Questions about DERIV measure type

[Neal Wood]

Hello Xyce Team.

I have a few questions pertaining to DERIV.

Question 1:

Am I correct that .MEASURE DC for multiple sweeps should only measure one (the first) value it encounters? E.g.,

* N-channel JFET output characteristics
vd d 0 dc 0
vg g 0 dc 0
vs s 0 dc 0
.model dut njf
j1 d g s dut
.dc lin vd 0 1 0.1 lin vg 0 -1 -0.1
.measure dc tmp1 deriv id(j1) at=1
.print dc format=gnuplot v(d,s) v(g,s) id(j1)
.end

TMP1 = 2.100000e-04 for AT = 1.000000e+00

Question 2:

Why does .MEASURE DC_CONT for multiple sweeps only measure one value? E.g.,

* N-channel JFET output characteristics
vd d 0 dc 0
vg g 0 dc 0
vs s 0 dc 0
.model dut njf
j1 d g s dut
.dc lin vd 0 1 0.1 lin vg 0 -1 -0.1
.measure dc_cont tmp2 deriv id(j1) at=1
.print dc format=gnuplot v(d,s) v(g,s) id(j1)
.end

TMP2 = 2.100000e-04 for AT = 1.000000e+00

Question 3:

Why is the use of WHEN for DC sweeps undocumented and why does it's behaviour differ from the previous examples for the case outlined below?

* N-channel JFET output characteristics
vd d 0 dc 0
vg g 0 dc 0
vs s 0 dc 0
.model dut njf
j1 d g s dut
.dc lin vd 0 1 0.1 lin vg 0 -1 -0.1
.measure dc tmp3 deriv id(j1) when v(d)=1
.measure dc_cont tmp4 deriv id(j1) when v(d)=1
.print dc format=gnuplot v(d,s) v(g,s) id(j1)
.end

TMP3 = 2.100000e-04 at VD value = 1.000000e+00
Measure Start VD Value= 0.000000e+00    Measure End VD Value= 1.000000e+00

TMP4 = 2.100000e-04 at VD value = 1.000000e+00
TMP4 = 1.900000e-04 at VD value = 1.000000e+00
TMP4 = 1.700000e-04 at VD value = 1.000000e+00
TMP4 = 1.500000e-04 at VD value = 1.000000e+00
TMP4 = 1.300000e-04 at VD value = 1.000000e+00
TMP4 = 1.100000e-04 at VD value = 1.000000e+00
TMP4 = 9.000000e-05 at VD value = 1.000000e+00
TMP4 = 7.000000e-05 at VD value = 1.000000e+00
TMP4 = 5.000000e-05 at VD value = 1.000000e+00
TMP4 = 3.000000e-05 at VD value = 1.000000e+00
TMP4 = 1.000000e-05 at VD value = 1.000000e+00
Measure Start VD Value= 0.000000e+00    Measure End VD Value= 1.000000e+00

Question 4:

Is there a plan to allow DERIV in a .PRINT statement, similar to HSPICE?

Thank you in advance.

Neal

[Neal Wood]

My mistake, WHEN is in fact documented for DC and DC_CONT sweeps. My main question is why is there a difference between the AT and WHEN examples? Thanks.