"It is just a zener diode, a resistor, an emitter-follower, some business (the other 20 transistors in very advanced configurations), and bob is your uncle! no worries" - Dave
if all you wanna do is show the basic working principle of the part that is a totally valid approach. also it would completely blow up the scope of this video to explain them. not to mention that all the protection features that the datasheet claims gotta be implemented somehow since the sure do not represent the inherent features of some semiconductor configuration....
@@urugulu1656 Yea, I think it was pretty clear this was not a critique on the subject's scope. It was highlighting the endearing presentation style that Dave uses. Tell me you don't get it without telling me you don't get it.
Dave, these JellyBean vids are GREAT, some of your JellyBeans were not common when I was new and studying to be an EE in the 70's they were more common when I went on to graduate study!!! Thanks
After spending the last three decades in the ethereal software world, in my retirement I have sensibly returned to the tangible world of electronics hardware. This series is exactly what I need. A perfect refresher. Also, it is comforting to see that most of these parts are the same as what I used a lifetime ago!
78 and 79 regulators have made power supply design so easy that they've overshadowed newer LDO and generally higher spec devices. Being around for so long and making life easy makes them worthy of nomination for the award for device of the century!
This series is great Dave. Jellybeans are exactly what I find when desoldering old boards. Us poor electronic nerds thank you. No component shortage is gonna stop me. Next, jellybean microcontrollers?
STM32 and clones are probably the closest you can get to an industry standard microcontroller, and yet, we see what component shortage has done to that line.
@@damny0utoobe and the now old stm32 F103 model is the only one quite a few times cloned, and rather poorly (even the GD32) when digging into advanced features and compatibility.
Takes me back - great series on these old 70's/80's chips. They made it easy to actually make something to sell - to start a company early (guitar effects/synthesizers). Many thanks - great channel. Cheers from Florida, USA
KEEP ON ROLLING with more videos like this!!! Top transistors, top logic families, etc!!! Please!!! Such info is critical for inexperienced engineers! This is MUCH more important than teardown videos! Anyone can do a teardown video (less experienced comments on design), but very few on YT can present such knowledge as you do! So spend less of your resources on teardowns please, and more on engineering tutorials!
in terms of not needing to put smoothing caps on the output of some of these regulators... Steve (SDG Electronics) did a recent video exploring this. With some test boards and different regulators etc. adding excessive capacitance where they are not needed is not always "harmless"... it can actually made the performance worse in some cases. Recommended to watch his video (SDG #212)
Will have to watch it. Yes, it's sometimes not that easy. Look at the TL431 stability app note linked in the description as a classic example. There is a unstable dead bad of capacitance values.
Oops ... I learned that if I correct my English by editing after you give a "heart by EEVblog" that I lose it. Thank you for the great series and cheers to you ...
I once saw a mid 1980s Japanese built 7805 in a Nintendo oscillate (best I could tell that was the problem anyway) with a failing output cap, it had a small ceramic right on the output and another electrolytic on that rail a few inches away, just the nearest electrolytic (so near that the 7805 cooked it) had gone high ESR and that was enough to get it going and cause weird video distortion.
If my memory serves me right, 78xx and 79xx should have the ceremonial 0.1 µF decoupling cap from input to ground, and from output to ground, and preferrably have a properly sized and designed ground to avoid various issues.
I might have seen one once, don't entirely recall. I experimented with that a few years back trying to make it happen for a video, couldn't do it. I always add the caps just in case.
The jellybean series was quite useful, because new designers and hobbyists do not know much about established standard components. Please, continue this series to cover other fields. E.g. transistors, mosfets, interface chips, ADC/DAC.
35 years ago at college as a project the whole class had to build fixed voltage bench PSU’s. +12, -12 and 5 volts out. First time I’d used a 7805. Used it to power all sorts of kit over the years.
One other 1970 era part I'm glad you mentioned is the TL431. It was designed as a voltage reference, but today finds heavy use as the feedback element in switching power supplies. It's made by the billions and most EE's haven't heard of it if they don't work on power supplies. The relevance of this video is it is a good complement to a LM317 / LM337 to get a far more accurate output voltage. You can get sub ±1% accuracy by pairing the two parts.
I had my electronics students built a half wave power supply with too small a filter capacitor. Then we added a 7405 voltage regulator. They could see about a 1 Volt peak to peak ripple on the input, but couldn't measure the output ripple because the noise was greater than the ripple. It was an impressive lab for my students. I love the 78xx regulators.
Stability is important! Remember . . . in the extreme case, lack of stability can apply the full input voltage to the output terminal. 12V input regulated (with stability) to 3.3V logic power can potentially blow up your logic chips in an unstable (or slipped scope probe) condition. Usually the highest tech (smallest device geometry, probably memory or MCU) part will act as a crowbar device to protect the rest of the parts on the board.
Sorry for my English! At the moment, we have a PSU, which so noisy, some MOBO can't even boot with it. The noise can be seen an oscilloscope in 10V/DIV easily on the 12V rail.
To protect expensive components from a bad regulator I use TVS (zener) with a PPTC fuse for voltage & current protection. if the Regulator exceeds the TVS limit, it clamps the voltage and the PPTC fuse limits the current to prevent the magic smoke genie. This can can also be used on the input (Source) side to protect against over voltage and reverse polarity on the regulator. FWIW: I never use linear regulators going from 12V (or higher) down to 3.3V\1.8V. I use simple switches which can operate with input supply voltages up to 60V (or 40V depending on which ones you use). Simple switchers can also supply a lot more current & are much more efficient. I use a PI filter on the output to address switching noise.
