Power supply conditioning...?

Questions relating to designing PCBs

Moderator: phalanx

smdFan
Posts: 250
Joined: Mon Jan 07, 2008 10:40 am

Power supply conditioning...?

Post by smdFan » Tue Oct 19, 2010 1:11 pm

I intend to supply power to AVR microcontroller using a 3V coin battery. I do not want to connect the +ve and -ve of coin battery directly to AVR MCU. Is there any IC that can be used in between to condition the power supplied to AVR?

Philipf
Posts: 84
Joined: Mon Jan 28, 2008 4:48 pm

Re: Power supply conditioning...?

Post by Philipf » Tue Oct 19, 2010 2:00 pm

Just use a resistor in series to limit the current, with a capacitor and inductor to clean the line.
Image

This should work.

smdFan
Posts: 250
Joined: Mon Jan 07, 2008 10:40 am

Re: Power supply conditioning...?

Post by smdFan » Tue Oct 19, 2010 2:14 pm

Philipf wrote:Just use a resistor in series to limit the current, with a capacitor and inductor to clean the line.
Image

This should work.
Thanks. I was thinking more along the lines of a ready made IC. By the way how do you determine the values of cap, res., and inductor? Is it somekind of filter?

tecoist
Posts: 160
Joined: Mon Aug 30, 2010 6:22 pm

Re: Power supply conditioning...?

Post by tecoist » Tue Oct 19, 2010 4:20 pm

What exactly did you want this "power conditioner" to do?

Unfortunately, the LCR filter suggested by phillipf is not helpful. You certainly do not need either inductance or resistance between the battery and your microcontroller.

If you'd like to see an example design, take a look at the SparkFun design for their Nordic Fob. Battery + AVR + radio = fun. Highly trained engineers selected those three capacitors and carefully placed them in the schematic for your enjoyment.

MichaelN
Support Volunteer
Posts: 1191
Joined: Fri May 08, 2009 6:14 pm
Location: Australia

Re: Power supply conditioning...?

Post by MichaelN » Wed Oct 20, 2010 4:06 pm

I agree with Teco - what are you actually trying to achieve?

If you need a regulated voltage (which a simple microcontroller won't need if you're using a single lithium cell), you should choose the lowest acceptable voltage the circuit will run on, then choose a LDO regulator with a tiny quiescent current (so you don't drain the battery too fast).

If you just want reverse-polarity protection, use a MOSFET "backwards" instead of a diode (which will introduce voltage drop). If you don't know what I mean by this, let me know & I'll post a diagram.

smdFan
Posts: 250
Joined: Mon Jan 07, 2008 10:40 am

Re: Power supply conditioning...?

Post by smdFan » Fri Oct 22, 2010 8:15 am

MichaelN wrote:I agree with Teco - what are you actually trying to achieve?

If you need a regulated voltage (which a simple microcontroller won't need if you're using a single lithium cell), you should choose the lowest acceptable voltage the circuit will run on, then choose a LDO regulator with a tiny quiescent current (so you don't drain the battery too fast).

If you just want reverse-polarity protection, use a MOSFET "backwards" instead of a diode (which will introduce voltage drop). If you don't know what I mean by this, let me know & I'll post a diagram.
Thanks for the info. My aim is

1. Provide a stable voltage to MCU. I have connected a cell battery to AVR but when I press the reset button, the voltage where it is supposed to be 0V jumps all over. What I mean I want the battery to provide clean and stable volage to MCU.

2. Prevent fast drainage of battery


Thanks

tecoist
Posts: 160
Joined: Mon Aug 30, 2010 6:22 pm

Re: Power supply conditioning...?

Post by tecoist » Fri Oct 22, 2010 9:54 am

Do you have bypass capacitors between Vcc and ground, physically near the AVR? Take a look at the Nordic Fob schematic to see an example.

What "voltage where it supposed to be 0V" are you measuring, and relative to what other point are you measuring it? What time scale are we talking about?

Generally speaking, batteries deliver very clean DC power, though with a little higher output impedance than would be ideal. Battery voltage does not, in general, "jump around." Because of the battery's output resistance and the inductance of the power supply traces (or wires), it's good to provide decoupling capacitors physically close to the microcontroller, but that's all the "conditioning" you should need to get the microcontroller to work reliably.

smdFan
Posts: 250
Joined: Mon Jan 07, 2008 10:40 am

Re: Power supply conditioning...?

Post by smdFan » Fri Oct 22, 2010 10:03 am

tecoist wrote:Do you have bypass capacitors between Vcc and ground, physically near the AVR? Take a look at the Nordic Fob schematic to see an example.

What "voltage where it supposed to be 0V" are you measuring, and relative to what other point are you measuring it? What time scale are we talking about?

