AD was comparing apples and oranges. The LTC6957-3 has CMOS outputs, which go from 0.2V to Vdd-0.3V (with a 3mA load). Perhaps they were looking at the LTC6957-1 which is LVPECL and would have a rather small swing. The LTC6957-3 should drive the AD9852 if its requirements are as noted below.
Bob LaJeunesse >________________________________ > From: d0ct0r <[email protected]> >To: [email protected] >Sent: Friday, April 25, 2014 1:33 PM >Subject: [time-nuts] CMOS level difference for LTC6957-3 and AD9852 > > > >To whom it may concern, here is the note regarding of CMOS level >difference for two product AD9852 from AD and LT6957 from Linear. > >My attempts to feed AD9852 directly from LTC6857-3 has failed. AD9852 >just ignored the REFCLOCK signal coming from LTC6957-3. So, I'll need to >think about some solution how to match the levels and do not impact the >source signal too much. > >(Received from tech. department of AD): >============================================================================ >... the logic levels for the AD9852 and LTC6957 >don't line up. From page 5 of the AD9852 datasheet, VIH minimum is >2.3V whereas the LTC6957 says it's VOH could be as low as 2.08V. This >means the LTC6957 could output 2.08V for a logic high, but our part >won't recognize it as it needs at least 2.3V to be a logic high. > >It's the same problem for low logic voltages. The AD9852 datasheet >says VIL max is 1V whereas the LTC6957 says it's VOL max is 1.63V or >1.67V depending on the temperature range of your part. So the LTC6957 >could output 1.63V for a logic low, but our part won't recognize it as >it needs a voltage below 1V to be a logic low. > >============================================================================= > > >--- >WBW, > >V.P. >_______________________________________________ >time-nuts mailing list -- [email protected] >To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >and follow the instructions there. > > > _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
