MJE1300X like MJE 13005 for higher voltage stuff and like BC547 for low power low voltage and TIPXX like TIP42C for lower voltage power applications like Linear PSUs IRFZ44N/IRF3205 for low-ish voltage stuff and 7N11 or similar for higher voltage stuff 7 stands for 700v and 11 for 11A. I don’t know about JFETs.
maybe not a shootout but more of a selection guide for young players. I needed something like this when starting out, uni doesnt teach component selection :(
Most standard used n-channel MOSFET is IRFZ44. At least that one Is one I have like 50 pcs and whenever going to build something around mosftets , it comes in consideration as a mosftet which going to be used.
Around 1964 I needed an opamp for special analog computing projects and designed one based on the one in the company's process controller. About a year later I replaced those approx. 99 components with an MC1433G for $30. About a year later I replaced that with a 741 for one dollar. A couple years later a 741 was, i forget, about a nickel or a dime. /Andyh
LM358 if you do not use the other half just use it as a buffer, and make it useful to drive longer lengths of cable. If you might need to use it simply connect the input, output and feedback using any cheap low value SMD resistors used elsewhere in the design. That way if you need it, you simply remove the 3 resistors, and have the uncommitted opamp available for the bodge wires, having pads already and no track cuts needed. Just be aware that input CMRR means 3V is not giving you much signal amplitude before it clips, and depending on who made it the output may or may not invert under large overload. Really you want 6V or better for good results.
Yes, rarely do I have the spare opamp. But if you do, yes, strap with removable resistors. Don't leave floating or tied hard whcih can draw excess current. Should have mentioned that in the video actually.
If my understanding is correct this is a win win for audiophiles and engineers because you can utilize negative feedback and not worry about inverting the original waveform because you can invert it back to the original polarity with an inverting buffer.
I am at a loss why do people not use the 2nd “unused” op amp in the feedback loop to tighten the phase shift in applications that could use that. Especially for audio. See James Wong’s ADI AN-107. If any one reading this hasn’t read that app note, it’d be a good idea to do so right away. It can pull some extra performance even out of an LM358. It’s just two extra resistors. Anywhere you could use a flatter response and less phase shift, this is the way to go. Especially useful in instrumentation/lab applications where you may want the amplifier to be as transparent as possible. Trip point for young players: this absolutely depends on all op amps used (2 or 3 depending on the degree of compensation) to be on the same semiconductor die. If you mix op amps in different packages, you’ll decrease the phase margin and the flatness of the transfer curve.
They are pretty nice, but drift and do have some shot noise, would not use them for "accurate" or long term measuring. For these voltages I would go with OPAx330. For most analogy input stuff they are a bit more care-free to use, but surely do cost a bit more. Availability I have not checked.
I just wanted to write that, but I see that you were faster. The third generation of OP-amps has several advantages, that are just priceless. I've been using MCP6001, which has rail-to-rail input and output, supply voltage range from 1.8V to 6V, is internally compensated to 1MHz and is really cheap and available.
The 600x was probably my favorite before I really knew what I was doing. It behaves ideally for Arduino projects, and if you need something comparable but better, there’s always the 602x.
Great video Dave, Thanks. Regarding the LM358/324 parts, we've always heard they are not good for audio. I figured this was audio snobbery. I didn't realize until recently (from a Bob Pease video) that this is because the output stage is Class B, not AB like most proper opamps. The output buffer has no bias. This is to save power. So the voltage amp stage needs to slew nearly 2V to switch the output from + to =. This causes lots of crossover distortion of an AC signal. But all is not lost: this can be fixed by using a pull-down or pull-up resistor on the output, making the output Class A.
Completely agree, I always avoided this opamp for audio, you can bias up the output stage to avoid the crossover but then you kill the low power performance, so why bother. Just use the TL07 series.
Underrated comment. Every now and then there is a question on some electronics forum about "LM358 and strange distortions" posted by novice. I feel like that knowledge should be more widespread.
Ah, if I saw your message before, well... I wouldn't have had the fun to discover it by myself. Testing a few opamp with super matched resistors to see what CMRR I could get without any dedicated IA, I was shocked by the crossover distorsion of the LM358. I came to the same conclusion looking at its functional diagram and came to the same solution : a pull-up was added in the process! Works like a charm.
I cannot remember how many of my favorite jellybean 555 timers I burned out as a kid. Sometimes they would fall from my desk to the carpeted floor. One time my mother came into my room to talk and she stepped on one of these chips. Those eight little pins plugged securely into her heel. Of course she made a big deal out of it. I had difficulty holding back my laughs, though I was horrified that it bled so much. I had wondered where that chip went. For a few days, she had eight little holes in her heel-turned-breadboard. I was more careful after that. I built all the circuits I found in the Radio Shack Engineering Notes booklets. I loved all the CMOS and TTL logic chips. But the 555 found it's way into nearly all my projects because it was able to provide adjustable frequency clock pulses and pulse width. Since I was a kid without an oscilloscope, I had to get my duty cycle right using an LED and then turn up the frequency. I wish I had spent more time understanding op-amps, but I was unwilling to do all the math. Digital, to me, meant computers and robotics. I pursued these more than analog circuits because they made more sense to me. The ubiquitous 555 was at the beginning of my hobby that later became my career as a controls engineer. But with microcontrollers that have built-in oscillators, the 555 is just not needed so much any more.
The NE5532 is a superb op-amp for audio circuits with low resistance values around it due to only 5nV/√hz. If resistance values go up, say +100k in filter circuits, the Johnson noise caused by input bias can outweigh the voltage noise. Then you want a FET like TL072. However if you look at products today coming out of big corps, the NJM4580/RC4580 (or equiv.) is more common than NE5532. It has slightly lower noise and is cheaper in bulk. The LM4562 was becoming very popular in high quality audio products, but got discontinued after TI bought out National. Today I found it's back in active production! Like an NE5532 with only 2.7nv/√hz.
They finally lower the price of LM4562 to more acceptable range, but I really doubt we'll see that widely used in mid range consumer grade audio device.
As some others have suggested, the MCP6002 is probably the go-to opamp for any applications that require rail-to-rail inputs as well as outputs. Its specs are comparable with most of the "jelly-bean" amps, and it's easy to specify for general purpose, even though it's possibly a bit more expensive than the LM358 which it comfortably outperforms in any low-voltage, low-power scenario. I'd also mention the MCP602 as a replacement for the LMV358 as it is generally similar, but with tighter and improved specs, and costs more or less the same.
Dave, I absolutely loved this video. As a certified Olde Phucke it warms my heart to see the components of my youth still being designed in AND useful 40 years later. I'm a 63 yr old engineer that spends more time in embedded software but your videos are near essential viewing imho
Alternate title: “TOP 5 OPAMPs you can salvage from your hoard of scrap electronics when you’re desperate” I’ll also mention the op-amp I use the most is the LMV321.
Wow - your video really takes me back. Please do a series on these old 70's/80's chips (jellybeans - hehe) ... the days in high school of designing and building analog guitar effects and synthesizers with those easy-to-work-with chips. They made it easy to actually made something to sell ... start a company early. Many thanks for the video. Great channel. Cheers from Florida, USA
Gosh. That's a blast from the past. Have been out of the industry for 30 years and assumed these classics would have dropped off the radar of today's designers. The 5532's had a habit of exploding in one product I worked on which ran them at +/-18V, which I seem to remember was near their max supply, but maybe that was just a bad batch - it was a long time ago!
