Simulating variable output impedance

[abzug]

From an email I got from Analog Devices:

"The output impedance of the AD1955 analog output has a range of 20k - 80k ohms, depending on the signal level. This filter value should change linearly (inverse, as well) with respect to the capacitor value."

How do I simulate this in LTSPICE?

[KSTR]

Ahem, this refers to the differential current outputs of the DAC?

Non-perfect current source could be simmed with a "resistor" parallel to a perfect current source. The value of the resistor must be made somewhat signal dependant, that means it must be constructed from behavioural sources, something along the lines of V=I(sense)*(20k+ 60k*F(signal)). Assuming that this F(signal) function is known...

Klaus

[sawreyrw]

abzug,

That quote from Analog Devices doesn't make any sense. The analog outputs are currents that depend on the digital input. The AD1955 should be used with an active current to voltage converter as shown on the data sheet. Doing this holds the voltage at the analog outputs constant.

Rick

[abzug]

Originally posted by sawreyrw
That quote from Analog Devices doesn't make any sense.

Sure it does. A perfect current source has infinite output impedance. A real one is finite, and in this case, it's signal-dependent.

The AD1955 should be used with an active current to voltage converter as shown on the data sheet. Doing this holds the voltage at the analog outputs constant.

That's why I'm using Hawksford's I/V. However, no I/V can do this with absolute perfection, and in practice there will be a tiny voltage variation as a result. I want to simulate THD down to 0.1 ppm so I cannot ignore these effects.