I'm looking for a switching regulator chip. Something inexpensive like
the MC34063, but not limited to 100 KHz because I need to reduce the
physical size of my layout, especially the inductor. Not having to use a
current sense resistor would be nice too.
The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
It doesn't have to be as cheap as the 34063, three times as much (US
prices) would be OK if the rest of it were fantastic.
>I'm looking for a switching regulator chip. Something inexpensive like
>the MC34063, but not limited to 100 KHz because I need to reduce the
>physical size of my layout, especially the inductor. Not having to use a
>current sense resistor would be nice too.
>
>The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
>
>It doesn't have to be as cheap as the 34063, three times as much (US
>prices) would be OK if the rest of it were fantastic.
AOZ1051 (Digikey), no diode needed, cheap ceramic caps, small inductor. That;s a 3A part - may be some others in the series that are cheaper for your power requirement.
>Any suggestions?
>
>Thanks,
>
>Bob
>
>--
>http://www.fastmail.fm - The professional email service
That's rather good for the money (US sourced).
$0.91/1 , $0.36/1000
Synchronous rectification.
0.8V feedback level - lower the better if using as a current source
(not in this case).
500 kHz
Reasonable switch Rdson - not stunning (70 mohm hi, 40 mohm low)
Cycle by cycle current limit (uses high side synchronous rectifier
MOSFET Rdson as Rsense so adds no losses that are not already there.
Thermally protected.
Short circuit protected.
Vinmin of 4.5V is a shame - too high.
Vin lockout at 4.1V rising and 3.7V falling.
This makes it poor to useless for LiIon 1 cell and useless for LiFePO4
1 cell and needs 5 x NimH to operate properly in spec.
They could easily have given it a lower startup voltage - even if it
ha to bootstrap its own supply before it could handle any load.
Internal output FETs Vgsth probably an issue.
Should give around 90% efficiency in envisaged application.
All the switching regulators I use have to be cheap (and small footprint, which may not apply to you), and for this I like the National (yeah, I'm not calling it TI on purpose :) Simple Switchers. They're a couple bucks or so each, but at anywhere from 500kHz to >1Mhz, the passives get really small, and hence cheaper, which very much offsets the price of the switching chip itself. And the number of external components are quite small.
> I'm looking for a switching regulator chip. Something inexpensive like
> the MC34063, but not limited to 100 KHz because I need to reduce the
> physical size of my layout, especially the inductor. Not having to use a
> current sense resistor would be nice too.
>
> The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
>
> It doesn't have to be as cheap as the 34063, three times as much (US
> prices) would be OK if the rest of it were fantastic.
>
> Any suggestions?
>
> Thanks,
>
> Bob
On Thu, Aug 2, 2012, at 09:23 AM, Mike Harrison wrote: {Quote hidden}
> On Thu, 02 Aug 2012 09:03:03 -0700, you wrote:
>
> >I'm looking for a switching regulator chip. Something inexpensive like
> >the MC34063, but not limited to 100 KHz because I need to reduce the
> >physical size of my layout, especially the inductor. Not having to use a
> >current sense resistor would be nice too.
> >
> >The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
> >
> >It doesn't have to be as cheap as the 34063, three times as much (US
> >prices) would be OK if the rest of it were fantastic.
>
> AOZ1051 (Digikey), no diode needed, cheap ceramic caps, small inductor.
> That;s a 3A part - may be some others in the series that are cheaper for
> your power requirement.
This will possibly fail on price, and availability looks not
marvellous * but otherwise looks superb.
(*Not listed by Digikey except evaluation board. Mouser, Arrow, ...
show $ but 0 stock)
IR3897 - lowest capability member of a family.
4A.
300 kHz - 1500 kHz.
1 <= Vin <- 21V.
Should be comfortably above 90% efficiency in your application.
> This will possibly fail on price, and availability looks not marvellous * but
> otherwise looks superb.
> (*Not listed by Digikey except evaluation board. Mouser, Arrow, ...
> show $ but 0 stock)
> IR3897 - lowest capability member of a family.
> 4A.
> 300 kHz - 1500 kHz.
> 1 <= Vin <- 21V.
> Should be comfortably above 90% efficiency in your application.
>
> www.irf.com/product-info/datasheets/data/ir3897m.pdf
>
> Family
>
> http://www.irf.com/whats-new/nr120628.html
>
>
> Russell
Another that may fail on price is Linear Technology (people seem to say their chips aren't the cheapest, but I haven't used them commercially so couldn't comment on that). They do have a pretty useful web page selector, and some of the chips have internal synchronous rectifier FETs to improve efficiency. A handful of the chips need very few external components.
> I'm looking for a switching regulator chip. Something inexpensive like
> the MC34063, but not limited to 100 KHz because I need to reduce the
> physical size of my layout, especially the inductor. Not having to use a
> current sense resistor would be nice too.
If you are willing to forgo any sophistication such as protection and
startup management then a 'roll your own' hysteretic buck converter
running at > 1 MHz with comparator, bipolar hi side switch (MOSFET
with extra driver components) and either external reference or added
reference may be able to be done. Using a darlington may (or may not)
make drive easier. Getting this working is simple. Getting it working
really well is 'a bit of an ask' overall and it depends on what volume
you want whether the playing is worth it, but using the smallest
available packages the end result could be quite tiny. If the 8V in is
stable and accurate enough (eg another regulated output) it may be
able to be used to provide the reference.