To be fair, designers these days are way more likely to be using a part from TI's OPA16xx range, perhaps an LM4562, an OPA2227, or if you're stingy, an NJM2068 or 4580. Douglas Self sure got his mileage out of 5532s though. The original Philips/Signetics part actually is long EOL, following a fire in a French Philips fab in 2003, and the various second-sourced 5532s differ somewhat in their characteristics (e.g. the common / cheap TI part has a more robust and stable output stage but substantially worse input common-mode distortion, as shown by Samuel Groner's testing). While no longer a truly all-round part, it is still hard to beat the output driving abilities of a 5532 at its price point.
The TL07x series suffer phase inversion when driven hard into a rail. This can cause them to latch up. I had a job many years ago where I had to replace TL074s in a design I was asked to fix because of the latch-up issue. I don't remember what I used to replace them.
@@MichaelWeaser A lot of modern CMOS chips are resistant to latch up, is that generally not done for opamps because it compromises the performance of the opamps?
I work mostly with audio and what I use almost always is the NE5532. So I agree with you, Dave!... I also use the LM4562 as the lowest noise OA I've found on the market. But they have a latch up problem with high impedance input resistors. So, they imply a bit more complicated anti-latch design, only to be used when super low noise is needed.
* lowest _voltage_ noise (in their price class) - they do have a whole bunch of input current noise due to what appears to be a degenerated bipolar input stage. LM4562s also seem to be super allergic to RF for some reason, way more than what you'd expect from GBW and slew rate alone (I can only guess that it might be their bias current cancellation at fault). Another bang/buck part for medium impedance line-level usage is the trusty JRC/NJR NJM2068 (now revamped as NJM8068). Noise almost as low as LM4562, just don't drive super load-impedance loads with it (avoid
MAA741 and MAA748, we had huge bag full of those horrile opamps at high school. Also high school hadn't money to buy components (at least we were told that), so all projects were based on 741 or 748 and old stock of Tesla resistors and capacitors. I remember how many circuits my classmates tried to reproduce weren't working and if you replaced 741 for ANYTHING ELSE, it started working perfectly :D
I like the MC34071/2/4 opamps, wide input voltage range 3-44v, fast slew rate, 13 V/us, 4.5Mhz bw, stable at large output capacitance. Good at driving output stages in power applications.
Great video! Audio engineer here. The reason that the NE5532 is the #1 Jellybean audio amp, is bandwidth and slew rate. To get 20khz full swing, you need a slew rate of at least 3v/us. The NE5532 is 9v/us and TL074 is 13v/us. In something like a microphone preamp, you want a maximum gain of 55 dB or so. This means the signal needs to be multiplied over 500 times. The NE5532 can do this in a single stage because it has a GBP of 10 mhz. The TL074 is really good for unity gain stages or stages with less than 40 dB gain since it has a GBP of only 3 Mhz. The LM358 and others with their slew rates of like 0.5 v/us and GBP of 1mhz can't really be used for audio.
Hi Dave I agree with your comment on the LM741. I'm in the rail industry and change of electronics happens to move really slowly in Rollingstock down here in Melbourne. We still use the LM741H on Westinghouse and GEC train equipment from the beginning of the 80's. They are still used in power and traction control control cards.
21:33 the external compensation of the 5534 (which is, of course, a single op-amp) came in handy recently when I wanted to deliberately limit the slew rate of a signal. Add a comically oversized compensation cap - job done. Another notable mention for audio applications is the NJM4556, when you need a bit more output current, it might be worth considering.
Do more of these! Great content and introduction to specifications of interest. Love to see this for FETs, BJTs, Diodes, Voltage Regulators, and Logic Gates,
My father was replacing a circuit on his Volkswagen Westphalia and it had an LM324 in it! Already had the part :p LM324 is a good one. Memorable pinout, single supply, decent bandwidth for hobby stuff.
Great idea for a series. One of the problems many of us have is identifying what I call “standard” go-to parts. To keep in stock for basic design use. I have a few that I’ve used before, but I like the “jelly bean”concept. Hope you do more of this.
I bought couple of hundred LM358s and NE5532s few years ago.. i'm never going to run out of op amps, those two really cover a wide range of needs. The 358 even has a decent output current. Got also a bunch of 072s but haven't found a good use for it, the 5532s do audio so well.. Billions of 072 are found inside millions of sound consoles.
My 2 rules of thumb when using op-amps: same inpedance on both inputs, resistor 1k between output and V-. This last trick turns op-amp in some sort of class-A amp, great for audio applications. I recently built a stereo headphone audio amplifier. LM324 for input stage, LM386 for output, applying those 2 rules gives excellent result
Some of my op-amps: LM124 / 324 quad bipolar GP LT1128 Ultra low-noise LT1010 High current (150 mA) unity gain buffer (TO220-5) LMC660 Femtoampere input bias CMOS NE5532 Aimed at audio. Widely used in mixing consoles MAX4328 Ultra-low input offset.
Cheer's for the OP-07. Invented by PMI, and it's successors the OP-77 and OP-177. Brings a tear to my eye having worked for them a few years before we were acquired by Analog Devices. Working for Analog Devices was also fun.
I was with you all the way up until I suddenly got hit with a wave of nostalgia and a realization of how old I'm getting. The 741 was the OP amp you got from Radio Shack back in the day. There was only the 741 and the quad. I think I still have a few from decades ago lingering around in the original Radio Shack packaging. Love these jellybean videos! Keep up the amazing work!
My KOSMOS electronics kit I had as a kid had an 741 as the opamp. I feel lucky to have gotten the last half decent generation (X1000-X4000) of those, before they dumbed them down into oblivion. What they offer today is a bad joke in my opinion. Im born in 1985, but grew up on my fathers electronics books from the 1970s in my "formative years", so .. yeah, I feel pretty old, too.
Great video series Dave! Really useful as someone who is self-taught on the hardware end and not always sure what's a good starting point when confronted with a massive list of parts to choose from even after filtering at a distributor's website
I'm by no means an EE, just a hobbyist but I'm also an electrician and do a lot of building automation. I felt really happy when I recently looked at some demolished equipment (laboratory fume hood controls) and knew what all the ICs were with the exception of the old early 2000s era microprocessor in each device. All TTL logic and cheapo quad comparators and op-amps.
I love this video. I know the jellybean op amps are already known by almost every engineer, but it is nice to look at the datasheets again. Looking at this makes me want to assemble an analog computer. Possibly a future project for my free time.
The NE5532 is my favorite, single Op-Amps in a package are used when you are amplifying minute dc Voltages, because they have offset pins to zero out the error, when an Op-Amp is operated at an elevated Voltage (and the other unit in a package would be unable to be usable) and to replace higher noise and drift single package. Like a NE5534 replacing a LM301. Ron W4BIN
Yup. Paul McGowan of "high end" audio company PS Audio was sure he could improve the performance of a vintage Studer desk he'd acquired by swapping in "better" opamp chips than the NE5532s with which it came -- he had to confess that there was no improvement whatsoever and had his techs reinstall the originals.
Coincidence. I've just recently started to build some LM 386 based mini guitar amps. 386's are amazingly versatile. With the right components you can make them sound like a Bassman or Ampeg or Marshall 800, Hi-Watt, you name it. I make them with or without volume or gain controls. I also get the N-4 version and run it at 12V instead of 9V. Bridge the 1 and 8 pins with a 10uf cap for a gain factor of 200. Or put a gain pot in there. The smallest and most simple one is 1 cap, 1 resistor, LM386, input jack, output jack. And that's it. About the size of a pack of cigarettes. It sounds awesome, they all do. Also making a _Walking Talk-Box._ A cut down funnel attached to a 3in speaker, tubing up to your mouth, clips onto your belt. Battery powered so you can walk anywhere and play anything from STP to Peter Frampton.
