ngspice / Support Requests / #32 Resistor Parameter Sweep

Holger Vogt - 2018-09-02

Please edit your post and mark your input file contents as 'code'

Holger

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Holger Vogt - 2018-09-02

What do you want to do?
Run a transient analysis with fixed RSENSOR
Run a dc sweep with varying resistor RSENSOR
Sweep the resistor RSENSOR, and each sweep step run a transient analysis

Last edit: Holger Vogt 2018-09-02

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Todd Gilbert - 2018-09-02
  
  Hi Holger,
  
  I want to run a dc sweep with the RSENSOR resistor, and monitor the v(12) and v(14) levels. I believe I'm not using the dc sweep correctly, as I see no change in those nodes. I left the transient analysis in the deck for troubleshooting purposes. Try removing the transient analysis asterisks, and running the transient analysis - try a couple different RSENSOR values (1k and 300k). You will see the resulting behavior with those two resistor values.
  
  Best Regards,
  
  Todd
  
  From: Holger Vogt h_vogt@users.sourceforge.net
  Sent: Sunday, September 2, 2018 9:25 PM
  To: [ngspice:support-requests]
  Subject: [ngspice:support-requests] #32 Resistor Parameter Sweep
  
  What do you want to do?
  Run a transient analysis with fixed RSENSOR
  Run a dc sweep with varying resistor RSENSOR
  Sweep the resistor RSENSOR, and each seep step run a transient analysis
  
  [support-requests:#32]https://sourceforge.net/p/ngspice/support-requests/32/ Resistor Parameter Sweep
  
  Status: open
  Group: v1.0 (example)
  Created: Sun Sep 02, 2018 06:43 PM UTC by Todd Gilbert
  Last Updated: Sun Sep 02, 2018 09:18 PM UTC
  Owner: nobody
  Attachments:
  
  nat_semi.libhttps://sourceforge.net/p/ngspice/support-requests/32/attachment/nat_semi.lib (422.8 kB; application/octet-stream)
  
  Hi,
  
  I need some assistance using the .dc command to perform a parametric sweep analysis on a resistor. I've included the spice deck below, and attached the library file. The v(12) and v(14) state should change when RSENSOR is varied somewhere between 10k and 270k, but the output doesn't show the change. If I manually select RSENSOR values and run transient analyses I see the output change states.
  
  Thanks,
  Todd
  
  .include nat_semi.lib LMC6484A/NS
  
  //////////////////////
  Air Separator Simulation
  //////////////////////
  
  190332 Sensor Board
  U1A
  XU1A 3 2 4 0 1 LMC6484A/NS
  VCC 4 0 12
  C73 2 0 .001u
  R130 1 2 261k
  R131 4 3 22.1k
  R132 3 0 22.1k
  R133 1 3 47.5k
  
  U1B
  XU1B 5 6 4 0 7 LMC6484A/NS
  R145 6 0 1k
  R146 6 7 4.99k
  R136 5 16 2.2M
  
  U1C
  XU1C 10 9 4 0 8 LMC6484A/NS
  R137 4 10 1k
  R138 10 0 866
  R139 7 9 619k
  C74 9 0 33p
  
  U1D
  XUlD 12 13 4 0 14 LMC6484A/NS
  R142 4 13 1k
  R143 13 0 2k
  R141 12 8 40.2k
  C75 12 0 39n
  
  Misc
  C76 16 15 1u
  R135 15 1 100k
  R98 4 14 10k
  RLOAD 14 0 10k
  *AIR-WATCH load
  
  NIVEAU-SENS Air Sensor
  v(14) state changes when RSENSOR transitions between 10k and 270k
  RSENSOR 16 0 10k
  *High=Saline Backfill, Low=no Saline Backfill
  
  .control
  tran 1us 11ms 10ms
  plot v(1)+48 v(7)+34 v(8)+20 v(12)+2 v(14)
  .endc
  
  .control
  dc RSENSOR 1k 300k 1k
  run
  print v(12) v(14) >> result.txt
  .endc
  
  .end
  
  Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/ngspice/support-requests/32/
  
  To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
  
  Related
  
  Support Requests: ~~#32~~
  
  alternate
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Todd Gilbert - 2018-09-02

Hi Holger,

I want to run a dc sweep with the RSENSOR resistor, and monitor the v(12) and v(14) levels. I believe I'm not using the dc sweep correctly, as I see no change in those nodes. I left the transient analysis in the deck for troubleshooting purposes. Try removing the transient analysis asterisks, and running the transient analysis - try a couple different RSENSOR values (1k and 300k). You will see the resulting behavior with those two resistor values.

