1. Home
  2. Electronics Design
  3. DC tracking regulator design

DC tracking regulator design

I'm working on a project where I need a voltage regulator that tracks the DC level of the input; it should remain a constant DC fraction, like

4/5ths or 2/3ds, below that input voltage while rejecting ripple. What makes this somewhat more difficult is that it needs to work at relatively high voltages, and over a fairly wide voltage range - say 50 to 150 volts. Load current is around 10 to 50mA-ish, voltage dependent.

I've come up with a few implementations, but they've all been unsatisfactory in some way. The naive implementation, if the fraction is close enough to unity (so that the 317's input to output voltage rating isn't exceeded at the upper voltage limit), is to use something like an LM317 as a power follower: use a voltage divider on the supply, divide it down and filter the crap out of it and then apply it to the

317's adjust pin. Zeners would have to be added across the 317 to prevent blowing it on startup. It's really not much of a regulator, though; the output ripple is completely dependent upon how well the divided down reference is filtered.

I've gone around in circles trying to think of a way to pre-regulate, but always seem to come back to the problem that any pre-reulator needs to track just above the output regulator, so it needs to know where the output is set, but the output is set based on the divider coming from the supply input, so the ripple gets injected into the pre-regulator as well.

Suggestions?

Reply to
bitrex
Loading thread data ...

.

could this work?

formatting link
d/64a179c7c75c7eda/

-Lasse

Reply to
langwadt

Not sure what that means. If it tracks the input, it should track the ripple. If it doesn't track the ripple, what function of the input voltage do you want to track sans ripple? peak? trough? average? rms? ??? You also want to track load droop?

Or maybe you need something less than what you asked for??

What

BAD, BAD idea. Power supplies get shorted no matter how hard you try.

It's really not much of a regulator, though; the

Reply to
mike

How much precision do you need? What sort of output impedance?

If your needs are not super tight, then divide the input voltage down to your magic 4/5ths, minus 1.5V or so. Put a honkin' big cap (with the right voltage rating, natch') on the ground leg of the divider. Apply the divider voltage to the base of a Darlington or the gate of an n- channel MOSFET of suitable ratings, with the input power applied to the collector or drain. Take your output from the emitter/source.

What could be simpler?

If you need tighter regulation or better output impedance, and you have or can make suitable power for an op-amp, then you can divide the input down within the input range of an op-amp, put a honkin' bigger (but lower voltage) cap on the ground leg, then make an extended-voltage amplifier by driving the base of a common-emitter stage with a suitable-voltage NPN, which in turn drives the base of that same MOSFET or darlington from before.

In the schematic below, you want to set Re and Rc so that when the opamp is putting out its maximum voltage Vout is lower than you'll ever want it to go. Rb should be something like 100 to 1000 ohms, and Cc stabilizes the op-amp against the poles introduced in your final amplifier.

+Vbig o-------o-------------------------------------o--------. | | | | | | | .-. | | Cc | | | | | | Rc | | || '-' | | .------||---. | ||-+ .-. | || | | ||--o-|___|----| Q1 o-----o------)-----|+/ Rb |>

| | | |/| | | | | | | | | | === | .-. |+ .-. GND .-. | | === | | | | | | /-\ | | | | Re '-' | '-' '-' | | | | | | | | === === === === GND GND GND GND (created by AACircuit v1.28.6 beta 04/19/05

formatting link 

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com
Reply to
Tim Wescott

I'd do it the clean way. Use a capacitance multiplier up front, similar to the link Lasse gave but one transistor usually suffices. Then do the proper opamp thing like in figure 6.47 in "The Art of Electronics" 2nd edition which I suppose exists on every bookshelf.

Yo, Winfield, when's the 3rd edition gonna be ready? Just kidding :-)

Oh, and always mind the potantial *KABLAM* scenarios. Things like what happens upon a sudden short? What if the two rails "touch"? What if large capacitors on the regulated outputs tried to backfeed everything after the power fails because ol' Leroy plugged in one too many space heaters?

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

Yup, that's how it's done, except for a few glitches:

IN+ and IN- on the opamp need to be flipped because Q1 inverts. Also, the resistor coming back from the output goes to the (new) IN+ input and not the opamp output.

There should be diodes to protect Q1 from reverse Vbe and (very important!) Q2 from exceeding Vgs. All it takes is a capacitive load and a nervous operator finger turning down the voltage a bit fast and ... tsk ... *BANG*

You might want to add a diode from Vout to the collector of Q1 to improve load change reactions. Pulls down faster that way.

Q1 as well as Q2 can be FETs or BJTs, doesn't really matter as long as there is enough base current available in case of BJTs.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

Uhh -- whoops.

