Op Amps
LM101/201/301
The 101 was developed by National Semiconductors in 1967, making it one of the oldest op amps, older than the 741. "LM" stands for "Linear Monolithic." Requires an external compensation capacitor. (National numbering: 1XX = Mil spec, 3XX = consumer grade)
µA741
The 741 was introduced by Fairchild in May of 1968. µ is the Greek letter mu and it is used as an abbreviation for micro (as in µF for micro farad). "µA" stands for "micro amplifier." Due to the lack of a "µ" character on most keyboards, a "u" is often substituted when looking for the original chip online. The amplifier is very similar to the LM101, but includes a 30pF capacitor internally which proved to be a popular decision with designers (one less component to deal with adds up!). The slew rate is given as 0.5V/µs typical. While this limited the 741's acceptance in pro audio designs, a poor slew rate appears to be a good thing when purposefully overdriving the chip for a distortion effect. LM741 was National's version, and you can many other variations of XX741 out there.
Notable uses:
MXR Distortion+ (1973/4?)
MXR Phase 90 (1973/4?)
Maestro MPS-2 Mini Phase Shifter (197?)
DOD 250 Overdrive Preamp (197?)
DeArmond Square Wave (197?)
LM108/208/308
The 108 was developed in 1969. This was notably at the time for having a respectable input impedance with bipolar inputs, but it has a slew rate even lower than the 741 at0.3V/µs typical. Limited slew rate/"slow" frequency response is a often cited reason for its popularity in distortion circuits. (National numbering: 1XX = Mil spec, 3XX = consumer grade)
Notable uses:
ProCo RAT (1978)
µA748
The 748 is an early single op amp released in 1969 as an answer to the LM101. Like the 101, these require an external compensation capacitor. This was originally available in 14 pin packages, with 6 unconnected pins. These crop up in the Gretsch effects and the ARP 2600. Modern packages will be 8 pin only.
Notable uses:
Gretsch Controfuzz and Expandafuzz.
LM118/218/318
In 1971 National released the LM318 with a 50V/µs slew rate as a sort of update to the X08 devices.
LM324
The 324 came out in 1972 and was a very popular quad op amp. It uses little power, can work off of low voltage power supplies, and you can give it input voltage down to the V- rail. The last feature makes this type of op amp not interchangeable with op amps that can't allow their inputs to come close to the power rail (TL0XX, for instance cannot get close to the V- rail) if the circuit requires such inputs. The low power and low voltage specs also mean this chip has worse noise and distortion performance than others, but these factors are often of no consequence in a guitar pedal.
Notable uses:
Common quad op amp in '70s and '80s battery powered effects.
LM358
General purpose dual op amp, as far as I know the dual version of the 324, also 1972. Like the 324, this op amp accepts inputs down to V-, so a substitute needs to be able to do the same.
Notable uses:
Electro Harmonix 4600 Small Clone
MC1458
This is two 741 op amps in a single package released not long after the 741. I generally only see this in early to mid '70s units. By the later '70s, the 1458 gave way to the very popular 4558 and TL072 dual op amps. Motorola first offered this part number as the MC1458. For some unknown reason (maybe cost?) Motorola's quad 741, the MC4741, didn't seem to get much use.
Notable uses:
Very common as a pre-4558 dual op amp.
Maestro PS-1 Phase Shifter (1971?) [Note: subbing "faster" op amps known to cause HF oscillation. Stick with the slow op amps, or be prepared to add caps or otherwise mod circuit!]
S. Hawk Tonal Expander
RC4558
Raytheon developed this device as the RC4558 in 1974. This is two "fast" 741 op amps in a single package. The 4558 is considered "fast enough" for general purpose audio work. The original Raytheon datasheet only shows a minor improvement in speed with a slew rate of 0.8V/µs. A 1976 Texas Instruments datasheet for the same device number has considerably better 1.1V/µs minimum, and 1.7V/µs typical, which are both major improvements over the 741 spec.
Notable uses:
Extremely common from the mid 70s through the 90s as a standard dual op amp.
Mutron/Dan Armstrong Blue Clipper (1976)
Ibanez Tube Screamer TS-808 (1979)
TL4558
Texas Instruments version of the 4558. Seen in some TS-808s with favorable reviews.
