Date: Sun, 29 Aug 1999 19:17:48 -0700 From: SuperBad MoFo Subject: Re: OP-Amps (long) At 4:33 PM 8/28/95, Anders Wikholm wrote: >Which OP-amps do you guys recommend? I intend to use them in >amps before and after an effect, so I don't want the >nice 741 distortion... I would recommend an LF411 (single) or LF412 (dual) op-amp for input/output buffering - they'll probably handle the capacitive loads of the cables as well as anything. I also use TL-071/TL-072/TL-074s (single/dual/quad) for summing amps & other internal functions...usually it just depends on which I can get cheaper - the 411s or 071s...there are better op-amps, but I think these have the best price/performance ratio. If you're replacing op-amps in an existing circuit w/ more modern op-amps, you probably want to cut the compensation cap (between pins 1 & 8 on singles) out of the circuit. Also, if you're dealing w/ an oscillator, you may wish to install a socket and listen to the replacement first - it may not oscillate over the same range (or not at all), or it may alter the sound too much. Just make sure you replace a single w/ a single (those duals in 8-pin packages can be confusing). I've also included some collected wisdom from AH on the subject of op-amps: >Date: Wed, 26 Oct 1994 14:02:29 -0700 >Subject: Re: op amps >From: Don Tillman > > From: Michael Trnka > Date: Wed, 26 Oct 1994 09:44:02 -0700 (PDT) > > Can anyone out there suggest a replacement for a 709 op amp? is used > commonly throughout " electronic music circuit guidebook" but seems to have > fallen into disuse in the last 21 years or so.... 741s are also called for a > lot and these can still be found but I here there were some replacements > suggested here in the last few weeks. Could someone please rehash for me? > >The 709 was the very first IC opamp produced in serious quantities. >It was truly horrible in every way, so substituting virtually any >opamp will give you several orders of magnitude better performance. > >The 709 required some external parts for "compensation" to keep >it stable at lower closed loop gains. The 741 was the improved model >that didn't need the compensation circuitry. It too is truly >horrible and should be avoided. > >I don't know the book, but I think it's safe to say that any reference >to a 709 or a 741 should be read as a generic op amp reference. For >replacement in an audio path use a low-noise low-distortion opamp >(ie., one that's actually intended for audio work). For control >paths, use one with low offset voltage. > > -- Don >Date: Wed, 26 Oct 1994 16:59:59 -0500 >From: EH%Ranger.dnet.dec.com@thais.lkg.dec.com (Eirikur Hallgrimsson) >Subject: Re: op amps > >At 4:30 PM 10/26/94, wrote: >> Can anyone out there suggest a replacement for a 709 op amp? > >A 748 replaces a 709 for all instances that I have seen in emusic circuits, >and actually, I think I've seen a more common part, the LM381 (38 >something, anyway). > >You should be able to replace the 709 with any pin-compatible >"uncompensated" op-amp. The 709 has a faster rise time due to having the >compensation capacitor of the 741 left off of the chip. My understanding >is that the 748/LM381 are similarly uncompensated. This is dated knowledge >from building modulars in the mid-seventies. Let's see if there are any >more recent votes. > >Eirikur Date: Wed, 26 Oct 1994 17:00:04 -0500 From: EH%Ranger.dnet.dec.com@thais.lkg.dec.com (Eirikur Hallgrimsson) Subject: Re: op amps O,btw, an uncompensated opamp, with tailored compensation as a part of the surrounding circuit design, will SOUND BETTER. This is due to the slew-rate-limiting imposed by the compensation (which is basically there to avoid feedback in the supersonic and low RF regions). Just think of all those 741's out there in analog gear. Handling 5 to 10 volt signals. Using compensated opamps like the 741 for line-level audio was debunked years ago. Sorry, I can't recall the name of the author. If line level signals are too much for these chips, just imagine the slewing problems when you multiply the signal level by 100 or 1000. Hmmm, early analog gear tends to sound good. It's the later stuff (post development of cheap opamps) that tends to be less satisifying.... >Date: Wed, 26 Oct 1994 23:46:22 -0700 >Subject: Re: op amps >From: Don Tillman > > Date: Wed, 26 Oct 1994 16:59:59 -0500 > From: EH%Ranger.dnet.dec.com@thais.lkg.dec.com (Eirikur Hallgrimsson) > >Some