Thats why Im using zener diodes as protection. In case of overvoltage some zener diodes do open circuit and some of them will make short circuit. Im using this with at least two parallel zener diodes.
@@guytech7310 Yes, I only use linear regulator for low-current applications. My stability warning also applies to switchers. An extreme case can be accidently activated if your probe shorts the feedback pin to ground . . . the output goes to maximum and things melt/smoke. Other user's comments about TVS or Zeners (which are "guaranteed to fail in short circuit mode") address my paranoid protection approach to address the stability failure mode.
These videos gave me so much clout in my lab. i'm a software guy, but even having a passing knowledge of standard parts got me a lot of street cred with the hardware guys. EEVBlog is probably responsible for me ever getting a job at all lol
5:52 I can imagine Dave answering a question on his exam at the day on how a 7805 works. "Well it's just a zener, a resistor and an emitter-follower doing some business and Bob's your uncle".
I've seen a 7809 regulator ran at 38 volts input voltage in a commercial product, and it held up fine for 10+ years, but the 7909 on the negative rail died, but still works for around 10 minutes until it goes open
I picked up an old powered breadboard - one of those big brick jobs from Global - that's got 5-way banana jacks for ground, +5, +15, and -15, and opened it up to check if it was all 78xxes or if the bipolar were adjustable, and it turned out the answer was yes. Plain 7805 circuit on the 5v supply, and 7805 and a trimmer on the +15 and -15. Definitely a surprise.
The longevity of these linear regulators is fantastic. Many parts come and go, yet these are still here. I design power electronics systems for a large multi-national, and one of our requirements is ± 24V for current sensors. The flux gate and closed loop hall current sensors up to about 300A can do ± 15V supply, but the big stuff into the high hundreds and thousands of amps requires ±24V. What makes it extra complicated is the hall and flux gate current sensors have a tight specification for the power supply rails. Today we use paired 7824 and 7924 regulators to post regulate the switcher supply and get our low ripple, high accuracy ±24V rails. Sadly, the 7924 has gone sole source and it's making the supply chain people nervous. We've considered using the LM317 & LM337 paired. I have an inclination that the LM337 will be longer lived, but... it is merely a hunch. In some ways I have the gut feeling that the LM337 may be the B52 of chips, and make it to the 100 year mark. Sadly, based on the fact that it is merely a hunch, we're going to high speed monolithic switching regulators. We'll see in a decade if it was the right call.
Top 5 jellybean rectifiers next. (I can actually only think of 3 - 1N700x, 1N4148/1N914, and 1N5817. A discussion of the dizzying array of alternatives might help some of us understand why there are so many different devices designed to do only one thing in almost exactly the same way.) If you need LM79xx, be sure you have some electrolytic or tag tant capacitors handy, as these regulators require much more bypassy capacitance than their positive voltage counterparts for stable operation. On that note, I've noticed several designs using LM78XX and LM79xx that have the same capacitor values on both positive and negative rails, presumably for BOM consolidation.
Really loving this series! I'm hoping for a massive video on the end so we can have some kind of general purpose Jellybean Box that can cover 80+% of the things people want to build.
The two jelly bean solutions I am frequently looking for are: 1. A simple "neutral reference" between rails, to avoid paying an op amp and 2 resistors to generate this, while losing sleep over whether there are any circuit conditions that will overload the op amp once the product is put into service. There ought to be a regulator designed to do this, but I haven't found it (yet). 2. Every design seems to involve switching analog signals around. A bilateral switch like the 4066 works, but it's made of super-static-sensitive CMOS. A 74HC4316 or 74HCT4316 seems to qualify, but they are certainly NOT available like jelly beans. Steve
The 431 is basically the entire reference and pass transistor control block of a linear regulator without the pass transistor. This is why the 431 is used so much in mains switching mode power supplies, as it can drive a linear optocoupler to close the feedback loop for the regulated output.
Yes, more top 5 things: Top 5 useful beans, ..... As an young player I get drowned on all the specs when I try to select components. I really like how you highlight the most significant spec limits.
The opamp in the LM431 isn't acting like a comparator switching its output on and off. It is acting like a linear high-gain differential amplifier with feedback.
Just ordered a bag full of 431s for the exact purpose described here. I have a lot of retro computers, from the Mac Classic and original PC to current gen. I always replace the capacitors in PSUs older than about 5-10 years, the threshold depending on how much I trust that brand's parts selection. Well, one time I didn't. It was a Delta ATX supply from the early early P4 days, with a heavy 5V rail and the 6-pin Aux connector for Athlon boards. I got three of those supplies, NOS, in one bid. I serviced two of them, and got lazy with the third. I ended up installing it in a box I use to image hard disks (it has a VIA Epia mini-ITX board and both PATA and SATA hot-swap bays.) One day it had just turned itself off, and I figured maybe we had a brown-out. Two months or so later, I tried to turn it back on and it wouldn't budge. The KVM it was plugged into was also acting kind of weird, so I pulled the case onto my bench. It was making a squealing noise, but some of my PSUs do that. I measured the +5VSB rail to see if it was working at all, and it was measuring FIFTEEN VOLTS. Yikes!! I figured that motherboard was toast, but it's actually alive and (I think) running fine. Can't believe that. Anyway, I pulled the cover on the PSU and the caps right next to the opto-isolation circuit were bulging from the top. Of course it's the caps. It's _always_ the caps. There was a 2222 transistor and a TL431 right next to it, so I figured I would go ahead and replace those too. I'm hoping to get some time today to replace parts and see if I can nurse it back to health. Shame to waste a good PSU on some dodgy caps... Always replace the caps! Even if they look OK, if it has a few miles on it, you're better off safe than sorry.
