Quad Opamp Suggestion

[amz-fx]

I have designed a new project around a quad opamp and the TL074 was the basic model that I used. I may have a couple MC33174 chips in DIP14 that I can try too.

Does anyone have a favorite quad opamp that I should try in the circuit?

Low noise is most important and most of the sound comes from the eq and clipping diodes.

Ideas?

Thanks, Jack

[Mark Hammer]

Hi Jack,

There's always the LM837, which is a quad LM833, functionally equivalent to four NE5534s in a chip.

Does anyone make a quad version of the NE5532?

[anotherjim]

Conventional wisdom is, if there are >100k resistors around an amp, a TL07x is a good choice for cost-versus noise.
Though you say low noise, LM324 can bring a sound of it's own.

[amz-fx]

Thanks for the suggestions!

I think I have some of the LM837 and the LM324 is my chip box, so I'll try them both.

I'm not aware of a quad 5532, and some of the high-end audio-specific National chips are not available in DIP. I could always make an SMT adapter but that's too much trouble just for playing around with the design. 

Best regards, Jack

[R.G.]

My experience is that the signal flow in a layout is much easier with two duals, not one quad, because the outputs of the opamps can all point in the same direction, not at four corners like all the common quads. Two duals fit end to end in a 16-pin space, two more pins than a quad, but the two extra pins are power, and no more taxing than power to a quad, in the middle of the chip.

What you get for the extra two pins is a simpler layout, better signal flow, and as you've noticed, a much wider choice of opamps to use. Unless "and it only uses one chip" is a big part of the design, I'd recommend two duals. High rate of return for an extra 0.1".

[Mark Hammer]

I'm finishing up a 6-stage version of a P90 but I'm using SIP-form 4558s for the phase shift stages.  If you think duals are more convenient for layout, try SIP-duals!  You have your choice of placing parts on either side of the chip.

[R.G.]

I'm finishing up a 6-stage version of a P90 but I'm using SIP-form 4558s for the phase shift stages.  If you think duals are more convenient for layout, try SIP-duals!  You have your choice of placing parts on either side of the chip.

Yeah, I've looked at SIP duals before. I have philosophical disagreements with the idea that everything should be smaller, no matter how small it is, but the SIP duals in combination with stand-up resistors makes for some incredibly compact layouts. In my private bucket of PCB parts, I keep a batch of "AXEND" parts - axial resistors standing on end, with spacings of 100mil, 150, and 200. If you're not done with the P90 clone, consider putting all the resistors on one side of the SIP dual, then having another resistors-and-SIP-dual right by its side. Easy to allot a standard space on the PCB to that, and then to iterate stages.

I think I didn't do more with SIP duals because although they're not too hard to find, the supply is chancier.

[ElectricDruid]

Hi Jack,

Back in the day, I used to like the LF444 because it let me build stuff that I could run off batteries. Discovering the TL064 was a step in this direction, but the LF444 goes even further. That's about the only "favourite quad" I've ever had.

But these days, and if you're after low noise specifically, TL074 is what I always use.

I second what others have said about duals being handy for layout., although occasionally, the mirror-reflection quality of the quad layout works really nicely.

Tom

[Ice-9]

Recently I had to use the SIP-BA4560N duals in a repair to a mixer which had been badly fried. It took around 30 of these to repair and I was happy they were SIP rather than DIP as that made removal a lot easier with a lot less chance of damage to the PCB's.

OP470 might be worth a look for a DIP low noise quad.

Link here to datasheet and spice models of OP470
http://www.analog.com/en/products/amplifiers/operational-amplifiers/high-voltage-amplifiers-greaterthanequalto-12v/op470.html#product-overview

[GGBB]

TI makes it pretty easy to find what you are looking for. As an example, here are all of TI's +9V compatible PDIP op-amps with noise (Vn) less than or equal to a TL074:

http://www.ti.com/lsds/ti/amplifiers/op-amps/op-amps-products.page#p480=4;4&p1261min=0.9;9&p1261max=9;105&p2954=PDIP&p7typ=0;18

[DISCLAIMER: I work for TI on the software team the developed this tool (though I didn't work on it myself).]

