How much is too much?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Sender

Well-known member
Joined
May 17, 2005
Messages
242
Location
USA
How much is too much of a voltage to drop for a regulator. IE. I have a 20-0-20 transformer that I would like to get -/+ 16 from. So I'm thinking after rectification I'll be getting 28 volts. Is it too much of a step down from 28 volts to 16? Am I going to need a MASSIVE heatsink??
 
The data sheet will give you a max operating voltage (remember to leave some margin for high line), the second consideration is power dissipation which will be the simple product of current out time voltage drop across regulator. Max power will depend on how heat sunk.

JR
 
[quote author="JohnRoberts"]The data sheet will give you a max operating voltage (remember to leave some margin for high line), the second consideration is power dissipation which will be the simple product of current out time voltage drop across regulator. Max power will depend on how heat sunk.

JR[/quote]

Well, I/O Voltage differential is 40V. So if a 20-0-20 transformer after regulation, puts out 28 volts, and the output is 16, that is a 12 volt differential. That's within 40. But I'm just concerned with it putting out too much heat with that much a drop.

As far as current. I'll be powering some API modules, about 6. So I'm thinking 1 amp, give or take.

So I'm a little lost on how to figure the power dissipation from there.
 
[quote author="Samuel Groner"]P = U*I, equals 12 Watt here... But I'm not sure if the APIs really eat that much current.

Samuel[/quote]

You're right, it should be about 100-200 milli amps, so that would work to be about to about 1.2 Watts. Im guessing just bolting them to the case shouse be fine
 
[quote author="Sender"]You're right, it should be about 100-200 milli amps, so that would work to be about to about 1.2 Watts. Im guessing just bolting them to the case shouse be fine[/quote]

Must insulate the metal tab, if there is one, on the regulator from the case. Otherwise it will work more as fireworks and less as a voltage regulator.
 
[quote author="tk@halmi"][quote author="Sender"]You're right, it should be about 100-200 milli amps, so that would work to be about to about 1.2 Watts. Im guessing just bolting them to the case shouse be fine[/quote]

Must insulate the metal tab, if there is one, on the regulator from the case. Otherwise it will work more as fireworks and less as a voltage regulator.[/quote]

Indeed, I just ordered some TO-220 insulator kits from Mouser.
 
Also. use a resistor before the regulator, and add a second reservoir cap stage, and you'll have a further R/C ripple-filtering stage, as well as relieving some of the heat load from the regulator...

Keith
 
Well, let's see. You want 16V at 1.2A from the regulator. But there are tolerances, so unless you use a trimpot on the regulator it could be as low as ~17V or as low as ~15.1. To figure out dropout, use the high number, 17V.

If it's a 317/337, allow for dropout voltage of 2.25V across the regulator, minimum. So the reg wants to see a minimum of 19.25V at its input. Allow a fudge factor of 0.2V (I'll talk about that later), so it really wants to see 19.45V at its input.

Consider that line voltage will be +/-10%. Your 20-0-20V tranny could be 18-0-18 on a hot summer day. That will have a peak voltage of about 23V, allowing 2.5V for diode bridge drops. With a 3300uF cap the DC voltage will be about 21.4V and the ripple will be about 0.87V. So your dropping resistor will be about (21.4 - 19.45)V / 1.2A = ~1.65 ohms. Use 1.6. With 1.2A passing through it, you'll dissipate about 1.2^2 * 1.6 = 2.3W, so use at least a 3W resistor (and 5W would be safer).

If you hang another 3300uF on the regulator side of that dropping resistor, you'll drop ripple to about 0.22V, close enough to your fudge factor to be okay. Bingo, you're set for worst case high regulator voltage and low wall voltage.

What about the reverse? If the line voltage is 10% high, without going through all the arithmetic the voltage at the regulator input will be 26.7V and the worst-case low regulator voltage will be about 15.1V, so you're dropping 26.7 - 15.1 = 11.6V across the regulator, at 1.2A, which is about 14W. That'll take a good-sized heatsink but it's doable with a 337. I'd use a TO3 case myself.

That's a bit of overkill, maybe, since it's very unlikely the regulators will really be that far from spec (and you can make them exactly on-voltage with a trimpot), but it'll give you a pretty bulletproof supply.

Peace,
Paul
 
My PSU calculator would have answered things as well: http://www.groupdiy.com/index.php?topic=21307

But there has never been much interest in this--people rather like to ask the same basic questions over and over again... :sad:

Samuel
 
[quote author="Samuel Groner"]My PSU calculator would have answered things as well: http://www.groupdiy.com/index.php?topic=21307

But there has never been much interest in this--people rather like to ask the same basic questions over and over again... :sad:

Samuel[/quote]

Thanks! Never saw that before, that's very helpful. There is only one input parameter I don't understand and that is Regulator Dropout Voltage, what exactly is this?
 
A regulator needs to maintain a certain voltage between its input and output to regulate properly. If, for example, you tried to get 16Vdc out of a 317 regulator being fed with 16.1Vdc, it wouldn't work. The amount of voltage necessary between input and output is called the Regulator Dropout Voltage.

How do you find it? You look on the regulator's data sheet; for most regulators the dropout voltage varies depending on how much current you're trying to draw.

Peace,
Paul
 
> My PSU calculator would have answered things as well

Bah. Real geeks don't need calculators. Note how Paul did it on a napkin. (Well, OK: the 4-digit precision on some of his steps does suggest a Curta or equiv.)

For everybody else, there's also Tangent's power supply parameter estimator. He's tricked it out with warnings when parameters are legal in general but dubious for common parts.

> people rather like to ask the same basic questions over and over again...

Right. Nothing ever got designed until the Internet was discovered.

> look on the regulator's data sheet

GASP! That's cheating!

You look on the innernet, find a place where the question was Asked and Answered last week, and ask it again.

That way you don't have to actually understand anything. And when it doesn't work, it isn't your fault.

Yeah, it's hot and sweaty here, and I'm annoyed at something else.... how did you guess?
 
[quote author="PRR"]> My PSU calculator would have answered things as well

Bah. Real geeks don't need calculators. Note how Paul did it on a napkin. (Well, OK: the 4-digit precision on some of his steps does suggest a Curta or equiv.) [/quote]

Actually a Casio scientific calculator, which I think cost about $16.00 at the drug store. But I also have a spreadsheet, similar in concept to Samuel's, which adds a few parameters such as regulator tolerances, use of a resistor before the first filter cap, use of a two-stage regulator, and use of a second transformer (so you can design with one transformer, like a wall-wart, outside the box and another one inside). I'm also more pessimistic about diode drops, I think. My sheet lets you calculate a few more parameters than Samuel's, but it doesn't look nearly as pretty. Particularly since his is vertically-organized, mine's horizontal, so that I can compare several setups at once. Flexible but unwieldy.

Peace,
Paul
 
[quote author="PRR"]

That way you don't have to actually understand anything. And when it doesn't work, it isn't your fault. [/quote]


Some of us are actually trying to understand things. :roll:
 
[quote author="Sender"][quote author="tk@halmi"][quote author="Sender"]You're right, it should be about 100-200 milli amps, so that would work to be about to about 1.2 Watts. Im guessing just bolting them to the case shouse be fine[/quote]

Must insulate the metal tab, if there is one, on the regulator from the case. Otherwise it will work more as fireworks and less as a voltage regulator.[/quote]

Indeed, I just ordered some TO-220 insulator kits from Mouser.[/quote]

and my good man, what is the part number for these insulators from Mouser?

thanks
Jay
 

Latest posts

Back
Top