JRC4558D Again.... So Where's The Magic Hiding?

[phaeton]

Seriousry....

I know we often laugh at the Funny Mojo Stuff that some people get all wound up about.  The infamous "JRC" is one of the hot topics, we all know this.  We all also know that it's not manufactured using rare elements mined from Titan or blessed by some $diety, or that it is even rare anymore (considering the NJR reissue and all).  Ibanez probably just chose to use it because it was a cheap and available dual op-amp that fit their noise/power/gain/production needs. Right?

So I go about expanding upon one of my old ideas.  I posted it here under the name "boiler", but I can't find the topic.  Doesn't matter, since it was fraught with errors.  Sorry, I don't have a schematic handy, but nonetheless, imagine a working device built around a dual op-amp. One op-amp has a gain variable from 0-10 (acts as a drive control) and it feeds the other op-amp that has the gain hard-wired at about 800 or so.  No real filtering or tone shaping- just tons of str8 up raw distortion, good bad and ugly all at the same time.  You'd think that swapping out various dual opamps in this circuit would illustrate their differences, right?  No tone-shaping stuff to 'taint' the results.  It doesn't sound as bad as you think.  In fact, I rather like it, for such a simple circuit.  However.....

Either I'm deaf or being too academic, but I can't really tell much difference between the Holy Grail JRC and a garden-variety TL082.  The JRC just has a little less gain and a little less noise, but that was sort of expected ('82s are typically regarded as decent but a tad noisy, right?).  I'm not just talking balls-to-the-wall distortion, either- backing the gain down from full tilt to 'medium' to 'just a little breakup' to 'clean' is all pretty much the same between the two chips.  My JRCs are the Real Deal, from Steve, with love.

So what gives?  Does the circuit need some sort of filtering for the differences in dual op-amps to become noticeable?

[KerryF]

Well I played around with Dual Op Amps in one of my BluesBreakers.  I had a stock TL072.  I tried 4558s, 5532s, TL072s again, etc...  I tried stacking and all.  I think I finally settled with a stack of 3 NE5532s.  It sounded a tiny bit more bluesy and light.  But the difference was so slight that it was barely noticable.

[R.G.]

Seriously -

There is not a well-know or even well researched and little known explanation.

As near as I can guess a good rationale goes as follows:
Early opamps were pretty dreadful as far as what happened if the inputs were not kept within the common mode range, or if the thing was driven too hard into clipping at the power supplies. They could invert the output and latch up, never to return until the power supply was turned off, and they could also oscillate at "edges" and have funny non-linear slewing on the front and back sides of fast-moving waveforms.

The Raytheon RC4558 was an early opamp intended to get better results on these "edge" conditions. It spawned a host of similar devices from other makers, including the JRC4558 from NJR. These opamps were much more orderly in how they responded to bad input conditions and overloads.

More modern opamps are even better about this.

In a diode-feedback clipping stage, the funny edge conditions never occur. The inputs are never driven out of the common mode range, and the outputs are not banged into the power supplies to recover from. But to my ear, and to the ears of others I've suggested this to, the more orderly and predictably that an opamp recovers from overload conditions, the smoother and less harsh it sounds in a diode clipper. I don't know why that is. It's just a correlation.

There is a huge temptation to a guy who's trying to make a living to say that there is some magic, otherwise unobtainable part that only he possesses that can make you sound good. Lots of people give in to that. Probably a lot of them believe it themselves. But you're right - the differences are small.

Try some all CMOS opamps and some OP275's. I think you'll like them too.

[phaeton]

Ahh..... so it's not so much that the RC4558 is so *magical* by today's standards, but compared to all the then-current opamps of the late 1970s it was.  I.E., if I go and pull this stunt with a pair of 741s i'll likely make myself ill. And thus the mythology starts, perpetuated throughout the ages, even if hoards and throngs and scads of op-amps since that one have incorporated the same or better stability improvements.

(And I'll try it with some 741s, just because....)

And thus, this is more or less the 'defining' quality of most modern op-amps?  IOW, they do all sound alike?

[Phorhas]

It also depends on the application. Not all of these modern OAs have the same characteristics at differant applications. If you build an input stage with eanough gain to distort  - using a 2M2 bias resistor, it's going to sound very differant when using a hot-HB equiped guitar  - if you swap differant chips. A 5532 has an input impedance of abhout 300k, when a JFET input device such as aTL072 or RC4558 has enough impedance to igrore for now.
So - using this hot guitar the 5523 will sound somewhat "muddy" or "middy" or less edgy, while the JFET input device may have better upper frequency response. It's a matter of semantics as well... the 5532 may also be named "warm" or "round" when the RC4558 might be called "grindy"

If you'll build aline driver  - tried to deliver a signal at 600ohm, you'll have better results with the 5523, as it is designed for this.

Well, you get the picture - it's also in the application

[brett]

Hi

A 5532 has an input impedance of abhout 300k

Actually, according to the datasheet that I read, the 5332 has a minimum input impedance of 30k and a "typical" value of 100k.
The tubescreamer uses an input buffer, so the op-amp resistance isn't an issue.  The main thing that I notice with a 5332 is the quietness.
In other circuits, such as the Rat, there's no buffer and the NE5334 causes the loss of some top-end due to the low impedance.  Some people like this (I think Jack Orman has suggested an 308 to 5334 change for the Rat)
cheers

[Phorhas]

Well, the Fairchild sheet sais that typical input resistance is 300K but whether it's 300k or 100k - the point still stands: the "sound" of a device derives from the relation between it's characteristics and the user's application of that specific device.

[Elektrojänis]

Actually, according to the datasheet that I read, the 5332 has a minimum input impedance of 30k and a "typical" value of 100k.
The tubescreamer uses an input buffer, so the op-amp resistance isn't an issue.