Best Regards,
Todd

//CODE//
.include nat_semi.lib LMC6484A/NS

//////////////////////
Air Separator Simulation
//////////////////////

190332 Sensor Board
U1A
XU1A 3 2 4 0 1 LMC6484A/NS
VCC 4 0 12
C73 2 0 .001u
R130 1 2 261k
R131 4 3 22.1k
R132 3 0 22.1k
R133 1 3 47.5k

U1B
XU1B 5 6 4 0 7 LMC6484A/NS
R145 6 0 1k
R146 6 7 4.99k
R136 5 16 2.2M

U1C
XU1C 10 9 4 0 8 LMC6484A/NS
R137 4 10 1k
R138 10 0 866
R139 7 9 619k
C74 9 0 33p

U1D
XUlD 12 13 4 0 14 LMC6484A/NS
R142 4 13 1k
R143 13 0 2k
R141 12 8 40.2k
C75 12 0 39n

Misc
C76 16 15 1u
R135 15 1 100k
R98 4 14 10k
RLOAD 14 0 10k
*AIR-WATCH load

NIVEAU-SENS Air Sensor
v(14) state changes when RSENSOR transitions between 10k and 270k
RSENSOR 16 0 10k
*High=Saline Backfill, Low=no Saline Backfill

.control
tran 1us 11ms 10ms
plot v(1)+48 v(7)+34 v(8)+20 v(12)+2 v(14)
.endc

.control
dc RSENSOR 1k 300k 1k
run
print v(12) v(14) >> result.txt
.endc

.end

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Todd Gilbert - 2018-09-02

Oops - didn't notice the asterisks had already been removed on the transient analysis.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- marcel hendrix - 2018-09-03
  
  I marked your file as code (using a string of three tildes above and below the section).
  
  -marcel
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Holger Vogt - 2018-09-03
  
  Todd,
  
  within this thread I now see 3 versions of your code. This is completely unnecessary!
  
  When you have a look at your text describing the code, you may detect that the sorcefourge text editor has modified it! For example asterisks disappear in front of the text line, so their comment function for ngspice gets lost, the code becomes thus unusable.
  
  When you your are logged in, you have several buttons on top of each post, for example the 'edit' button (in form of a pencil). When pushing this button, a text window opens with your previous post, however potentially already modified.
  
  Then delete you original lines describing the ngspice input code. Then put in anew the complete code (from your ngspice text editor). Then mark this code text with your mouse. Then hit the </> botton above your text window to mark this text as code (i.e. the sourceforge text editor will not modifiy this part of the text, and write it out in a special typewriter font.
  
  If you fail to do so, any 'code' in a normal text window will probably become unreadable, useless for directly putting it to ngspice (e.g. for testing).
  
  Holger
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Todd Gilbert - 2018-09-03

Thanks. Below is the result.txt file - v(12) and v(14) do not change. I've also attached transient analysis results using RSENSOR = 10K and RSENSOR = 300k demonstrating the circuit's behavior. I believe there is something wrong with how I'm capturing the v(12) and v(14) values in the dc sweep analysis.

-Todd

*include nat_semi.lib DC transfer characteristic Mon Sep 3 05:35:17 2018

Index res-sweep v(12) v(14)