NEVER depend on internal overtemperature shutdown (ALWAYS mount on sufficient heatsink to avoid the problem in the first hand)! Some of these kick in at 175°C TJ, etc. Ouch!!! And some datasheets clearly state "if you repeatedly activate the internal overtemp shutdown, you are degrading the chip"! And what will chip performance be right before shutdown???
Great video, tons of practical information! Can’t really use a shunt voltage reference as a power regulator unless you’re prepared to run its current at the maximum current draw of your circuit. This means not low power. Love those TL431’s, even saw one with the opto isolator built in.
Brilliant. More of these great vids the better. I often only need small amounts of items at a time so jellybeaning is going to be a great way of bulking out an order to avoid extra P&P costs, almost free components if you want to look at it that way. 👍 Thanks.
Using TL431 with LM317/337 together is great for low noise and stable linear power supply. Only critical point to achive that is to right matching capacitors in PSU design.
Hi Dave, love the channel. I built a variable voltage PSU, used it to charge a battery but it popped, magic smoke. Apparently, the work around is to insert a BY127 or similar between the output leg and positive output terminal of the PSU, being constructed.
I was a junior design engineering back in the 1980s and all if these parts were industry standard back then. It is staggering they are still being used.
Fantastic series Mr Jones keep em coming! come-a-gutsa Verb. (Australian colloquialism) To be thrown off a moving object at high speed, arms and legs flailing, landing on a hard surface and sustaining multiple wounds and broken bones.
I have notice that a lot of people pick up LM317 with around output current up to 1.2-1.5 A. But what most people seems to have is LM317 from China with much lower max output current. I have a bunch at 0.1A. Took me as an young player a while to realise.
10:50 To use the 317 at low load currents, replace the 240 (R1) resistor with one of 120 Ω and choose the value for the voltage adjustment resistor (R2) accordingly. In this way a minimum of 10 mA current will flow at all times.
I love this series and hope to see more like this, I hope Dave makes a similar series for design. something like favorite/jellybean topologies to explain those most useful and recyclable circuit designs and blocks.
Oh yeah, the 2N2222 and it's variants have been around for ages. The first time I saw a circuit for it was in the second or third "Book of Radio and Electronics" by Alfred Morgan.
Thanks a lot for this video series. It might be useful and very interesting to overview jellybean transistors. Maybe split in some categories like BJT, JFET etc.
Thanks for this series, it is very informative. You did not mention that the 1117 has a maximum input voltage of 15 volts. Also there is an LM317HV available from TI that has a maximum input voltage of 60 volts. It is only available in TO-3, TO-39 and TO-220 packages. There are a few 5 volt regulators with 50 volt input but they are hard to get. A video on comparators would be good.
@@invisiblekincajou The crux is in power-on, where the output capacitor may still be at ~0 V while the input already is near nominal value. So unless you can bring up the input voltage very gently, I wouldn't be betting on that!
The 1117 is the only regulator I have around for general use, I use a ceramic output cap with a 1R series resistor to meet the ESR requirement, better than using tantalums.
I HAVE had the output of 7805's go into osculations when the load cap is open. I don't know if it was brand of chip or what. The output measured around 4v, if I remember right. The scope showed a large amount of noise on the 5v rail. I replaced the cap and everything started working right.
I've seen it to, didn't have a working scope at the time, but like you said replaced the first electrolytic on the 5V rail (which had an ESR of around 140 ohms) and the unit behaved again.
µA723 variations are quite common in older adjustable power supply designs. While not super easy to use and practical by modern standards, they are still good for low-noise supplies, beating even a bypassed LM317 and making it close to LT304x terrain.
My favorite to. I used one to design a 1500 Volt Regulated Screen Grid Supply for a big mutha transmitting tube. As Aretha would say the supply is Rock Steady !
The LM340, was National Semiconductor's original number, for their competing part to Fairchild's uA78xx. But that was back in the 70s. Its all evolved into the generic 78xx. The stabilizing caps were needed on older parts. I had them oscillate, without caps, back in the early 80s. They DO change the part design, as the years progressed, without changing the number. Any 78xx built in the last 25 years, the caps are optional.
Quite a few years ago, I wanted to have very long battery life for my back-then-new Asus EEE mini laptop. I had an external 12v sealed lead-acid battery (as in a home alarm system) and I used a circuit that I made myself with a LM317 regulator, to make the 9.5V the laptop charge port asked for. It worked for a short time, then the output shut off. My problem was that I didn't know about the minimum 3v difference between input and output voltage at that time!
These are great, and if you run out of jellybean ones I'd like to see some videos with random deep dives into familiar or obscure component type hunts for the ultimate specs.
1117 regulators are common because you can take USB 5v in and lower it to 3.7v to power 3.3v (and lower) microcontrollers. I see them all the time in HDMI adapters, ARM devices, microcontrollers, etc.
Reverse Protection Diode for 78xx, as at 15:20, can be eliminated in many cases by using 79xx instead, and regulating the 'ground' rail instead of the positive rail.
Dave, these "top 5 jelly bean" component videos are very interesting and useful (& educational to us "young players," regardless of our ages). Also, the thumbs down function is still working at the moment (2123 EST, 18 Nov 2021), so _maybe_ the backlash had the desired effect?