friendly disagreement here guy. > > A 748 replaces a 709 for all instances that I have seen in emusic circuits, > and actually, I think I've seen a more common part, the LM381 (38 > something, anyway). > >I disagree; I think the circuit context is real important for any kind >of performance. When an old schematic said "709" it really meant "the >only reasonably priced opamp in existance right now -- the fact that >an amplifier on a chip can do this at all is real impressive". > > You should be able to replace the 709 with any pin-compatible > "uncompensated" op-amp. The 709 has a faster rise time due to having the > compensation capacitor of the 741 left off of the chip. > >Actually the 709 was around way before the 741. (That's sort of like >saying the Telecaster was created for folks who didn't want a wang bar >on their Strat.) Actually getting the compensation on chip was the >741's claim to fame. (Just like the wang bar... hey, that's a good >analogy.) > > O,btw, an uncompensated opamp, with tailored compensation as a part of the > surrounding circuit design, will SOUND BETTER. This is due to the > slew-rate-limiting imposed by the compensation (which is basically there to > avoid feedback in the supersonic and low RF regions). > >No! An uncompensated opamp will only sound better under certain >conditions and only if all other things are equal (and they almost >never are). A 709 is far slower than any compensated amp introduced >over the past ten years (except for a couple optimized for ultra low >power consumption). > > Just think of all those 741's out there in analog gear. Handling 5 to 10 > volt signals. Using compensated opamps like the 741 for line-level audio > was debunked years ago. Sorry, I can't recall the name of the author. If > line level signals are too much for these chips, just imagine the slewing > problems when you multiply the signal level by 100 or 1000. > >Certainly true. > > Hmmm, early analog gear tends to sound good. It's the later stuff (post > development of cheap opamps) that tends to be less satisifying.... > >Nope, most of the early analog gear sounds good either because it uses >no opamps at all in the signal path, or possibly in spite of the fact >that there's a 741 there. For instance, the Moog Modular has only a >few opamps in the audio paths of only a few modules, and only in >places where it doesn't make much difference. And the ARP Odyssey has >no opamps in the audio path (with a possible exception of the potted >VCF module). > > -- Don >From: "Haible J. ZFE BT SE 42" >Subject: AW: Re: op amps >Date: Thu, 27 Oct 94 11:00:00 PDT > >>slew-rate-limiting >> [...] >>Just think of all those 741's out there in analog gear. Handling 5 to 10 >>volt signals. Using compensated opamps like the 741 for line-level audio >>was debunked years ago. Sorry, I can't recall the name of the author. If >>line level signals are too much for these chips, just imagine the slewing >>problems when you multiply the signal level by 100 or 1000. >> >>Hmmm, early analog gear tends to sound good. It's the later stuff (post >>development of cheap opamps) that tends to be less satisifying.... >> > >This is a very interesting point IMO. I haven't thought about it very much >before, but let me think about it: >Take the famous SEM-filter. (I hear them cry: Not again! - but nevertheless >:->) >It surely handles signals in the range of some volts. >All its outputs are buffered with 741's. >The slew rate of these 741's will surely affect signals with short rise >times, >such as pulse and saw waves commonly used in synthesizers. >The question is: *how* does this affect the sound? >First throw: slowing down edges means reducing higher harmonics. > ( => Is this the reason for the SEM's "smooth" sound ??? ) >Second throw: Slew limiting in a closed feedback loop (very strong >feedback in case of a unity gain buffer!) will cause saturation effects >in stages previous to the slew-limiting stage. This is a source of >further distortion. ( => "smoothing" assumption above is rather >questionable! ) >All in all it seems to be rather clear that 741's change the sound. But I am >not sure whether these effects are part of the "famous vintage sound" or >not. >In my currently devellopped SEM type filter module I substituted the >FET + 741 buffers with TL074 (FET input, and much higher slew rate) opamps. >Do I have to rethink this? Will it ruin the