[ElectricDruid]

[DISCLAIMER: I work for TI on the software team the developed this tool (though I didn't work on it myself).]

Wait for all the feature requests any moment now...

T.

[Passaloutre]

Please pardon my novice question, as it's only tangentially related to the thread topic.  I'm fairly new to both JFETs and opamps.

It seems that JFETs are often used in circuits for their "tube-like" response. I also often see it mentioned that the TL074  has a JFET input stage. You can probably see where I'm going with this. Would a TL074 lend a more "tube-like" sound than other opamps? For a particular comparison, how about the MC33174 I just put in my Echo-Matic?

[EBK]

Please pardon my novice question, as it's only tangentially related to the thread topic.  I'm fairly new to both JFETs and opamps.

It seems that JFETs are often used in circuits for their "tube-like" response. I also often see it mentioned that the TL074  has a JFET input stage. You can probably see where I'm going with this. Would a TL074 lend a more "tube-like" sound than other opamps? For a particular comparison, how about the MC33174 I just put in my Echo-Matic?

No, JFET op amps are essentially quite linear by design.

[anotherjim]

I don't believe there's any tube sound in any op-amp. That hasn't stopped many claiming that if there is a J-fet or CMOS op-amp involved. Thanks to advertising hype, many non-technical folk also believe it. But an op-amp has huge open-loop gain and with negative feedback, it remains extremely linear unless either the inputs are overdriven or the output clips. There is no soft-clip stage before the limits.

[Rixen]

I've spent many hours pouring over op-amp data sheets, looking at noise figure, bias currents, supply current, operating voltage, package availability, IO range etc. As well as availability and cost.

Hard to beat the TL07x family for battery operation, when all other factors are considered.

I started listing them all in a spreadsheet for comparison, will try to remember to post this.

[EBK]

I use op amps because they are predictable and magical at the same time. I have the knowledge to understand op amps deeper (in fact, I'm positive that I was required to do so at some point in my education), but I can reduce my attitude toward different op amps to this: TL07x is a "good" op amp.  More expensive ones might be "better" if I ever determine that I need something better (while also recognizing that rail-to-rail and auto-biasing op amps exist), and I will blindly trust anyone who explicitly recommends a specific op amp for whatever I happen to be building. 

[ElectricDruid]

I think they're magical because they're predictable! That alone is enough!

BJTs vary significantly. FETs are even worse. Germanium trannys have to be hand-picked.

Op-amps...you can take even a 741 and a pile of 5% resistors and it still does what you tell it every single time. Now that is magic.

Personally, I don't want a super-special circuit that only I have. I want to design something that everyone can bang together with the bits they have at hand *and it still works and sounds great*! That's where the cool stuff is for me.

Tom

[Rixen]

I should probably set this up in Google docs to allow collaboration/completion. Basically anything in a red cell indicates a parameter that can be regarded as 'worse' than a TL07x (used as a benchmark) and where attention may be needed (interesting that the NJM4558 as used in Tube Screamers is quite borderline for use on a 9V battery regarding operating voltage)

[PRR]

> NJM4558 as used in Tube Screamers is quite borderline for use on a 9V battery

I believe it "degrades gracefully" down to 4V. It has an internal 7V regulator to set all the currents, so needs 8V to "meet specs". Below 8V, currents are lower than designed, but everything still works. Somewhere like 4V performance is way off, or stages starve. Bad for precision measurement systems. Maybe very tolerable in something called "screamer". It may scream a little lamer, or a little harsher; may not matter.

The numbers for '741 should be taken with salt. I remember when it was "the 40V process", but for many decades most chips on that process have carried 36V "Absolute Max" ratings. I suspect that when they sold thousands of chips (days before '741), used for a few years, a 1 in a million failure looked small. But '741 and children sold millions and millions, in jobs running for decades. After they had a fat handful of dead chips returned, they notched the Ab Max rating down to keep us out of trouble.

[Rixen]

link to Google docs version, open to view, but not edit (at the moment):

https://docs.google.com/spreadsheets/d/1VjHWtAaqLJBZi7CKDw5SlzluXgqv__cc42lV0K1jMvQ/edit?usp=sharing