Power supply advice

Hi all. I need to design a power supply capable of producing about 150mA at 10V from an input of about 4V to 15V. To be more specific the input is about 4V to 8V when the unit is operating from internal batteries (when efficiency is important) and about 12V to 15V when operating from an external supply (when efficiency is less important). The device it's powering consumes a fairly constant current and its voltage tolerance is fairly wide - this power supply is replacing an unregulated supply direct from 8 NiCD cells - but it needs to be as smooth as possible (it's an FM transmitter PA stage). Cost and board area used aren't super-critical, though of course the cheaper/smaller the better (as always). Reliability is important and it needs to be able to operate over something like -10C to 40C external ambient temperature. The components used need to be readily available in small quantities in the UK (Digikey is OK too- they ship to the UK now). One more requirement- I need to be able to switch off the output of the supply from a 3.3V logic level output of a microcontroller.

I'm coming from the software and MCU side of things and am not an expert on switch mode power supply design. The only ones I've built in the past have been simple buck regulators copied pretty much exactly from the manufacturer's example circuit. Some possibilities I've considered are:

  • Modify the Tx to operate from a voltage higher than the maximum input voltage (say 18V) and just use a boost mode switcher to get up to that level. I asked the guy who designed the Tx about this and he said he 'wasn't too happy' about the idea. I'm not sure why. This may still be a possibility if I can persuade him to do it.
  • Use a boost mode switcher to get up above the maximum input voltage followed by a buck mode switcher to get down to 10V. This is a possibility but it seems inelegant.
  • Use a SEPIC mode switcher. This possibility sounds interesting but I haven't managed to find much information about how to do it. The few documents I've found about them on the net quickly descend into complicated switch mode theory which goes over my head. I have two huge books about switchmode power supply design and neither of them even mentions SEPIC (in fact I was disappointed to find that they contain very little about low voltage DC-DC stuff).
  • Use a boost mode switcher when operating from batteries and a linear regulator when the external supply is connected. I think there are some technical problems with this one- when the external power is connected the battery side of the boost mode switcher will get higher than 10V (due to the fast-charge circuit) and conduct through the diode in the switcher to the Tx, bypassing the linear regulator, unless I use something like a relay which switches over to the linear regulator when external power is applied. I would also need to put a FET or something in the linear side to allow the MCU to turn the output on/off.

Any advice on the best way to go about this would be appreciated.

--
------------ Alex Holden - http://www.alexholden.net/ ------------
If it doesn't work, you're not hitting it with a big enough hammer
Reply to
Alex Holden
Loading thread data ...

I forgot to mention: this part needs to work in such as way that the output is always off unless the MCU explicitly tells it to power up. The external power may be present even when the MCU is powered down.

--
------------ Alex Holden - http://www.alexholden.net/ ------------
If it doesn't work, you're not hitting it with a big enough hammer
Reply to
Alex Holden

The simplest approach generic is to use a switchable linear regulator into the high side of a switching boost regulator like so: View in a fixed-width font such as Courier.

P-CH +------+ +--------------+ | v | | | --- LINEAR | | | 5V REG | +-[r3]--+ +--------+ | | | | | | Vin>--+-----------|in out|--+------+ | | | | |+ | | | | gnd | === | | | +--------+ | 100u )|| | | | gnd )|| L | | gnd )|| | | )|| | | | SD | | +--|>|--+----+->out | | boost swtchr | | |+ | | +---------------+ | | === +------------|vin swtch|-+ | | | | | | [r1] gnd | | |--- | | [r3] +----|lbo fb|---------+ | | | | +------------|lbi | [r2] | | gnd | | [r4] +---------------+ | | | | gnd--+---------------------+----------------+------> gnd

Most switching regulators intended for portable applications will have low battery in/out or lbi/lbo signals, and you use these to switch the linear regulator in and out. In your case you would select r3/r4 so that Vin

Reply to
Fred Bloggs

Thanks, that makes sense.

--
------------ Alex Holden - http://www.alexholden.net/ ------------
If it doesn't work, you're not hitting it with a big enough hammer
Reply to
Alex Holden

May I suggest a chip something similar to the MAX710

formatting link
or MAX1672?
formatting link
- Maxim has a wide variety of "regulators-on-a-chip." Hopefully you'll surf up a good match.

I had a similar problem until I decided to toss out the regulator and power the circuit directly from lithium-ion cells, just putting all hardware into sleep mode when not in use.

A li-ion cell is 4.1 or 4.2v fully charged (depending on composition) and about 2.5/2.75v fully discharged. This conveniently works well with the LF series of PIC microcontrollers. Li-ion AAA cells can be rated for 1A/h or more (max steady-state current draw like 100mA.) 2.5-4.2v is quite a wide operating range, but they also have an excellent temperature range and good cycle characteristics with no memory effect. In short, lithium gives lots of pros. If you do use li-ion cells, beware of series charging - a charge equalizer will be needed to prevent "burn-out." There are many li-ion charger IC's on the market, some even have built-in buck/boost controllers. I'm working with a simple LM3622 linear charger for an automotive application-

formatting link

-M

Reply to
Mark Jones

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.