Help with a boost converter

[vargoal]

Hi,

I want to make a boost converter for my arduino board. I wanted it to be able to power an xbee as well. The arduino uses around 100-200 mA depending on how many sensors I have plugged in. The xbee uses another 200-300 mA so I wanted the power supply to work up to at least 500 mA. After looking through many data sheets I found the LT1038B(http://cds.linear.com/docs/en/datasheet/1308abfb.pdf). I was looking to see if anyone that has experience with boost converters to tell me if this would work for my application. Also advice about if the typical application circuit they have would work or if I need to tweak the values of the inductors and caps. Thanks for any feedback you have.

[rx8pilot]

I have done a number of boost converters. The learning curve was heavily weighted on PCB layout. The Linear part you linked is probably not a bad start - there are many. I have the most experience with TI's offerings but have used all sorts of Linear parts with great success.

In general, the schematics are simple for a boost converter made with a controller. This is especially so with integrated MOSFET. Look for some layout resources on SMPS PCB layout. It is CRITICAL to get the switch node as low inductance as you can. That is accomplished by having the inductor and pass diode as close as possible to the SW pin. In addition, I would add an RC snubber option that you can populate if it is noisy.  The snubber can reduce noise ( generally at the expense of efficiency).

On the FB resistors, I would also add a pad for a small value cap to have the option to control noise at the expense of transient response. The two FB resistors should be very close to the FB pin. On the output, consider the option of having some small value caps to filter the higher frequencies that larger caps will simply pass through.

Take a look at the PCB layouts on various demo boards, there is a lot to be learned there. My first few boost converters had poor performance that was entirely due to the PCB layout.

Measuring and evaluating an SMPS is an art and a science all its own.

[vargoal]

I have an initial board layout for this boost converter(see attachment). I tried to layout the components according to the LT1308 datasheet. Any feedback or questions welcome.

[vargoal]

I revised the board a little. Notable changes include a lot larger traces for the high power areas. Also I made it a little bigger.

[rx8pilot]

The second revision is a major improvement over the first.

If you can, I would consider putting a simple RC snubber on the SW node. You don't have to use it, but if the circuit is too noisy a snubber can really help. you may also consider an optional cap on the FB pin to filter HF noise that can cause instability. It would be a fairly small value at most. It does not hurt to have a few options that will otherwise be difficult after the PCB is fabbed.

What you have should work reasonably well. Use a low inductance probe setup when you test it to get a noise signature under load. Look for ringing on the SW node that can cause problems with downstream circuits, especially analog. You just have to measure it right or you may see phantom ringing from poor measurement technique causing you to chase your tail.