Op Amp vs Single Transistor Amps

[dknguyen]

Hi. Cost aside (maybe not even that), what are the pros and cons of single transistor amplifiers vs op-amps? My ultrasonic amplifier I did have 3 cascaded op amps acting as amplifiers, but I've been reading and it mentioned very low signal levels and that single transistor amps have lower noise.

 

[Torben]

Hi. Cost aside (maybe not even that), what are the pros and cons of single transistor amplifiers vs op-amps? My ultrasonic amplifier I did have 3 cascaded op amps acting as amplifiers, but I've been reading and it mentioned very low signal levels and that single transistor amps have lower noise.

A single transistor amp is temperature sensitive. While another matched transistor can be used to compensate, it's often easier to get better compensation by just using an op-amp. Not sure whether that's always such a big deal though; the ultrasonic amp I built (basically this one but with an active filter instead of passive) uses a single-transistor preamp and it seems to work fine in any conditions I've used it in. It's probably more of an issue for RF work, I would think.

Many audio types won't touch an op-amp with a ten foot pole. Whether their claims are justified or not, many prefer discrete JFET circuits or even valves. You can probably charge more if you build it discrete (or with tubes) and call it "boutique", even though its measured performance may be identical.


Torben

 

[dknguyen]

Yeah a transistor pre-amp is probably enough to get the signal beyond the noise and then use op amps after. THere seems to be a lot of erronous data (or at least claimed erronous by others) on those audio forums so it's hard to tell. I don't really care about the gain being constant with temperature either...just needs to be big enough so the comparator won't be confused.

 

[audioguru]

A fairly good opamp has much lower distortion than a transistor or two.
Tests were made with a 3V output and plenty of negative feedback for a voltage gain of 3:
A single transistor has 0.2% to 1.5% distortion.
Two transistors have 0.02% to 0.3% distortion.
A fairly good opamp has 0.0005% distortion.
The distortion of the transistor and two transistors is the lowest when it inverts. Without negative feedback the 40% distortion at a high output from a transistor is horrible.

You won't hear noise from a low noise opamp.

 

[dknguyen]

A fairly good opamp has much lower distortion than a transistor or two.
Tests were made with a 3V output and plenty of negative feedback for a voltage gain of 3:
A single transistor has 0.2% to 1.5% distortion.
Two transistors have 0.02% to 0.3% distortion.
A fairly good opamp has 0.0005% distortion.
The distortion of the transistor and two transistors is the lowest when it inverts. Without negative feedback the 40% distortion at a high output from a transistor is horrible.

You won't hear noise from a low noise opamp.

DIstortion includes noise right? How does an op amp which is made of a hundred transistors have the same (or less) distortion (or noise?) than a single transistor? Or does it have to do with optimization during the manufactureing process for the transistors in an op amp?

 

[Torben]

DIstortion includes noise right? How does an op amp which is made of a hundred transistors have the same (or less) distortion (or noise?) than a single transistor? Or does it have to do with optimization during the manufactureing process for the transistors in an op amp?

I think in this sense, distortion and noise are separate. Distortion as I understand it would include artifacts introduced by the system's inability to perfectly recreate the input signal at the output (crossover distortion, flattening wave tops when there isn't enough headroom, etc). Noise would include extra stuff superimposed on the signal like Johnson noise, which is actually created within the resistors, and maybe EM interference would also be considered to be noise.

I'm not sure about other sources of distortion or noise, or even if I'm drawing the right distinctions here, but that's the way I currently think of it.


Torben

 

[audioguru]

0.0005% distortion in an opamp is down 106dB. The noise level is down 120dB or more so the noise isn't in the measurement.

The input transistors in a low noise opamp have very low noise. They have plenty of gain so they determine the noise level of the entire opamp.

 

[dknguyen]

I see. SO all those audiophiles are subjective fear-mongerers then? EIther that or they can't find op amps strong powerful enough, though I doubt it.

Op amps it is then.

 

[audioguru]

I see. SO all those audiophiles are subjective fear-mongerers then? EIther that or they can't find op amps strong powerful enough, though I doubt it.

Op amps it is then.

