Searching \ for '[EE] Divide-by-two switching supply.' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: massmind.org/techref/method/math.htm?key=divide
Search entire site for: 'Divide-by-two switching supply.'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Divide-by-two switching supply.'
2009\05\04@213737 by Forrest W Christian

flavicon
face
Ok... an interesting question (at least to me)...

I have a device (product) which is typically powered from a lead-acid
battery array..  Usually either +12 or +24VDC.   As part of this device,
it meters the input voltage to provide battery voltage status.

I occasionally run into someone who needs to run off of a +48VDC (or
worse -48VDC) array.   I could just regulate it down to +12 or +24V but
that would basically eliminate the ability to measure the input voltage.

The currents are high enough (an amp or two) that I'd really prefer not
to use a linear regulator in this application.

What I would really like to do is to create a switching supply which
will take the +48VDC in, and step it down to +24VDC... but instead of it
being regulated to 24VDC, have it be regulated to half of the input
voltage.    That way, the voltage monitoring remains intact, but reads
exactly half of the actual input voltage.

To add insult to injury, it would be really nice if this was an isolated
switcher - so that I could take -48VDC and produce +24VDC out of it.

I'm planning on digging into some of the Microchip app notes as far as
using a PIC in the SMPS applications (seems like a logical way to
proceed)...  but really would rather use a more 'traditional' approach.

Ideas?

-forrest

2009\05\04@215258 by peter green

flavicon
face

> I occasionally run into someone who needs to run off of a +48VDC (or
> worse -48VDC) array.   I could just regulate it down to +12 or +24V but
> that would basically eliminate the ability to measure the input voltage.
How about putting a jumper in the link to the battery monitoring
equipment. Then those using a seperate regulator before your device
would pull the jumper and wire it seperately to the battery bank.


2009\05\04@215933 by Richard Prosser

picon face
2009/5/5 Forrest W Christian <spam_OUTforrestcTakeThisOuTspamimach.com>:
{Quote hidden}

>

2009\05\04@220247 by Forrest W Christian

flavicon
face
peter green wrote:
> How about putting a jumper in the link to the battery monitoring
> equipment. Then those using a seperate regulator before your device
> would pull the jumper and wire it seperately to the battery bank.
>  
Mechanically not really practical.  Essentially I have a 12 screw
terminal strip, four of which are used for power inputs (A and B
redundant).  Not really room to change... plus the other design is
done.    The measurement is done on the power inputs themselves.

So, really I'm faced with either just regulating it... and measuring the
voltage another way (which is definitely possible), or doing something
creative with scaling the voltage input to the device.  Which means the
divide by half regulator.

-forrest


2009\05\04@222446 by Xiaofan Chen

face picon face
On Tue, May 5, 2009 at 9:35 AM, Forrest W Christian <.....forrestcKILLspamspam@spam@imach.com> wrote:
> I occasionally run into someone who needs to run off of a +48VDC (or
> worse -48VDC) array.   I could just regulate it down to +12 or +24V but
> that would basically eliminate the ability to measure the input voltage.
>

You can still do that (regulating the voltage down). To measure the input
voltage, use a resistor divider and feed that to your voltage monitoring
circuit or an extra circuit.

--
Xiaofan http://mcuee.blogspot.com

2009\05\04@233250 by Forrest W Christian

flavicon
face
Xiaofan Chen wrote:
> You can still do that (regulating the voltage down). To measure the input
> voltage, use a resistor divider and feed that to your voltage monitoring
> circuit or an extra circuit.
>  
Let me see if I can clarify this...

I have an input on a device, which isn't really able to be changed at
this point.

It can handle 9-32V input or thereabouts.   This two pin input is *both*
the source of power for the device, and the spot which is metered.  
Again, this is not able to be changed.   Designed, and in-production
device.   And the current draw varies quite a bit in some configurations
(relays being turned on and off, etc.), so it might be 1/4A at one
moment, and a full amp or more the next.

So... a simple resistor divider isn't really going to work to take say
18-64V and divide it to stick it into this device.  Plus, as the current
draw changes, the voltage would change, which is not good.   Secondly,
the up to 32V drop at an amp or so is 32 watts.  Not pretty either.  
The voltage drop/current thing also applies to *any* linear regulator
style device on the front end.  I can't be wasting 32watts not doing
useful work, not to mention the heat issues.   32W is a lot of heat to
dissipate.  

I realize I could regulate this down to a fixed 24VDC with a switching
regulator.  I do have this option, and will probably end up doing that,
and measuring the voltage in some other way.  Just not as 'neat' of a
solution from the end-user's standpoint.

