Maker Pro
Maker Pro

3.3V from 36V-48V?

L

Leon Heller

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

Thanks for any suggestions.

Leon
 
B

budgie

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

I'd personally be looking at something like an MC34063. Check the data sheet.
Depending on your transformer size vs the inductor, this may have as small a
footprint.
 
L

Leon Heller

Jan 1, 1970
0
budgie said:
I'd personally be looking at something like an MC34063. Check the data sheet.
Depending on your transformer size vs the inductor, this may have as small a
footprint.

It's 40V max. 8-(

Leon
 
M

Mike Harrison

Jan 1, 1970
0
It's 40V max. 8-(

Leon
Nat Semi do -HV suffix versions of many of their simple switcher range, with inputs up to 60V.
 
J

J M Noeding

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

Thanks for any suggestions.

Leon

you didn't inform about polarities

-jm
 
R

Roger Hamlett

Jan 1, 1970
0
Leon Heller said:
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

Thanks for any suggestions.
Look at:
http://www.romanblack.com/smps.htm
You would need to change the zener used, and alter the bias components,
together with ensuring that the transistors can take the higher voltages,
but in terms of total component cost, and overall efficiency, a design
based on this will be very hard to beat.

Best Wishes
 
J

Jacek Bogusz

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

Why not to use step down converter? You can choose one (i.e.) at
http://www.national.com/appinfo/power/webench/. In my opinion, your idea is
very complicated. You don't need output isolated from input, so why to make
transformer if you can use simple inductor? Check for example LM2594HV-3.3
application note. All you need is 2 capacitors, Schottky diode, 270uH
inductor and LM. Maximum input voltage is 60V.

Jacek
 
J

John Jardine

Jan 1, 1970
0
Leon Heller said:
small

It's 40V max. 8-(

Leon
I confirmed that 40V max voltage yesterday. Blew 5 MC34063's in a row,
looking for a non existant 'wrong' connection. Well, after all, the
equations *said* that a step-up to 100V was OK, so who am I to read the
datasheet? :)
 
J

J M Noeding

Jan 1, 1970
0
Why not to use step down converter? You can choose one (i.e.) at
http://www.national.com/appinfo/power/webench/. In my opinion, your idea is
very complicated. You don't need output isolated from input, so why to make
transformer if you can use simple inductor? Check for example LM2594HV-3.3
application note. All you need is 2 capacitors, Schottky diode, 270uH
inductor and LM. Maximum input voltage is 60V.

Jacek
in most cases 48V means -48v, and 3.3v means +3.3V although you have
to experience it once you are testing it in practice at the customer.
Perhaps communication with customer is somewhat odd. Some takes an
artificial way out it - planning an inverter for positive voltage, and
with -53.3V input, +3.3V happens to be -50V with all the funny things
to happen, and of course - fuses on the ground side

-jm
 
S

Spehro Pefhany

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.

Thanks for any suggestions.

Leon

An LM5007 in a 4mm x 4mm LLP package will do it with one SMT inductor.

Best regards,
Spehro Pefhany
 
B

Bob Wilson

Jan 1, 1970
0
I need to convert a supply voltage between 36V and 48V to 3.3V as
efficiently as possible. The 3.3V system will probably be using an MSP430
and hence will take very little current. What I'm proposing to use is a
simple two-transistor astable oscillator powered directly from the supply
with a step down toroidal transformer (centre-tapped primary) feeding a
rectifier and smoothing capacitor, followed by a 3.3V LDO. Is there a better
way to do it? Cost/size need to be as low/small as possible.


A simple flyback would probably be a far better proposition here. I think
you have your head in "AC transformer mode". DC-DC conversion is not done
the same way that AC-AC conversion is.

Bob.
 
B

Bob Wilson

Jan 1, 1970
0
It's 40V max. 8-(


So use a 50 cent 384x switchmode controller, and drive a separate switching FET
capable of handling the voltage. This way, you will not need a post regulator.

The key point is that you want this thing to be "as efficient as possible".
Your suggested method will not be very efficient.

Bob.
 
B

Bob Wilson

Jan 1, 1970
0
It's 40V max. 8-(


Another method that I forgot to mention in an earlier post, will work for low
currents. Forget the transformer altogether, and drive a capacitive charge pump
with the output of your oscillator. The output of the cap then drives a bridge
rectifier, where 2 of the 4 diodes are substituted by Zeners. This forms a
bridge rectifier, that inherently clamps the output voltage to the value of the
zeners.

As with any shunt regulator, there has to be series current limiting. In this
case, the "limiting" is the capacitive charge pump which acts as a constant
current source. The larger the cap, and the higher the frequency, the greater
the current.

Bob.
 
L

Leon Heller

Jan 1, 1970
0
Bob Wilson said:
A simple flyback would probably be a far better proposition here. I think
you have your head in "AC transformer mode". DC-DC conversion is not done
the same way that AC-AC conversion is.


What about the old blocking oscillator? I worked on a system incorporating
one 40 years ago.

Leon
 
R

Rich Grise

Jan 1, 1970
0
A simple flyback would probably be a far better proposition here. I think
you have your head in "AC transformer mode". DC-DC conversion is not done
the same way that AC-AC conversion is.

Bob.

Actuallly, a flyback isn't all that great a selection for a step-down.
If it doesn't need isolation, a buck converter can do it, albeit I'm not
sure if the available chips go up to 48V. in. And there are off-the-shelf
DC-DC converter modules available just about anywhere.
 
C

craig

Jan 1, 1970
0
For such current, use a small buck converter from national
in comes in SO-8 package, and you need a small inductor, and
cap, and shottky diode.
The national web site (power designer) web software will even
design the circuit you need. The circuits are very efficient.

-Craig
 
Top