# Can I use digital pots for my feedback resistors to a buck-boost regulator?

I'm working on a design, and this will be my first attempt at a buck-boost regulator.

I need the output to be variable because my application calls for different voltages under certain conditions.

The buck-boost regulator I'm considering is the LT8705

Most of the reference designs I've seen have used either a fixed output, or in some cases, I've seen where a potentiometer was added but never a digital one.

I haven't yet decided on specific digital pot, but let's assume I decide to use this MCP4021/2/3/4 to replace the 392k and 10k resistors connected to the FBOUT pin on the right side of the image shown above. I haven't calculated what exact resistors I'll need so for the purpose of this question, assume that the digital pot's resistance is in range.

Max current through the resistors is 2.5mA. My max output voltage would be 28V. If I set my pot to be greater than 11.2kohm, the current would be < 2.5mA through it.

$$R_{min}= \frac {28V_{max}}{2.5mA_{max}} = 11.2k\Omega$$

If this is true, then my next question is what considerations should I be aware from the following snippet from the datasheet

..set by an external feedback resistive divider carefully placed across the output capacitor.

What would 'carefully placed' entail ? Would placing this IC near the regulator potential disrupt its behaviour ? Any additional feedback would be appreciated.

Thanks!

• if you haven't looked at that already, look at the PCB layout recommendation on pp. 35-36 in the LT8705 datasheet. – Nick Alexeev Dec 8 '13 at 6:01
• @NickAlexeev yup I did. Looked a design app note from them too. Couldnt find the answer to my question about digital pots / digital pot layout with this regulator. – efox29 Dec 8 '13 at 6:29
• The datasheet for that (and most, I think) digital pot requires that the voltage on the pot terminals be between Vss and Vdd, with a maximum Vdd of 5.5 volts - if your desired output voltage is above 5.5 volts, you can't use this digital pot. – Peter Bennett Dec 8 '13 at 8:03
• I'd be very wary of this. It's the sort of thing that may work perfectly well most (or all) of the time but if it gave problems I'd not be surprised. It's usual to place a small capacitor across the upper feedback resistor (from Vout to FB pin). Deep-ending on other factors this may cause immense improvements to transient response. If the digital pot has characteristics which vary dynamically with Vout transients (perhaps due to noise getting into switch control ccts or whatever) then it may have the effect of a semi randomly variable AC coupling across the upper FB R. Murphy says this is ... – Russell McMahon Dec 8 '13 at 11:33
• ... more likely to produce bad results than good ones. | General layout of components in the loop response governing sections of switching regulators is often important. The aim is to get signals to where they should go and avoid signals coming from unintended places and to not load or modify signals in unintended way. Digital pots offer more ways of getting this wrong than do most 2 terminal passive resistors. -> So - try it, watch it carefuilly, don't let it fool you. -> Report back. – Russell McMahon Dec 8 '13 at 11:36