LTspice 26.0 Current not found for .meas OK in viewver

Hi esabard38V2 ,

Will take a look.

mike

Hi esabard38V2 ,

Circuit doesn't run for me. Time step too small.

mike

Hi!

Strange, it runs very quickly on my side : 

Options:ITL1=800 ITL2=400 ITL4=800 gmin=1e-9 abstol=1e-10 reltol=1e-3 vntol=1e-6
solver = Alternate
Maximum thread count: 22
tnom = 27
temp = 27
method = modified trap
gmin = 1e-09
abstol = 1e-10
Compress1stOrder = true
Compress2ndOrder = true
Direct Newton iteration succeeded in finding operating point.
Total elapsed time: 0.086 seconds.

You may can delete the voltage source, or change it to 0V.

Hi!

I deleted some unused components.

Problem comes from : [Lib file]

.SUBCKT VDRH05B035 1  2 PARAMS: TOL=0
X7  1 2 VDR_BASE_MODEL  PARAMS: T = {1+TOL/100}  C= 560pF  L=8nH  B1= 2.00338249979319
+  B2=0.0720452771473565  B3=-0.000422278414133302  B4=0.0245178498406733
.ENDS
*$

Into [Lib file]

.SUBCKT VDR_BASE_MODEL 1 2  PARAMs:  B1=1 B2=1 B3=1 B4=1  L=1h C=1pf  T=1
C_C1         2 4  {C}  TC=0,0 
V_V1         4 5 0Vdc   
[...]
.ENDS

and symbol file : Varistor_Vishay_VDR_VDS_MLV_VDRS_VDRH_VDRUS.asy [Symbol1]

PINATTR PinName A
PINATTR SpiceOrder 1
PIN 16 96 NONE 0
PINATTR PinName B
PINATTR SpiceOrder 2

for Varistor_Vishay_VDR_VDS_MLV_VDRS_VDRH_VDRUS - Copie.asy [Symbol2 for test]

PIN 16 16 NONE 0
PINATTR PinName 1
PINATTR SpiceOrder 1
PIN 16 96 NONE 0
PINATTR PinName 2
PINATTR SpiceOrder 2

The simulation can work even if the symbol PinNames do not match the SUBCKT definition in the library.
The curves in the viewer display correctly, but the .meas commands do not work when the PinNames of the symbol differ from those in the symbol and the .lib file.

Current_Meas (2).zip

What is the expected behavior?

Thank you, Emilien

Hi esabard38V2 ,

I'll try it again.

mike

Hello  MStokowski 

Have you try the file above?

Another exemple here, but instead of .meas, I used the current from U2 in BV source. It's not working. Explanation are included in the file.

BUG_Current_Comp.zip

Thank you !

Hi Emilien,

I'm afraid this is a regression bug introduced by some recent optimizations. It'll be fixed in the next release.

Unfortunately, the only workaround is not to use the pin names but the port names of the actual sub-circuit. ("Ix(U9:1)" in your case)

Best Regards,
Mathias

Hello Mathias,

Thank you rou answer. It's works with the workaround for measurement or using the expression in BV source.

Bye