The 386 is both a solid, basic part and, as you said, very easy to work with. I expected Dave to include it in the video. But, the more I thought about it, the more I realized that it was probably excluded due to the fact that the 386, though popular, isn’t really a general-purpose opamp, in the same sense as the others mentioned. Have fun with your new amp. Interestingly, I’ve also recently started designing my first audio amplifier. Although I have designed plenty of digital and analog systems, I have never I had the need to include anything more than a buzzer in any of them. I suppose that’s because of my work centers around embedded devices, not ones that get all that close to humans. Audio is a whole, fascinating world unto itself. Should be fun.
@@a1nelson Probably the very best feature of a 386 for guitar amp is the fact it nearly duplicates the tonal characteristics of a tube-driven amp. Specifically, the dynamic sensitivity while playing softly has a cleaner sound and articulation, just like a tube amp. And the harder it's played, the distortion comes through strong, just like a tube amp. And does it without increase in gain or volume. And that's a big issue because all other solid state circuits can't replicate that. Ok, maybe some, but not at such a meager price point. Thank you for your reply
I've been playing for 6 years now, and I'm also an ee junior at uni, so im gaining traction on learning more about this stuff and I'm curious how the op-amp type changes the sound? I've pondered the idea of trying to build my own amps, to experiment with sound, especially for my bass. I'm gonna have to try the simple amp idea you mentioned. thanks!
@@bobbyquinn4952 Hmmm........bass...that's a good Idea. I'll have to do a little homework and see what the signal frequency parameters of a 386 is capable of and what the pickups on a bass produce. You would certainly need to scale up power, and maybe two 386's to help spread the load.
Curious: are you referring to their potential use for one of your videos? (Which are absolutely fantastic!) So, are you talking about breadboarding? Or, hobbiest market aside, are there modern applications where a DIP is still the best option? Personally, I try to leave human-facing connectors, power MOSFETs and higher power relays as the _only_ TH components in my designs (plus some specialty parts, like transformers). Maybe I’m taking SMT too far? Edit: added clarification re uncommon (for me) TH parts.
@@spakecdk One that I have been using a lot more lately is the LMC6482 - mainly because I like the r-r input and output, and I happen to have a tube of them. But, they aren't cheap. There are likely better/cheaper solutions for R-R in/out in a dip package with a wide power supply range available now - haven't looked in a while... The MCP6002 is a nice choice for low supply voltage applications.
Hey Dave, could you make a "101" video on unity gain stability you've mention @ 4:13? We've had those lectures at the uni, but we've never been given an example of an application where you'd need to take care about the stability while designing circuitry; i.e. how you could accidentally make the amp go haywire. I only remember it has something to do with a pole at 0 dB. .... students are fed too much data in a way too short period of time, often without having prerequisite knowledge, and they're fed raw, inapplicable data.
the most simple practical rule is to break the loop, insert a (virtual) frequency response analyzer there, and make a bode plot. Look at the bode plot. Find the point where phase shift hits 180deg, and see if the amplitude is reliably below 0db at that point and beyond. If it is, the loop is stable. Additionally, note the phase where the amplitude crosses the 0db line - (180-that) is the phase margin. If phase margin is low, the system will likely ring for quite a while before settling. This is not the comprehensive stability criterion as far as i know, as some weirdo systems that fail it can still be stable or some systems that pass may still be unstable, but for opamp feedback design, it's a reasonable rule. As for poles/zeros, it is tricky. I recommend watching a series of control systems lectures by Brian Douglas for that. BTW, poles and zeros are specified by their complex frequencies (s-plane coordinates), i don't think they have any dB property (i can't say i'm fluent with this stuff, though, maybe i'm wrong).
Oh YEAH! This is a good topic! Thanks Dave. I'm to the point in my education where I can start to disagree with you on a few things but these videos are exactly the sort of thing you don't learn in school!
@@j1952d Possibly because you're looking at years and years of experience in sifting through potential parts to find one best suited to do a particular job, often in tension with a commercial need to keep cost low and availability high. Most folks with that sort of experience aren't going to be teaching at universities.
@@RexxSchneider Oh, yea. When I went, there were teachers with lifetimes of experience that they were prepared to pass on. They're all dead now, and I'm not far behind .
@@j1952d I didn't have quite your experience. When I was at university in the late 1960s, the head of the faculty was Professor AHW Beck, who wrote one of the definitive textbooks on "Thermionic Valves: Their Theory and Design". This was at a time when I had moved on from germanium to silicon transistors and was experimenting with early opamps like the 709. Nobody on the staff had seen an ic opamp and I remember having a crowd of demonstrators around one of my prototypes of a primitive capacitive proximity sensor. I was able to get a lot of advice on ECC83s and EL84s, but not much on which opamp was best suited for an application. :(
The Art of Electronics (textbook from Harvard professors Paul Horowitz & Winfield Hill) has tables of jellybean parts for all the common component types. Plenty of other universities use it.
I hope you do more videos like this. My jellybean OP amps are exactly what you covered. My favorite jellybean part, is the TL431. It's in almost every SMPS for a reason.
Dissapointing to learn this after I made a USB powered voltage controller, it does 0-3.5 volts output. Chose the 358 because I didn't need a high speed (audio rate) response time.
Another famous weak point of the LM358/LM324 (and long-time regret of its designer) is its output stage, which cannot switch from sourcing to sinking without crossover glitches... if you find you must use one for audio (despite the awful noise level), Class A bias the thing by all means.
I really like this format. Maybe more like this, jellybean mosfets ect. It's Definitely more advanced than "what is an Op amp" but really helpful for people who have already done school but don't have a ton of experience under their belt.
Loving this Jellybean series! Would be great if you included a comparison table with description, benefits and cost per dozen for each part covered. Would save a lot of time for a lot of people
More of this stuff Dave please. I really enjoyed it. Bring more sheets of other chips and comment on them. Not just OpAmps. More talk about the ground reference and rail to rail headroom. You the man!
Another good reason for using dual over single OpAmps is logistics. As a manufacturer, or even a hobbyist, it's much better to stock a single part than try and maintain a stockpile of two different part numbers.
LMV321 comes in (at least) two footprint versions of SOT23-5 and they are NOT interchangeable. Check your files before soldering, some vendors sell only one variant, some have two
You might want to mention LM358 has issues with crossover distortion on the output, and so the output should be pulled down sufficiently for certain applications, especially AC coupled outputs, where the output both sources and sinks current.
Before watching the video, 4 in the list are obvious to me: LM358 (general purpose, low frequency, low cost, single supply), OP07 (low offset voltage, 'precision', great for raw sensor front end), TL071 (high input impedance, okay for audio), NE5532 (standard audio opamp, relatively low distortion in audio band for the cost, quite alright at anything). I think these will be the same for 99% of engineers.
As expected, all the celebrities are in the video. At 21:35 though, the 5534 is a single part, not quad! It is also in PDIP package, but only a single opamp because it exposes the compensation capacitor pins, and is only stable with feedback gains of 3x +. On the plus side, it does have sliiightly better noise though. But, if you want better noise than the 5532 you'd be looking at other parts anyways, so kind of an obsolete opamp at this point in time in my opinion.
I was taught opamp theory in the Navy in 1974 by a Senior Chief Petty Officer who was in the design group to the Grumman A-6E Intruder fire control systems. Milspec ceramic DIP cases and "flat packs". I still use these "jellybeans" along with the venerable 555 timer.
I was taught using 741's... all I used in two years of analog classes. I did use some others (rail to rail capable) in later years but at least in education, it seems the 741 still has a use.