Hello Dave. I have to be honest here.. This is one of the videos i had to pause and start it again.. I'm planning to make power supply for my precision voltage source and i wanted to use Lm317 but as i see it it looks like there are other options i didn't consider before.. I think i will try to use that LM431 or maybe that Ref 1 ..I have to study datasheets just to be sure..but Thanks for video. I hope you will take more time to go over all these components.. Things like Opamps ..as you would say that could be trap for young players..there are too many options ..and it is had to say wht is the best.. Haha now i noticed your video had over 44 minutes..i didn't even notice that..I had to save it for later use..for sure !
Simple switchers like the LM2679 or LM2596, More amps, more efficient, higher supply voltages. At this point I only use Linear regulators for low power 3.3V\1.8V needs, or for very low power, ultra-low noise (ie ADC, LNA, precision voltage references). Issue with TL431 is possible drift from temperature, unless you use temperature compensated components. A lot of devices can be under significant temperature swings from 0C\32F to 50C\122F.
A major difference between TL431 and LM4040 is that the 431 family has the internal bandgap reference voltage referred to the anode while the 4040 family have the internal reference referred to the cathode. Makes the 431 a bit like a high gain NPN BJT with a very precise Vbe and the 4040 a bit like a high gain PNP with a very precise Vbe!!
Hi Dave, i really like the jelly bean series. so i say keep them coming! maybe you can also make a serie for different applications/uses/possibilities for these jelly bean components. that would be awesome!
10 million jelly points goes to you.. Great video and presentation. You know we all have those cheap chinese V/A meters in our attics. Lately, I realized they display different measurements. Since you have covered voltage precision components in your video, wondering if you care to make another video about calibrating multimeters/AM meters DIY version. Cheers,.
I've thrown in 2x the LM317 into every toolkit (parts, tools, breadboards,...) i made for our interns/students, but until now they all were too lazy to build it onto their breadboards and all used external lab power supplies for their different voltages 😅 Now thats fine for first setups looking for smoke, but i find it a lot nicer if you only hook up one supply.. but building a board without cross wiring all and building a ratsnest seems to be the starting hurdle.. which makes it extremely hard to then find the wiring errors and issues.
IMO, the craziest thing about the LM317 is that it has no ground pin: it's adjust pin is 1.25V relative to Vo (1.25V/240ohm = 5.2mA, so Vo = R[adjust-gnd] * 5.2mA + 1.25V). So with a maximum input-to-output voltage of 40V, you *could* get 5V out of a 45V source, or a maximum 37V from a 77V input.
At very low efficiency unless your load is very tiny. Linear regulators basically are just an adjustable resistor to create a voltage drop to the target voltage. better option is to use simple switchers like the LM2679 or LM2596 for large voltage differences between the supply source and the output voltage.
I'd add LP295x for low-power applications
i believe you got interesting story about it
7:00
The great Mike has spoken.
Yep, video was already way too long.
@@EEVblog references should have been on separate one so you can cover the subject of regulator a little better imho.
"It is just a zener diode, a resistor, an emitter-follower, some business (the other 20 transistors in very advanced configurations), and bob is your uncle! no worries" - Dave
Trivial stuff for Dave hahah
if all you wanna do is show the basic working principle of the part that is a totally valid approach. also it would completely blow up the scope of this video to explain them. not to mention that all the protection features that the datasheet claims gotta be implemented somehow since the sure do not represent the inherent features of some semiconductor configuration....
@@urugulu1656 Yea, I think it was pretty clear this was not a critique on the subject's scope. It was highlighting the endearing presentation style that Dave uses. Tell me you don't get it without telling me you don't get it.
Dave, these JellyBean vids are GREAT, some of your JellyBeans were not common when I was new and studying to be an EE in the 70's they were more common when I went on to graduate study!!! Thanks
After spending the last three decades in the ethereal software world, in my retirement I have sensibly returned to the tangible world of electronics hardware. This series is exactly what I need. A perfect refresher. Also, it is comforting to see that most of these parts are the same as what I used a lifetime ago!
78 and 79 regulators have made power supply design so easy that they've overshadowed newer LDO and generally higher spec devices. Being around for so long and making life easy makes them worthy of nomination for the award for device of the century!
This series is great Dave. Jellybeans are exactly what I find when desoldering old boards.
Us poor electronic nerds thank you. No component shortage is gonna stop me. Next, jellybean microcontrollers?
do those even exist? I mean it all comes down to programming and companies such as Arduino tried to solve this problem of integration.
@@D4no00 yes, 8051
STM32 and clones are probably the closest you can get to an industry standard microcontroller, and yet, we see what component shortage has done to that line.
@@ignaloidas they're too modern to be jellybean
@@damny0utoobe and the now old stm32 F103 model is the only one quite a few times cloned, and rather poorly (even the GD32) when digging into advanced features and compatibility.
Takes me back - great series on these old 70's/80's chips. They made it easy to actually make something to sell - to start a company early (guitar effects/synthesizers). Many thanks - great channel. Cheers from Florida, USA
KEEP ON ROLLING with more videos like this!!! Top transistors, top logic families, etc!!! Please!!!
