TS5A3359: Low simulated (spice) off resistance

Tool/software:

Hi,

We are using the mentioned part as a gain switch, however its performance is much worse in simulation that in reality. Could there be a bug in the off resistance model? The problem is reproducible in LTspice using both the PSPICE and the TINA-TI models.

At 100 Hz in the AC analysis the off resistance is approximately 1/3 MΩ. Corresponding to 1 MΩ Roff between COM and each of the NO0, NO1, and NO2. Based on the off state leak current and bandwidth specs I would expect Roff to be at least two orders of magnitude higher.

Could you please check the models?

TIA

/Asger

Netlist:

* gain-switch.asc
V1 VCC 0 3.3
V3 in 0 1.65 AC 1
XU3 gain_a gain_b in VCC N001 N002 N003 0 TS5A3359
V4 gain_a 0 {gain_aa}
V5 gain_b 0 {gain_bb}
R1 out N001 500k
R3 out N003 5k
R2 N002 out 50k
R5 0 out 50
.step param n 1 4 1
.param gain_aa table(n,   1, 0,    2, 3.3,   3, 0,       4, 3.3)
.param gain_bb table(n,   1, 0,    2, 0,      3, 3.3,    4, 3.3)
; calculate resistance at 100 Hz, display result in (mag, phase)
.option meascplxfmt=polar
.meas AC res FIND V(out,in)/I(R5) AT 100
.ac dec 10 20 100k
.lib TS5A3359.lib
.backanno
.end

From the .log file
Measurement: res
  step	v(out,in)/i(r5)	at
     1	(377349,-8.13902°)	100
     2	(252343,-4.43533°)	100
     3	(46147.8,-0.604821°)	100
     4	(4960.34,-0.0610923°)	100