Notabel uses:
Boss OD-1 Overdrive (1977)
Ibanez TS-808 Tube Screamer (1979)
JRC4558
Japan Radio Corp version of the RC4558. Very common and found in all kinds of gear. Vintage chips are sought after for putting inside any type of Tube Screamers as these are considered the best "sounding" ICs for the pedal (YMMV!). Slew rate is spec at 1V/µs typical, considerably lower than the '76 TI spec.
Notable uses:
Ibanez TS-808 Tube Screamer (1979)
TA75558P
Toshiba device that is infamous for being used in TS9s, and subsequent reports that these TS9s sound inferior. I've heard one or two sources claim this is some sort of high fidelity chip, but I cannot find any real evidence to back that up. The Toshiba datasheet lists audio applications, but they are not advertised as "high fidelity." The slew rate is listed as a mere 1V/µs, which is comparable to the JRC4558 spec.
Notable uses:
Ibanez TS-9 Tube Screamer (1982)
JRC2043DD
Again I'll just mention this odd ball due to its appearance in the TS9, and analogman's claim that these can sound really bad in a Tube Screamer. The only datasheet I've found is from 2003, so it may not be vintage accurate, but it does list a fast 6V/µs slew rate, which is an extreme change over the early slew rates. Maybe once again we are seeing that there is something to using a low slew rate chip in distortion circuits.
Notable uses:
Ibanez TS-9 Tube Screamer (1982)
NE5532
The original device was released by Signetics in the 1970s. This is considered a high fidelity, yet cheap, dual op amp. Its primary drawback is a large quiescent power consumption, so it is generally not included in battery powered designs.
I include it here because it is often sited as a "good" op amp, and thus it is not uncommon to see this substituted for the pedal's original op amps in attempt to improve the pedal. How well this works is unknown to me (and would surely depend upon the specific circuit), but I doubt it is too noticeable, and the "good" performance of the 5532 is generally measured using a 30V (+/-15) or higher power supply. At 9V (+/-4.5V), the performance specs may not be as nice.
Notable uses:
"Upgrade" chip option
NE5534
Appeared in 1978 (according to Preface for Self on Audio 3rd Ed). This is the single op amp package of the 5532. I rarely notice the single package option. Like most single op amps, 2 pins are used for trimming dc offset which is not a useful feature in +9V powered guitar effects.
LM386
Power op amp. This is a specialized op amp that can drive a speaker. It can be used as a regular op amp and I see it used for overdrive circuits from time to time, usually in boutique or home brew designs. The famous "Smokey" amps and other milliwatt practice amps can be constructed around this small chip The Smokey amp's website says it has been around since the early '80s, which is about as close as I can get to dating the LM386's release.
Notable uses:
Zinky Smokey Amp (198?)
Lovepedal Black Magic (199?)
FET Input Op Amps
General characteristics:
High input impedance. Noise performance can be worse than BJT inputs. Inputs sensitive to death from static electricity. Technical hurdles delayed the development of FET input op amps. They are rarely encountered until the very late 70s and early 80s, very popular from then on.
TL0XX Family
This family of op amps from Texas Instruments is the most common FET op amp I personally see. This line was released in 1978.
The TL022 low power op amp was released earlier, as early as 1973, but this is was bipolar input device. The 022 op amps appear in Moog designed Maestro pedals.
The common through hole devices follow a TL0XX system where the first X denotes a subclass of the chip, and the second X tells you the number of amps in the package.
TL071, TL072, and TL074 are the "low noise" chips that are very popular in audio circuits. "7" indicates low noise, and the final number indicates you can get single, dual, or quad packages.
TL06X are the "low power" chips that are fairly common in battery powered audio circuits of the '80s and '90s. These have lower current draws at the expense of noise and distortion performance.
TL08X are the "low offset" chips that are selected for superior dc performance. This chip is more likely to appear in a voltage control circuit than an audio path.
TL05X is an updated/improved version to replace both type "7" and "8" chips. They are expensive; I've only encountered them in an SSL recording console so far.
LFXXX Family
This is a family of FET input op amps from National Semiconductor. The LF series began in 1974.
