Real Men Don't Use Opamps

[Mooly]

And how many OpAmps has the music been through before it got to your amp 🙂 Nothing you can do about that !

 

[mlloyd1]

Re: Real Man Don't Use Opamps

assuming you include preamps in this comment, there are some manufacturers that might take exception. current mark levinson preamps, for example. pretty much cost no object, but uses opamps in the audio handling circuitry.

Of course, this will beg the question of who determines what means "true reference quality amplifiers".

for what it is worth ...

mlloyd1

Op-amp has never been considered as an option for building true reference quality amplifiers.

 

[Onvinyl]

And how many OpAmps has the music been through before it got to your amp 🙂 Nothing you can do about that !

There is no logical link between this statement and the sound quality of opamps.
regards,
Rüdiger

 

[Salas]

Try a single ended current sourced 2SK170 Fet with no feedback instead of an op amp gain stage. Or a linear triode in same configuration. Compare to any op amp in music replay applications. Audio is peculiar.

 

[Onvinyl]

Hi salas,
how much gain can you tweak out your proposal and stilll get good sound? Do you have a schematic?

Rüdiger

 

[Salas]

You can achieve a wide gain selection. See here for JFETs.

I also attach two typical cases of medium and high μ triodes. G=18 & 79.

 

[Salas]

...

Of course you can resistor load the Fet or triode and have the simplest of amplifier stages if gain is enough.

 

[unclejed613]

add a few volts of 50, 60,100 or 120hz ripple to those power supplies and try it again........i tried it with the second schematic, and added a rather optimistic 5V of 120 hz ripple to the 350v supply. at the output the signal to noise(in this case, hum) ratio was only 48db. the ripple factor of the power supply is 15db, so the actual PSRR of this circuit is only 33db. most op amps are in the >100db PSRR range.

the ripple factor was calculated as 5v ripple compared to 350v of a possible rail-to-rail swing.

 

[Salas]

These are known issues. They need regulated supplies. We are talking musical performance, not the versatility vs an op amp. And if you simulate resistor loaded, you get even worse PSRR for single ended circuits like that.

 

[Panelhead]

Opamps used in real gear

I used to think that opamp usage was an indicator of poor/cheap design. Never purchased equipment that used them if reasonable options with discrete circuits were available.
I now think it is possible to build very good circuits using them. One example of an extreme high end amp is the ASR Emitter II. I have never heard or seen one, so it is all second hand. But the press and owners say it is extremely good sounding. The outputs are discrete, the input stage is a cheap and dated opamp. Does not seem to limit the Emiitter II. They were selling for about 25K the last time I looked.
Like the real designers here will chime in with; power supply, layout, and optimising the parameters will make or break either type of build.

George

 

[Salas]

For CDPs I see no reason why not the last op amp after IV can't be a Pass B1 style buffer . Very simple, very effective, superb sounding.

 

[unclejed613]

i have seen a lot of successful (i.e. stable, very low distortion and noise, very low output impedance and good reliability) amplifiers that use op amps for input stages. the toughest bug to overcome is stability. most of the op amps today have internal compensation which can cause problems with output stages that are much slower than the op amp. look at AN-47 on www.linear.com for a real good explanation of this. one of the things that many op amps posess that makes excellent amplifiers is very high open loop gain

 

[lineup]

.... AN-47 on www.linear.com for a real good explanation ....

thanks, unclejed613 😎
http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1028,P1219,D4138

AN47 High Speed Amplifier Techniques

This application note, subtitled "A Designer's Companion for Wideband Circuitry,"
is intended as a reference source for designing with fast amplifiers. Approximately 150 pages and 300 figures cover frequently encountered problems and their possible causes. Circuits include a wide range of amplifiers, filters, oscillators, data converters and signal conditioners. Eleven appended sections discuss related topics including oscilloscopes, probe selection, measurement and equipment considerations, and breadboarding techniques.

 

[benchtester]

using op amps

Unclejed - Thanks for the AN-47, looks really good.


Apex Microtechnology (now owned by Cirrus) make high power/voltage op amps and has a wealth of information in app notes and even have seminar manuals posted on-line.

http://apex.cirrus.com/en/products/apex/app_notes.html

Their PA-05 and PA-19 power op amps can be used for audiophile power amps. BTW they have pin-outs for external compensation.

Contrary to some previous posts, there are some op amps designed for high end audio. National Semiconductors have a new line of op amps designed for audio. They have "LME" in the part number prefix.
http://www.national.com/analog/audio/high_performance

(Also contrary to comments, a number of the National engineers are also audiophiles.)