Wish I had this video just few months earlier! I figured I'd probably do away with a jelly bean product - rail to rail was ideal but a bit of headroom wouldn't hurt, I thought. Unfortunately local provider showed LM741 as the most popular... As a result, the knob was only effective on like lower 30% but the difference was *just* enough to tell, so user would have the impression that they had control over the whole range. It wasn't any critical part, so that was enough 😁
Dave, I hope to be buried with an LM-324! As a junior engineer in the late 70's, early 80's, I both cut my teeth and quickly adopted the LM-324 and used it for almost every product that I was commissioned to design for one really unique reason: To me, the real power of the LM-324 is the fact that you have 4 OP-amps in the package which allows you to effortlessly increase the performance of many types of circuits (especially precision comparators) by using 2 of the OP-amps as buffers by tying the (-) to the output and using the (+) as the input. This allows you to not load, say a comparator, by isolating its input from the external circuit. Conversely, you can use the second buffer as output-isolation from the load, so again, you don't upset or change the operating point of a comparator. Additionally, you'll find using these buffers (in & out) will aid in the precision of your 10X, 100X etc. amplifiers. So, this is why I completely skipped over the LM358 because I just got such better performance with not a whole lot of additional real-estate! Anyway, we're definitely on the sam page for 'Jellybean' Op-amps! Thanks Dave!
thanks you remembered us our first projects , now we need some jellybean transistor and CMOS or TTL logic gate and counter and processor like old microship or atmel finely we mustn't forget the famous 555 and the Tl 431 .
My biggest rant is on data sheets is when they list both an A and B version of the part but fail to highlight the differences forcing one to do a line by line examination of the data sheet in an attempt to figure it out.
I wish I could give this comment 1000 thumbs up. The odious practice of not proving “delta” information (between versions within a datasheet, or between “sister” models in a series) drives me nuts. I’ve seen some where people gave up and asked the manufacturer (in their forums) what was different between two versions, and the manufacturer wrote back with the difference - _which wasn’t in either datasheet._ 🤬🙄
And with the situation in the semiconductor business over the last what 15 years, with mergers and acquisitions - you can't trust one manufacturer over the other to have proper good datasheets either. (Not mentioning different segments inside the same company)
@@tookitogo Connectors are so bad about this. Why do these two parts look identical but have different part numbers? Well one has little drain holes in the bottom so they don't collect fluid when the board is washed after assembly. Also, we don't make that kind anymore. Of course!
In want mooooooorrreeeeee! My favorite OPAMPs: Precision stuff, Rail-to-Rail, low voltage(5.5V max), low power consumption: OPA2333 = TLC2333 Not so precise, Rail-to-Rail, low voltage (5.5V max), low power consumption: OPA2348 Very low bias current (0.2pA), Rail-to-Rail, 3-16V Supply, fairly low power consumption: LMC6482 My fav cheap low noise Audio OPAMP: RC4580 My favorite higher current and capacitor driving OPAMP: LM8272 (Good for Rail Splitters) My favorite higher current driving low voltage OPAMP: OPA2322 My former favorite +-5V low power audio opamp: TLC2262/TLC2272 Just to name a few, got so many more...
Thanks for this tutorial! I didn't realize the 4558 chip was essentially a pair of 741 chips in an 8 pin DIP. For audiophile work the Burr-Brown op amps are nice. In 8 pin DIP the OPA2134 is something to consider, although there are newer ones with even better specs available in SMD packages only. It's tempting to get an adapter to try plugging an SMD variant into the DIP sockets on one of my guitar pedals just to see how it responds to being driven into the nonlinear region.
@JM Coulon If the design is an EQ without coupling capacitors, the bias current of the 5532 can thrash the track, creating the dreaded permanent scratch. That is really the only reason I use it, otherwise I just use the 5532 if input bias current is not an issue. I’ve found it to be the best compromise of noise, cost, and power consumption if low bias current is a hard requirement.
@JM Coulon Yes, the Burr Brown stuff is pricey. $4 or so for a single OP2134, which I liked only marginally better than a CA32040. The 5532 is also a solid choice. Could you name a couple of the current audiophile op amps for me? They are probably all SMT-only.
Watched this video to find a better op amp than my "go to" LM324. Left with LMV324. Good enough, I guess, and I don't have to change anything else. Earned a like and maybe a sub.
I think having a Top 5 shootout list for BJTs, MOSFETs & JFETs would be an interesting video too.
MJE1300X like MJE 13005 for higher voltage stuff and like BC547 for low power low voltage and TIPXX like TIP42C for lower voltage power applications like Linear PSUs
IRFZ44N/IRF3205 for low-ish voltage stuff and 7N11 or similar for higher voltage stuff 7 stands for 700v and 11 for 11A.
I don’t know about JFETs.
I agree.
maybe not a shootout but more of a selection guide for young players. I needed something like this when starting out, uni doesnt teach component selection :(
Yes that is a good idea.
+
A "jellybean components you should know" series sounds OUTSTANDING. Please do it!
A series about "standard" parts would be super awesome! Next MOSFETs pls! :D
Most standard used n-channel MOSFET is IRFZ44. At least that one Is one I have like 50 pcs and whenever going to build something around mosftets , it comes in consideration as a mosftet which going to be used.
I think it is 2N7000 for signal applications, and IRF540, IRF9540 for power
Around 1964 I needed an opamp for special analog computing projects and designed one based on the one in the company's process controller. About a year later I replaced those approx. 99 components with an MC1433G for $30. About a year later I replaced that with a 741 for one dollar. A couple years later a 741 was, i forget, about a nickel or a dime. /Andyh
Yes please, do a series about it. I would love to know the jellybean components that professionals consider for their design.
LM358 if you do not use the other half just use it as a buffer, and make it useful to drive longer lengths of cable. If you might need to use it simply connect the input, output and feedback using any cheap low value SMD resistors used elsewhere in the design. That way if you need it, you simply remove the 3 resistors, and have the uncommitted opamp available for the bodge wires, having pads already and no track cuts needed.
Just be aware that input CMRR means 3V is not giving you much signal amplitude before it clips, and depending on who made it the output may or may not invert under large overload. Really you want 6V or better for good results.
Yes, rarely do I have the spare opamp. But if you do, yes, strap with removable resistors. Don't leave floating or tied hard whcih can draw excess current. Should have mentioned that in the video actually.
If my understanding is correct this is a win win for audiophiles and engineers because you can utilize negative feedback and not worry about inverting the original waveform because you can invert it back to the original polarity with an inverting buffer.
@@HazeAnderson Yep, inverting is really useful for lots of reasons, and having another to get back to original polarity is often necessary.
I am at a loss why do people not use the 2nd “unused” op amp in the feedback loop to tighten the phase shift in applications that could use that. Especially for audio. See James Wong’s ADI AN-107. If any one reading this hasn’t read that app note, it’d be a good idea to do so right away. It can pull some extra performance even out of an LM358. It’s just two extra resistors. Anywhere you could use a flatter response and less phase shift, this is the way to go. Especially useful in instrumentation/lab applications where you may want the amplifier to be as transparent as possible.
Trip point for young players: this absolutely depends on all op amps used (2 or 3 depending on the degree of compensation) to be on the same semiconductor die. If you mix op amps in different packages, you’ll decrease the phase margin and the flatness of the transfer curve.
For simple, low voltage (
They are pretty nice, but drift and do have some shot noise, would not use them for "accurate" or long term measuring. For these voltages I would go with OPAx330. For most analogy input stuff they are a bit more care-free to use, but surely do cost a bit more. Availability I have not checked.
I just wanted to write that, but I see that you were faster. The third generation of OP-amps has several advantages, that are just priceless. I've been using MCP6001, which has rail-to-rail input and output, supply voltage range from 1.8V to 6V, is internally compensated to 1MHz and is really cheap and available.
