JFet vs bipolar

[rickpt]

Some of you (myself included) like so much the NE5534 because it's lowest distortion opamp ever made, it even beats the AD797 in a null test. take the OPA627, it has 20dB more distortion at 20khz than the NE5534 or the AD797(these NE haters guys don’t like the Ad797 too, can you see why?)

if you do the null test to the OPA627 you can clearly see where the frequency compensation(1khz aprox.) point is and there is where distortion rises skyward! and there were the NE really shines, it has a brilliant compensation scheme that maintains the open loop gain linear over the audio band, and the distortion down as well.

And by the way, add more 10db of distortion to the OPA627 if you are driving a 1k load...region were the Ne5534 and the AD797 really shines because distortion is almost the same with load or without

you guys can say all you want but the NE5534 is an audio engineering masterpiece and its performance isn’t equaled by any monolithic device ever made and some of us now this!...

(I can show the various null results of various audio opamps if you guys want)

Best regards

Ricardo

[Elso Kwak]

Originally posted by rickpt

(I can show the various null results of various audio opamps if you guys want)

Best regards

Ricardo

Yeah, Ricardo,
Show us your null results. I am looking forward to it!
Ciao,

[rickpt]

to the end of this week you will have them, i have to convert all the voltage values to dB and that takes time...

The forum members that dont now this type of mesurement, search the forum and also search the hafler XL280 excelinear manual. there is very relevant info on the null test in this forum...
A test invented by a genius named David Hafler.

the opamps measured were the:

NE5534(texas)
NE5534(signetics)
OPA627BP
AD797
OPA604AP
OP27GP(PMI)
OPA134PA
TL071

Best Regards

[john curl]

I second Elso's input. In principle, the AD797 should run rings around the 5534. Please understand, I have used the 5534, since 1977, when I was first sampled by Signetics. It was a serious improvement in audio IC op amp design and performance. However, we found that IF you bypassed the input stage and substituted a fet input stage, we could make a better sounding device. We (my tech actually) made them for Dave Wilson for his $100,000 speaker system's active equilizer. Dave paid $80 each, 15 years ago, I believe.
The AD797 is similar to the 5534, but significantly improved, at least in principle. I'm sure that the designer of the AD797, Scott Wurcer, knows the ins-and-outs of the 5534. He would not have made an inferior device on purpose, at least.

[PRR]

> The AD797 is similar to the 5534

The spec-sheet puffery says otherwise:

"The new architecture of the AD797 was developed to overcome inherent limitations in previous amplifier designs. Previous precision amplifiers used three stages ..... The AD797 on the other hand, uses a single ultrahigh gain stage to achieve dc as well as dynamic precision."

The way they draw it, I was very puzzled for a while. Mostly it is a semi-conventional bipolar op-amp, though it takes great advantage of the "Complementary Bipolar (CB) process" (good PNP and NPN on the same die) which was not available to the 5534 designer. But the bootstrapped current mirror and the compensation it allows are pretty wacky, at least novel, to these old eyes.

On gross-specs, it has noise voltage similar to or better than 5534. Input current, rather high on 5534, has been reduced on the 797, but still can't be confused with a TL071. Output current is almost/not-quite as high as 5534. Gain at low-Z loads is better. Max rated supply voltage does not match the 5534's exceptional +/-22V, which may not matter to most people but is sometimes handy. (I'm not sure there is a difference in the silicon, or just in the specs. Certainly normal +/-18V chips do not explode at 18.1 or even 20 volts.)

Others have noted the 5534's unhappiness when you feed it DAC spikes. My reaction is: "Don't DO that if it hurts!" The 5534 was designed for an older mellower world without digital gizmos, and we always put some thought into out-of-band crap-control before signals hit an active junction. Yes, for typical crap-levels, a bipolar is wiped-out and most FETs just take it without distress. And not dealing with crap-control may lead to better sound. But that is one of the trade-offs we face.

[rickpt]

the null test results

cheers

Ricardo

[Jocko Homo]

Might be a radically new concept to some.

Mandatory to the rest of us. Even with JFETs, although they tolerate it more easily.

Jocko

[Jocko Homo]

A friend recently tried to sell one of his D/A converters (built by me) that he has an extra one of.......

Tried to sell it to a guy who just spent $$$$$ on having all of his system cryogenically treated. (I have no idea why....waste of money that he could have spent on me!)

Anyway, he declined to buy the D/A box. Said that he now heard "too much detail", and that "frightened" him.

No explanation why some people like to spend $$$ on stuff, yet want it to sound all mucked up.

IOW, if JFETs have "too much detail", maybe something else is wrong with your system.

Uh, change that "maybe" to "probably".

Jocko

[dimitri]

The art of electronic engineering is to put the right component in a right place within a budget, not to populate your board with the most expensive, most recent so-called "technological achievements". Don’t choke the cat with cream!

… and you can find helpful Scott Wurcer design note AN348 here:

http://www.analog.com/UploadedFiles/...01060AN348.pdf

and ad797 story here:

"An Operational Amplifier Architecture with a Single Gain Stage and Distortion Cancellation," AES Preprint 3231

[johnferrier]

Originally posted by dimitri
and ad797 story here:

"An Operational Amplifier Architecture with a Single Gain Stage and Distortion Cancellation," AES Preprint 3231

Hi Dimitri,

Do you have an online source for the ad797 story? Google didn't help me.



JF