Op Amp Upgrades - Meta

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

dandeurloo

Well-known member
White Market Member
Joined
Jan 20, 2008
Messages
1,136
Location
Minneapolis, MN
I guess I am wondering why we don't have some sort of a chart that has common upgrades to newer chips.  Kind of a starting point of what will work.  Is this a good idea or a bad idea?

So it would have common upgrades for chips like tl074, ne5534 and so on. Thoughts?
 
bad idea... explain you thoughts. 

I thought it would be cool so people could have a reference of possibilities to experiment with in different projects.  Obviously, it wouldn't be 100% that chip upgrades will sound great in each application but could be a good resource.

The other reason it could be good is it would allow guys like myself to test different chips in the same circuits.  That way i would get a better understanding of the chips and their own sound.  positive and negative.

Just kicking the idea around. 
 
dandeurloo said:
bad idea... explain you thoughts. 

To say: opamp "A" is an upgrade to opamp "B" is just a too broad statement since it all depends on the circuit which was designed to work with opamp "B".

 
radiance said:
dandeurloo said:
bad idea... explain you thoughts. 

To say: opamp "A" is an upgrade to opamp "B" is just a too broad statement since it all depends on the circuit which was designed to work with opamp "B".

agreed.  you can swap them but the end result may cause all kinds of sonic issues. Best to just ask on a case by case/project by project
 
dandeurloo said:
The other reason it could be good is it would allow guys like myself to test different chips in the same circuits.  That way i would get a better understanding of the chips and their own sound.  positive and negative.

Hey Dan,

I'm with Radiance on this one, for a couple of reasons

a) We don't need to perpetuate the myth that drop in chip upgrades are a panacea.  If you don't know what you are doing and why you may well cause more trouble with instability and oscillation.

b) It will attract a crowd of "what chip to make my behringer sound rad" folks.  

c) The most important one - Every circuit is different and has it's own unique opportunities and pitfalls.  I think it would be far better if we concentrate on understanding why a certain chip works in a certain circuit and why another may be better.  I've been studying this pretty intently lately for an eq project I'm working on and questions like why use a fet or bipolar opamp, what are offsets and when and where do they matter, compensation issues, drive capability and then the least important stuff like distortion specs/noise specs  :) etc etc etc.  

Sorry if it seems like I'm raining on your parade, it's not my intention at all.  I have a thread coming in the next week that will highlight some issues around opamp choices for an EQ but I've spent many weeks studying to understand the issues before I post it.  Hopefully it will help highlight some of the typical questions I listed above.

Cheers,
Ruairi






 
Selecting opamps is an exercise in compromise. 

CMRR, Offsets, THD, Current drive, GBW, Slew rate, NF, Supply voltage, package, cost - it's a balancing act, where some parameters are more important than others for a particular application.

No single opamp is perfect, except the ideal textbook opamp, which is 100% perfect in performance and costs $0.00
 
Just to be contrary, it might be nice to have a good general reference listing all the better modern opamps in a spread sheet format with all the pertinent specifications so comparisons could be made. I sure haven't kept track of them all.

As a circuit designer, I don't appreciate the swap out opamp du jour mentality, that believes just using newer parts will make dramatic improvements in a circuits performance. There is a little more to design than buying parts. The short target list offered of TL07x and NE553x are relatively fast (10V/usec) and relatively quiet, (3 uV or .5uV), so those tired old parts, were once considered new and wonderful (showing my age), as we happily swapped them in for 741s, 301, 709, whatever... decades ago.

These days there is less low hanging benefit to pick up, while there can be some marginal improvement in difficult circuit applications.  Looking at the seriously better new parts, you would just about need to start from scratch to get the full benefit of their uber low noise and strong drive capability. Dropping them into an old socket, without changing the circuit values around them, wouldn't do them full justice. 

JR
 
JohnRoberts said:
As a circuit designer, I don't appreciate the swap out opamp du jour mentality, that believes just using newer parts will make dramatic improvements in a circuits performance.

