Prev: Moto 68HC11 dev kit
Next: MC6850 - help with pin outs for the PLCC part
From: Joe G (Home) on 3 May 2010 07:06
"MooseFET" <kensmith(a)rahul.net> wrote in message
news:13f54e44-a9d1-4e79-a9e0-27f52509e7a4(a)a39g2000prb.googlegroups.com...
On May 1, 8:36 pm, "Joe G \(Home\)" <jo...(a)optusnet.com.au> wrote:
> Hi All,
>
> My radio device requires the following Vin Specifications
>
> Vin range 3v2 to 4v5 nominal 3v6
> Current 2amps peak
> Ripple less than 250mV up to 10KHz,
> Ripple less than 40mV from 10kHz to 100kHz and
> Ripple less than 5mW greater than 100KHz

Assuming no ripple on the input and a switcher running at
100KHz, the ripple values below 100KHz are all easily met.

The best design is likely to be a fairly simple bucker.

To get the low ripple, you need to be careful with where the
ripple currents end up flowing.


L1 ---------- L2 L3
IN-+--))))---+---! Switcher !---+---))))--+--))))--+--
! ! ---------- ! ! !
===C1 ===C2 ! ---D1 ===C3 ===C4
! ! ! ^ ! !
! ! ! ! ! !
! ---------+---+---- ! !
! ! GND1 ! !
! --------------+ !
! GND2 ! !
---+--------------------------------------+--------+----

You need to bring the ground end of C2 and D1 together directly
Move the parts around until they almost touch

The next priority is to get C3 to ground near the same point.

If this is a multilayer PCB, a GND1 can be an internal layer that
on other parts of the PCB is a trace layer. The whole idea here is
to keep the ripple currents local to the circuit.

If this is a stand alone item, the GND2, input and output should
all be in one small area. Don't put them on opposite ends of the
PCB as would seem intuitive.

The connection between C1 and C4 should be also as short as you can
make it.

L1 and L3 are basically large RF beads. They need to be lossy.


All the capacitors are actually collections of capacitors as
needed to get the ESR and ESL numbers low enough.

>
> From the outside world I would like to supply my radio devices with a
> voltage range 9 - 36V
>
> What can you suggest for regulator options?
>
> Should a use a two step approach.... a pre-regulator to drop the voltage
> down to 9-12V then a very low ripple DC-D regulator.
>
> Should I add DC capacitive filtering to achieve the ver low ripple
> requirements
>
> What can you suggest?
>
> Thanks in advance.
>
> Joe
>> Many thanks for your good email... I see you have used a sort of PI
>> filter
Regs
Joe


From: Joe G (Home) on 3 May 2010 07:11
> Google on "buck regulator".
>
> ...Jim Thompson
> --
Yes... but after googling buck regulator... the search results give too
many buck regulator options.

Only some buck regulators offer up ripple specifications - Linear Tech is
one.

I was after an "executive summary" an pointer in the right direction...
which is why I made my question quite specific.

Some else has replied with a detail answer in few lines.... I many need a PI
filter and careful PCB design.

Regs
Joe



From: MooseFET on 3 May 2010 09:39
On May 3, 4:11 am, "Joe G \(Home\)" <jo...(a)optusnet.com.au> wrote:
> > Google on "buck regulator".
>
> >                                        ...Jim Thompson
> > --
>
> Yes...   but after googling buck regulator... the search results give too
> many buck regulator options.
>
> Only some buck regulators offer up  ripple specifications - Linear Tech is
> one.

Many of the other makers provide the information you need to work out
the ripple.
You need to do some looking at those data sheets and math on what you
find.

If you do a design, you will likely find that the ripple is determined
by the
capacitors.

> Some else has replied with a detail answer in few lines.... I many need a PI
> filter and careful PCB design.