vintage SEM sound? >(I didn't want to make the sound "brighter", "less muffled" or something >like this. I just wanted the original SEM sound with cheap today's >components ... ) >Any opinions? Don? Anybody else? > >JH. >Date: Thu, 27 Oct 1994 10:36:51 -0700 >Subject: Re: AW: Re: op amps >From: Don Tillman > > From: "Haible J. ZFE BT SE 42" > Date: Thu, 27 Oct 94 11:00:00 PDT > > Take the famous SEM-filter. > It surely handles signals in the range of some volts. > All its outputs are buffered with 741's. > The slew rate of these 741's will surely affect signals with short rise > times, > such as pulse and saw waves commonly used in synthesizers. > The question is: *how* does this affect the sound? > >There's a lot going on inside an opamp; the signal makes it's way >through a series of 7 or so transistors, each adding some >nonlinearity, there's the compensation, so the nonlinearities change >with frequency, and there's the feedback loop which changes the >timbre of all the nonlinearities. Boiling all that down to one spec, >like say a slew rate, is missing the point in a big way. > >The slew rate is a worst-case kind of TIM distortion in much the same >way that clipping is a worst-case kind of nonlinearity. So just >because the stage isn't slewing doesn't mean that it's not having >a major effect on the sound. > > All in all it seems to be rather clear that 741's change the sound. > But I am not sure whether these effects are part of the "famous > vintage sound" or not. > >I don't think so. I've played instruments though mixers that used >741s and through effects boxes where in bypass mode the signal still >went though a 741 buffer stage, and each time the effect on the sound >was blatent to me; the sound becomes thin, brittle, lifeless, no >balls, like all the soul has been sucked out. It's enough to convince >me that 741s should be banished from the face of the earth forever. > > In my currently devellopped SEM type filter module I substituted the > FET + 741 buffers with TL074 (FET input, and much higher slew rate) opamps. > Do I have to rethink this? Will it ruin the vintage SEM sound? > >I think the TL074 would be a noticeable improvement. You could always >socket the thing and spend a rainy day swapping chips and taking >notes. > > >Nope, most of the early analog gear sounds good either because it uses > >no opamps at all in the signal path, or possibly in spite of the fact > >that there's a 741 there. > > I tend to have the same opinion, but I start questioning in now. > I just hated 741's. (BTW, I also used to hate discrete transistor amps, but > I've corrected myself in this!) I always used TL07x series, being far > superior in slew rate etc. > >Note that a 741 is a specific device that can only be used in a >specific way (ie., with tons of feedback). "Discrete transistors", on >the other hand, can mean any number of a dozen different types of >devices in literally millions of different circuit configurations. >Just as discrete circuitry allows one to craft a really pretty >sounding machine, it also allows one to build a cheezy fuzzbox. > > -- Don >From: "Haible J. ZFE BT SE 42" >Subject: op amps again >Date: Thu, 27 Oct 94 11:51:00 PDT > >> Hmmm, early analog gear tends to sound good. It's the later stuff (post >> development of cheap opamps) that tends to be less satisifying.... >> >>Nope, most of the early analog gear sounds good either because it uses >>no opamps at all in the signal path, or possibly in spite of the fact >>that there's a 741 there. > >(Sorry for this follow-up, but I haven't seen this one before I sent out >my two cents about this thread) > >I tend to have the same opinion, but I start questioning in now. >I just hated 741's. (BTW, I also used to hate discrete transistor amps, but >I've corrected myself in this!) I always used TL07x series, being far >superior in slew rate etc. >But the assumption we always make is: >They didn't have anything better than xxx then. If they were to design it >today, >they surely would use yyy. This is absolutely true, IMO. >*BUT* over the years we all learned to love the sound of devices that use >xxx. Now we assume, if we rebuild it today and use yyy, we would love >the sound even more, automaticly. And this assuption is definetly *false*. >It may work in 90%, especially when things like noise or hum are involved. >But with distortion (any kind of distortion ...) it wouldn't