Such info is critical for inexperienced engineers! This is MUCH more important than teardown videos! Anyone can do a teardown video (less experienced comments on design), but very few on YT can present such knowledge as you do! So spend less of your resources on teardowns please, and more on engineering tutorials!
in terms of not needing to put smoothing caps on the output of some of these regulators... Steve (SDG Electronics) did a recent video exploring this. With some test boards and different regulators etc. adding excessive capacitance where they are not needed is not always "harmless"... it can actually made the performance worse in some cases. Recommended to watch his video (SDG #212)
Will have to watch it. Yes, it's sometimes not that easy. Look at the TL431 stability app note linked in the description as a classic example. There is a unstable dead bad of capacitance values.
These "jellybean" lists are great! I hope to see a decent list of these in your video library. Thank you for posting! 🙏
This series is very helpful! Please keep 'em coming!
Jellybean micros?
Yes more plz. Makes my life and job easier.
I had used a TL431 and a power transistor to make a very simple variable electronic load that self-compensates for the temperature.
Oops ... I learned that if I correct my English by editing after you give a "heart by EEVblog" that I lose it. Thank you for the great series and cheers to you ...
I've seen 78xx regs oscillate with no output cap
I've seen 7805s oscillate when the output cap got the freezer spray treatment.
The circuit itself may require bypass caps.
I once saw a mid 1980s Japanese built 7805 in a Nintendo oscillate (best I could tell that was the problem anyway) with a failing output cap, it had a small ceramic right on the output and another electrolytic on that rail a few inches away, just the nearest electrolytic (so near that the 7805 cooked it) had gone high ESR and that was enough to get it going and cause weird video distortion.
If my memory serves me right, 78xx and 79xx should have the ceremonial 0.1 µF decoupling cap from input to ground, and from output to ground, and preferrably have a properly sized and designed ground to avoid various issues.
I might have seen one once, don't entirely recall. I experimented with that a few years back trying to make it happen for a video, couldn't do it. I always add the caps just in case.
The jellybean series was quite useful, because new designers and hobbyists do not know much about established standard components.
Please, continue this series to cover other fields. E.g. transistors, mosfets, interface chips, ADC/DAC.
35 years ago at college as a project the whole class had to build fixed voltage bench PSU’s. +12, -12 and 5 volts out. First time I’d used a 7805. Used it to power all sorts of kit over the years.
I would like to see a video on extreme components with unusual specs.
I could do this myself but it is more entertaining to see Dave do it.
How about that plasma rifle in the 40 watt range?
-Terminator movie from 1984 (if you didn't get the reference)
One other 1970 era part I'm glad you mentioned is the TL431. It was designed as a voltage reference, but today finds heavy use as the feedback element in switching power supplies. It's made by the billions and most EE's haven't heard of it if they don't work on power supplies.
The relevance of this video is it is a good complement to a LM317 / LM337 to get a far more accurate output voltage. You can get sub ±1% accuracy by pairing the two parts.
YESSS DAVE!!! Awesome idea, top jellybean components! Absolutely loved the op-amp one, keep them coming! You're a real treasure sir! 🙂
I had my electronics students built a half wave power supply with too small a filter capacitor. Then we added a 7405 voltage regulator. They could see about a 1 Volt peak to peak ripple on the input, but couldn't measure the output ripple because the noise was greater than the ripple. It was an impressive lab for my students. I love the 78xx regulators.
Stability is important! Remember . . . in the extreme case, lack of stability can apply the full input voltage to the output terminal. 12V input regulated (with stability) to 3.3V logic power can potentially blow up your logic chips in an unstable (or slipped scope probe) condition. Usually the highest tech (smallest device geometry, probably memory or MCU) part will act as a crowbar device to protect the rest of the parts on the board.
Sorry for my English! At the moment, we have a PSU, which so noisy, some MOBO can't even boot with it. The noise can be seen an oscilloscope in 10V/DIV easily on the 12V rail.
To protect expensive components from a bad regulator I use TVS (zener) with a PPTC fuse for voltage & current protection. if the Regulator exceeds the TVS limit, it clamps the voltage and the PPTC fuse limits the current to prevent the magic smoke genie. This can can also be used on the input (Source) side to protect against over voltage and reverse polarity on the regulator.
FWIW: I never use linear regulators going from 12V (or higher) down to 3.3V\1.8V. I use simple switches which can operate with input supply voltages up to 60V (or 40V depending on which ones you use). Simple switchers can also supply a lot more current & are much more efficient. I use a PI filter on the output to address switching noise.
Thats why Im using zener diodes as protection. In case of overvoltage some zener diodes do open circuit and some of them will make short circuit. Im using this with at least two parallel zener diodes.
@@guytech7310 Yes, I only use linear regulator for low-current applications. My stability warning also applies to switchers. An extreme case can be accidently activated if your probe shorts the feedback pin to ground . . . the output goes to maximum and things melt/smoke. Other user's comments about TVS or Zeners (which are "guaranteed to fail in short circuit mode") address my paranoid protection approach to address the stability failure mode.
@@guytech7310 well clearly you are putting every thing in a soldered shut tin can or you just dont have to deal with emc
These videos gave me so much clout in my lab. i'm a software guy, but even having a passing knowledge of standard parts got me a lot of street cred with the hardware guys. EEVBlog is probably responsible for me ever getting a job at all lol
5:52 I can imagine Dave answering a question on his exam at the day on how a 7805 works. "Well it's just a zener, a resistor and an emitter-follower doing some business and Bob's your uncle".
Haha, good one.
