anyone have a suggestion for a very simple, efficient 6.3V regulator that can

[zachary vex]

supply 300-600mA and uses a minimum of parts?  it has to work from an 8-30VDC supply. 

i'm familiar with a lot of adjustable voltage regulators, and ways to beef up their output, but i'm just wondering if there's some new technology or a novel approach i haven't seen for this specific application (obviously it's for heaters on a tube.)

[gez]

Just use a 5V regulator and stick a couple of Si diodes in series with the ground connection.  Boss use this to get 9.2V from an 8V regulator.

[gez]

PS  If you're worried about the large drop across the regulator(s) a small value resistor in series with the input can drop some of the voltage and prevent max power ratings being exceeded.

[zachary vex]

thanks, gez!  excellent approach... i've seen it before, but had forgotten about this one.

[gez]

http://www.datasheetarchive.com/semiconductors/download.php?Datasheet=251995

7805 Would be ideal.

[zachary vex]

8^)  simplest, most common 5VDC regulator out there, combined with a pair of the simplest, most common diodes...

thanks again, gez!  three components... a trifecta of sorts!  i love it!

[puretube]

use 2 tubes, and hookup the filaments in series... 

[Sir H C]

going from 8V to 6.3 with a 7805 bumped up ~1.4 volts might squash out the regulator and have it in drop-out.  These usually use at least a darlington output so two Vbe drops.  Better to look at an LDO, I know National makes some that should work here.  Look for ones that are for the automotive industry, they have to handle over 40 volts if it is to go on the battery line.

[Paul Perry (Frostwave)]

If the input can vary from 8 to 30V, and you are draining up to 600ma, and you want efficiency of better than 20%, it's a switchmode supply or nothing.
Sorry, but that's the fact.
On the bright side, if all you are doing wiht it is running filaments, you can get by with a pretty choppy current. I would imaginea 555 monostable, switching the current thru the heater via a power miosfet, and then look at the voltage across a small value current sensing resistor in series with the heater, and average it via a RC smoother, and run that to a comparator that switches the 555 on again.
What I'm trying to say, is to run the heater from the 8 to 30V, but to switch it on & off so that you average 6.3 across it.
R.G. is the king of switch mode supplies,maybe he or someone else has a thought. I don't see it being a high count problem, and I don't see inductors involved. I do see problems wiht radiated noise, but nothing one couldn't solve.

[Sir H C]

For high efficiency when bucking 30 volts to 6.3 volts, you want a switching regulator.  Linear Tech has a lot of good ones that can do this with very high (>90%) efficiency.  Here is one:

http://www.linear.com/pc/productDetail.do?navId=H0,C1,C1003,C1042,C1032,C1082,P2090

Because remember a linear regulator at 30 volts in and 6.3 volts out can not be more efficient than 21%.

[Joe]

something like this maybe?

[Sir H C]

something like this maybe?

Problem is with 30 volts in 6.3 volts out and 1/2 amp, you have about 12 watts going through the transistor and a 2N3904 can't take that.  The only way to get efficiency with a large delta between input and output voltage is a switching regulator.

[gez]

A power trannie (I've used MOSFETs in these type of circuits) can be subbed in, but output voltage will vary with supply...unlike a 'proper' regulator.

As mentioned in my previous post, large voltage drops across a regulator can be compensated for by a small series resistor to the input.  For valve heaters the current draw is constant and easily measured/guesstimated so this approach is quite effective

[Sir H C]

For efficiency, any resistive device just dropping the voltage is a big hit to the efficiency.  That is why switching regulators exist, they can do large bucks and boosts and still give you greater than 90% efficiency.  In the schematic shown, if R2 is changed with a 7V zener you can get pretty good regulation.

[The Tone God]

There are 6.3v and 12.6v LDO and switching regulators that can put out 800mA to 1.5A. I have used a few in some tube projects.

PQ6RD083: 6.3v, 800mA, LDO

NJM2396F63: 6.3v, 1.5A, LDO

PT5048, PT5548: 12.6v, 1A, switching

Other options include running at 12v using a standard 7812 which would run the heaters within the tolerence (check the datasheet of course) or if 6.3v is the preference you can use a 7805 and float the ground slightly as was suggested. You have to watch the potential for noise with the switcher and float options. I have had some bad noise problems before.

Andrew

[Sir H C]

8-30 volt supply means no 12.6V heaters unless you go buck-boost.  Those are real pains in the @$$.

[The Tone God]

8-30v is quite a large range. I don't know if that is a suggested range or the actual range. If 15v and above is to be the availible average then 12.6v is an option otherwise the switcher is the best option if going 12.6v but then you have whole new set of problems to deal with at that point. Ick. Anyways just suggesting options.

Andrew

[puretube]

use 2 tubes, and hookup the filaments in series... 

ooops - missed the 8V minimum input 

in this case: use a 90V....240V switching wallwart,
and get out a constant 12 to 15V to work from;
(and do the starved plate thing with that, too  )

[toneman]

Hey Z,
check out the LTM4600.
one chip one resistor very efficient
6  -> 28VDCin....

Tone

[zachary vex]

There are 6.3v and 12.6v LDO and switching regulators that can put out 800mA to 1.5A. I have used a few in some tube projects.

PQ6RD083: 6.3v, 800mA, LDO

NJM2396F63: 6.3v, 1.5A, LDO

PT5048, PT5548: 12.6v, 1A, switching

Other options include running at 12v using a standard 7812 which would run the heaters within the tolerence (check the datasheet of course) or if 6.3v is the preference you can use a 7805 and float the ground slightly as was suggested. You have to watch the potential for noise with the switcher and float options. I have had some bad noise problems before.

Andrew

hey Andrew, that NJM part looks excellent... it would certainly take care of the LDO problem.  thanks!