LF353, LF411, LF412, LF444.
Burr Brown and Analog Devices
These companies offered various high performing op amps over the years. The price of these specialized chips is usually several orders of magnitude higher than the common chips. I can't recall any pedal effect that came stock with op amps from these companies, but tinkerers will replace their common op amps with high performance ones to try and improve the overall sound. I don't have much experience with this. I have swapped some chips by request, and I can't recall ever being impressed by the results. As we are discussing op amps specifically for guitar effects here, it is worth noting that the historically prized chips in famous pedals appear to actually have poor specs.
Single Supply vs. Bipolar (Dual or "+/-") Supply
Assuming you are making a capacitor coupled audio circuit, the short answer to the op amp power supply question is: it doesn't matter. The op amp does not care if you are calling your power supply +9V or +/-4.5V. Similarly, a +/-15V power supply looks the same as +30V to the op amp. It doesn't care; neither should you.
One way to screw this up in a guitar effect is to declare the Vref rail "ground" by connecting it to the metal chassis and/or the input/output jack sleeves. Here you are truly creating a +/-4.5V system out of a 9V supply, but you cannot daisy chain this system unless you daisy the entire +/-4.5V connections throughout the chain. The only advantage or reason why this could come up is if you needed a dc coupled input/output for your effect (perhaps for control voltage purposes). This scheme does work as long as you don't try to daisy chain it. (The more common approach would be to just use the +9V and then derive a -9V rail from a charge pump.)
The key to a typical 9V pedal is that the input and output jacks see 0V as the negative battery/power source terminal. This allows for daisy chaining. Other than the odd split 9V situation described above, the other situation that messes up daisy chaining is when the input and output jacks see 0V as the positive battery or power source terminal. This pops up with germanium transistor based fuzz pedals, and the result of daisy chaining is a short circuit between the negative and positive power supply terminals. This is what is meant by germanium transistor circuits having a "positive ground;" this means the supply is truly -9V.
What if you are not doing a simple audio circuit? Then the single supply vs. bipolar can matter. If you want dc coupled inputs that normally sit at 0V, and you want the output to be able to swing both positively and negatively, then you need a bipolar supply. For many digital circuits, only a V+ and 0V are necessary, and typically that is all you will need. "Single supply" type op amps (like the LM324 and LM358) work in these circuits because they will accept 0V inputs with 0V on the V- terminal. Not all op amps (TL0XX comes to mind) will work that way.
Absolute Maximum Power Supply
For typical audio amplifier work, you connect the most positive voltage to the V+ pin, the most negative voltage to the V- pin, and a reference voltage (VRef) to your non-inverting inputs. Each op amp's datasheet will specify an absolute maximum voltage between the V+ and V- pins. Confusingly, some datasheets may give this as a "+/-" value, while others will be specified in just a single voltage. This is just an arbitrary choice by the datasheet author. +/-15 = 30V and vice versa. Most op amps are "recommended" to be run at +/-15V (30V), but the absolute max is typically +/-18V (36V). High voltage op amps are available that can withstand greater than 36V across the power supply pins.
A simple way to reduce confusion is to plot the voltages out on a number line. Plotting the numbers this way, you can visually see that +/-15V = 30V, or any +/-xV = 2x.
Further, VRef will be at the midpoint of the number line. For any symmetrical +/-V supply, VRef = 0V, and for any single supply, VRef = V+/2.
Typical non-inverting amplifiers will have a relatively high impedance resistor connect the non-inverting pin to the VRef rail. Typical inverting amplifiers will make a direct connection from the non-inverting pin to the VRef rail (sometimes it is better to insert a resistor between the non-inverting pin and the VRef rail, typically a resistance that is equal to a resistor in the inverting circuit - see any op amp manual for more details).
Single supply systems require coupling caps to connect input jacks (normally held to 0V) to op amp inputs (normally held at VRef or V+/2) and op amp outputs (again at VRef) to output jacks (back to 0V reference). This causes perhaps the most confusion, because it is a somewhat arbitrary decision in the circuit design as to whether op amp stages will be coupled together using only the VRef rail or the signal might go back to a 0V reference as it passes through panel controls like potentiometers.