Parallel connecting of big capacitors with smaller capacitors for optimising the sound??

[Cqwet Dbdfte]

The OPA1677 Fig.6-9 THD%+N curves looks like it could be a fixed ~1uV input noise until it gets suppressed by other non-linearities at around 3V rms out. This may be the low frequency random noise from Fig. 6-2. If it was some monotonic noise it would be bad.
Simplified schematic Section 7.2 shows a true input differential topology with output stage having gain with a common source connection.
Nice part.

 

[ZHOUJUN]

I think the selection of these capacitors is based on precise calculations， Changing the capacity or connecting them in parallel will have an impact on the circuit.

 

[Cqwet Dbdfte]

I think the selection of these capacitors is based on precise calculations， Changing the capacity or connecting them in parallel will have an impact on the circuit.

Power supply bypass cap calculations are usually done by the accountants for a precise amount of profit.
Yes it will have an impact. Less unwanted ACV in the power supply. The closer to zero ACV, the better in power supplies.

However, If a voltage regulator circuit is used with a lot of gain, the wrong amount of capacitive loading can set it into oscillation.
Caps may need a series resistor to stabilize the loop. Some LDO regulator chips advertise stability with (Low ESR) ceramic caps, some do not.

Coupling caps are high pass networks, so only a minimum value need to be selected for a required low frequency roll-off for a load resistance connected, which is a rough calculation.

 

[thor.zmt]

If you don't mind sharing, am curious about what other opamps were tested and the opinions on those?

Full range of TI audio units, JRC Muses and Audio standard parts, some AD I forgot.

Generally the latest units did better, J-Fet over Mosfet in and Fet over Bipolar.

The main outlier was OPA1652 which ended up most preferred. In a separate test we compared it on adapters to OPA627 and OPA637 which in my books are still are still the gold standard for sound quality. It was very close, to my ears indistinguishable.

It should be noted that the circuits did not stress the Op-Amp's, moderate loads, levels. More stresses may change the outcome.

Since the big test the OPA1656 has been introduced which objectively performs even better at a modest premium and even better objective performance and the industrial OPA2156 which is not just re-symbolised but slew rate enhanced.

For mass production or mass modification of existing gear full of 5532 & 074 my suggestion is OPA1678/OPA1679 (Dual/Quad) due to VFM.

For small scale DIY I'd go with OPA1656 or OPA2156, depending if slew rate or HF distortion are of greater concern.

The JRC units, 553X, TL07X/TL08X are not competitive, some AD IC's scored highly but are very expensive.

I observed some loose correlation with (internal) circuit design.

Double Differentials with J-Fet input (which in discrete are a staple of legendary "good sound" Hifi gear from Luxman and the various DIY Kaneta designs) tend to end up in top.



Folded cascodes come next and are close.



These structures for input lock out all top positions.

Classic "Lin/Self" style circuits with differential in loaded by current mirror and SE VAS are all clustering near the bottom, especially Bipolar input ones.



Outputs including different structures like diamond or traditional follower vs. rail to rail outputs showed no obvious correlation.

No obvious correlation with THD or other objective parameters was found.

Thor

 

[thor.zmt]

Simplified schematic Section 7.2 shows a true input differential topology

Folded cascode, actually.

with output stage having gain with a common source connection.

Nowadays called "Rail to Rail"output.

Cheap Op-Amp's come a long way from 5532/07X/08X.

Thor