DIY - High voltage linear regulator based on the LM317 - Part 1
Vložit
- čas přidán 26. 07. 2024
- #108 In this video I start looking into building a high voltage linear regulator. I want a supply that can work with high voltages (above 300V) so that it can be used in vacuum tube circuits or other similar applications but I also want it to be as integrated as possible. At the same time, I want to be able to built the circuit as both a positive voltage regulator as well as a negative voltage regulator. This all brings me to the LM317/LM337 pair.
More in this mini-series:
Part1: • DIY - High voltage lin...
Part2: • DIY - High voltage lin...
Part2b: • DIY - High voltage lin...
Related videos:
Capacitance multiplier - • Electronics Tutorial -...
HV supply - • Projects Revisited - H...
Datasheets:
www.ti.com/lit/ds/snvs775l/sn...
ww1.microchip.com/downloads/en...
References and further reading
- The Art of Electronics; Horowitz, Hill; Cambridge; Cap 9.13.5 C
Special Thanks to all my supporters on Patreon! Especially @afiskon and Ralf B.!
If you liked this video be sure to check out my other videos and you can also subscribe to be up to date with all the new ones!
If you want to support the creation of more and better videos please consider checking out: / feszelectronics - Věda a technologie
Dude, you are an excellent teacher. No offence to the other creators but in my mind you're one of the best electronics channels on this platform. Using LTSpice to simulate things in almost all your video's is the icing on the cake. Awesome stuff.
Now that's what I'm talking about.
Can't wait for part 2 and more...
Just what I needed for my vacuum tube amplifier...all these voltage regulators extra usefull
Awesome video and great and simple idea! 👍🏻
One little comment. On audio circuits noise and ripple is more important than exact voltage.
Especially with some tubes.
So a very simple capacitance multiplier would probably work just fine.
I have played with a similar idea but with a tracking switching regulator (as a LAB PSU) in front to reduce heat dissipation.
nice one waiting for the second part
Quite good analysis, some points; as MOSFET's tend to fault on short condition, I would advise against using them, think of Bipolars as built like tanks, mosfets like a small french car, if you do insist on using them, make sure you are using MOV's etc.
2) simulation great, BUT for the new players out there, always build prototypes first, there are a few simulation traps which may cause your circuit to behave different to reality, 3) TRIPLE check HV electrolytic polarity, those things explode, sometime in life-changing injury ways.
I enjoyed the video. I have built many HV linear regulators, both tube type and solid state. Some were my design. I have always steered away from the floating LM317 topology because a brief short circuit will likely fry the LM317. Having current limit with short circuit protection is hard to do. The vacuum tube versions are a lot more bulky, but are more forgiving when it comes to overloads and shorts. Jim Wd5JKO
I guess the major advantage that this circuit will have over tubes is that you can easily make it on the negative rail also; with tubes you don't really have a "PNP" version, so if you make a differential supply, it won't be that symmetrical
@@FesZElectronics MOSFETs in HV situations, make me VERY nervous, they are fragile little critters, and invariablly always result in short-circuit fault condition.
This is exactly what I was looking for! Great video! Thank you for including LTspice simulations, etc.
😎👍
Nice job bro !!! Can't wait for part 2
Have a look at page 19 of: Power Supply Cookbook, 2nd. Edition (Marty Brown, Newnes 2001) ...using the ancient uA723 with a pass transistor for a 350V supply.
Great work bringing a circuit that expands the use of the wonderfully ancient LM317 to light! I’m likely to use something similar to get a low-noise 400V rail for my Geiger counter.
But there’s two issues I have.
Firstly, while using a MOSFET to decrease the zener’s quiescent current with respect to the maximum output current is good, it should only be done with a MOSFET rated for linear circuits. Power/switching MOSFETs have multiple switching elements within each transistor, and during low overdrive / linear operation they become thermally unstable. There is a real possibility for thermal runaway as hotter channels take more current, particularly with high Vds and low Vgs. See Fairchild’s AN4161. Hope there are high-voltage analog MOSFETs, though analog MOSFETs in general are pretty uncommon. If not there are reasonably high-voltage monolithic BJTs or Darlingtons, for high-voltage audio amplifiers. IGBTs might suffer the same problem with thermal instability, though likely to a lesser extent, so they might be worth meddling with if they don’t oscillate or something.
