wide range power regulation methods?

I'm wondering how to go about making an efficient but safe voltage regulator that can take anything from 5~50VDC, and supply clean 3.3VDC to a basic low-current Arduino project. The typical regulators on Arduinos tend to support a textbook range like 5~12, stretched to 20V. If I knew I'd be given at least 40~50V, I could easily do a naive 1:4 voltage divider but that would be wasting a lot of power to heat, and won't allow smaller supplies like 5VDC.

Well this one will allow up to 30vdc input with adjustable output.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300362642650&ssPageName=STRK:MEWAX:IT

Lefty

Well this one will allow up to 30vdc input with adjustable output.

Related Question: I didn't see a datasheet on that eBay page; How do you adjust it? :P

look up a lm317 or lm138

http://www.national.com/mpf/LM/LM317.html#Overview

upto 37v in 1.2 to V+ out, and you can even get the 317 at radioshack (if you need about the same amperage as a 7805, 138 for upto a max of 5 amps)

Osgeld, I found the LM317HV (high voltage) variant on that site, thanks for the link! The HV type can accept from 3V~60VDC, and output a respectable 1.5A anywhere from 1.2V up. I'll have some more datasheet reading to do, but it looks good!

Related Question: I didn't see a datasheet on that eBay page; How do you adjust it?

I suspect by turning the adjustment screw on the blue trim pot shown on the board. The ad says it's build using the LM2596S-ADJ chip, so that's where I would look for a datasheet. It's a switcher so it should have very good efficiency specs, much better then any 78XX or other linear voltage regulator.

Lefty

I suspect by turning the adjustment screw on the blue trim pot shown on the board.

Man; I need to look more carefully before I post! ;D I've just been asking stupid questions galore today :o!

Osgeld, I found the LM317HV (high voltage) variant on that site, thanks for the link! The HV type can accept from 3V~60VDC, and output a respectable 1.5A anywhere from 1.2V up. I’ll have some more datasheet reading to do, but it looks good!

even better, good eye!

The HV type can accept from 3V~60VDC, and output a respectable 1.5A anywhere from 1.2V up. I'll have some more datasheet reading to do, but it looks good

As a variable voltage regulator the LM317 family has a wide input/output range, but if you compare input/output differential voltage to current output capability it really isn't much different from the fixed regulators. As such I can't really see this as a solution to the first post issue (large variation on input voltage).

I don't have a good answer myself (other than heat-sink, switching regulator and so forth). There are a some good harware guys on this forum however and I would love to see some feedback on a best-practice approach to a low-cost efficient wide voltage input range power regulator (e.g. 500mA).

What about adding a single switching transistor between the input voltage and the regualtor (using a zener to limit output voltage) - is this viable or is this asking for trouble?

Once you start pushing the spec of anything you start pushing the price as well. So for a wide input range it has to be a switching regulator but you will not be able to get 5V out for 5V in. The minimum for 5V out with such a wide input range is likely to be 8V or so. Switch mode supplies are not actually in amateur territory for home construction as a PCB layout is almost mandatory, and it is not just any old layout but a good one that is needed.

Here is one I came across googling:- http://www.watt-power.co.uk/catalogue/-p-233.html

Grumpy_Mike - What is your view on adding a transistor/zener pre stage to regulate the input voltage fed to a standard regulator (e.g. 7805).

As below, but 12V would be 9-50 and 5V should be 7V. Emitter goes to regulator (edit: picture is also not including the required regulator input capacitor - e.g. 47/100uF).

BenF wrote:

As a variable voltage regulator the LM317 family has a wide input/output range, but if you compare input/output differential voltage to current output capability it really isn't much different from the fixed regulators. As such I can't really see this as a solution to the first post issue (large variation on input voltage).

I'm not sure what you mean. Are you saying that this part will NOT supply 3.3VDC*1.5A if given 5V or 50V as an input? Or are you saying it will do so, at a horrible inefficiency above and beyond the output current demand?

If I had to bend the specification at all, I would say it's critical that the input allow anything from 9V~45V. At the worst, I could accept a low-voltage logic switch (not connected to the high voltage) that switched it from "big input" to "small input." These solutions that accept only 32VDC are not adequate for the kinds of inputs I want to support.