The 600x was probably my favorite before I really knew what I was doing. It behaves ideally for Arduino projects, and if you need something comparable but better, there’s always the 602x.
Definitely worthy of an honorable mention.
@@TrickyNekro What do you mean by "accurate"? Something like 0.1%? The OPAx330 is at least 10 times as expensive, it's a different amp all together.
Great video Dave, Thanks. Regarding the LM358/324 parts, we've always heard they are not good for audio. I figured this was audio snobbery. I didn't realize until recently (from a Bob Pease video) that this is because the output stage is Class B, not AB like most proper opamps. The output buffer has no bias. This is to save power. So the voltage amp stage needs to slew nearly 2V to switch the output from + to =. This causes lots of crossover distortion of an AC signal. But all is not lost: this can be fixed by using a pull-down or pull-up resistor on the output, making the output Class A.
Completely agree, I always avoided this opamp for audio, you can bias up the output stage to avoid the crossover but then you kill the low power performance, so why bother. Just use the TL07 series.
A better than TL07x is 4558. Better than 4558 is 5532. Better than 5532 is opa1678 but it's not jellybean.
Underrated comment. Every now and then there is a question on some electronics forum about "LM358 and strange distortions" posted by novice. I feel like that knowledge should be more widespread.
Nice explanation. Thanks.
Ah, if I saw your message before, well... I wouldn't have had the fun to discover it by myself. Testing a few opamp with super matched resistors to see what CMRR I could get without any dedicated IA, I was shocked by the crossover distorsion of the LM358. I came to the same conclusion looking at its functional diagram and came to the same solution : a pull-up was added in the process! Works like a charm.
I cannot remember how many of my favorite jellybean 555 timers I burned out as a kid. Sometimes they would fall from my desk to the carpeted floor. One time my mother came into my room to talk and she stepped on one of these chips. Those eight little pins plugged securely into her heel. Of course she made a big deal out of it. I had difficulty holding back my laughs, though I was horrified that it bled so much. I had wondered where that chip went. For a few days, she had eight little holes in her heel-turned-breadboard. I was more careful after that. I built all the circuits I found in the Radio Shack Engineering Notes booklets. I loved all the CMOS and TTL logic chips. But the 555 found it's way into nearly all my projects because it was able to provide adjustable frequency clock pulses and pulse width. Since I was a kid without an oscilloscope, I had to get my duty cycle right using an LED and then turn up the frequency. I wish I had spent more time understanding op-amps, but I was unwilling to do all the math. Digital, to me, meant computers and robotics. I pursued these more than analog circuits because they made more sense to me. The ubiquitous 555 was at the beginning of my hobby that later became my career as a controls engineer. But with microcontrollers that have built-in oscillators, the 555 is just not needed so much any more.
The NE5532 is a superb op-amp for audio circuits with low resistance values around it due to only 5nV/√hz. If resistance values go up, say +100k in filter circuits, the Johnson noise caused by input bias can outweigh the voltage noise. Then you want a FET like TL072. However if you look at products today coming out of big corps, the NJM4580/RC4580 (or equiv.) is more common than NE5532. It has slightly lower noise and is cheaper in bulk.
The LM4562 was becoming very popular in high quality audio products, but got discontinued after TI bought out National. Today I found it's back in active production! Like an NE5532 with only 2.7nv/√hz.
@JM Coulon For low signal level sources you probably want a FET though, no? Not to load the source down.
I missed the 5532 as well in this presentation
@JM Coulon Point taken! Not all low voltage output are low current outputs.
Lm4562 = lme49720
They finally lower the price of LM4562 to more acceptable range, but I really doubt we'll see that widely used in mid range consumer grade audio device.
As some others have suggested, the MCP6002 is probably the go-to opamp for any applications that require rail-to-rail inputs as well as outputs. Its specs are comparable with most of the "jelly-bean" amps, and it's easy to specify for general purpose, even though it's possibly a bit more expensive than the LM358 which it comfortably outperforms in any low-voltage, low-power scenario.
I'd also mention the MCP602 as a replacement for the LMV358 as it is generally similar, but with tighter and improved specs, and costs more or less the same.
Dave, I absolutely loved this video. As a certified Olde Phucke it warms my heart to see the components of my youth still being designed in AND useful 40 years later. I'm a 63 yr old engineer that spends more time in embedded software but your videos are near essential viewing imho
Alternate title: “TOP 5 OPAMPs you can salvage from your hoard of scrap electronics when you’re desperate”
I’ll also mention the op-amp I use the most is the LMV321.
Wow - your video really takes me back. Please do a series on these old 70's/80's chips (jellybeans - hehe) ... the days in high school of designing and building analog guitar effects and synthesizers with those easy-to-work-with chips. They made it easy to actually made something to sell ... start a company early. Many thanks for the video. Great channel. Cheers from Florida, USA
Gosh. That's a blast from the past. Have been out of the industry for 30 years and assumed these classics would have dropped off the radar of today's designers. The 5532's had a habit of exploding in one product I worked on which ran them at +/-18V, which I seem to remember was near their max supply, but maybe that was just a bad batch - it was a long time ago!
To be fair, designers these days are way more likely to be using a part from TI's OPA16xx range, perhaps an LM4562, an OPA2227, or if you're stingy, an NJM2068 or 4580. Douglas Self sure got his mileage out of 5532s though. The original Philips/Signetics part actually is long EOL, following a fire in a French Philips fab in 2003, and the various second-sourced 5532s differ somewhat in their characteristics (e.g. the common / cheap TI part has a more robust and stable output stage but substantially worse input common-mode distortion, as shown by Samuel Groner's testing). While no longer a truly all-round part, it is still hard to beat the output driving abilities of a 5532 at its price point.
You sure sparked some old neurons deeply buried in my brain. And I remember when OP07 was $$$$! Thanks.
On a related note: Bob Widlar is a fucking legend!
Was a f*@%ing legend. He was also a legendary functioning alcoholic. He died from a heart attack at age 53.
@@MrJohnBos dying doesn't stop you from being a legend
@@oleksiishekhovtsov1564 One might even say it's a prerequisite :-)
@@MrJohnBos he helped shape the world around you and changed all our lives. His drinking habits do not change that
The TL07x series suffer phase inversion when driven hard into a rail. This can cause them to latch up.
I had a job many years ago where I had to replace TL074s in a design I was asked to fix because of the latch-up issue. I don't remember what I used to replace them.
Any reason why they haven't made latch up resistant versions of that part?
@@NiHaoMike64 all FET input op amps basically have the issue. It has to do with the input using FET transistors.
@@MichaelWeaser A lot of modern CMOS chips are resistant to latch up, is that generally not done for opamps because it compromises the performance of the opamps?
@@MichaelWeaser There are plenty of fet input opamps which do not have phase inversion
I work mostly with audio and what I use almost always is the NE5532. So I agree with you, Dave!...
I also use the LM4562 as the lowest noise OA I've found on the market. But they have a latch up problem with high impedance input resistors. So, they imply a bit more complicated anti-latch design, only to be used when super low noise is needed.
* lowest _voltage_ noise (in their price class) - they do have a whole bunch of input current noise due to what appears to be a degenerated bipolar input stage. LM4562s also seem to be super allergic to RF for some reason, way more than what you'd expect from GBW and slew rate alone (I can only guess that it might be their bias current cancellation at fault).
Another bang/buck part for medium impedance line-level usage is the trusty JRC/NJR NJM2068 (now revamped as NJM8068). Noise almost as low as LM4562, just don't drive super load-impedance loads with it (avoid
MAA741 and MAA748, we had huge bag full of those horrile opamps at high school. Also high school hadn't money to buy components (at least we were told that), so all projects were based on 741 or 748 and old stock of Tesla resistors and capacitors. I remember how many circuits my classmates tried to reproduce weren't working and if you replaced 741 for ANYTHING ELSE, it started working perfectly :D
I have few tens of Tesla's MAA501...504 in metal cans. They look awesome but work mediocre.