Secondly, I think it’s feasible to use something like a TL431 along with a pass-transistor to get a high-voltage floating regulator, in a simpler way than your circuit. It might not have as much ripple rejection, but it’s better in terms of part-count. Would make a good replacement for the zener-based part of the 317 circuit anyhow. Actually, considering the TL431 can be used as a comparator, it might be possible to add some positive feedback to it and make a high-voltage self-resonant buck converter, not that it would be better than a linear solution if you want a low-noise low-current supply. The TL431 is a very versatile IC.
Hello James! I did consider using the TL431 for this project (see e2e.ti.com/support/power-management/f/196/t/233943?Negative-LDO-for-100V-to-90V ) - this schematic uses an extra transistor as a cascode to limit the voltage on the TL431.
The main reason I wanted to use the LM317 is to try to take advantage of its internal overcurrent and over temperature features - more on this in part 2. These features do not exist in the TL431 and need to be built separately.
I'm not sure if there will be any thermal runaway issues - the Vgs will only vary in a small quantity in comparison to the voltages being used. On the other hand local hotspots in the die could appear - we'll see. Anyway, thanks for the AN, I will check it out!
@@FesZElectronics Also. Funny enough a LM317 can be cheaper as well in most cases.
Well the overcurrent and overtemp features will only matter so long as the LM317 is stressed more than the pass transistor, which it almost certainly won’t. The LM317’s current limit will likely be way over what will cause the pass transistor to cook at such high voltages. Could thermally bind the two packages together I guess, but I’d recommend a PTC thermistor for that instead.
I’ve never actually seen the thermal runaway of a MOSFET myself, so I think I’ll make a rectified-mains source-follower with an IRF840 and see if it lasts.
@@Scrogan I have seen thermal runaway with MOSFETs, but definitely not in these kind of circuits.
@ James Barclay: I do agree with your comment about "switching"MOSfets used in linear mode. There is another interesting paper on that matter at www.infineon.com/dgdl/Infineon-ApplicationNote_Linear_Mode_Operation_Safe_Operation_Diagram_MOSFETs-AN-v01_00-EN.pdf?fileId=db3a30433e30e4bf013e3646e9381200
When building my own HiVoltage PS (400V 100 mA or 200 mA) I tested 3 flavors of "switching" Mosfets: STP4NK80Z: works fine up to 100 mA, STW28N60DM2: fatal crash(twice) & FQA7N80C_F109: works fine up to 200 mA.
Tracking pre regulator
Hats off, dude
Excellent
However, this circuit is only useful for testing ~led strings~ / the only example that comes in mind now ..... since they are usually in series and consume very little current.. in practice, this circuit is worthless. Let's say we want to add some power to it. could we add a bypass p=channel transistor to boost the current and also add a bc457 to limit the current :? thus, i believe we would have a practical and useful setup. I find interesting that we can use lm317 as a high voltage regulator, but the current should be higher, though I understand you used 20mA as an example. I look forward to your next video.. You sound like a very good teacher building people`s thinking and using very good didatics.. congratulations. such an innate teacher.
Excellent information.
very interesting
Great video
Nice video, looking forward to part 2! Have you seen the LR8 by Supertex?
I'm happy you enjoyed it! Well I did mention the LR8 by Microchip (I think Microchip bought over Supertex so it should be the same component)
Hi Fes is this topography capable of driving loads requiring up to 2 amps?
Nice video, as always. One question was left unanswered. Do you need a high voltage linear regulator for some particular application or building it out of curiosity?