Yes to both.
From: George Herold on 3 May 2010 22:06
On May 2, 12:32 pm, MooseFET <kensm...(a)rahul.net> wrote:
> On May 1, 8:36 pm, "Joe G \(Home\)" <jo...(a)optusnet.com.au> wrote:
>
> > Hi  All,
>
> > My radio device requires the following Vin Specifications
>
> > Vin range  3v2 to 4v5    nominal  3v6
> > Current  2amps peak
> > Ripple less than 250mV up to 10KHz,
> > Ripple less than 40mV from 10kHz to 100kHz and
> > Ripple less than 5mW greater than 100KHz
>
> Assuming no ripple on the input and a switcher running at
> 100KHz, the ripple values below 100KHz are all easily met.
>
> The best design is likely to be a fairly simple bucker.
>
> To get the low ripple, you need to be careful with where the
> ripple currents end up flowing.
>
>        L1          ----------        L2        L3
>  IN-+--))))---+---! Switcher !---+---))))--+--))))--+--
>     !         !    ----------    !         !        !
>    ===C1     ===C2      !       ---D1     ===C3    ===C4
>     !         !         !        ^         !        !
>     !         !         !        !         !        !
>     !          ---------+---+----          !        !
>     !                       !        GND1  !        !
>     !                        --------------+        !
>     !       GND2                           !        !
>  ---+--------------------------------------+--------+----
>
> You need to bring the ground end of C2 and D1 together directly
> Move the parts around until they almost touch
>
> The next priority is to get C3 to ground near the same point.
>
> If this is a multilayer PCB, a GND1 can be an internal layer that
> on other parts of the PCB is a trace layer.  The whole idea here is
> to keep the ripple currents local to the circuit.
>
> If this is a stand alone item, the GND2, input and output should
> all be in one small area.  Don't put them on opposite ends of the
> PCB as would seem intuitive.
>
> The connection between C1 and C4 should be also as short as you can
> make it.
>
> L1 and L3 are basically large RF beads. They need to be lossy.
>
> All the capacitors are actually collections of capacitors as
> needed to get the ESR and ESL numbers low enough.
>
>
>
>
>
> > From the outside world I would like to supply my radio devices with a
> > voltage range 9 - 36V
>
> > What can you suggest for regulator options?
>
> > Should a use a two step approach.... a pre-regulator to drop the voltage
> > down to 9-12V then a very low ripple DC-D regulator.
>
> > Should I add DC capacitive filtering to achieve the ver low ripple
> > requirements
>
> > What can you suggest?
>
> > Thanks in advance.
>
> > Joe- Hide quoted text -
>
> - Show quoted text -

This is great Moosefet. (Though I have no plans to build switchers at
the moment.. I love the circuit discussion.)

Can/Have you told SED where you work?

George H.



From: MooseFET on 4 May 2010 23:12
On May 3, 7:06 pm, George Herold <ggher...(a)gmail.com> wrote:
> On May 2, 12:32 pm, MooseFET <kensm...(a)rahul.net> wrote:
>
>
>
> > On May 1, 8:36 pm, "Joe G \(Home\)" <jo...(a)optusnet.com.au> wrote:
>
> > > Hi  All,
>
> > > My radio device requires the following Vin Specifications
>
> > > Vin range  3v2 to 4v5    nominal  3v6
> > > Current  2amps peak
> > > Ripple less than 250mV up to 10KHz,
> > > Ripple less than 40mV from 10kHz to 100kHz and
> > > Ripple less than 5mW greater than 100KHz
>
> > Assuming no ripple on the input and a switcher running at
> > 100KHz, the ripple values below 100KHz are all easily met.
>
> > The best design is likely to be a fairly simple bucker.
>
> > To get the low ripple, you need to be careful with where the
> > ripple currents end up flowing.
>
> >        L1          ----------        L2        L3
> >  IN-+--))))---+---! Switcher !---+---))))--+--))))--+--
> >     !         !    ----------    !         !        !
> >    ===C1     ===C2      !       ---D1     ===C3    ===C4
> >     !         !         !        ^         !        !
> >     !         !         !        !         !        !
> >     !          ---------+---+----          !        !
> >     !                       !        GND1  !        !
> >     !                        --------------+        !
> >     !       GND2                           !        !
> >  ---+--------------------------------------+--------+----
>
> > You need to bring the ground end of C2 and D1 together directly
> > Move the parts around until they almost touch
>
> > The next priority is to get C3 to ground near the same point.
>
> > If this is a multilayer PCB, a GND1 can be an internal layer that
> > on other parts of the PCB is a trace layer.  The whole idea here is
> > to keep the ripple currents local to the circuit.
>
> > If this is a stand alone item, the GND2, input and output should
> > all be in one small area.  Don't put them on opposite ends of the
> > PCB as would seem intuitive.
>
> > The connection between C1 and C4 should be also as short as you can
> > make it.
>
> > L1 and L3 are basically large RF beads. They need to be lossy.
>
> > All the capacitors are actually collections of capacitors as
> > needed to get the ESR and ESL numbers low enough.
>
> > > From the outside world I would like to supply my radio devices with a
> > > voltage range 9 - 36V
>
> > > What can you suggest for regulator options?
>
> > > Should a use a two step approach.... a pre-regulator to drop the voltage
> > > down to 9-12V then a very low ripple DC-D regulator.
>
> > > Should I add DC capacitive filtering to achieve the ver low ripple
> > > requirements
>
> > > What can you suggest?
>
> > > Thanks in advance.
>
> > > Joe- Hide quoted text -
>
> > - Show quoted text -
>
> This is great Moosefet.  (Though I have no plans to build switchers at
> the moment.. I love the circuit discussion.)
>
> Can/Have you told SED where you work?


I can say where but I can't talk about the what so the where doesn't
matter much. I don't talk about the where so I am not tempted to
get into "whats" that I'd rather not.

I have also consulted from time to time on the side. On some of those
I have been asked not to talk about the where.


First  |  Prev  | 
Pages: 1 2
Prev: Moto 68HC11 dev kit
Next: MC6850 - help with pin outs for the PLCC part