The LM324 is my go-to op-amp for the last 30 years.
For audio the TL 074 it's slightly better and is pin to pin compatible with LM324
I like the MC34071/2/4 opamps, wide input voltage range 3-44v, fast slew rate, 13 V/us, 4.5Mhz bw, stable at large output capacitance. Good at driving output stages in power applications.
Great video! Audio engineer here. The reason that the NE5532 is the #1 Jellybean audio amp, is bandwidth and slew rate. To get 20khz full swing, you need a slew rate of at least 3v/us. The NE5532 is 9v/us and TL074 is 13v/us. In something like a microphone preamp, you want a maximum gain of 55 dB or so. This means the signal needs to be multiplied over 500 times. The NE5532 can do this in a single stage because it has a GBP of 10 mhz. The TL074 is really good for unity gain stages or stages with less than 40 dB gain since it has a GBP of only 3 Mhz. The LM358 and others with their slew rates of like 0.5 v/us and GBP of 1mhz can't really be used for audio.
What about the current protection that the TL074 has? He didn't mention that aspect... does come up in your world?
The LM4562 has become my jellybean part for audio. Not as cheap as the 3352, but it impresses me everywhere that I use it.
Interesting, I have a few of those lying around... I have been buying the OPA1679 as my "normal" opamp.
I prefer OPA2134 when I want hi quality opamp
what is special about them?
Thank you. Our club loves this component walk about. We watch it several times in fact. Thank You soooo much from Bill Fischer
I'll join the choir, this should totally be a series! Could totally use one of these for FETs, linear + switching regulators, etc.
Hi Dave I agree with your comment on the LM741. I'm in the rail industry and change of electronics happens to move really slowly in Rollingstock down here in Melbourne. We still use the LM741H on Westinghouse and GEC train equipment from the beginning of the 80's. They are still used in power and traction control control cards.
Chips, dips and jellybeans.... sounds delicious.
21:33 the external compensation of the 5534 (which is, of course, a single op-amp) came in handy recently when I wanted to deliberately limit the slew rate of a signal. Add a comically oversized compensation cap - job done. Another notable mention for audio applications is the NJM4556, when you need a bit more output current, it might be worth considering.
For more info on the 5532 and its quasi successor the LM4562, read chapter 4 of Small Signal Audio Design by Douglas Self.
Do more of these! Great content and introduction to specifications of interest. Love to see this for FETs, BJTs, Diodes, Voltage Regulators, and Logic Gates,
Great video Dave. The series for jellybean parts sounds like a winner mate. Thanks.
My father was replacing a circuit on his Volkswagen Westphalia and it had an LM324 in it! Already had the part :p LM324 is a good one. Memorable pinout, single supply, decent bandwidth for hobby stuff.
Having the power and ground in the center helps with isolating the amps from each other.
Great idea for a series. One of the problems many of us have is identifying what I call “standard” go-to parts. To keep in stock for basic design use. I have a few that I’ve used before, but I like the “jelly bean”concept. Hope you do more of this.
I bought couple of hundred LM358s and NE5532s few years ago.. i'm never going to run out of op amps, those two really cover a wide range of needs. The 358 even has a decent output current. Got also a bunch of 072s but haven't found a good use for it, the 5532s do audio so well.. Billions of 072 are found inside millions of sound consoles.
My 2 rules of thumb when using op-amps: same inpedance on both inputs, resistor 1k between output and V-. This last trick turns op-amp in some sort of class-A amp, great for audio applications. I recently built a stereo headphone audio amplifier. LM324 for input stage, LM386 for output, applying those 2 rules gives excellent result
Some of my op-amps:
LM124 / 324 quad bipolar GP
LT1128 Ultra low-noise
LT1010 High current (150 mA) unity gain buffer (TO220-5)
LMC660 Femtoampere input bias CMOS
NE5532 Aimed at audio. Widely used in mixing consoles
MAX4328 Ultra-low input offset.
Thanks once again Dave!
Here is one more for the "top 6" list:
TLC27x
(1=single, 2=dual, 4=quad)
low Ib input bias current
Absolutely! The 271 can even be tuned for desired performance
555 timers!! I just designed a PWM by a LDR that actually works. 😂😂😂 Love the video Dave!!
I love this type of video, its the knowledge people don't tell you and you pick up from years of experience.
Cheer's for the OP-07. Invented by PMI, and it's successors the OP-77 and OP-177. Brings a tear to my eye having worked for them a few years before we were acquired by Analog Devices. Working for Analog Devices was also fun.
Good video, I love the 358 and 324's. Great parts. You can often design around the limits in many cases.
EE student here. It’s like wisdom is flowing into my brain lol I love these series.
I was with you all the way up until I suddenly got hit with a wave of nostalgia and a realization of how old I'm getting. The 741 was the OP amp you got from Radio Shack back in the day. There was only the 741 and the quad. I think I still have a few from decades ago lingering around in the original Radio Shack packaging. Love these jellybean videos! Keep up the amazing work!
My KOSMOS electronics kit I had as a kid had an 741 as the opamp.
I feel lucky to have gotten the last half decent generation (X1000-X4000) of those, before they dumbed them down into oblivion. What they offer today is a bad joke in my opinion.
Im born in 1985, but grew up on my fathers electronics books from the 1970s in my "formative years", so .. yeah, I feel pretty old, too.
I'm a EE in the USA , I've never heard the term "Jelly Bean Component" until I started watching Dave :)
I love the LM358! I was so shocked and pleased a few years ago when I realized its output can go to ground. It's been my go-to OPAMP ever since.
Great video series Dave! Really useful as someone who is self-taught on the hardware end and not always sure what's a good starting point when confronted with a massive list of parts to choose from even after filtering at a distributor's website
I'm by no means an EE, just a hobbyist but I'm also an electrician and do a lot of building automation. I felt really happy when I recently looked at some demolished equipment (laboratory fume hood controls) and knew what all the ICs were with the exception of the old early 2000s era microprocessor in each device. All TTL logic and cheapo quad comparators and op-amps.
Great top Dave, my favorite OPAMP the OP07 undoubtedly!
EEVBLOG school of engineering. your channel is the best campus for EE.
The TL071 has the offset voltage nulling pins, very useful in many cases!
I love this video. I know the jellybean op amps are already known by almost every engineer, but it is nice to look at the datasheets again. Looking at this makes me want to assemble an analog computer. Possibly a future project for my free time.
The NE5532 is my favorite, single Op-Amps in a package are used when you are amplifying minute dc Voltages, because they have offset pins to zero out the error, when an Op-Amp is operated at an elevated Voltage (and the other unit in a package would be unable to be usable) and to replace higher noise and drift single package. Like a NE5534 replacing a LM301. Ron W4BIN
Re the 5532 everyone has listened to the result of hundreds of them in domino style stuffed into the desk of every studio since about 1975.
Yup. Paul McGowan of "high end" audio company PS Audio was sure he could improve the performance of a vintage Studer desk he'd acquired by swapping in "better" opamp chips than the NE5532s with which it came -- he had to confess that there was no improvement whatsoever and had his techs reinstall the originals.
@@editorjuno Interesting. Thanks Bruce.
Spot on!. This is info every circuit designer (Pro or not) should know!.