I've seen a 7809 regulator ran at 38 volts input voltage in a commercial product, and it held up fine for 10+ years, but the 7909 on the negative rail died, but still works for around 10 minutes until it goes open
I picked up an old powered breadboard - one of those big brick jobs from Global - that's got 5-way banana jacks for ground, +5, +15, and -15, and opened it up to check if it was all 78xxes or if the bipolar were adjustable, and it turned out the answer was yes. Plain 7805 circuit on the 5v supply, and 7805 and a trimmer on the +15 and -15. Definitely a surprise.
This series is so helpful to the beginner and hobbiest. BIG THUMBS UP from the Old Dart.
When you said TL431 I smiled from ear to ear, yes it's a great part. For some reason I've never used the 4040/4041.
Thank you for the series
I like these top 5 ones! Gives a good base for newbies to look in to with regards to some common parts.
The longevity of these linear regulators is fantastic. Many parts come and go, yet these are still here. I design power electronics systems for a large multi-national, and one of our requirements is ± 24V for current sensors. The flux gate and closed loop hall current sensors up to about 300A can do ± 15V supply, but the big stuff into the high hundreds and thousands of amps requires ±24V. What makes it extra complicated is the hall and flux gate current sensors have a tight specification for the power supply rails. Today we use paired 7824 and 7924 regulators to post regulate the switcher supply and get our low ripple, high accuracy ±24V rails.
Sadly, the 7924 has gone sole source and it's making the supply chain people nervous.
We've considered using the LM317 & LM337 paired. I have an inclination that the LM337 will be longer lived, but... it is merely a hunch. In some ways I have the gut feeling that the LM337 may be the B52 of chips, and make it to the 100 year mark.
Sadly, based on the fact that it is merely a hunch, we're going to high speed monolithic switching regulators. We'll see in a decade if it was the right call.
Top 5 jellybean rectifiers next. (I can actually only think of 3 - 1N700x, 1N4148/1N914, and 1N5817. A discussion of the dizzying array of alternatives might help some of us understand why there are so many different devices designed to do only one thing in almost exactly the same way.)
If you need LM79xx, be sure you have some electrolytic or tag tant capacitors handy, as these regulators require much more bypassy capacitance than their positive voltage counterparts for stable operation. On that note, I've noticed several designs using LM78XX and LM79xx that have the same capacitor values on both positive and negative rails, presumably for BOM consolidation.
Really loving this series! I'm hoping for a massive video on the end so we can have some kind of general purpose Jellybean Box that can cover 80+% of the things people want to build.
The two jelly bean solutions I am frequently looking for are:
1. A simple "neutral reference" between rails, to avoid paying an op amp and 2 resistors to generate this, while losing sleep over whether there are any circuit conditions that will overload the op amp once the product is put into service. There ought to be a regulator designed to do this, but I haven't found it (yet).
2. Every design seems to involve switching analog signals around. A bilateral switch like the 4066 works, but it's made of super-static-sensitive CMOS. A 74HC4316 or 74HCT4316 seems to qualify, but they are certainly NOT available like jelly beans.
Steve
The 431 is basically the entire reference and pass transistor control block of a linear regulator without the pass transistor. This is why the 431 is used so much in mains switching mode power supplies, as it can drive a linear optocoupler to close the feedback loop for the regulated output.
Yes, more top 5 things: Top 5 useful beans, .....
As an young player I get drowned on all the specs when I try to select components. I really like how you highlight the most significant spec limits.
The opamp in the LM431 isn't acting like a comparator switching its output on and off. It is acting like a linear high-gain differential amplifier with feedback.
I learn my PPM's at 22:50 ! I love this jellybean series I hope for loads more....cheers !
Just ordered a bag full of 431s for the exact purpose described here. I have a lot of retro computers, from the Mac Classic and original PC to current gen. I always replace the capacitors in PSUs older than about 5-10 years, the threshold depending on how much I trust that brand's parts selection. Well, one time I didn't.
It was a Delta ATX supply from the early early P4 days, with a heavy 5V rail and the 6-pin Aux connector for Athlon boards. I got three of those supplies, NOS, in one bid. I serviced two of them, and got lazy with the third. I ended up installing it in a box I use to image hard disks (it has a VIA Epia mini-ITX board and both PATA and SATA hot-swap bays.) One day it had just turned itself off, and I figured maybe we had a brown-out. Two months or so later, I tried to turn it back on and it wouldn't budge. The KVM it was plugged into was also acting kind of weird, so I pulled the case onto my bench. It was making a squealing noise, but some of my PSUs do that. I measured the +5VSB rail to see if it was working at all, and it was measuring FIFTEEN VOLTS. Yikes!! I figured that motherboard was toast, but it's actually alive and (I think) running fine. Can't believe that.
Anyway, I pulled the cover on the PSU and the caps right next to the opto-isolation circuit were bulging from the top. Of course it's the caps. It's _always_ the caps. There was a 2222 transistor and a TL431 right next to it, so I figured I would go ahead and replace those too. I'm hoping to get some time today to replace parts and see if I can nurse it back to health. Shame to waste a good PSU on some dodgy caps...
Always replace the caps! Even if they look OK, if it has a few miles on it, you're better off safe than sorry.
NEVER depend on internal overtemperature shutdown (ALWAYS mount on sufficient heatsink to avoid the problem in the first hand)! Some of these kick in at 175°C TJ, etc. Ouch!!! And some datasheets clearly state "if you repeatedly activate the internal overtemp shutdown, you are degrading the chip"! And what will chip performance be right before shutdown???
Yeah i always got the feeling that the protection circuits on these regulators couldn't be trusted.
Thanks for confirming my suspicions!