(The inputs are model aircraft lithium polymer, wired anywhere from two to twelve in series [2S to 12S] and this device must tap across the whole battery in parallel with the aircraft instead of just pulling from a few of the cells.)

If you put 50V in for 3.3V out at 1.5A you are dissipating 70W ( (50-3.3) * 1.5). For a TO-220 without a heatsink the temperature rise is 70DegC per W. 70degC per W * 70W = 4900degC rise. The fire will occur after about 100DegC rise ;)

For 1.5A out you should get a switching supply. A 10:1 range of input voltage is going to be hard to find.

(* jcl *)


www: http://www.wiblocks.com twitter: http://twitter.com/wiblocks blog: http://luciani.org

I'm not sure what you mean.

jluciani explained it well - it is the difference between input and output voltage that is your problem. If you look at the LM317 datasheet, there is a diagram that plots differential voltage versus output current. When the difference is higher than about 12V there is a sharp drop in current output capability.

The small schematic I referenced will convert a standard regulator (7805/LM317 doesn't matter) based power supply to a switched supply. Whenever the input voltage to your regulator exceeds the zener voltage, the transistor will turn off. When the voltage drops (across the capacitor on the regulator input side) the transistor will restore power. If you choose a zener voltage close to the min input voltage requirement of your regulator you will maximize effciency.

I haven't tried this myself, and it would be interesting to hear from the hardware guys what the negatives might be.

Okay, that makes sense. I think I can understand 4900 degrees C as being a problem. I can support a heat sink, as long as it's smaller than a kitchen sink.

How smooth is the handoff in the "switching" supply, when the zener diode overflows? Could I append two regulators and a switching diode like this, such that high voltage is cut down by one stage when present, but low voltage skips the first stage?

The switching regulator does not remove the amount of power you need to dissipate. The power is now split between two devices instead of one. Instead of 4900 DegC in one device you have 2450DegC in two. Still a BIG problem.

If you require a 10:1 range in input voltage (and 5W output) you really need a switching power supply.

(* jcl *)

The fire will occur after about 100DegC rise

No junction temperatures are quite happy up to about 130C

The switching regulator does not remove the amount of power you need to dissipate.

Yes it does. With a switching regulator the power you need to dissipate is controlled by the current and the voltage across the switching device, which is normally a FET and a very low resistance.

With a series regulator the power you need to dissipate is controlled by the current and the voltage difference between the regulated and unregulated supply.

This is a huge difference.

The fire will occur after about 100DegC rise

No junction temperatures are quite happy up to about 130C

You're assuming an ambient temperature of 25-30DegC. Since we are not sure about the packaging I was allowing a little margin which is why I mentioned about 100DegC rise. I would not use a power device above 80-85% of absolute maximum rating.

Although I would hate to encourage linear regulation of 5W+ supply with a 10:1 input voltage range -- using a FET as a series-pass element would give you additional power dissipation since you can get FETs with maximum Tj's of 175DegC.

(* jcl *)

How smooth is the handoff in the "switching" supply, when the zener diode overflows?

This will depend on the switching characteristics of the transistor. Voltage across the regulator input is maintained by a capacitor. There will be a ripple on input to the regulator (frequency will depend on load), but then again this is what regulators are good at.

Could I append two regulators and a switching diode like this, such that high voltage is cut down by one stage when present, but low voltage skips the first stage?

Not with this simple circuit and also you don't really need (or want) this as long as you're on battery power. Two regulators means twice the loss, half the efficiency.

Benefits of using a switching transistor on a standard regulator would include the following:

  • High efficiciency (irrespective of input voltage)
  • Tolerant across large variations in input voltage
  • Lower cost than a commercial switching regulator

Benefits of using a switching transistor on a standard regulator would include the following:

  • High efficiciency (irrespective of input voltage)

The switching transistor in the diagram is a series pass element (as is a linear regulator). The efficiency is necessarily dependent on change in input voltage.

The efficiency of a series pass circuit is Vout / Vin (ignoring quiescent currents which will be very low compared to the output current in this application).

You can spread the voltage drop across many series elements but your total voltage drop (and efficiency) is unchanged.

(* jcl *)