 

[lineup]

Re: using op amps

Contrary to some previous posts, there are some op amps designed for high end audio. National Semiconductors have a new line of op amps designed for audio. They have "LME" in the part number prefix.
http://www.national.com/analog/audio/high_performance

(Also contrary to comments, a number of the National engineers are also audiophiles.)

Quite right you are, benchtester.
See this post of mine, that nobody seems to have read 🙁
I really do not understand how some people browse this foum.
They do miss a lot of great information posted 🙁

Originally posted by lineup
😎

LME49860 as mentioned in first post is a real killer!
The absolute max Supply Voltage is 2 x 23 Volt. Recommended is max 2 x 22 Volt.

Very good data for audio.
==========================================
Input bias current: 10 nA (max 72 nA)
Input noise: 2.7 nV/Hz
Slewrate: 20 V/uS
GainBW: 55 MHz
PSRR & CMRR: better than 120 dB

Data at +- 22 Volt.
==========================================
Maximum input Voltage: +21.0 / -20.8
Maximum output Voltage: +21.0 / -21.0 (into 2 kOhm)
Maximum output Current: 30 mA (minimum into 600 Ohm)
==========================================

I prefer to use SINGLE Op-Amps. The single version is LME49870
Product pages with Datasheets:
http://www.national.com/pf/LM/LME49860.html
http://www.national.com/pf/LM/LME49870.html

PS.
I downloaded the spice model for LME49860 (same for LME49870)
I ran some circuit tests.
The results corresponds rather well with figures found in datasheets.
Like maximum Output Current/Voltage and Gain BandWidth
So, I would say, it is a useful spice model.

LME49860 Spice ZIP-file
http://www.national.com/models/spice/audio/LME49860.zip


Regards, Lineup
http://www.diyaudio.com/forums/showthread.php?postid=1567678#post1567678

 

[audio-kraut]

I have a phono preamp - transistor based - by naim, compared to several op amp based phono preamps of studio extraction, and have found that with similar amplification ratio the studio pre amps are about 20dBs better regarding noise as the "naim" unit - and the naim was run with batteries, while the studio preamps were run with their original powersupplies.

Tonally I could not tell the difference (other than one being more annoyingly noisy).

 

[EUVL]

Just my 2 cents :

1) I have heard an ASR Emitter 2 with additional battery power supply. It has opamp frontend and Toshiba MOSFETs at the output. I am still building my own CLass A power amps ever since.

2) If my memory works, someone pointed out that the Burson opamps came from a source in China. You may probably be able to get them cheaper.

3) My own discrete opamp based on Pass article is much simpler, with only 7 FETs, 2 resistors, and optionally one cap, but I could not make them for less than 10 bucks, including PCBs, connector pins, ....., etc.

http://www.diyaudio.com/forums/showthread.php?postid=1471015#post1471015

4) I know of no IC opamp which works in rich Class A for all 3 (or more) stages. You can of course bias the output stage with a CCS diode, but that gives you 2mA of Class A. Especially for a JFET output stage, I'd rather have at least 10mA or more.

5) I agree with John and Charles, no global feedback sounds better. You may say it is prejudice. So be it.


Patrick


PS I work in the semicon industry, so I should really be supporting IC's. And of course I have sampled most of the "audio" application opamps, including the LME49860, AD797, OPA637, ..........

 

[Wavebourn]

4) I know of no IC opamp which works in rich Class A for all 3 (or more) stages. You can of course bias the output stage with a CCS diode, but that gives you 2mA of Class A. Especially for a JFET output stage, I'd rather have at least 10mA or more.

I know one with excellent output stage, but input noise is huge (very old design)

 

[lineup]

4) I know of no IC opamp which works in rich Class A for all 3 (or more) stages. You can of course bias the output stage with a CCS diode, but that gives you 2mA of Class A. Especially for a JFET output stage, I'd rather have at least 10mA or more.

thanks Patrick

This is one very main difference of the option you have with today's discrete vs. op-amp:
Class A output stage

I know one higher idle current power amp chip = TDA2050
This works more in a Class AB way.
But far from Class A.


Now, how come those new and real hi-fi chips can deliver so good in Class AB low idle currents?
Because of high gain + high feedback factor.
The output difference vs. a discrete Class A can be:
More 3rd harmoincs distortion vs. 2nd harmonics.



But then again, being fair .. those 3rd harmonics of a good modern audio op-amp
are at such ridiculously low levels
.. it would not upset your ears one tiny bit, my friend 😎
If your mind would think so .. this is what you get!!!!!!!!!!
because of subjective self bias deception of sound impressions.

 

[jean-paul]

Harris made some fast opamps that were biased in class A in the past.