Great video, tons of practical information! Can’t really use a shunt voltage reference as a power regulator unless you’re prepared to run its current at the maximum current draw of your circuit. This means not low power. Love those TL431’s, even saw one with the opto isolator built in.
Please more Jellybeans! Very valuable information for old young players like myself!
I just love this, "Jellybean Component series", thanks! Dave
Wow; I remember when the 78XX series when they first came out, they were amazing. Still around to this day...
I enjoyed these videos of jellybean components-please continue this series!
Brilliant. More of these great vids the better. I often only need small amounts of items at a time so jellybeaning is going to be a great way of bulking out an order to avoid extra P&P costs, almost free components if you want to look at it that way. 👍 Thanks.
Before switching supplies (70s) were popular, the 723 regulator was in most every linear regulated supply. I still see them in linear supplies today.
I love the series! Please continue it ;)
Using TL431 with LM317/337 together is great for low noise and stable linear power supply. Only critical point to achive that is to right matching capacitors in PSU design.
My jellybean switching regulator, is the MC34063. Multiple manufacturers, and was (is ?) used in car USB chargers, in all those dollar store units.
Great video. It has rekindled my interest in building a 431 based crowbar circuit.
Hi Dave, love the channel. I built a variable voltage PSU, used it to charge a battery but it popped, magic smoke. Apparently, the work around is to insert a BY127 or similar between the output leg and positive output terminal of the PSU, being constructed.
You're a good man Dave! I'm sure EE shops are struggling for parts and it's even harder as a newbie trying to design my first pcb.
I finally feel like part of the community.
"You have absolutely, guaranteed, heard of these." Success.
LM317 fan speed controller, the best crude and run-away fan speed controller :D.
puolup is a very famous resistor manufacturer... I hear about puolup resistors ALL THE TIME.
Yep, their everywhere !
I was a junior design engineering back in the 1980s and all if these parts were industry standard back then. It is staggering they are still being used.
For the LM317, just use the voltage feedback divider to provide your minimum 3.5mA load current.
Fantastic series Mr Jones keep em coming!
come-a-gutsa Verb. (Australian colloquialism) To be thrown off a moving object at high speed, arms and legs flailing, landing on a hard surface and sustaining multiple wounds and broken bones.
Jellyvean switching regulators/pwm generators please! Great series btw.
That is easy just any Unitrode part (now Texas instruments parts start with UC)
I have notice that a lot of people pick up LM317 with around output current up to 1.2-1.5 A.
But what most people seems to have is LM317 from China with much lower max output current.
I have a bunch at 0.1A. Took me as an young player a while to realise.
10:50 To use the 317 at low load currents, replace the 240 (R1) resistor with one of 120 Ω and choose the value for the voltage adjustment resistor (R2) accordingly.
In this way a minimum of 10 mA current will flow at all times.
I love this series and hope to see more like this, I hope Dave makes a similar series for design. something like favorite/jellybean topologies to explain those most useful and recyclable circuit designs and blocks.
TOP 5 jellybean switching converter? I think 34063 will be in the list but I want to know MORE!
LM2679 OR LM2596
Awesome
Would appreciate transistors and power transistors next
Now the top 5 jellybean transistors 😊
Great video, thanks!
Oh yeah, the 2N2222 and it's variants have been around for ages. The first time I saw a circuit for it was in the second or third "Book of Radio and Electronics" by Alfred Morgan.
Thanks a lot for this video series. It might be useful and very interesting to overview jellybean transistors. Maybe split in some categories like BJT, JFET etc.
Thanks for this series, it is very informative.
You did not mention that the 1117 has a maximum input voltage of 15 volts.
Also there is an LM317HV available from TI that has a maximum input voltage of 60 volts.
It is only available in TO-3, TO-39 and TO-220 packages.
There are a few 5 volt regulators with 50 volt input but they are hard to get.
A video on comparators would be good.
I thought 317 have no "maximum input voltage" because they are floating, and referenced to ground
@@invisiblekincajou The crux is in power-on, where the output capacitor may still be at ~0 V while the input already is near nominal value. So unless you can bring up the input voltage very gently, I wouldn't be betting on that!
@@PileOfEmptyTapes good point
The 78xx is also remarkably quiet. Check out Marco Reps video, as quiet as a battery!
The 1117 is the only regulator I have around for general use, I use a ceramic output cap with a 1R series resistor to meet the ESR requirement, better than using tantalums.
LM317 for me - adjustable and reliable ! along with the 555 timer it is a basic must have component .
I'd also add MC34063 and LM2596 for switching mode DC/DC converters, but that's a bit more complicated :)
Yup, the National "Simple Switcher" IC family.
I HAVE had the output of 7805's go into osculations when the load cap is open. I don't know if it was brand of chip or what. The output measured around 4v, if I remember right. The scope showed a large amount of noise on the 5v rail. I replaced the cap and everything started working right.
I've seen it to, didn't have a working scope at the time, but like you said replaced the first electrolytic on the 5V rail (which had an ESR of around 140 ohms) and the unit behaved again.
I think the LM723 is useful when we need a completely customizable regulator and I have a huge (old) stock on them
Or the µA 723. I missed that one in the walkthrough.
µA723 variations are quite common in older adjustable power supply designs. While not super easy to use and practical by modern standards, they are still good for low-noise supplies, beating even a bypassed LM317 and making it close to LT304x terrain.
My favorite to. I used one to design a 1500 Volt Regulated Screen Grid Supply for a big mutha transmitting tube. As Aretha would say the supply is Rock Steady !