You're right, and that's what I get for just spewing out stuff without engaging the brain fully. The stabilizing cap still needs to go to the op-amp's - input, but then the whole thing does, indeed, need to be flipped around because of the phase change of the common-emitter stage.

And I hadn't considered the protection diodes at all; I'll have to remember that in case I ever do one of these for real.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com
Reply to
Tim Wescott

Hi guys, thanks for your input! estroys the ripple performance. It's a similar idea to what I had in mind with the LM317, except the LM317 packages up the opamp and pass device. Vo will track 1.25 volts above the voltage on the adjust pin. Of course the filter capacitor tied to the junction of the voltage divider will have to be a high voltage type; as Mr. Wescott mentioned the above circuit doesn't have that disadvantage.

I've found the trouble with that circuit is the need for the protection diode - on startup the 317 will see the full voltage from its input to output and needs some kind of zener across it for protection. Unfortunately in simulation the zener capacitance then d

Reply to
bitrex

I'll try that again:

Hi guys, thanks for your input!

It's a similar idea to what I had in mind with the LM317, except the LM317 packages up the opamp and pass device. Vo will track 1.25 volts above the voltage on the adjust pin. Of course the filter capacitor tied to the junction of the voltage divider will have to be a high voltage type; as Mr. Wescott mentioned the above circuit doesn't have that disadvantage.

I've found the trouble with that circuit is the need for the protection diode - on startup the 317 will see the full voltage from its input to output and needs some kind of zener across it for protection. Unfortunately in simulation the zener capacitance then destroys the ripple performance. I'll experiment with the above circuit and see what I get.

Reply to
bitrex

If you absolutely have to do it with a low parts count, take a look at this device:

formatting link

Digikey has them, abour $2 in singles, goes to well over 100V differential.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

look up Peak detector circuits and use that as the biasing signal in a regulator circuit. Something like a Emitter follower type coming from a fixed Higher voltage supply on the collector side.

This is the closes I can comprehend from your description. Jamie

Reply to
Jamie

Really dumb: voltage divider, bypassed, and a Darlington follower.

I've seen 317s boosted, bootstrapped from their own output voltage. Just force the 317's output to be, say, 5 volts above its input. That's not hard to do, zener and a mosfet follower maybe. But you'll still have overload issues.

--

John Larkin, President
Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation
Reply to
John Larkin

Oops. Input above output.

--

John Larkin, President       Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Reply to
John Larkin

I came up with the following idea for the circuit last night:

formatting link

It seems to have really good performance wrt ripple in simulation. The second LM317, U2, is configured as an active choke. D1 and M1 are to protect the devices against overvoltage on startup and shutdown; in a real circuit D1 would be like a 35 volt zener and M1 a high voltage pfet.

Reply to
bitrex

How about this, all discrete, depletion mosfet:

formatting link 

--

John Larkin, President       Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Reply to
John Larkin

Doing it with discretes is cool and would eliminate the start up issue, but I don't know if that circuit will have that much better performance than using a '317, which is essentially a comparator with a pass device. I do see the advantage of being able to use a low voltage capacitor on the left there rather than a high voltage type in an LM317 based circuit, so a bigger filter time constant would be available.

With a 2 second time constant the original circuit without any post regulation or active choke had pretty poor ripple performance - at 1kHz down about -25 dB.

Reply to
bitrex

ke

hat

0

ent.

on

,
s
e
s

- Hide quoted text -

The more traditional dual LM117 preregulator looks like so with an N- channel depletion HV MOSFET taking the brunt of the volatge drop. Without foldback current limiting, expect to heat sink the MOSFET for something like 10W. Also, the 117's typically need minimum loading of about 3ma.

Please view in a fixed-width font such as Courier.

. . . .----------------|--+--- '-+-|I O|--+--|I O|---+-->

. | | | | | A | | . | N- | ------- ------- | . | | | | . | | +--[R3]-+ . | | | | . +---------------[R1]---+------+--|>|--+ . | | | | | . | | | | | . | | | [R2] | . | | | | | . =3D=3D=3D =3D=3D=3D =3D=3D=3D =3D=3D=3D . | | | | | . | | | | | . | | . | | | . --- --- --- --- --- . . . . .

Reply to
Fred Bloggs

tracks

ripple.

fraction

something

it

317

how

pre-regulate,

an

to

have

input

MOSFET

ever

=20

=20

|

|=20

=20

___

|=20

=20

=20

=20

=20

=20

| | =20

|

=20

=3D=3D=3D GND GND GND =20

and

and

You might want a DC load resistor from Vout to ground as well.

?-)

Reply to
josephkk

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.