To be honest, I don't have a particular use case in mind, but I am open to suggestions :D
@@FesZElectronics Will it be of interest to change the requirements to something like 24V 30+A? This is something commonly used for high power (300-1500 W) RF amplifiers and it might be a little problematic to buy such a power supply. Another option is 13.8 V 30A which is typically used for 100W transceivers, but there are plenty of these on the market and they are not too expensive.
On second though I don't know much about tube amplifiers, maybe unlike the solid state ones they require a much higher voltage. I know that 1000+W power amplifiers is one of few real application for tubes left these days. I've just checked my notes. The "The ARRL Handbook for Radio Communications 2019. Vol 3" book explains how to build a modern RF amplifier based on Russian tubes GI-6B (ГИ-6Б) or GI-7B (ГИ-7Б). There should be similar information in previous editions. ARRL Handbook is published annually and digital copies are not difficult to find online.
This is a great book! It will take me some time to look over it, but on the surface it looks very well written. I guess to make vacuum tubes useful in todays world, you will only use them in the final amplifier stage, so they will be working in the >1KV supply range. I guess that would require a different supply circuit; so would the high current versions you mentioned. I may look into that at some point.
The supply I am currently working on would be most useful for something like a vacuum tube audio amplifier or anything that needs high voltage but as clean as possible in the
which voltage source during frequency simulation?
how to add
v7 AC I 200??
Does D7 add noise to the output?
part 2 please :-)
Don't worry, its in work :D
Can i input 230 AC and take 60 v
Granted they only provide a few mA, On Semi has the NCP785 and NCP786 linear regulators (up to 450V). I use the smaller NCP781 (up to 150V input and 100mA) because its ultra low quiescent current for low power application.
I think I came across these IC's when I was preparing but the issue was that these are low output voltage components - the NCP785 has fixed 3.3 or 5V output and the 786 and 781 has a maximum of 15V. Its interesting to point out though that these components have a similar internal structure as the circuit I built - there is a high voltage pre-regulator stage, with a basic Zener and Mosfet and then a post-regulator stage with a precision reference and error amplifier.
Anyway, if low quiescent and low output is needed, than these parts are far superior to anything built with the LM317.
@@FesZElectronics Indeed, your design is much more flexible concerning the output voltage
I assume we could just make it with a TL431. Most precise for minimum component count and your dropout voltage would also be just the voltage of the pass transistor. Of course extend the voltage drive range with a transistor before the final pass transistor
Here is a high voltage implementation with the TL431: see e2e.ti.com/support/power-management/f/196/t/233943?Negative-LDO-for-100V-to-90V
@@FesZElectronics that's awesome perfect... But somehow how I can't grasp in my mind how they thought about the negative voltage regulator. With and upside down TL431? (That part number I can't remember). It would all fine and dandy but I can't quite put my finger on how would the schematic look like besides the NPNs changed to PNPs and that's all
Somewhere in the text they mention the TL4051 - that is supposed to be the inverted version of the TL431; Just be careful of the 10V maximum voltage drop limit on the 4051 - more info in the datasheet though
@@FesZElectronics well I'll better falstad it :D. I need regulateable negative V. Its a one time adjust and forget it but when the disposable part of the device is replaced it needs to readjusted. The range would be -15 down to -100.
Csá nyugati Szlovákiából ;)
What simulation software is this?
I use LTspice in all my videos, this one makes no exception.
As a ham radio operator I would also add a requirement of having low noise or RFI generation.
Well being a linear regulator, it should excel at this point by default. Also having the 2 stage circuit improves on noise performance since both stages contribute to the reduction of various noise frequencies.
@@FesZElectronics Thank you so much for the response...I've had linear power supply on my list a things build! Looking forward to many more of your videos! 73 de KN4FTT
you have a slav accent :) jestes z polski?
:D no, I'm from Romania, close by in the region
@@FesZElectronics btw, great job ;)
Hello! When we increase C6: pulse response of the regulator LT317 (on/of a load) might be quite slow - milliseconds.I my opinion.