I would much rather put a 'switching voltage divider' on the front
end.   Then I don't need any intelligence at all, or any special
connections, or anything like that.   Just say 'plug your 48VDC in here,
and connect this to your existing device'.   Voltages get halved, and
simple.   If it wasn't for the heat issue, I'd just do this with some
sort of linear solution.  Instead I need a switcher  due to the currents
involved.

-forrest

2009\05\05@061755 by Alan B. Pearce

face picon face
>I'm planning on digging into some of the Microchip app notes as far as
>using a PIC in the SMPS applications (seems like a logical way to
>proceed)...  but really would rather use a more 'traditional' approach.

It would be hard to do using a 'normal' switchmode chip, but would work
using a PIC. Instead of using an internal fixed reference as the output
voltage reference, use half the input voltage as the reference (or more
correctly, a suitably divided form of it, to get to PIC suitable voltages).

I would start by looking at the circuit that Olin has published through this
list before, using a PIC 10F200 series device, and see if that is adaptable.
OTOH the Microchip app notes may be more amenable to the voltage/power
levels you are looking at.

2009\05\05@065925 by Djula Djarmati

flavicon
{Quote hidden}

You can build your own, some switching regulators have the Vref output
available but you have to find a chip that has no buffer between
internal and external Vref.

You can sample the input voltage and force the Vref lower or higher
(higher is better since the Vref output driver won't fight it) as your
input voltage changes.

I would start with older chips like UC3842 which would be a nice fit for
small isolated flyback.

Djula




2009\05\05@075803 by olin piclist

face picon face
Forrest W Christian wrote:
> What I would really like to do is to create a switching supply which
> will take the +48VDC in, and step it down to +24VDC.
> but instead of
> it being regulated to 24VDC, have it be regulated to half of the input
> voltage.

So go ahead.  What's the problem?

> To add insult to injury, it would be really nice if this was an
> isolated switcher - so that I could take -48VDC and produce +24VDC
> out of it.

That makes it more complicated.  Now you can't just 50% PWM into a inductor
to get 1/2 the input voltage.  Since your voltages are low, the easiest way
is probably a center tapped primary with fixed input drive.  The turns ratio
then dictates the input to output voltage ratio, except there will be diode
drops on the output side.  Using a H bridge instead of a center tapped
primary makes the transformer smaller at the price of having 4 drive
transistors instead of 2, with 2 of them high side.  Similarly, you get get
a bit more efficiency due to less diode drops with a center tapped secondary
at the expense of a larger transformer.

In either case, this is pretty simple stuff.  What exactly is the question?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\05\05@083603 by Dave Tweed

face
flavicon
face
Djula Djarmati wrote:
> > I have an input on a device, which isn't really able to be changed at
> > this point.
> >
> > It can handle 9-32V input or thereabouts. This two pin input is *both*
> > the source of power for the device, and the spot which is metered.
> > Again, this is not able to be changed. Designed, and in-production
> > device. And the current draw varies quite a bit in some configurations
> > (relays being turned on and off, etc.), so it might be 1/4A at one
> > moment, and a full amp or more the next.

[snip]

{Quote hidden}

You guys are way over-thinking this.

Think aobut it ... what happens if you drive the switch in a buck
regulator with a fixed 50% duty cycle control waveform?

As long as you pick a coil suitable for the range of load currents
expected (coil current needs to be continuous), the output voltage will be
50% of the input voltage. If you use synchronous rectification -- a second
FET that can sink as well as source current -- then it works all the way
down to zero load current.

This is the exact technique used in commercial products such as the
Vicor "V*I chip" VTM (voltage transformation module); the only difference
is that the ratio there is typically 24:1 or 32:1 instead of the 2:1 that
you are looking for.

Actually, they do offer a 2:1 version, the V048F240T012-CB
(see http://www.vicr.com/products/vichip/evalboards/prm_vtm/).
It's even isolated, so it meets all of your requirements, even if it is
overkill in the power-handling department (300+ W @ 96% efficiency).

-- Dave Tweed

2009\05\05@093041 by Michael Rigby-Jones

flavicon
face


> -----Original Message-----
> From: piclist-bouncesspamKILLspammit.edu [.....piclist-bouncesKILLspamspam.....mit.edu] On
Behalf
> Of Forrest W Christian
> Sent: 05 May 2009 02:35
> To: Microcontroller discussion list - Public.
> Subject: [EE] Divide-by-two switching supply.
>
> Ok... an interesting question (at least to me)...
>
> I have a device (product) which is typically powered from a lead-acid
> battery array..  Usually either +12 or +24VDC.   As part of this
device,
> it meters the input voltage to provide battery voltage status.
>
> I occasionally run into someone who needs to run off of a +48VDC (or
> worse -48VDC) array.   I could just regulate it down to +12 or +24V
but
> that would basically eliminate the ability to measure the input
voltage.
>
> The currents are high enough (an amp or two) that I'd really prefer
not
> to use a linear regulator in this application.
>
> What I would really like to do is to create a switching supply which
> will take the +48VDC in, and step it down to +24VDC... but instead of
it
> being regulated to 24VDC, have it be regulated to half of the input
> voltage.    That way, the voltage monitoring remains intact, but reads
> exactly half of the actual input voltage.
>
> To add insult to injury, it would be really nice if this was an
isolated
> switcher - so that I could take -48VDC and produce +24VDC out of it.
>
> I'm planning on digging into some of the Microchip app notes as far as
> using a PIC in the SMPS applications (seems like a logical way to
> proceed)...  but really would rather use a more 'traditional'
approach.
>
> Ideas?
>