They are paranoid.
They take an OPA2134 opamp with 0.00008% distortion which is way below what anybody can hear, and they add a DC load so its output operates in class-A to reduce the distortion!

 

[Torben]

I see. SO all those audiophiles are subjective fear-mongerers then? EIther that or they can't find op amps strong powerful enough, though I doubt it.

Op amps it is then.

Well, the tube advocates like them because of the distortion, not in spite of it. Tubes seem to distort in a more "musical" fashion than transistors; I think it's something to do with the harmonics created when they overload. Op-amps can create a much cleaner signal path, but after so many decades of the tube sound, lots of people don't want to give it up.

I know that in my last band, the guitarist's amp (Fender Hot Rod DeVille 2x12; a tube amp) would overdrive with a really nice sound. However, every so often the tubes get old and either go microphonic or just die outright; the distortion in the last hours of the tubes' life was absolutely fantastic. Sad really; it always sounded best just before the tubes would have to be replaced.

But otherwise, there does seem to be more emotion than science in the arguments you see on those sites.


Torben

 

[Leftyretro]

"But otherwise, there does seem to be more emotion than science in the arguments you see on those sites. "

Yes that sure is the case in my dealing with several audio sites. That coupled with the old " a little knowledge is a dangerous thing " adage. Of course music is an emotional thing and I don't fault their desire to try and make technical facts fit their beliefs and myths

 

[Nigel Goodwin]

I see. SO all those audiophiles are subjective fear-mongerers then? EIther that or they can't find op amps strong powerful enough, though I doubt it.

Fear mongers or not, but it's likely that every CD (and album) in your collection has already passed through thousands of opamps in the recording studio.

Generally they are idiots who like to spend money on something that can't be measured, and can't be heard (blind tests prove that).

 

[Torben]

Just found a bit of interesting reading on the topic, at least as it relates to audio preamps:

https://liutaiomottola.com/PrevPubs/Preamp/preamps.htm

Summary: a double-blind study was performed, with a nearly perfect split right down the median. More research would be needed for a really good data set, but this is interesting anyway.


Torben

 

[dknguyen]

Same thing for violins (modern ones vs antique ones), and then there are people who are utterly unconvinced and will unequivocally (right context?) claim they can tell the difference. A blind study was able showed there was little difference and that when they were lied to that one was a certain kind the opinion could be shifted 180 degrees.

 

[audioguru]

The guitar preamp test is flawed. The FET had a 10M input resistor but the opamp had two 220k resistors in parallel (only 110k ohms combined). Experts say that a few megohms is the minimum that should be used.

Many electric guitar players want the sound to be very distorted.

 

[Nigel Goodwin]

The guitar preamp test is flawed. The FET had a 10M input resistor but the opamp had two 220k resistors in parallel (only 110k ohms combined). Experts say that a few megohms is the minimum that should be used.

Yes, 1Mohm is usually the minimum impedance for a guitar input, with more being common.

This only applies to passive instruments, an active guitar can happily feed a lower impedance.

 

[audioguru]

Hi Ron,
Your reply about the different input impedance of the two preamps disappeared.

The pickup has a certain sound quality when it is very lightly loaded with 10M. Its resonances are hightened. It makes a different sound when it is loaded with a lower resistance.

The output level from a guitar pickup is frequently a couple of volts, much higher than noise from a resistor. The FET has a gain of about 2.

 

[Roff]

Hi Ron,
Your reply about the different input impedance of the two preamps disappeared.

The pickup has a certain sound quality when it is very lightly loaded with 10M. Its resonances are hightened. It makes a different sound when it is loaded with a lower resistance.

The output level from a guitar pickup is frequently a couple of volts, much higher than noise from a resistor. The FET has a gain of about 2.

Yeah, I misread your post, and thought you were talking about the upper limit of resistance. After reading Nigel's follow-up, I saw the error of my ways, so I deleted the post. No one was supposed to see it.

 

[audioguru]

The moderator on another website keeps deleting my posts when he disagrees with me. Everybody sees them long before he deletes them and even after they are deleted.

 

[Speakerguy]

I can't wait til I start my new job and have access to the latest Audio Precision measurement gear. I think I'll breadboard up some of the stupider things I've seen on the net and post the results.