work. >Just set "Tubes" for "xxx" and transistors for "yyy". Nobody originally >wanted >an electric guitar to be distorted by an amp; they just wanted to make it >louder ... (well, the example is over-used, I know.) You have given another >example Yourself, setting "discrete transistors" for "xxx" and opamps for >"yyy" >when You talked about the Moog. I highly agree in this today; see above. >I am not sure at all what the ultimate truth is for "741" vs. "high slew >rate". >But I am quite sure that it's not as easy as > "Higher slew rate => Better performance in a synthesizer". >We're not doing HiFi here - what we are into is *creating* sound. > >JH. >From: "Haible J. ZFE BT SE 42" >Subject: AW: Re: AW: Re: op amps >Date: Fri, 28 Oct 94 09:56:00 PDT > >> All in all it seems to be rather clear that 741's change the sound. >> But I am not sure whether these effects are part of the "famous >> vintage sound" or not. >> >> [...] >> >> In my currently devellopped SEM type filter module I substituted the >> FET + 741 buffers with TL074 (FET input, and much higher slew rate) >opamps. >> Do I have to rethink this? Will it ruin the vintage SEM sound? >> > >>I don't think so. I've played instruments though mixers that used >>741s and through effects boxes where in bypass mode the signal still >>went though a 741 buffer stage, and each time the effect on the sound >>was blatent to me; the sound becomes thin, brittle, lifeless, no >>balls, like all the soul has been sucked out. It's enough to convince >>me that 741s should be banished from the face of the earth forever. > >Hmm, I hoped to make it clear that speaking of mixers and everything >that is intended for "High Fidelity", i. e. *reproducing* complex signals, I > >would banish the 741 even more than You. >My question was about this 741 in circuits that *generate* sound, handling >not-so-complex waveforms like saws and pulses. >I don't think at all that a SEM sounds "brittle, lifeless, no balls, like >all the >soul has been sucked out" - I don't believe You have this opinion either. > >>I think the TL074 would be a noticeable improvement. You could always >>socket the thing and spend a rainy day swapping chips and taking >>notes. > >Yes, this would have been the first thing I would try. Sadly, I have omitted >the FETs, and the circuit would have great offset problems with 741-type >opamps .... > >JH. >Date: Thu, 27 Oct 1994 10:13 EST >From: MARSHALLR@opsusa.sms.siemens.com (RICHARD MARSHALL--ASTON, PA) >Subject: re: op amp suggestion > >Try and get ahold of the Analog Devices OP-275 (dual) and/or OP-176 (single) >op amps. They are fast, quiet, and sound great, (this is the series used by >Tim Smith of Weyer/Smith labs when they POOG old gear). > >RM >From: mlloyd@ihlpm.att.com (Michael Lloyd +1 708 713 5497) >Date: 27 Oct 94 15:14:00 GMT >Subject: Re: Op Amps, LM741, etc. > >Just to offer a slightly different opinion, I'd like to mention an >internally compensated opamp that I use that works and sounds just fine. > >For folks looking to not use 741's anymore, I'd like to recommend >the Analog Devices AD711 (single) and the AD712 (dual). It has >pretty good specs for some key parameters for analog gear design. > >Here are typical numbers for the lowest grade (ie. cheapest grade): > >Slew Rate: 20V/uS >Settling Time: 1uS to 0.01% (Note: Fast slewing with long settling time > is not very useful! This was a problem with > older fast devices) >Unity Gain Bandwidth: 4MHz >Open Loop Gain: 400V/mV >Offset Voltage: 0.3mV > >It has served me well in a variety of home project applications (audio, >gain stages, filters, oscillators, power supplies, instrumentation, etc.), >offers a good performance balance between precision, speed and >cost and is very easy to use. I don't know how much they cost now, but I >bought a bunch (~100) from Analog Devices back in 1988 (I think) for >somewhere in the $60.00 neighborhood. I would guess by now the >price is much lower, but who knows what price distributors ask for 'em. > > > >Michael >mlloyd1@mcs.com > >Disclaimer: This message has nothing to do with my employer, I have no >stock in Analog Devices, these are my opinions from research at home. >Date: Thu, 13 Jul 1995 16:29:37 +0059 (EDT) >From: "Craig O'Donnell" >Subject: Re: analogue V1 #369 > >Anything ? If it has 8 pins, is an op amp: it will work. they are all, >or mostly all pin for pin compatible with 741s. It