distortion, and without overall negative feedback.
The minimum power supply voltage required to obtain this power is 21V.

Pics will follow

[Bigun]

Hi Elvee,

You must be powered by Oktoberfest beer, I can't keep up with all of your inventions

[wakibaki]

Cool Elvee. You need to get out more.

w

[Elvee]

I run on pure, high quality biofuel, and there is not one square centimeter of carpet left in my home.

Here are the pictures of the prototype and the test set-up.
Measuring the THD proved quite challenging, not only because it is very low, and the circuit is a bit noisy, but also because the output is floating.

I had to use a 1:1 transformer, which served to subtract the output from the input.
With 12Vpp input, the residue caused by phase shifts, imperfections in the amplifier and transformer was 70mVpp. That's 44.5dB down.

The residue was measured on a spectrum analyzer (I didn't include a screenshot of the output waveform, it is just a boring sinewave).

The dominant harmonic is the third (cursor), at -53.3dB.
The global THD is about -51dB. Taking the subtractor into account, this put the actual THD at -95dB or 0.0018% (18ppm).
This is quite surprising, because the sim predicted a much higher figure of 0.0080%, and moreover, this figure also includes the distortion of the test transformer, which is subjected to the full amplitude of the signal.

This means the actual figure is even lower, particularly for the odd components.
The Tringlotron topology is basically asymmetric, and tends to produce more even harmonics.
The transformer is a very high quality one, salvaged from a professional Philips equipement, but it certainly contributes significantly to the THD figure.
The true distortion of the Chtringlunator is probably under 10ppm, and this is at the max power of 500mW.
At more reasonable levels, it is even lower.

[Elvee]

Some further details:

Unlike the class B regulator chip amp, this one has no slew-rate issues: it can provide its full power on the whole audio spectrum, and well beyond.

And though its performances are no better than a discrete tringlotron, they are much less dependent on accurate matching and thermal tracking of the output devices.

[Bigun]

The golden question - how does it sound ?

[Elvee]

Originally Posted by Bigun

The golden question - how does it sound ?

I'm afraid I'm the least qualified individual to answer that question, because,
a) I have tin ears,
b) I am the creator of the circuit, and therefore I cannot be objective on the subject. It's like asking a man if he is a good lover or a good driver
I will just venture so far as to say it sounds normal...
For the rest, you'll have to build one for yourself.

[jkeny]

How easily can this be changed to an I-V converter with zero impedance input - I'm thinking of something to use as output stage for a Sabre DAC?

I see John Broskie has done his own config - a Triadtron & used it as I-V

[Elvee]

This topology cannot be used as such for that purpose.

But the tringlotron principles can be applied to an I/V converter.

Here is an example: it is not very elaborate, and the performances are modest, but it shows the way.
With some minor modifications, the transistors could be replaced by voltage regulators.

[Zigis]

You say headphone amp? See this:
24V Aikido Line Stage and Headphone Amplifier

I use Broskie's LM317 output stage with different tube stage, and powered tube heater with lower LM317 in the same time.
I can't say it sound "normal", it sound really good!