Coincidence. I've just recently started to build some LM 386 based mini guitar amps. 386's are amazingly versatile. With the right components you can make them sound like a Bassman or Ampeg or Marshall 800, Hi-Watt, you name it. I make them with or without volume or gain controls. I also get the N-4 version and run it at 12V instead of 9V. Bridge the 1 and 8 pins with a 10uf cap for a gain factor of 200. Or put a gain pot in there.
The smallest and most simple one is 1 cap, 1 resistor, LM386, input jack, output jack. And that's it. About the size of a pack of cigarettes.
It sounds awesome, they all do.
Also making a _Walking Talk-Box._ A cut down funnel attached to a 3in speaker, tubing up to your mouth, clips onto your belt. Battery powered so you can walk anywhere and play anything from STP to Peter Frampton.
The 386 is both a solid, basic part and, as you said, very easy to work with. I expected Dave to include it in the video. But, the more I thought about it, the more I realized that it was probably excluded due to the fact that the 386, though popular, isn’t really a general-purpose opamp, in the same sense as the others mentioned. Have fun with your new amp.
Interestingly, I’ve also recently started designing my first audio amplifier. Although I have designed plenty of digital and analog systems, I have never I had the need to include anything more than a buzzer in any of them. I suppose that’s because of my work centers around embedded devices, not ones that get all that close to humans. Audio is a whole, fascinating world unto itself. Should be fun.
@@a1nelson Probably the very best feature of a 386 for guitar amp is the fact it nearly duplicates the tonal characteristics of a tube-driven amp. Specifically, the dynamic sensitivity while playing softly has a cleaner sound and articulation, just like a tube amp. And the harder it's played, the distortion comes through strong, just like a tube amp. And does it without increase in gain or volume. And that's a big issue because all other solid state circuits can't replicate that. Ok, maybe some, but not at such a meager price point.
Thank you for your reply
I've been playing for 6 years now, and I'm also an ee junior at uni, so im gaining traction on learning more about this stuff and I'm curious how the op-amp type changes the sound? I've pondered the idea of trying to build my own amps, to experiment with sound, especially for my bass. I'm gonna have to try the simple amp idea you mentioned. thanks!
I think the LM 386 can have 10% distortion, but maybe you desire that.
@@bobbyquinn4952 Hmmm........bass...that's a good Idea. I'll have to do a little homework and see what the signal frequency parameters of a 386 is capable of and what the pickups on a bass produce. You would certainly need to scale up power, and maybe two 386's to help spread the load.
Too bad that the LMV358 isn't readily available in a DIP8 package though.
Curious: are you referring to their potential use for one of your videos? (Which are absolutely fantastic!) So, are you talking about breadboarding? Or, hobbiest market aside, are there modern applications where a DIP is still the best option?
Personally, I try to leave human-facing connectors, power MOSFETs and higher power relays as the _only_ TH components in my designs (plus some specialty parts, like transformers). Maybe I’m taking SMT too far?
Edit: added clarification re uncommon (for me) TH parts.
What is your favourite (not necessarily) jelly bean op-amp?
@@a1nelson Yeah -I was mainly thinking of breadboard and prototyping where TH components are generally easier to work with.
@@spakecdk One that I have been using a lot more lately is the LMC6482 - mainly because I like the r-r input and output, and I happen to have a tube of them. But, they aren't cheap. There are likely better/cheaper solutions for R-R in/out in a dip package with a wide power supply range available now - haven't looked in a while... The MCP6002 is a nice choice for low supply voltage applications.
Just use a MCP602 instead for prototyping.
Hey Dave, could you make a "101" video on unity gain stability you've mention @ 4:13?
We've had those lectures at the uni, but we've never been given an example of an application where you'd need to take care about the stability while designing circuitry; i.e. how you could accidentally make the amp go haywire. I only remember it has something to do with a pole at 0 dB.
.... students are fed too much data in a way too short period of time, often without having prerequisite knowledge, and they're fed raw, inapplicable data.
Positive feedback with a gain greater than one. It will oscillate.
the most simple practical rule is to break the loop, insert a (virtual) frequency response analyzer there, and make a bode plot. Look at the bode plot. Find the point where phase shift hits 180deg, and see if the amplitude is reliably below 0db at that point and beyond. If it is, the loop is stable. Additionally, note the phase where the amplitude crosses the 0db line - (180-that) is the phase margin. If phase margin is low, the system will likely ring for quite a while before settling.
This is not the comprehensive stability criterion as far as i know, as some weirdo systems that fail it can still be stable or some systems that pass may still be unstable, but for opamp feedback design, it's a reasonable rule.
As for poles/zeros, it is tricky. I recommend watching a series of control systems lectures by Brian Douglas for that. BTW, poles and zeros are specified by their complex frequencies (s-plane coordinates), i don't think they have any dB property (i can't say i'm fluent with this stuff, though, maybe i'm wrong).
I like your Slavic Symbol.
Oh YEAH! This is a good topic! Thanks Dave. I'm to the point in my education where I can start to disagree with you on a few things but these videos are exactly the sort of thing you don't learn in school!
Lovely, this kind of things they don't and can't teach at uni.
Can't? Why not?
@@j1952d Possibly because you're looking at years and years of experience in sifting through potential parts to find one best suited to do a particular job, often in tension with a commercial need to keep cost low and availability high. Most folks with that sort of experience aren't going to be teaching at universities.
@@RexxSchneider Oh, yea. When I went, there were teachers with lifetimes of experience that they were prepared to pass on. They're all dead now, and I'm not far behind .
@@j1952d I didn't have quite your experience. When I was at university in the late 1960s, the head of the faculty was Professor AHW Beck, who wrote one of the definitive textbooks on "Thermionic Valves: Their Theory and Design". This was at a time when I had moved on from germanium to silicon transistors and was experimenting with early opamps like the 709. Nobody on the staff had seen an ic opamp and I remember having a crowd of demonstrators around one of my prototypes of a primitive capacitive proximity sensor. I was able to get a lot of advice on ECC83s and EL84s, but not much on which opamp was best suited for an application. :(
The Art of Electronics (textbook from Harvard professors Paul Horowitz & Winfield Hill) has tables of jellybean parts for all the common component types. Plenty of other universities use it.
I love 5532s for audio, it's my favourite sounding
I'm old enough to remember the LM709 as a jellybean op amp!
a)
b)
c)
4)
Stuff like this is why we love you 😂😂
I hope you do more videos like this.
My jellybean OP amps are exactly what you covered.
My favorite jellybean part, is the TL431. It's in almost every SMPS for a reason.
Haha! The TL431 is a not a bad op-amp if you can live with the fixed (and precise) offset voltage!
Love TL072, many analog audio devices (from home up to stage equipment) used them.
Important note about LM358: it's not rail-to-rail. Supplying it from 3.7V you can expect only 3.7-1.5=2.2V output max.
Dissapointing to learn this after I made a USB powered voltage controller, it does 0-3.5 volts output. Chose the 358 because I didn't need a high speed (audio rate) response time.
Another famous weak point of the LM358/LM324 (and long-time regret of its designer) is its output stage, which cannot switch from sourcing to sinking without crossover glitches... if you find you must use one for audio (despite the awful noise level), Class A bias the thing by all means.
I've come to think of "rail-to-rail" as a marketing gimmick, those words are mostly lies if you read the datasheet carefully.
Ask me how I know :P
@@charlesdorval394 There’s also input rail to rail and output rail to rail. I think most rail to rail op amps can’t drive a low impedance load.
I really like this format. Maybe more like this, jellybean mosfets ect. It's Definitely more advanced than "what is an Op amp" but really helpful for people who have already done school but don't have a ton of experience under their belt.
Loving this Jellybean series! Would be great if you included a comparison table with description, benefits and cost per dozen for each part covered. Would save a lot of time for a lot of people
Super helpful! I think I've unintentionally kept the LM358 and TL07x in my habitual go-to op-amps, but didn't realize they were so ubiquitous.