Generally speaking, batteries deliver very clean DC power, though with a little higher output impedance than would be ideal. Battery voltage does not, in general, "jump around." Because of the battery's output resistance and the inductance of the power supply traces (or wires), it's good to provide decoupling capacitors physically close to the microcontroller, but that's all the "conditioning" you should need to get the microcontroller to work reliably.
Thanks for the info. I guess what I am trying to find out if it is at all good idea to hook 3V coin battery directly to MCU, of course with all the decoupling caps. Yes. I had used decoupling caps. Thanks

MichaelN
Support Volunteer
Posts: 1191
Joined: Fri May 08, 2009 6:14 pm
Location: Australia

Re: Power supply conditioning...?

Post by MichaelN » Fri Oct 22, 2010 5:31 pm

smdFan wrote:I guess what I am trying to find out if it is at all good idea to hook 3V coin battery directly to MCU, of course with all the decoupling caps. Yes. I had used decoupling caps. Thanks
Connecting a coin battery directly to a micro is done all the time - it allows very low power, since it avoids the quiescent current requirements of a regulator.

Personally, if there was any chance of the coin cell being connected incorrectly, I'd add a logic level MOSFET to provide reverse-polarity protection. You can either use an N-channel or P-channel FET. Note the drain & source connections, which are "back to front" compared to normal:
Image

smdFan
Posts: 250
Joined: Mon Jan 07, 2008 10:40 am

Re: Power supply conditioning...?

Post by smdFan » Wed Oct 27, 2010 9:14 am

MichaelN wrote:
smdFan wrote:I guess what I am trying to find out if it is at all good idea to hook 3V coin battery directly to MCU, of course with all the decoupling caps. Yes. I had used decoupling caps. Thanks
Connecting a coin battery directly to a micro is done all the time - it allows very low power, since it avoids the quiescent current requirements of a regulator.

Personally, if there was any chance of the coin cell being connected incorrectly, I'd add a logic level MOSFET to provide reverse-polarity protection. You can either use an N-channel or P-channel FET. Note the drain & source connections, which are "back to front" compared to normal:
Image
MichaelN,

Thanks a lot for the information. I appreciate it. Can you also suggest a small smd type of MOSFET for this purpose? Thanks

Regards,

SMDFan

tstalcup
Posts: 16
Joined: Tue Jun 08, 2010 2:53 pm

Re: Power supply conditioning...?

Post by tstalcup » Wed Oct 27, 2010 10:58 am

Just to be clear - the source and drain of the p-mosfet shouldn't be connected together, right?

tecoist
Posts: 160
Joined: Mon Aug 30, 2010 6:22 pm

Re: Power supply conditioning...?

Post by tecoist » Wed Oct 27, 2010 1:10 pm

That's correct. Shorting the source and drain together would turn the mosfet into an funky capacitor, and wouldn't do much in the way of polarity protection.

If you are looking for a surface-mount P-channel device for this application, the Si4403BDY would work. You want the threshold voltage to be low enough to turn the sucker on (and of course you need to be able to block about 2x the battery voltage and pass as much current as you need, with a reasonably low Rds(on)).

MichaelN
Support Volunteer
Posts: 1191
Joined: Fri May 08, 2009 6:14 pm
Location: Australia

Re: Power supply conditioning...?

Post by MichaelN » Wed Oct 27, 2010 3:01 pm

tstalcup wrote:Just to be clear - the source and drain of the p-mosfet shouldn't be connected together, right?
Oops - well spotted! I also screwed up the labeling of "S" and "D" for the N-channel MOSFET in the original. With mistakes like that, it's surprising my PCBs actually work!

Here's a corrected version:
Image

JonChandler
Posts: 226
Joined: Fri Jun 17, 2005 1:05 pm
Contact:

Re: Power supply conditioning...?

Post by JonChandler » Fri Oct 29, 2010 4:40 am

I came across the circuit suggested by MichaelN sometime back and decided to try it out.

I used the only MOSFET available at RadioShack. For some reason they only carry an N-channel, type IRF510. A P-channel blocking the high side would have been my choice, but I tried what was available. The circuit couldn't be simpler - I soldered it together on literally a scrap of perf board with a terminal block. I tested it mainly at 5 volts, with a few measurements at 10 volts, and measured the voltage drop across it vs. load current. I used 50 mA steps from 50 mA to 300 mA and 100 mA steps from there up to a total of 1 amp. It looks like I dry-labbed this, but the plot below shows my results.
MOSFET Vdrop.jpg
The voltage drop for small currents is very low. Above about 400 mA with this part, a series Schottky diode might result in the same performance. With careful selection of the MOSFET, the useful operating range could probably be extended.

And by the way, yes, the circuit is effective at blocking voltage when connected backwards.

The data sheet shows the same data I measured in Figure 1. The second line from the bottom is for a 5 volt supply.
mosfet figure 1.jpg
This is definitely recommended for applications with a current draw below a few hundred mA. A search of data sheets may show even better options.
Jon

JonChandler
Posts: 226
Joined: Fri Jun 17, 2005 1:05 pm
Contact:

Re: Power supply conditioning...?

Post by JonChandler » Fri Oct 29, 2010 4:49 am

Here are my test results for a 10V supply.
mosfet Vdrop @ 10v.jpg
Jon

Post Reply