2
\$\begingroup\$

So... I'm trying to build an EMG sensor for biofeedback purposes, but I'm having loads of trouble both on the breadboard and on the CAD side as well. I'm using a M1-processor mac mini which could be a problem regarding ngspice.

Please note that I'm a newbie to circuit design, KiCad and SPICE simulation.

KiCad schematic

This is my circuit. If there's something blatantly wrong with it (besides not adding bypass caps, but I don't think I'll need them for simulation), I wouldn't really be surprised. The SPICE model for INA128 I'm using is from TI: https://www.ti.com/lit/zip/sbom764

The choice of INA128 is basically because it's the part I have, but most likely not the one I'll finally end up using. It's just a DIP-8 package that's easier to do on the breadboard.

noise and sine outputs

Now... if I remove the instamp complex from the schematic and just run the simulator, I get a nice noise and a nice noisy sine wave like I wanted.

If I understand the INA128 correctly, it has a gain of 1+50k/[gain resistor], which means putting a 47ohm resistor there should give me a gain of a bit more than 1000.

Alternate pin configs

Here I should have specified hopefully the correct pin numbers.

BUT... when I then run the simulation, I get this wonderful error:

ngspice error

Mind you, this is the same error I got when I tried to do ngspice's own example of working with an opamp: http://ngspice.sourceforge.net/ngspice-tutorial.html#OpAmp

A google told me that this has something to do with ngspice itself, so I finally managed to install ngspice via homebrew (don't get me started on xcode), saved the SPICE Netlist created by KiCad to a separate file:

KiCad SPICE Netlist

and ran ngspice from the command line, and got THIS wonderful message:

joonas@Celephais-2 SPICE files % ngspice instamp.cir 
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/spice2poly.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/spice2poly.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/spice2poly.cm' (no such file), '/usr/local/lib/spice2poly.cm' (no such file), '/usr/lib/spice2poly.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/spice2poly.cm couldn't be loaded!
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/analog.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/analog.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/analog.cm' (no such file), '/usr/local/lib/analog.cm' (no such file), '/usr/lib/analog.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/analog.cm couldn't be loaded!
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/digital.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/digital.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/digital.cm' (no such file), '/usr/local/lib/digital.cm' (no such file), '/usr/lib/digital.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/digital.cm couldn't be loaded!
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtradev.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtradev.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtradev.cm' (no such file), '/usr/local/lib/xtradev.cm' (no such file), '/usr/lib/xtradev.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtradev.cm couldn't be loaded!
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtraevt.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtraevt.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtraevt.cm' (no such file), '/usr/local/lib/xtraevt.cm' (no such file), '/usr/lib/xtraevt.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/xtraevt.cm couldn't be loaded!
Error opening code model "/opt/homebrew/Cellar/ngspice/36/lib/ngspice/table.cm": dlopen(/opt/homebrew/Cellar/ngspice/36/lib/ngspice/table.cm, 0x0002): tried: '/opt/homebrew/Cellar/ngspice/36/lib/ngspice/table.cm' (no such file), '/usr/local/lib/table.cm' (no such file), '/usr/lib/table.cm' (no such file)
Error: Library /opt/homebrew/Cellar/ngspice/36/lib/ngspice/table.cm couldn't be loaded!
******
** ngspice-36 : Circuit level simulation program
** The U. C. Berkeley CAD Group
** Copyright 1985-1994, Regents of the University of California.
** Copyright 2001-2020, The ngspice team.
** Please get your ngspice manual from http://ngspice.sourceforge.net/docs.html
** Please file your bug-reports at http://ngspice.sourceforge.net/bugrep.html
******

No compatibility mode selected!

warning, model type mismatch in line
    s11 41 42 1 41 sp
warning, model type mismatch in line
    s12 4 43 1 41 sm
warning, model type mismatch in line
    s21 45 46 2 45 sp
warning, model type mismatch in line
    s22 4 47 2 45 sm

Circuit: KiCad schematic

Error on line 0 :
  a$poly$e.x1.x1.egnd %vd [ vcc 0 vee 0 ] %vd ( x1.x1.99 0 ) a$poly$e.x1.x1.egnd
 MIF-ERROR - unable to find definition of model a$poly$e.x1.x1.egnd
    Simulation interrupted due to error!

Now... this is also similar to the error I got when trying to run ngspice's opamp example from the command line and THAT I managed to get plotted using MacSpice... but that didn't help here:

MacSpice 32 -> source /Users/joonas/Documents/KiCad/SPICE\ files/instamp.cir

Circuit: .title KiCad schematic

Note: Found one or more Spice-2 'poly' devices; will convert to Spice-3
      'B' sources. Use 'listing physical expand' to check the result(s).
On line 6:
    v5 noise net-_v5-pad2_ trnoise(20u 0.5u 0 0)
    Error: Unknown parameter (trnoise).
On line 325 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    .model 1:1:jx pjf(is=15.00e-12 beta=31.58e-3 vto=-1)
    Missing level parameter - using level=1
On line 427 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    .model 1:2:jx pjf(is=15.00e-12 beta=31.58e-3 vto=-1)
    Missing level parameter - using level=1
On line 179 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    s:1:11 1:41 1:42 noise 1:41 1:sp 
    Warning: Unable to find model '1:sp' for this device - using default.
On line 185 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    s:1:12 vee 1:43 noise 1:41 1:sm 
    Warning: Unable to find model '1:sm' for this device - using default.
On line 193 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PS-- hit return for more, ? for help -- 
PICE_AIO/INA128.LIB" included at line 2:
    s:1:21 1:45 1:46 sine 1:45 1:sp 
    Warning: Unable to find model '1:sp' for this device - using default.
On line 199 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    s:1:22 vee 1:47 sine 1:45 1:sm 
    Warning: Unable to find model '1:sm' for this device - using default.
On line 215 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    .model 1:sp vswitch(ron=10 roff=100e3 von=.7 voff=1) 
    Error: Unknown model type vswitch - ignored.
On line 217 of "/Users/joonas/Documents/KiCad/Projects/EMG simulator/INA128_PSPICE_AIO/INA128.LIB" included at line 2:
    .model 1:sm vswitch(ron=10 roff=100e3 von=-.7 voff=-1)
    Error: Unknown model type vswitch - ignored.
MacSpice 33 -> run
On line 6:
    v5 noise net-_v5-pad2_ trnoise(20u 0.5u 0 0)
    Warning: Source 'v5' has no DC value specified, 0V assumed.
On line 5:
    v3 sine noise sin(0 100u 2k)
    Warning: Source 'v3' has no DC value specified, using value at transient time = 0s.
On line 24:
    .tran 10u 2m
    Warning: Unspecified TRAN parameter, TMAX will default to 4e-05s.
On line 24:
    .tran 10u 2m
    Warning: Basic iteration method failed to find operating point; iteration limit (100) exceeded.
    TRAN: initial timepoint: Trouble with node 'net-_r1-pad2_'.
    Note: Looking for operating point by Gmin stepping:
        Trying Gmin = 1.0000E-02  succeeded
        Trying Gmin = 1.0000E-03  succeeded
        Trying Gmin = 1.0000E-04  succeeded
        Trying Gmin = 1.0000E-05  failed
    TRAN: initial timepoint: Trouble with node 'net-_r1-pad1_'.
    Warning: Standard Gmin stepping failed.
    Note: Try 'set gminsteps = 1' to enable adaptive Gmin stepping.
    Note: Starting source stepping:
-- hit return for more, ? for help -- 
        DC sources scaled to   0.0000%  succeeded
        DC sources scaled to  10.0000%  failed
    Warning: Source stepping failed.
    Note: Try 'set srcsteps = 1' to enable adaptive source stepping algorithm.
    TRAN: initial timepoint: Trouble with node 'net-_r1-pad1_'.
Error: Iteration limit reached.
Error: run simulation(s) aborted.

To begin with, apparently MacSpice has trouble with trnoise. So what the bleep have I done wrong, is there any way to make this work (so if I've understood correctly I would get a sine-ish wave with peaks at +/- 100 mV at output)? I'd like to get the simulation working so I'd have at least some hope for making it work on the breadboard.

\$\endgroup\$
4
  • \$\begingroup\$ You have ideal 0 Ohm impedance for all sources which is unrealistic. \$\endgroup\$
    – Tony Stewart EE75
    Jan 24, 2022 at 18:54
  • \$\begingroup\$ Thanks for pointing that out. It’s certainly something I would need to take into account when I actually get simulation working, but currently I don’t think setting impedance values would affect the main problem at hand. \$\endgroup\$
    – Joonas Joensuu
    Jan 25, 2022 at 16:19
  • \$\begingroup\$ In Falstad, this you can't even put two caps in parallel without a fault because dV/dt would make Ic infinite in circulating currents. A higher CM impedance permits better attenuation. I suggest you add some R now \$\endgroup\$
    – Tony Stewart EE75
    Jan 25, 2022 at 16:32
  • \$\begingroup\$ So with Transmission Line theory applied to simulations, 0 Ohms differential merges even and odd mode impedances for CM analysis. I'm not sure but I see math or declaration faults with this. \$\endgroup\$
    – Tony Stewart EE75
    Jan 25, 2022 at 16:40

2 Answers 2

Reset to default
2
\$\begingroup\$

This seems to be a ngspice installation problem, when code models are not found. And you have to set the compatibility to "LTSPICE and PSPICE" in Eeschema or set ngbehavior=ltps in .spiceinit.

There may be a bug in the circuit, as the line

X1 noise sine VCC VEE output 0 Net-_R1-Pad1_ Net-_R1-Pad2_ INA128

only invokes the pure sine of V3 between + and - input. Changing it to

X1 0 sine VCC VEE output 0 Net-_R1-Pad1_ Net-_R1-Pad2_ INA128

will yield something like (ngspice-36, Windows 10) this plot:

output plot

\$\endgroup\$
1
\$\begingroup\$

LTSpice

Got the Netlist running under LTspice, which I just needed to reinstall.

Edit: Apparently LTSpice processes noise simulation differently from ngspice, so that bit failed, but that's more related to ngspice and the instamp as such worked as expected.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ If you need (white) noise in LTspice then you need a behavioural source with V=white(time*x)*y, where x means a minimum of twice the bandwidth of interest. The levels are -0.5 V ... 0.5 V, by default, so choose y appropriately. \$\endgroup\$
    – a concerned citizen
    Jan 25, 2022 at 17:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.