A series about "TOP 5/10" most used components would be so useful for a lot of people.
More of this stuff Dave please. I really enjoyed it. Bring more sheets of other chips and comment on them. Not just OpAmps. More talk about the ground reference and rail to rail headroom. You the man!
Another good reason for using dual over single OpAmps is logistics. As a manufacturer, or even a hobbyist, it's much better to stock a single part than try and maintain a stockpile of two different part numbers.
"One good OP amp is worth 1000 microprocessors" - Bob Pease
Great idea. This could be a series...Will be...Is?
That's the idea, if there is enough interest.
LMV321 comes in (at least) two footprint versions of SOT23-5 and they are NOT interchangeable. Check your files before soldering, some vendors sell only one variant, some have two
You might want to mention LM358 has issues with crossover distortion on the output, and so the output should be pulled down sufficiently for certain applications, especially AC coupled outputs, where the output both sources and sinks current.
Yep I have a few LM324 and 358 just lying around!
Tl07# is generally my initial go-to opamp for the majority of my applications, and then I adjust as needed :P
yes please a list on all jellybean components (bjt, mos & jfet, and other assorted ICs)
Before watching the video, 4 in the list are obvious to me:
LM358 (general purpose, low frequency, low cost, single supply), OP07 (low offset voltage, 'precision', great for raw sensor front end), TL071 (high input impedance, okay for audio), NE5532 (standard audio opamp, relatively low distortion in audio band for the cost, quite alright at anything). I think these will be the same for 99% of engineers.
As expected, all the celebrities are in the video.
At 21:35 though, the 5534 is a single part, not quad! It is also in PDIP package, but only a single opamp because it exposes the compensation capacitor pins, and is only stable with feedback gains of 3x +. On the plus side, it does have sliiightly better noise though. But, if you want better noise than the 5532 you'd be looking at other parts anyways, so kind of an obsolete opamp at this point in time in my opinion.
I was taught opamp theory in the Navy in 1974 by a Senior Chief Petty Officer who was in the design group to the Grumman A-6E Intruder fire control systems. Milspec ceramic DIP cases and "flat packs". I still use these "jellybeans" along with the venerable 555 timer.
I just learned something today: Op-amp have more tendency to oscillate at UNITY GAIN! Who knew? (apparently, I didn't) (4:12)
Great information, thanks for posting Dave!
I was taught using 741's... all I used in two years of analog classes. I did use some others (rail to rail capable) in later years but at least in education, it seems the 741 still has a use.
Wish I had this video just few months earlier! I figured I'd probably do away with a jelly bean product - rail to rail was ideal but a bit of headroom wouldn't hurt, I thought. Unfortunately local provider showed LM741 as the most popular... As a result, the knob was only effective on like lower 30% but the difference was *just* enough to tell, so user would have the impression that they had control over the whole range. It wasn't any critical part, so that was enough 😁
The input of TL07x goes up to Vcc, that's why it's useful for high-side current sensing.
Dave, I hope to be buried with an LM-324!
As a junior engineer in the late 70's, early 80's, I both cut my teeth and quickly adopted the LM-324 and used it for almost every product that I was commissioned to design for one really unique reason:
To me, the real power of the LM-324 is the fact that you have 4 OP-amps in the package which allows you to effortlessly increase the performance of many types of circuits (especially precision comparators) by using 2 of the OP-amps as buffers by tying the (-) to the output and using the (+) as the input. This allows you to not load, say a comparator, by isolating its input from the external circuit. Conversely, you can use the second buffer as output-isolation from the load, so again, you don't upset or change the operating point of a comparator. Additionally, you'll find using these buffers (in & out) will aid in the precision of your 10X, 100X etc. amplifiers. So, this is why I completely skipped over the LM358 because I just got such better performance with not a whole lot of additional real-estate! Anyway, we're definitely on the sam page for 'Jellybean' Op-amps! Thanks Dave!
I just built up a bunch of boards using an LM358. In college, I also liked to use the LM324 op-amps.
thanks you remembered us our first projects , now we need some jellybean transistor and CMOS or TTL logic gate and counter and processor like old microship or atmel finely we mustn't forget the famous 555 and the Tl 431 .
Never knew about lmv358 and lm358 noise specs, thanks Dave!
My biggest rant is on data sheets is when they list both an A and B version of the part but fail to highlight the differences forcing one to do a line by line examination of the data sheet in an attempt to figure it out.
I wish I could give this comment 1000 thumbs up. The odious practice of not proving “delta” information (between versions within a datasheet, or between “sister” models in a series) drives me nuts. I’ve seen some where people gave up and asked the manufacturer (in their forums) what was different between two versions, and the manufacturer wrote back with the difference - _which wasn’t in either datasheet._ 🤬🙄
And with the situation in the semiconductor business over the last what 15 years, with mergers and acquisitions - you can't trust one manufacturer over the other to have proper good datasheets either. (Not mentioning different segments inside the same company)
@@tookitogo Connectors are so bad about this. Why do these two parts look identical but have different part numbers? Well one has little drain holes in the bottom so they don't collect fluid when the board is washed after assembly. Also, we don't make that kind anymore. Of course!
In want mooooooorrreeeeee!
My favorite OPAMPs:
Precision stuff, Rail-to-Rail, low voltage(5.5V max), low power consumption: OPA2333 = TLC2333
Not so precise, Rail-to-Rail, low voltage (5.5V max), low power consumption: OPA2348
Very low bias current (0.2pA), Rail-to-Rail, 3-16V Supply, fairly low power consumption: LMC6482
My fav cheap low noise Audio OPAMP: RC4580
My favorite higher current and capacitor driving OPAMP: LM8272 (Good for Rail Splitters)
My favorite higher current driving low voltage OPAMP: OPA2322
My former favorite +-5V low power audio opamp: TLC2262/TLC2272
Just to name a few, got so many more...
Thanks for this tutorial! I didn't realize the 4558 chip was essentially a pair of 741 chips in an 8 pin DIP. For audiophile work the Burr-Brown op amps are nice. In 8 pin DIP the OPA2134 is something to consider, although there are newer ones with even better specs available in SMD packages only. It's tempting to get an adapter to try plugging an SMD variant into the DIP sockets on one of my guitar pedals just to see how it responds to being driven into the nonlinear region.
I’ve used OPA2134s to upgrade TL072 and they work and sound pretty good to my ears. I’ve been using them in my Soundcraft 2400.
@JM Coulon If the design is an EQ without coupling capacitors, the bias current of the 5532 can thrash the track, creating the dreaded permanent scratch. That is really the only reason I use it, otherwise I just use the 5532 if input bias current is not an issue. I’ve found it to be the best compromise of noise, cost, and power consumption if low bias current is a hard requirement.
I always thought the LM1458 was the disk version of the LM741
*dual version
@JM Coulon Yes, the Burr Brown stuff is pricey. $4 or so for a single OP2134, which I liked only marginally better than a CA32040. The 5532 is also a solid choice. Could you name a couple of the current audiophile op amps for me? They are probably all SMT-only.
Great video. Would be nice to see more of these for other areas. Would also be nice to revisit a series of go to circuits for us beginners.
LM 741 is the craziest, but still my favourite. Especially LM 741H for high temperature applications.
Thank you Dave. Very interesting, as usual!
CMOS FETs are the beast techhnology, no question. But regular BJTs will serve in a pinch
Huge thanks! Please do more.
Watched this video to find a better op amp than my "go to" LM324.
Left with LMV324. Good enough, I guess, and I don't have to change anything else. Earned a like and maybe a sub.