How precise does the voltage halving have to be?  You could perhaps use
a traditional flyback switcher, but feedback from the 24v side would
have to be optically isolated, and rather than compare against a fixed
reference, compare against a fraction of the input voltage.  With an
opto-isolator in the feedback path, getting an accurate voltage ratio
without some kind of calibration might be problematic.

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2009\05\05@113110 by olin piclist

face picon face
Michael Rigby-Jones wrote:
> How precise does the voltage halving have to be?  You could perhaps
> use a traditional flyback switcher, but feedback from the 24v side
> would have to be optically isolated,

As I mentioned and Dave Tweed pointed out explicitly, you don't need
feedback.  A divide by 2 can be done open loop.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\05\05@225300 by Forrest W. Christian

flavicon
face
Dave Tweed wrote:
> Think aobut it ... what happens if you drive the switch in a buck
> regulator with a fixed 50% duty cycle control waveform?
>  
If the answer is you get 50% of the input.. then I have a widely
distorted picture of a buck regulator.   I guess I need to do some more
research, and need spend some time on a breadboard....

Because my mental image of a buck regulator indicates that a 50% duty
cycle into a buck regulator (aka switch into an inductor and capacitor
'filter'), will rarely, if ever produce 50% of the input DC voltage on
the output.  The actual output voltage will vary based on load.

So... I'll run off and play with this.. and a simple flyback converter
as well.

-forrest

2009\05\06@023736 by Richard Prosser

picon face
2009/5/6 Forrest W. Christian <EraseMEforrestcspam_OUTspamTakeThisOuTimach.com>:
> Dave Tweed wrote:
>> Think aobut it ... what happens if you drive the switch in a buck
>> regulator with a fixed 50% duty cycle control waveform?
>>
> If the answer is you get 50% of the input.. then I have a widely
> distorted picture of a buck regulator.   I guess I need to do some more
> research, and need spend some time on a breadboard....
>
> Because my mental image of a buck regulator indicates that a 50% duty
> cycle into a buck regulator (aka switch into an inductor and capacitor
> 'filter'), will rarely, if ever produce 50% of the input DC voltage on
> the output.  The actual output voltage will vary based on load.
>
> So... I'll run off and play with this.. and a simple flyback converter
> as well.
>
> -forrest

You could try the simple push-pull converter also. Two switching
devices, a transformer and a resistor & cap or 2..

RP

2009\05\06@074558 by olin piclist

face picon face
Forrest W. Christian wrote:
>> Think aobut it ... what happens if you drive the switch in a buck
>> regulator with a fixed 50% duty cycle control waveform?
>>
> If the answer is you get 50% of the input.. then I have a widely
> distorted picture of a buck regulator.

Apparently.  But keep in mind that, as Dave mentioned, this is only true if
running in continuous mode.

> The actual output voltage will vary based on load.

Only to the extent the inductor, switching element, and diode aren't perfect
and add series impedence.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2009\05\06@104055 by Herbert Graf

picon face
On Tue, 2009-05-05 at 20:52 -0600, Forrest W. Christian wrote:
> Dave Tweed wrote:
> > Think aobut it ... what happens if you drive the switch in a buck
> > regulator with a fixed 50% duty cycle control waveform?
> >  
> If the answer is you get 50% of the input.. then I have a widely
> distorted picture of a buck regulator.   I guess I need to do some more
> research, and need spend some time on a breadboard....
>
> Because my mental image of a buck regulator indicates that a 50% duty
> cycle into a buck regulator (aka switch into an inductor and capacitor
> 'filter'), will rarely, if ever produce 50% of the input DC voltage on
> the output.  The actual output voltage will vary based on load.

Actually, if operated in continuous mode (current through the inductor
never reaches zero) the output will be pretty much exactly 50% of the
input (I'm ignoring things like voltage drops across the transistor).

In discontinuous mode yes, the voltage will vary based on load, but
depending on the load it might be very easy to guarantee you always stay
in continuous mode.

TTYL

More... (looser matching)
- Last day of these posts
- In 2009 , 2010 only
- Today
- New search...