On Thu, Mar 15, 2012 at 11:40 AM, Werner Almesberger <wer...@almesberger.net > wrote:
> > - the TPS76301 data sheets shows 1 uF input caps but clarifies > that no or only >= 47 nF caps are needed. They recommend larger > caps if there are large transients and the regulator is several > inches from the power supply. > > We have 100 nF. Given that our (external) power supply is very > far away, perhaps C234 and C145 should be increased ? U23 is > also far from C154, which may double as buffer for the 5 V > rail. > C234 yes is far from 5V rail about 3.74 inches, C145 is close to C154. Since Bom has 1uF/0402. Let C145 as 1uF directly, C234 goes for 10uF, how do you think ? > - U12 drops 5-1.8 = 3.2 V. If we assume the "High K" scenario > with an ambient temperature of 70 C, then we can burn 305 mW. > The maximum current would thus be 305/3.2 = 95.3 mA > Your analysis is reasonable. If we set @ 25 C, then 555mW, the maximum current would be 555/3.2 = 173.4 mA, but won't get there since it's output only 150 mA. @ 39 C, it's the temperature I measured for U12 after run 30 minutes, PD(max) = (125 - 39) / 180 = 477.8 mW, so maximum current = 477.8/3.2 = 149.3 mA > > 1V8 feeds (only) the video decoder, and there the digital core > and the PLL. Depending on what it does, the core's typical > current is specified as 90-106 mA and the PLL's as 11-12 mA. > > Thus we have Iout(max) = 95 mA vs. Iin(typ) = 118 mA. > > Should we worry ? > I cut L18 to measure my ADV7181C's IPVDD current( pin 31 ) It's typically around 11.4mA which meets to IPVDD CVBS input sampling at 54 MHz for typ. 11mA spec. The digital core 1.8V supplied by pin 24 for IDVDD, I can't measure it practically. But from rc1's http://en.qi-hardware.com/wiki/Milkymist_One_RC1_Reports#Measurement_of_Voltage.2FCurrent_during_Standby.2FQuiescent_Mode The IAVDD measured which also meets ADV7181B's spec. so can be also expectedly IDVDD must be kept less or equal than 106 mA for sure. Actually yes your concern is right that we should worry this LDO. But a next scenario is that shall we put M1 in ambient temperature of 70 C ? If we assume M1 I think to be about common use product instead of experimental platform then it's in safe for 149.3 mA. > > - the fuses (F2 and F3) may be a bit weak: at 60 C, their hold > current is only 1.5 A, which we could exceed on both 5 V rails > if high-current USB devices are connected. Not sure if we have > much of a choice, though. > Need more data for it, but will still meet same scenario like above. > - TI show on the first page of the PTH04000W data sheet that Rset > should be rated at 0.1 W. Our Yageo RC0402FR-072K32L only does > 1/16W. But TI relent to 0.05 W on page 7. > Nice catch on TI's goof, It's being safe to change as 0603 as like: RC0603FR-072K32L http://search.digikey.com/us/en/products/RC0603FR-072K32L/311-2.32KHRCT-ND/ Thanks, - adam
_______________________________________________ http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org IRC: #milkymist@Freenode