The zener reference inside the 723 tends to be a lot less noisy than bandgaps.
The LM340, was National Semiconductor's original number, for their competing part to Fairchild's uA78xx. But that was back in the 70s. Its all evolved into the generic 78xx. The stabilizing caps were needed on older parts. I had them oscillate, without caps, back in the early 80s. They DO change the part design, as the years progressed, without changing the number. Any 78xx built in the last 25 years, the caps are optional.
Quite a few years ago, I wanted to have very long battery life for my back-then-new Asus EEE mini laptop. I had an external 12v sealed lead-acid battery (as in a home alarm system) and I used a circuit that I made myself with a LM317 regulator, to make the 9.5V the laptop charge port asked for. It worked for a short time, then the output shut off. My problem was that I didn't know about the minimum 3v difference between input and output voltage at that time!
Great series! Love hearing what you have to say
Wonderful. Very comprehensive, love it.
Excellent! I like these "Jelly Bean" lectures, Its Hobby Grade stuff. I can walk away with something here. Thanks
These are great, and if you run out of jellybean ones I'd like to see some videos with random deep dives into familiar or obscure component type hunts for the ultimate specs.
To think at the time Atari 2600 came around 44 years ago, the 7805s they used would've been quite the marvel in low cost integrated technology
29:53 - Circle and line drawn over Shunt Regulator circuit - Now I know how Prince got his logo!
1117 regulators are common because you can take USB 5v in and lower it to 3.7v to power 3.3v (and lower) microcontrollers. I see them all the time in HDMI adapters, ARM devices, microcontrollers, etc.
Reverse Protection Diode for 78xx, as at 15:20, can be eliminated in many cases by using 79xx instead, and regulating the 'ground' rail instead of the positive rail.
When you have a to largely varying Voltage, and if the differences allows it you could try to feed such regulator by a usual zener setting.
Love these Jellybean vids Dave, love to see more if possible, Thanks for all you do.
Useful thank you for the explanation and comparing those.
Dave, these "top 5 jelly bean" component videos are very interesting and useful (& educational to us "young players," regardless of our ages). Also, the thumbs down function is still working at the moment (2123 EST, 18 Nov 2021), so _maybe_ the backlash had the desired effect?
Hello Dave. I have to be honest here..
This is one of the videos i had to pause and start it again.. I'm planning to make power supply for my precision voltage source and i wanted to use Lm317 but as i see it it looks like there are other options i didn't consider before..
I think i will try to use that LM431 or maybe that Ref 1 ..I have to study datasheets just to be sure..but Thanks for video.
I hope you will take more time to go over all these components.. Things like Opamps ..as you would say that could be trap for young players..there are too many options ..and it is had to say wht is the best..
Haha now i noticed your video had over 44 minutes..i didn't even notice that..I had to save it for later use..for sure !
This is such a great series, Keep the videos coming ❤️
Simple switchers like the LM2679 or LM2596, More amps, more efficient, higher supply voltages. At this point I only use Linear regulators for low power 3.3V\1.8V needs, or for very low power, ultra-low noise (ie ADC, LNA, precision voltage references).
Issue with TL431 is possible drift from temperature, unless you use temperature compensated components. A lot of devices can be under significant temperature swings from 0C\32F to 50C\122F.
Jellybean series has been awesome, Dave! Since you brought it up, maybe "jellybean" comparators next?
A major difference between TL431 and LM4040 is that the 431 family has the internal bandgap reference voltage referred to the anode while the 4040 family have the internal reference referred to the cathode. Makes the 431 a bit like a high gain NPN BJT with a very precise Vbe and the 4040 a bit like a high gain PNP with a very precise Vbe!!
Another fantastic video. Maybe a video on jellybeans sets to buy for your lab would be something.
Hi Dave, i really like the jelly bean series. so i say keep them coming!
maybe you can also make a serie for different applications/uses/possibilities for these jelly bean components.
that would be awesome!
10 million jelly points goes to you.. Great video and presentation. You know we all have those cheap chinese V/A meters in our attics. Lately, I realized they display different measurements. Since you have covered voltage precision components in your video, wondering if you care to make another video about calibrating multimeters/AM meters DIY version. Cheers,.
And a trap for young players - 79xx has different pinout than 78xx as well as Vin instead of GND on the tab. Don't ask how I know that.
Did you also blow it up? :)
2 parts on the same heatsink....
I've thrown in 2x the LM317 into every toolkit (parts, tools, breadboards,...) i made for our interns/students, but until now they all were too lazy to build it onto their breadboards and all used external lab power supplies for their different voltages 😅
Now thats fine for first setups looking for smoke, but i find it a lot nicer if you only hook up one supply.. but building a board without cross wiring all and building a ratsnest seems to be the starting hurdle.. which makes it extremely hard to then find the wiring errors and issues.
IMO, the craziest thing about the LM317 is that it has no ground pin: it's adjust pin is 1.25V relative to Vo (1.25V/240ohm = 5.2mA, so Vo = R[adjust-gnd] * 5.2mA + 1.25V). So with a maximum input-to-output voltage of 40V, you *could* get 5V out of a 45V source, or a maximum 37V from a 77V input.
At very low efficiency unless your load is very tiny. Linear regulators basically are just an adjustable resistor to create a voltage drop to the target voltage. better option is to use simple switchers like the LM2679 or LM2596 for large voltage differences between the supply source and the output voltage.
Nice vid again thanks dave! Wow REF02. Saw that on a german site and used it for ADCs. Nice :)
Excellent series, keep em comming!