Chris and Mark, On Mon, Oct 6, 2014 at 8:29 PM, Chris Zhong <zyw at rock-chips.com> wrote: > > On 10/01/2014 06:52 AM, Mark Brown wrote: >> >> On Tue, Sep 30, 2014 at 06:57:51PM +0200, Heiko St?bner wrote: >>> >>> Am Dienstag, 30. September 2014, 09:43:47 schrieb Doug Anderson: >>>> >>>> On Tue, Sep 30, 2014 at 2:44 AM, Chris Zhong <zyw at rock-chips.com> wrote: >>>>> >>>>> - REGULATOR_LINEAR_RANGE(700000, 0, 63, 12500), >>>>> + REGULATOR_LINEAR_RANGE(712500, 0, 63, 12500), >>>> >>>> I have no problem with this (and the current "dts" file I see don't go >>>> this low anyway), but it doesn't match the datasheet I have right now >>>> which says "0.700 V" is the min. >>> >>> Then it would be problem. >>> The original statement above says, that a register value of "0" equals >>> 700mV, >>> so changing this in this way would mean everything runs 12.5mV above the >>> target voltage. >>> And I can confirm, that the current datasheet lists 700mV as equallying >>> register value 0. >> >> Has anyone got a voltmeter? > > The latest datasheet has fixed this error. > The register value of "0" equals 712.5mV, so the max value can reach 1.5v > but not 1.4875v. OK, that sounds like this is the right change, then. To Mark's point, I measured voltage using the builtin voltage measurement feature this board (not 100% sure the accuracy / precision guarantees on them, but...): Asked for 1.300 volts on vdd_arm (dcdc1). * Before: got 1.326 * After: got 1.310 Asked for 1.3 on vdd_gpu (dcdc2): * Before: 1.326 * After: got 1.314 Asked for 1.0 on vdd_gpu: * Before: 1.038 * After: got 1.026 Asked for .8 on vdd_gpu * Before:.836 * After: got .822 It almost sounds like we're still running a little too high for voltages, but this gets us in the right direction I guess. -Doug
