Simple discrete voltage regulator / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Simple discrete voltage regulator

Move Thread

LAN_403

robert

Sat Dec 16 2006, 11:05PM

Registered Member #188 Joined: Thu Feb 16 2006, 05:18PM
Location:
Posts: 67

Appearantly, many seem to have forgotten that many things can be built without highly-integrated ICs nowerdays...

So, just if anyone is interested, here is the real simple voltage regulator.
It has negative feedback (so the output voltage stays constant with Uin/Iload changes) and current limiting, but that should be abvoius anyway.
I dont have to explain how it works?

In case anyone asks, i DO NOT claim to have invented this, this circiut can be found in almost ANY textbook on electronics (so you should all know it already).

Input terminal is "V+", output isnt marked but its the point where you expect it, the R2/R3 junction.

ShawnLG

Sun Dec 17 2006, 12:30AM

Registered Member #286 Joined: Mon Mar 06 2006, 04:52AM
Location:
Posts: 399

This one is not as stable as the one above, but it is more simple. I have used it years ago before I started using the usual 74## series.

Dr. Slack

Sun Dec 17 2006, 09:02AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

It's always good to "get back to basics". We often do not require a "perfect" regulator, so can often use simpler circuits to do just the job we need. The five major specs we often require are
a) stability (in the not oscillating sense)
b) independance from line changes
c) independance from load changes
d) independance from temperature
e) attenuation of line noise (like b, but an AC spec rather than DC)

Robert's circuit scores surprisingly well on temperature stability IF the voltage of the zener is chosen such that its tempco is the inverse of T2 VBE tempco, 6.2v IIRC at some mid current. However zener tempco is current dependant to some extent. Choosing a different voltage zener, or replacing it with a silicon diode, or nothing, will result in significant tempco. Unfortunately the line stability, as the current through R1 flows in the zener, is not good. A more stable arrangement will bias the zener from the regulated output voltage, rather than the input voltage, though this does introduce a start-up diffiuclty which can be solved with a bleed resistor. The load stability is also poor, due to the low loop gain.

The voltage adjust passes the full R3 current through the wiper. When the wiper bounces, it introduces a positive voltage transient on the output, which is not acceptable in a lab adjustable power supply. A less contact-resistance prone variation would use it as a potentiometer with the wiper to T2 base, replacing R3 and R4, though bounce would still create a positive transient. Reversing the direction of bias current to being sunk rather than sourced at the wiper (perhaps with a PNP emitter follower to ground, with its associated tempco) would be needed to make the transient negative.

Shawn's very simple circuit (with yet lower gain) is absolutely adequate if the regulator is between a bridge rectifier power supply and the rail to a bunch of opamps or CMOS that are happy with small rail variations.

There is an even simpler version than Shawn's, by replacing the Zener with a capacitor, that performs only "line cleaning", attenuation of noise on the input, if the load can tolerate large rail variations.

PSUs that are very stable to input, load and temperature changes are sufficently complex that the modern integrated way is usually the best way to go. However, in a fully integrated device, the main pass element heats the reference elements, so a discrete design which seperates the heat and the reference will always be better when the highest stability is required.

Many modern regulator circuits have such high loop gains to acheive good DC accuracy for load and line changes that they are somewhat twitchy on stability in the not oscillating sense. Most require some minimum output load C, and many even specify the minimum ESR (not maximum, so electrliyitcs are OK, ceramics are not !) that it will be stable with. And many are not particularly good in the attenuation of line noise, check their graphs if you want to use them to kill audio band noise, and put a good-old passive filter after them to clean up.

Wolfram

Sun Dec 17 2006, 04:34PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Are these simple regulators good for higher voltages? I have been looking for a good discrete regulator that can handle an input voltage of at least 800V. I have some tube based 0-500V PSUs, but they are awfully big (and not too stable, thanks to the old capacitors), and I want to make a new one based on a solid-state regulator.

robert

Sun Dec 17 2006, 05:06PM

Registered Member #188 Joined: Thu Feb 16 2006, 05:18PM
Location:
Posts: 67

This is possible for sure, if you use parts with adequate voltage ratings.The current limiting transistor doesnt need to withstand high voltages however.
Just the power dissipation gets a little high, i highly recommend the use of a darlington output transistor to keep the base current to a minimum.

Im using it with parts chosen for Uo 300v and Is 0.2A to charge a capacitor.
Parts are BU208 with some other 400v/1A NPN as darlington output stage, current limiting transistor being a BC337, feedback transistor again a 400v/1A NPN and a 6.2v zener (minimum tempco).
Works well, even if its not perfectly stable but its fine for my application.

Steve Conner

Sun Dec 17 2006, 11:14PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Mikkel, I designed a regulator years ago that took 570V in and gave regulated 475 and 360V rails.

I've also designed one that regulates +/-110V down to +/-65.

They all use combinations of the techniques posted in this thread. The 475V one is really just Shawn's simple circuit in disguise, using a TL431 cascoded with a HV MOSFET as a 475V adjustable "Zener". The +/-65V one is quite like robert's, just substituting a LTP for his single transistor error amp. I killed a lot of transistors getting this to work, but never blew a single one in the higher voltage circuit. (just as well, since the 1000V TO-220 MOSFETs were pricey)

Both have been running fine for several years now, and the 475V circuit even survived a near-short caused by gassy tubes once. I trust it enough that I played some big gigs with the amp that has it in.

robert, what is the safe operating area of the BU208 like? If I ever built that regulator again, I was thinking of using an IRFP460, but BU208s would look more at home with tubes, due to their "Retro" TO-3 packages

HV Enthusiast

Mon Dec 18 2006, 09:49PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

wrote ...

Are these simple regulators good for higher voltages? I have been looking for a good discrete regulator that can handle an input voltage of at least 800V. I have some tube based 0-500V PSUs, but they are awfully big (and not too stable, thanks to the old capacitors), and I want to make a new one based on a solid-state regulator.

A better solution for a higher voltage linear regulator is to use a stacked MOSFET arrangement. You can do this with BJTs as well. Horowitz's Encylopedia of Electronic Circuits shows several examples of this.

For the drive, you simply drive the lower FET and use a capacitor / resisitive divider up the series chain to control the other MOSFETs in fashion.

I've used this topology on several occasions to design high voltage linear regulators (up to 3000V) for use to regulate the beam voltage on high powered TWTs.

Wolfram

Tue Dec 19 2006, 06:35PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

I went looking for some HOTs at work yesterday, it seems that most HOTs have a maximum collector emitter voltage of 800V and most of the plastic packaged ones have a max power dissipation of 45-50W at 25 degrees C. There are also other things to worry about, like secondary breakdow. I think MOSFETs would be more suitable for me. As most of the common MOSFETs have too low voltage ratings for this application, I think EVRs idea sounds interesting. EVR, is this the circuit that you're talking about

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob