Vraptor 16 Watt Ultra low distortion and noise amp(0.000088%THD)

[Monstercore]

There were questions about the amp I showed in another thread.

M

Post in thread 'Choctaw - 10 Watt Amplifier, 1 Opamp + 2 MOSFET'

2024-09-26 2:37 pm

Yes I have balanced preamp.
20 till 30Watt is enough, I need it for 108db tweeter.

Now using 8 x parallel LME49600 with OPA1611 composit.

Next build should outperform this one on distorsion and noise(0.000088%/1W/8ohm)

• Monstercore

This is an amp as described in the wire thread. I use newer opamp like OPA1611 and special low noise regulator from Koifarm https://www.diyaudio.com/community/threads/an-other-lm78xx-regulator.403825/page-3#post-7494937

Max output is 16W.
Distortion = 0.000088% at 1W to 8ohm.

Gerber files for PCB ordering in the attachment.

Below the used regulator boards from Koifarm. You can also use your own +/-15V power supply.

Member LVQ have fun building this.

 

[UniQ]

Hi MC, nice design!
I am just curious did you encounter any problems running so many buffers in parallel?
Have you been able to check the temperature of each buffer, if so did you find any exceptional temperature difference between them?
Usually running op-amp's in parallel is often quite problematic, at least due to DC offset but sometimes also due to potential stability issues, however these are buffers with gain just below 1 for practical use (ex. 0,95x @ 100 Ohm load) so may be more forgiving and not causing much problem at all, I think I would still have added a small resistor to the output for each buffer, say 1 Ohm, the effective output impedance will still be much lower due to the global feedback.

There are some interesting discussions on TI support regarding paralleling buffers.
https://e2e.ti.com/support/audio-group/audio/f/audio-forum/745822/lme49600-paralleling-devices
https://e2e.ti.com/support/amplifie...m/273601/buf634-in-parallel-bw-pin-connection
https://e2e.ti.com/support/amplifie.../opa593-maximum-number-of-parallel-amplifiers

 

[Monstercore]

In this forum is a very long thread called “the wire”.

https://www.diyaudio.com/community/...ance-lpuhp-16w-power-amplifier.204069/page-39

https://www.diyaudio.com/community/...-performance-headphone-amplifier-pcbs.179298/

Here are a lot of different designs with lme49600 in parallel, up to 16 or 32 pieces. All without the need of resistors in the output. I did not encounter overheated chips. Daily in use.

 

[tomchr]

Is that measured or simulated THD?

Tom

 

[Monstercore]

Measured with Cosmos ADC/APU/DAC at 8ohm 1Watt output.

 

[TNT]

Didn't know OP1611 came in that package,,, what is e.g. U6 in your pictures above?

//

 

[Monstercore]

U2 till U9 are LME49600.

 

[Mark Tillotson]

I am just curious did you encounter any problems running so many buffers in parallel?
Have you been able to check the temperature of each buffer, if so did you find any exceptional temperature difference between them?
Usually running op-amp's in parallel is often quite problematic, at least due to DC offset but sometimes also due to potential stability issues, however these are buffers with gain just below 1 for practical use (ex. 0,95x @ 100 Ohm load) so may be more forgiving and not causing much problem at all, I think I would still have added a small resistor to the output for each buffer, say 1 Ohm, the effective output impedance will still be much lower due to the global feedback.

The output impedance is high enough to wire in parallel, from the datasheet you can work out the gain drops by about 0.05 when the load drops from 1k to 100R, implying output impedance of about 5 ohms. Since the output offset is +/-60mV max, that's about 12mA maximum current flow between outputs for paralleled devices in the absolute worst case with one +60mV and the other -60mV - this 0.12V difference is dropped across both 5 ohm output impedances in series, 0.12V/10ohms = 12mA. For a 250mA current buffer this isn't much distress...

 

[MagicianT]

I concerned about safety of opa1611. DS doesn't specified max input differential voltage, but circuits shows two diodes, likely 1-2V at max.
Since it's power amplifier, and dV /dt is quite high when power switch off, than two electrolytics attached to opposite rails would "tear apart" input of opa1611. Input protection is necessary, simple as two resistors or more advanced.

 

[Mark Tillotson]

I concerned about safety of opa1611. DS doesn't specified max input differential voltage, but circuits shows two diodes, likely 1-2V at max.
Since it's power amplifier, and dV /dt is quite high when power switch off, than two electrolytics attached to opposite rails would "tear apart" input of opa1611.

True, it is at risk of being fried. A contributary factor may be that C4 is too large a value, its only needed for the high frequency stability/performance I think.
You can also protect the OPA1611 inputs using back-to-back Schottky diodes which will conduct before the internal protection diodes.

The internal diodes protect the input transistors from reverse Vbe transients that destroy BJT low-noise performance. All low noise audio opamps should have such a feature.