Unrecognised parameters when passed through symbol

Dear fellows,

since my last QSPICE update:


I started having some troubles with parameters passed through symbols.
For example, here is a case of RC network defined in these two symbols. The capacitances are parametrized to “scaleF” and its value is given by the parameter “comp”

Here the subcircuits called by the allegedly problematic symbols:

.subckt 2lpcb_2p5by2p5EP EP Tamb

.param Rth1=2.88014 Rth2=0.33827 Rth3=8.09952 Rth4=9.7919
.param Rth5=13.74416 Rth6=10.58544 Rth7=13.66308
.param scaleF=1
.param Cth1=0.01785/scaleF Cth2=0.00692/scaleF Cth3=0.05294/scaleF
.param Cth4=16.09951/scaleF
.param Cth5=3.24026/scaleF Cth6=0.2653/scaleF Cth7=3.25879/scaleF

Rth1 EP n1 Rth1
Cth1 EP n1 Cth1
Rth2 n1 n2 Rth2
Cth2 n1 n2 Cth2
Rth3 n2 n3 Rth3
Cth3 n2 n3 Cth3
Rth4 n3 n4 Rth4
Cth4 n3 n4 Cth4
Rth5 n4 n5 Rth5
Cth5 n4 n5 Cth5
Rth6 n5 n6 Rth6
Cth6 n5 n6 Cth6
Rth7 n6 Tamb Rth7
Cth7 n6 Tamb Cth7

.ends 2lpcb_2p5by2p5EP

.subckt 4by4QFN nPH nPL nEP

.param Rmutual=0.24629 RL1=0.54579 RL2=2.46275 RL3=0.45178
.param RH1=RL1 RH2=RL2 RH3=RL3
.param scaleF=1
.param Cmutual=0.00819/scaleF CL1=0.00515/scaleF CL2=1.67e-4/scaleF CL3=3.8e-5/scaleF
.param CH1=CL1 CH2=CL2 CH3=CL3

R10 nPH nPL Rmutual
C10 nPH nPL Cmutual

R1 nPL nL1 RL1
C1 nPL nL1 CL1
R2 nL1 nL2 RL2
C2 nL1 nL2 CL2
R3 nL2 nEP RL3
C3 nL2 nEP CL3

R4 nPH nH1 RH1
C4 nPH nH1 CH1
R5 nH1 nH2 RH2
C5 nH1 nH2 CH2
R6 nH2 nEP RH3
C6 nH2 nEP CH3

.ends 4by4QFN

So far this technique worked great for me.
Have I been doing something unorthodox so far and maybe I have been caught :slight_smile: ?
Thanks in advance for your help!

Forgot to add this detail: this behavior seems to be weirdly random.
Sometimes everything works as usual, sometimes it doesn’t…

March 16


February 11

There is no difference, no failures either

Hi Belov, thanks a lot for sharing.
Could you also show the symbol properties window ?
Also, can you show me how you pass the parameter “scaleF” from the schematic into the symbol?
I’d like to see if I made a mistake in how I package the symbols.

I am trying different methods, including yours, and I still get the problem…and it is intermittent.
So frustrating…
By the way, I just noticed you are using a QSPICE release from more than a month ago. Maybe the issue is really in the latest update?

@Engelhardt sorry to bother you, have you or your team recognized something odd with parameters passing through symbols ?

Is there a way to get back to a previous QSPICE release ?

I think this is how to replicate your bug. The parameter cannot be passed into symbol through instance parameter. The formal method is to email this to Mike for review. I guess it may related to this update : 03/15/2025 Further improved the speed of evaluating user-defined parameters.
Mike email can be found in Qspice > Help > About
[I checked a previous version and can run this schematic]

Next time, please upload your schematic and library (rename with .txt and can upload library). This will save us time in replicating the problem.

subcktlib.qsch (3.2 KB)
subcktlib.txt (1.1 KB)

1 Like

thank you so much for verifying this.
I will keep your advice for next time I’ll seek for help.

Both versions work, see above, on February 11 and March 16

@belov You didn’t update Qspice64 and Qspice80. Can see they are still old version in your rev history. But your GUI QUX.exe is updated. Not sure why. This issue possibly related to SPICE engine instead of GUI program.

I used the full version as well, it didn’t affect the results. I have several versions of QUX to test.

But your Qspice64 and Qspice80 are month ago. If you update your Qspice completely, and load the sch and library I upload, you should replicate this issue.

QUX.exe is just GUI program. It converts sch into netlist. What really run the netlist is Qspice64 or Qspice80.exe


solvers last

Problem with QUX https://forum.qorvo.com/t/fatal-error-realloc-cant-allocate-91964-bytes/21862.

1 Like

Last version

I think this issue is very confusing. Let based on only the latest version for discussion.
@Egrana is correct that it nature is unpredictable. So far, my duplication method is that, if it can run successfully, change all these parameters name, for example, change them all to “xyz” or “abc” etc… rerun and this warning may appear.

1 Like

Apparently, there are problems with the new version after all.

1 Like

thank you both for the verification… I started to think I was allucinating :rofl:

Only “func” works

The latest release from today solves this issue.