probably even says so >on the packaging, or go look at that big book down at the electronics >store. >>> > >Well, a TL062 / TL072 / TL082 will be a pin-for-pin replacement as will >the NE5532. The 5532 is probably the best of this lot. Not ALL 8-pin op >amps have the same pinout as the 741!!! I think the 062 is a low-power >(ie, lower voltage) 082, and the 072 is a low-noise 082 more or less on >par with the 5532. > >Radio Shack has the 082. >From: Haible_Juergen#Tel2743 >Subject: AW: Re: opamps and HF >Date: Mon, 17 Jul 95 09:30:00 PDT > >>I havn't really thought about that but that could be correct. However due >>to the Hi-impedance nature of the JFET input stages of these op-amps, they >>are more succeptable to stray pickup of all kinds. Hi-Z plus hi gain and >>you're asking for trouble. I once built a touch switch that was based on > >No. It just depends on You, the designer, how high- or low-impedance >Your inputs are. Just choose the resistor You need. Even with a standard >741, the external input resistor should be way smaller than the input >impedance >of the device. You *can* use higher resistors with FET opamps (with all the >need >of taking care for stray effects, noise etc.), but nobody forces You to do >this. > >>However. The best op-amps for audio circuits I've found are the afore >>mentioned LM833s from National. (If you need dual) or NE5534s if you need >>a single. > >But they would do more AM demodulation than a FET opamp ... (If this theory >is true after all, but as I said, I'm not sure about it) > >>If you need a quad then there's really only one choice the >>LF347/TL084 (they're virtually the same thing) > >I don't know about the LF347, but the LF35x opamps have a great advantage >over the TL0xx: They can handle larger capacitive loads. (Important for >driving >shielded cables!) > >JH. >From: Ryckebusch@aol.com >Date: Wed, 9 Aug 1995 21:37:56 -0400 >Subject: Re: opamp replacement > >The absolute best audio opamp is the AD-797, if you can find them. All the >suppliers have a 8 week lead time and frown when you only want a few. They >have an ungodly 9Part Per Million THD spec and will sink or source >50Milli-amps. I am currently building a 32 input board that uses them as >summing amps. The OP-275 from Analog Devices is also fantastic. It is pin >compatible with NE5532 etc. the AD-797 is pin compatible mwith 741 etc. > >Jules Ryckebusch Date: Thu, 10 Aug 1995 10:50:12 -0700 From: ftom@netcom.com (Tom May) Subject: Re: Op Amp upgrade chips shockad@primenet.com sez: >The most important factor in deciding a replacement op amp spec-wise is the >SLEW-RATE. Look in any cross reference guide for the manufacturer of the >chip you decide on-they almost always have a section with pin out info and >some specs. It depends on the application. For an op-amp in the signal path, yes, a certain minimum slew rate is necessary. But for a S+H circuit it is most important to have a high input impedance. For a DC voltage follower a low input offset voltage is a good thing. Many different kinds of op-amps are made because of the variety of applications. (Also, for an externally compensated op-amp the slew rate will be determined largely by the external compensation capacitor and not so much by the op-amp itself.) And Jules Ryckebusch (Ryckebusch@aol.com) sez: >The absolute best audio opamp is the AD-797, if you can find them. All the >suppliers have a 8 week lead time and frown when you only want a few. They >have an ungodly 9Part Per Million THD spec and will sink or source >50Milli-amps. Beware that people in the audio and electronics newsgroups are always mentioning that THD is an over-hyped spec (at least in power amplifiers), that there are other distortions and undesirable effects caused by the fact that your signal is going through so many non-linear components, and that the negative feedback loop you have to wrap around an op-amp is a bad topology. The theory and all looks really great, but remember that real op-amps are not ideal. For audio, how it sounds is more important than specs (although you can narrow down your selection of parts that are likely to do the job based on the specs). That is especially true for sound generation as opposed to reproduction. As shockad goes on to say, and I agree: >I should say that it is possible to over do >741 replacement in >certain components-they can add some of that elusive charecter. Tom.