• State Resolved
  • Replies 1 reply
  • Subscribers 16 subscribers
  • Views 38 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

TS5A3359: Low simulated (spice) off resistance

Asger Overby
Prodigy 50 points
Part Number: TS5A3359
Other Parts Discussed in Thread: TINA-TI

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:

Fullscreen
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
* 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
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Fullscreen
1
2
3
4
5
6
7
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
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

  • +1
    TI__Mastermind 29638 points

    Hi Asger,

    I'm seeing lower than expected ROFF values as well. In PSPICE I get some strange results that fail to converge and in TINA frankly I get even lower resistance than you're seeing. 
    It's an older model and typically ROFF isn't something modeled. It really isn't something we spec out in general. You can kind of get of given the off isolation. If you know it's at 100Hz, you can say that Off-Iso will be closer to -85dB. At 5V at the input, I calculate about 281uV on the drain side.
    I would trust that value as it's from actual lab data vs the model.

    Long story short, I don't think it's necessarily a bug but rather something just not considered for the model itself. Definitely wouldn't expect that performance from the mux. 
    Thanks,
    Rami