• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Can op amps now be as good as discrete circuitry?

I am trying to imagine a unit being designed where the op amp to be used is not a known factor during the design process.

"We are done, now for the op amp. Bring out the "Wheel Of OP Amps" and give it a spin, let's see what random one we will be using!"
May I direct you to units where op-amp rolling is a selling point?
 
Lord, I'm starting to despise certain parts of this godforsaken industry for spreading FUD and general crap, this then spouted to tech-ignorant end-users by equally or more tech-ignorant dealers, who sell what they're either told to, or what gives them the most profit. YES, I was a 'dealer' once, but I made sure to get some half decent product knowledge and bore the pants off any manufacturer I was able to speak to - the extra 'knowledge' I learned often getting me into trouble...
starting?

STARTING?

In all honesty I can't bring myself to despise it. It's silly, insignificant and laughable. There's enough in the world really worth despising to occupy me.
 
It is possible to build an operational amplifier with discrete parts or all in one on a planar chip.

The cost and effort to select and match discrete parts is not within the reach of a boutique manufacturer. On a chip, today, you can get better matching, lower noise, and the cost to implement things like current mirrors for a gain stage is trivial in comparison to discrete.

I've never taken apart a discrete block module like a Hardy, Sparkos, 2550, etc. but I would bet they have sets of dual matched transistors on the same chip in the design. The original purpose of those block op amps was high gain and they drove the relatively low impedance primary of an output transformer to provide voltage gain. Planar chip op amps of today's quality were not yet available in about 1970 when block op amps for audio emerged. They was the basis of Neil Muncy's Suburban Sound mixer, and a slew of API mixers.

You can even go to a chip op amp maker like Analog Devices and have them bin out select parts for a price using their test equipment. It is also possible to do laser trimming on a chip in a wafer probe tester for even more precision matching of things like a resistor ladder. I doubt that is done much today.
 
Last edited:
If you're copying text it helps to give a link to the source, both for attribution and context.

The video has some other funny remarks too, e.g. "the more components on a signal path the worse it sounds". Gotta get me a Burson crystal set.

Also funny that an audio vendor should have such poor audio on the video.
 
Fine, let's do this..

Advantages​


  • Custom circuitry depending on application
I already mentiond the thousands of types of opamps. Pretending they are all one and the same is a red hearing
  • Significantly reduces component count to better preserve signal integrity
I really doubt that you end up with less components if you want to achieve the same lever of performance of an integrated circuit.
  • Component are top-quality audio-grade components
Whatever that means
  • 1%-matched metal-film resistors and silver mica capacitors
1% matching is a joke to laser trimmed integrated circuits
  • Extremely temperature stable
So are opamp
  • Each transistor is chemically optimized for its application: NPN or PNP
So are opamps, that these cannot be mixed on an IC is nonsense.
  • Each transistor tested and matched before hand-soldering onto the PCB
Yes, that is mandatory for proper operation. I already mentioned laser trimming.. This is not a good argument.

As for hand building, I always find it baffling that this is an argument that something is better. Machines can do things a thousand times more precise than people. Do you want a human to align the lenses and mask of the TSMC NA EUV machine by hand to expose the wafer of your latest CPU? I think not!
 
So Burson says
They also say that integrated circuits are made on a "silicone dice"

1725629071058.png

I guess that explains a lot!
 
Do they say use anything, or do they have a list? Where does the list come from?
Picking the first I thought of: https://fosiaudio.com/pages/2-channel-amplifier-v3 - nothing beyond making a point of 'swappable op-amps.' From what I remember that's typical of other manufacturers too. Nothing about not using x,y,z because they will be unstable in this application, and no list of op-amps tested and known stable. They probably were actually designed with a specific op-amp in mind, despite what the marketing guff says.
 
Fine, let's do this..

I already mentiond the thousands of types of opamps. Pretending they are all one and the same is a red hearing

I really doubt that you end up with less components if you want to achieve the same lever of performance of an integrated circuit.

Whatever that means

1% matching is a joke to laser trimmed integrated circuits

So are opamp

So are opamps, that these cannot be mixed on an IC is nonsense.

Yes, that is mandatory for proper operation. I already mentioned laser trimming.. This is not a good argument.

As for hand building, I always find it baffling that this is an argument that something is better. Machines can do things a thousand times more precise than people. Do you want a human to align the lenses and mask of the TSMC NA EUV machine by hand to expose the wafer of your latest CPU? I think not!
It's sold on the same concept as haute couture... There, it's haute couture in sound... so welded by hand, it's like the hand sewing of a Hermès bag. ..

It's stupid marketing that does a lot of harm to high fidelity...

It should be noted that an old designer like Peter Walker from Quad used AOPs in his Quad 405 and that the circuits of the Quad 44 preamp were designed like this... in the late 1970s-early 1980s... It would be moreover interesting that ASR can measure this preamp
 
Famous French blogger Jipihorn once measured and analysed an AOP from Burson :



 
Late to the party, but since 'better' is not defined, any statement goes and has zero meaning.
Unless the meaning is 'better for the balance sheet'.
Then Burson clearly wins.

Jan
 
The jelly bean op amp can outdo the discrete one because stray R, L & C are greatly minimized due to tiny component size. For more capability, a discrete unity gain buffer can be used inside the bean's feedback loop to provide greater power output. Then you get the best of both worlds.
 
Also late, I think that implementing Burson discrete OP-Amp into a real circuit device needs also some engineering knowledge. Implementing a IC-OP-Amp need much more engineering knowledge. Not to forget that most DAC use IC-OP-Amps and have excellent measured data. So all depends on the application goals like low-noise, very high bandwidth, low-distortion, out power and impedance and so on. This makes the comparison between the two kinds discrete vs. integrated not easy and can to my opinion not be generalized. Take that design which works best for a specific application.
 
Back
Top Bottom