Near ZERO IMD buffer amp

[wenye]

Hi

I just finish my buffer amp modular design and build a sample.
I use audioprecision's system two (2322) to measure it, I got execllent result! In IMD testing I got a near zero(measurement limit of testing equiment)! Other testings results like THD+N, noise floor, lineartly,etc. were show execllent performance!

I think there are no commercial open-loop buffer amp IC or Modular can reach this performance! Even if the well know good buffer amp as BUF-03, OPA-633, BUF-634, etc. are not able to reach such high performance!

Here is a CCIF IMD vs amplitude testing result:

[wenye]

This is SMPTE IMD vs amplitude testing result:

[wenye]

THD+N residual spectrum @ 1KHz 0dBV input:

[wenye]

THD+N vs frequency

[wenye]

Close view of the open-loop buffer amp modular

[wenye]

Modular and testing board:

[eLarson]

congratulations. Are there any details that you care to share?

[wenye]

Thank you ELarson,

It's a open-loop buffer with a non-feedback distortion reduce technology ( a bit like pre-distortion, but different ). I spend around half a year on it's research and design......


Thanks to open-loop(non-feedback) buffering technology it got execllent IMD performance. And also thanks to my non-feedback distortion reduce technology, even if non good matching between transistors, we also got a result in THD+N testing...

Of course, "good matching good performance", still a law...

[Terry Demol]

Originally posted by wenye
Thank you ELarson,

It's a open-loop buffer with a non-feedback distortion reduce technology ( a bit like pre-distortion, but different ). I spend around half a year on it's research and design......


Thanks to open-loop(non-feedback) buffering technology it got execllent IMD performance. And also thanks to my non-feedback distortion reduce technology, even if non good matching between transistors, we also got a result in THD+N testing...

Of course, "good matching good performance", still a law...

Hi Weyne,

Good to see someone else out there doing open loop buffers

The measurements look great but at what load? I would suggest
re-doing them with a 600ohm load to give the buffer some real
stress and then see what harmonics are being generated.

Further, if you have designed an open loop buffer then it would
appear that you would drive it from an open loop front end. This
means a finite driving impedance and not the low Z OP from AP
test set. So I do all my AP buffer measurements WITH a 20k
resistor between AP and buffer AND into a 600 ohm load from
1kHz up to 20kHz and look at spectrum on FFT. This is a "real
world" stress test.

So far the best I get is about 0.0008% at 2V RMS into 600ohm
load driven by 20k source Z. This is pretty flat to 10kHz and
consists of H2 (most) and H3 (less) only (rest buried). At 5V RMS
-> 600ohms it rises to 0.0015%.

This, of course will say not too much of how it will sound but is
a good optimising tool for a given sonically preferred topology.


Cheers,

Terry

[wenye]

Hi Terry,

I'm happy to see there are someone doing open loop buffers and having success on it.

All of these results were measured under 1K resisitant loading. I can not understand why you suggest me to re-do it by a 600ohm load? As my knowledge, 600ohm as a standard balance I/O impedance in pro-audio system. So, the equivalent loading of each output buffer(hot or cold pin) is 300ohm. Then we can evaluate the same system performance as 600ohm balance load in a singal-end system using 300ohm load.

And you mention about the word "real-world". So, I think the resistant load can not give us a result like "real-world" condition. We also need to considering the load capacitance. Even if 100pF capacitance can make a big difference result. Especially in amplifing system with NFB. Okay, I will re-do these test with capacitance load, thank you for remind me this metter

Correct, I would drive it from an open loop front end still under design. Maybe the impendance is 1K maybe is 10K or more... But, I really interest in what performance difference when driven impendance increase from 20ohm to 20Kohm.... Hum... I'll try it too... Could you share us you testing result difference of this change?

0.0008% at 2Vrms is great, but the result for what? THD+N vs frequency? or something else?

Will do a listenning test after all performance... maybe in this week-end, maybe latter...