As I know you know, driving a MOSFET at that frequency is going to
need a fast driver or super-optimised luck. You'd be OK at 100 kHz
with little more than an NPN/PNP pair but for 1 MHz plus you are going
to need an extra transistor and some extra componentry to get driver
speed. That's why a single bipolar or perhaps darlington (with its own
problems) may be easier.
A possible out exxists IFyou can find a comparator that is fast
enough, cheap enough and with enough drive power. If driving a single
bipolar high side switch an LM339 (or 1 per pkg variant) may do quite
well. It will slew high to low on a 5V supply in about 100 ns typical
and low to high in about double that. The 6 mA guaranteed, 16 mA
typical on Vdd=5V is not good enough to drive a FET gate capacitor
fast enough but may be OK for a bipolar.
It would take minimal bench (or SPICE) time to trial an LM339,
BC807-40 or your favourite high-Beta equivalent (500 mA+ testbed) , a
1N5817, a TL431 and not too much else. With 6mA guaranteed drive you
need a beta of about 160 at 1A which is achievable. (Zetex will always
oblige). This is looking like about 4 x SOT23 or smaller actives (1
pkg comp, PNP, 431 if needed, schottky) plus passives to taste.
Inductor is "small". Be careful not to inhale the 0201 resistors :-).
On Thu, Aug 2, 2012 at 12:03 PM, Bob Blick <spam_OUTbobblickTakeThisOuTftml.net> wrote:
> I'm looking for a switching regulator chip. Something inexpensive like
> the MC34063, but not limited to 100 KHz because I need to reduce the
> physical size of my layout, especially the inductor. Not having to use a
> current sense resistor would be nice too.
>
> The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
>
> It doesn't have to be as cheap as the 34063, three times as much (US
> prices) would be OK if the rest of it were fantastic.
>
> Any suggestions?
>
> Thanks,
>
> Bob
There's a relatively new microchip switcher IC that does buck mode. I
was going to look up the part number but I can't get the microchip
website to work.
Perhaps someone already mentioned the *3063 which is the "advanced" 34063.
www.onsemi.com/pub_link/Collateral/NCP3063-D.PDF
$0.68 at 100 qty.
> Perhaps someone already mentioned the *3063 which is the "advanced" 34063..
> www.onsemi.com/pub_link/Collateral/NCP3063-D.PDF
> $0.68 at 100 qty.
Unfortunately the 3063 is just a coat of paint and new wallpaper. The
basic core is largely unchanged :-(.
They say 150 kHz which the 34063 could do at a stretch.
The graphs say you can get maybe 300+ kHz from it.
Still the high 1.25V reference. Still the poor darlington output (used
to get switching "speed" as a non saturating and thus lossy switch.)
LM339 still looks attractive :-).
> I'm looking for a switching regulator chip. Something inexpensive like
> the MC34063, but not limited to 100 KHz because I need to reduce the
> physical size of my layout, especially the inductor. Not having to use a
> current sense resistor would be nice too.
>
> The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
>
> It doesn't have to be as cheap as the 34063, three times as much (US
> prices) would be OK if the rest of it were fantastic.
You did not say up or down converting. The sense resistor is optional but highly
recommended:
MCP1623/4 boost for single battery op. of mcus.
AL1603 as above but from Diodes
AP1533SG buck 300kHz 1.5A 8sop: ~$0.4
TL/LM2575 buck I don't have to explain this one
AP1609 boost counterpart of AL1603 300kHz
There are many many others out there. I shun 34063 because it requires bulky
external parts, but there is a new 34063 variant which uses a higher switching
frequency and is otherwise identical. I just can't find it now, it may be
CMOS inside, even. All the above are available from Farnell/Newark and are under
$1 in 1s.
I'm looking for a switching regulator chip. Something inexpensive like
the MC34063, but not limited to 100 KHz because I need to reduce the
physical size of my layout, especially the inductor. Not having to use a
current sense resistor would be nice too.
The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
I don't have prices but the Micrel range of switching chips might be worth looking at.
The MIC4576 can pump out 3A runs at 200kHz, same pinout as LM2576 and is available in fixed 3v3 or adjustable 1v23 - 33v versions, input is 4 -36v.
For 1MHz and 2MHz there are the MIC2204,2202 (max 5v5 in though) or many other variants with 4v5 - 26v in @ 600kHz - http://www.micrel.com
> > The basic operating parameters will be 8V in, 3.3V out, at 1 amp.
> >
>
> You did not say up or down converting. The sense resistor is optional but highly
> recommended:
Umm, I would have thought that 8V in, 3v3 out would define up or down pretty well ... ;))
<alan.b.pearce <at> stfc.ac.uk> writes:
> Umm, I would have thought that 8V in, 3v3 out would define up or down pretty
Heh I missed that part, I was under pressure. Anyway each of the regulators I
posted has an ADJ version, and I tend to stock that and not the fixed version.