30 miles into my bike ride last weekend, I spotted a mutual friend of ours out for a walk, so I paused for a chat. -I'm reminded of his signature of old:

[quote author=Bob Katz]
There are two kinds of fools in this world:  One who says "This is old and therefore good", and another who says "This is new and therefore better!"
[/quote]
 
As someone who knows next to nothing of "swapping chips", is it safe to say that you could say things like x chip needs n input impedance and if you want to use y chip, you need to change your resistors on the legs of the chip to be blah blah blah?

In other words, even though chips aren't really "drag and drop" is there a standard method that could be used to calculate what you'd want to do to switch from x to y?
 
I don't want to make this sound like speaking latin in the church so common people can't understand.

Taken very simply, you can generally replace a JFET input opamp, with a newer, better JFET opamp and not make anything worse, but it might not make things much better. Likewise, a bipolar opamp should be easily updated with a newer bipolar opamp. The general consideration is input bias current. FET input opamps have less bias current, but modern bipolar opamps are much better at canceling out this bias current, so some advanced modern bipolars might work in a Bifet or JFET input socket (or not).  

Opamps are general purpose gain blocks with huge open loop gain. The feedback network applied around an opamp trades a bunch of that open loop gain to deliver a more precise closed loop gain as defined by that feedback network. Beyond a point, adding more open loop gain, has a diminishing effect on linearity. Likewise, wrt opamp speed, if the opamp is already faster than the signal being passed, substituting a faster opamp will not make a difference, because the audio signal isn't trying to move that fast.

Even these comments are a gross simplification, and better is usually better. My suggestion about needing to redesign feedback values was for the specific case of some modern ultra low noise opamps. When the opamp noise gets that low, the thermal noise of the feedback network now becomes a limiting factor. So to get the full S/N capability of the new opamp, one needs to substitute smaller value resistors that make less self noise. But now everything draws more current, so a marginal layout for stability may no longer be stable and power supplies could run out of current.

I don't want to make a blanket dismissal of all component upgrades, and products designed before what I consider truly audio capable opamps were available, can surely benefit, but perhaps ironically the 553x and TL07x listed as targets to swap out are what I consider the newer generation of audio capable opamps, with speed well in excess of what is needed for audio and decent noise performance. The older opamps than those were truly marginal and more difficult to apply for good results.  

For difficult applications (like high gain), or where the original execution was marginal.. (like using a TL07x for an output stage), there may be some room for improvement. Otherwise when properly applied the 553x and TL07x should not be the limiting factor for signal integrity.

So I'm sorry, there is no simple set of guidelines. The opamp is a general purpose part, that can be made to perform a very large universe of very different applications, so any simple rules would have to be limited to very specific applications.

IMO many such swaps are a waste of time and money, but that means not all are.  So YMMV.      

JR
 
I have a less detailed and thoughtful answer - something from my own experience, chip rolling is fun, so nobody can stop you from doing it, advice or otherwise. But the resulting sonics vary indeed - some places I swapped the chips, it was great, the sound went from crap to good. Then there were some devices where swapping chips didn't improve the sound (to my surprise) and after much listening I actually put the original chips back inside...
YMMV considerably.
 
There are some cases where chip-swapping has a justification, but only after you have done your homework.
You must know why you can't drop a 1nV/sqrtHz in place of a TL0, you must know why you can't replace a low-noise bipolar by a JFET in a transforme-input mic pre.
But after having done the initial selection, you may be left with several good-looking candidates and then only experimentation will help you make your decision.
I particularly remember having spent numerous hours experimenting with the opamps used in the so-called "gyrators" in graphic EQ's, where my objective selection suggested the RC4559, JRC2043, JRC2068 and MC33078. I ended up spec'ing the 4559, for a number of reasons, among which availability, cost and PSRR were dominant.
 
Good stuff here guys.

Another reason for the meta could be for novices/noob like me who are trying to design a new small circuit. It would be nice to know what op amps could be prototyped in the circuit.  You know basically build the circuit and it supporting case on Sockets so one could easy change a few values along with the resistors and caps around it to see which one might sound the best in the design.  I was really just thinking of it as useful for modding older chips with modern ones all though that could be handy as well.

 
I would like to say at this point that JR's post was brilliant. :eek:

PS. The Crown PSA-2 is doing just fine with the upgrades and renovation..
 

Latest posts

Back
Top