Hmmm... The number probably depends on whose 5532 data sheets you are looking. ON-semiconductors has minimum 30k and typical 300k. Or are you actually referring on something else than 5532 (it seems that every post on this thread has a different part number, but I quess everyone means 5532 as that is quite common part).

But anyway... Actually, even if you discard the input buffer, you can't just take the input impedance for an opamp from the datasheet because the number in the datasheet is usually the open loop input impedance. In allmost all practical audio circuits with opamps, negative feedback is used. Negative feedback pushes the input impedance up and makes the output impedance lower. Usually both figures change quite radically.

Hey... check out these articles if you haven't allready: http://sound.westhost.com/articles.htm

There is a lot of good info in there about opamps and lots of other stuff too. You might disagree on some stuff thet the guy writes, but he documents the stuff well so the readers have good basis to decide what to believe and even do some testing themselves. The viewpoint is mostly hifi, but most of that information is aplicable to effects and a lots of other stuff too.

[bancika]

As for any other mojo stuff my opinion is that most vintage components are used just because they were available/cheap at the moment.

[MartyMart]

There really isn't any "magic" though I was led to believe this about four years ago !
I quite often use 5534's/5532's and these are even used in SSL recording preamps ! how they
benefit a distortion circuit is negligable, othr than deaining the battery faster ...

Dont forget the humble 741 - I LOVE it in a Dist+

MM

[Eric H]

A 5532 has an input impedance of abhout 300k, when a JFET input device such as aTL072 or RC4558 has enough impedance to igrore for now.
So - using this hot guitar the 5523 will sound somewhat "muddy" or "middy" or less edgy, while the JFET input device may have better upper frequency response. It's a matter of semantics as well... the 5532 may also be named "warm" or "round" when the RC4558 might be called "grindy"

Your point is good, however, the 4558 is bipolar-input --not FET.

-Eric

[jonathan perez]

JRC4558D Again.... So Where's The Magic Hiding?

under the 4th leg 

[Meanderthal]

 MartyMart brings up a good point- the old 'obsolete' opamps with horrible slew rate and noise specs. It seems that when used in a distortion these tend to... jeeze, how do I put it? Add or bring out or create or let through some kind of harmonics that seem to sound... 'better'. I used an old salvaged LM307(NOT 308) in my Distortion+ build, and to my ears it sounds even(slightly) better than a 741. As I recall, I think the specs for this showed it to be a particularly crappy opamp, but the 'crap' seems to be adding something. Maybe it's the noise creating the illusion of a fuller sound(kinda like a synth patch) or maybe it has something to do with the slew rate, I really don't know. But for distortions, I get the impression that the modern clean efficient opamps by comparison sound a bit sterile...

I think the 'magic' is in the 'crap'.

Of coures, the instant you use one for a 'clean' application like a mixer or compressor the opposite would apply- using the 'good' ones will definitely sound better!

[Elektrojänis]

I think the 'magic' is in the 'crap'.

I think that is not very suprising when we are trying to create distortion... That is exactly what the designers of the chips are trying to minimize.

Magic = Crap = Distortion

Well... Seriously... Low slew rates might have something to do with it... Lower slew rate should give out less high frequencies when driven hard.

[Gus]

Lets take the TS

I have a number of TSs and have worked on a few for others.  I have three 7s chip(different) and caps changes(all the same in all three) stock other values outside unmarked except for a piece of tape marked a,b or c that day

Turn the tone full treble
Turn the distortion full

Guitar TS(a) TS(b) TS(c) amp

I leave the room and let someone try the TSs I tell them to change the TS order at least 3 times 

Untill the gain and tone is maxxed the differences can be small even maxxed the change can be small and only noted when switching between TSs in a short time span.

I believe the tone opamp section might be the IC part that makes more of the IC based difference going along with R.G.s post about the "edge".

One time we did this my friend had two old good sounding real 808s There is something with the good 808s(some 808s just  sound like good 9s) I don't think it is all the opamp.   

[phaeton]

Speaking of the 741, I pulled the same stunt with a couple of them.  I was expecting to hear some of these 'edge oscillations' that RG talked about.  Instead, I was very pleasantly surprised.  The 741 is a winner for this application, especially with a BMP type tone control in between them.


Magic == crap.   

[Mark Hammer]

First, I seem to recall something from Jack Orman a half dozen years back that suggested the differences between the JRC chip and other dual ops could easily be mimicked by means of a 1k fixed resistor in series with the diode pair.  It had something to do with the internals of the JRC.  Hopefully Jack will chime in and clarify this.  The good news, assuming this is true, is that a) the JRC fanatics aren't crazy or in need of religious conversion to see the error of their ways, and b) you don't necessarily need a JRC4558D to produce the desired subtlties that people rave about.

Second, I can't remember a point where Ibanez is reputed to have released a TS that had the JRC4558D but NOT the changes on the output resistors.  In effect, the properties of that issue confound semiconductor and passive output components.  That is not to say that the chip plays NO role, but rather that we don't seem to do as many collective experiments with changing the ouput components alone as we do with changing the chip alone.

I, for one, would be curious to hear from anyone who has taken a stock Soundtank TS-5 (or 10 or whatever it is), and changed the output components but NOT the chip.

[puretube]

I`ve got an over 30 years experience in preferring µA709 rather than the "741"...

[MartyMart]

I`ve got an over 30 years experience in preferring µA709 rather than the "741"...

I dont suppose you'd be willing to swop a few of your uA709's for something ... would you ?
( sorry, dont have the right symbol ! )

MM.

[puretube]

round metal cans, or the (later) DIPs    


btw:

( sorry, dont have the right symbol ! )

"Alt Shift" & "m" ...