Settling Time MAX5875

Question

Hello, 
 
 we are using the MAX5875 as a high speed and high accuracy DAC. But we have problems to meet our requirements because of a fairly long settling time. 
 After a step of the output, the voltage rises within a couple of nanoseconds. But it overshoots the desired value by approximately 9mV (at 5V full scale output) and then it takes hundreds of microseconds to slowly decrease and reach the final value. Strange thing: This error seems to be independant of the step size - whether you step full range or just 100mV, there is always this 9 mV overshoot. 
 
 All supply and reference voltages are stable and do not show any change during the step of the DAC values. Do you have an idea what could be the reason for this behaviour?

The schematic of the output: 
 
 Full Scale Range is 20 mA and REFIO is an external 1.2V reference voltage source.

red and yellow: Analog output signal zoomed in: 
 
 Analog Output zoomed out:

PMH2023 · Accepted Answer

Hello,

DAC output signal will also contain some amount of clock feedthrough that remains signal independent. Since the settling response is being limited by the LTC6269 having only a 350 MHz -3dB when operating as buffer.............perhaps one can add some amount of capacitance (i.e. 5 pF) across each of the 50 ohm load resistors that limits this high frequency energy from getting into the op amp inputs. Also..........unclear if the LTC6229 is operating off of +5V supply while trying to generate a voltage output that swings 5 V thus not having any headroom (even if it is rail to rail).

Electronic Design Services - PartnerZone

Products Mentioned

Settling Time MAX5875