0 1.000000e+03 1.199869e+01 1.199831e+01
1 2.000000e+03 1.199869e+01 1.199831e+01
2 3.000000e+03 1.199869e+01 1.199831e+01
3 4.000000e+03 1.199869e+01 1.199831e+01
4 5.000000e+03 1.199869e+01 1.199831e+01
5 6.000000e+03 1.199869e+01 1.199831e+01
6 7.000000e+03 1.199869e+01 1.199831e+01
7 8.000000e+03 1.199869e+01 1.199831e+01
8 9.000000e+03 1.199869e+01 1.199831e+01
9 1.000000e+04 1.199869e+01 1.199831e+01
10 1.100000e+04 1.199869e+01 1.199831e+01
11 1.200000e+04 1.199869e+01 1.199831e+01
12 1.300000e+04 1.199869e+01 1.199831e+01
13 1.400000e+04 1.199869e+01 1.199831e+01
14 1.500000e+04 1.199869e+01 1.199831e+01
15 1.600000e+04 1.199869e+01 1.199831e+01
16 1.700000e+04 1.199869e+01 1.199831e+01
17 1.800000e+04 1.199869e+01 1.199831e+01
18 1.900000e+04 1.199869e+01 1.199831e+01
19 2.000000e+04 1.199869e+01 1.199831e+01
20 2.100000e+04 1.199869e+01 1.199831e+01
21 2.200000e+04 1.199869e+01 1.199831e+01
22 2.300000e+04 1.199869e+01 1.199831e+01
23 2.400000e+04 1.199869e+01 1.199831e+01
24 2.500000e+04 1.199869e+01 1.199831e+01
25 2.600000e+04 1.199869e+01 1.199831e+01
26 2.700000e+04 1.199869e+01 1.199831e+01
27 2.800000e+04 1.199869e+01 1.199831e+01
28 2.900000e+04 1.199869e+01 1.199831e+01
29 3.000000e+04 1.199869e+01 1.199831e+01
30 3.100000e+04 1.199869e+01 1.199831e+01
31 3.200000e+04 1.199869e+01 1.199831e+01
32 3.300000e+04 1.199869e+01 1.199831e+01
33 3.400000e+04 1.199869e+01 1.199831e+01
34 3.500000e+04 1.199869e+01 1.199831e+01
35 3.600000e+04 1.199869e+01 1.199831e+01
36 3.700000e+04 1.199869e+01 1.199831e+01
37 3.800000e+04 1.199869e+01 1.199831e+01
38 3.900000e+04 1.199869e+01 1.199831e+01
39 4.000000e+04 1.199869e+01 1.199831e+01
40 4.100000e+04 1.199869e+01 1.199831e+01
41 4.200000e+04 1.199869e+01 1.199831e+01
42 4.300000e+04 1.199869e+01 1.199831e+01
43 4.400000e+04 1.199869e+01 1.199831e+01
44 4.500000e+04 1.199869e+01 1.199831e+01
45 4.600000e+04 1.199869e+01 1.199831e+01
46 4.700000e+04 1.199869e+01 1.199831e+01
47 4.800000e+04 1.199869e+01 1.199831e+01
48 4.900000e+04 1.199869e+01 1.199831e+01
49 5.000000e+04 1.199869e+01 1.199831e+01
50 5.100000e+04 1.199869e+01 1.199831e+01
51 5.200000e+04 1.199869e+01 1.199831e+01
52 5.300000e+04 1.199869e+01 1.199831e+01
53 5.400000e+04 1.199869e+01 1.199831e+01
54 5.500000e+04 1.199869e+01 1.199831e+01
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Todd Gilbert - 2018-09-03

Here is the attachment referenced in the previous message.

transient analysis plot 10k and 300k.doc

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- marcel hendrix - 2018-09-04
  
  Your circuit is an oscillator and at least 3 of the 4 opamps
  are completely saturated all of the time. Also,
  the main capacitor C76 has a triangular wave on it There is a cyclic
  steady state, but .dc (opens capacitors, shortcircuits inductors) is
  not designed for that.
  What .DC will return when Rsensor is varied is the voltages in
  either of the two saturated states, ~0 or ~12V. Not what you want.
  
  -marcel
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Todd Gilbert - 2018-09-04

Hi Marcel,

Thank you for your careful analysis of the circuit. It is an unorthodox circuit. Not sure who designed it, but it was done several years ago. I'm attempting to simulate because of EOL component changes. I won't bore you with the details of the circuit, but there are a few items of importance: 1) NIVEAU-SENS is a sensor that is being modeled with a resistor (RSENSOR); 2) there is a performance zone where the node 8 pulse width is varying resulting in a relative node 12 voltage change; and 3) once this node rises above the node 13 8VDC threshold, node 14 switches to a constant 12VDC. Please review the .doc transient analysis plots in an earlier thread, which demonstrate the changes in output based on two RSENSOR values (10k and 300k).

Since I expect node 12 to vary in an unknown RSENSOR range, I'm attempting to perform a dc sweep. My dc sweep output shows no variance - v(12) and v(14) are 12V, regardless of the RSENSOR sweep value (see earlier thread). That lead me to wondering if I'm utilizing the dc sweep function correctly. What are your thoughts?

-Todd

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- marcel hendrix - 2018-09-04
  
  I am quite sure DC analysis is not possible here.
  
  You probably can get the information you need by
  replacing Rsensor with a Voltage-controlled resistor
  (this is a documented feature of NGSPICE's R and in the
  manual). You control the new Rsensor with a very slowly
  rising rampvoltage (see PULSE) and do a transient
  analysis of the modified circuit. A plot of V(12)'s
  and V(14)'s average value (if necessary add a low-pass
  filter) versus the ramp voltage, or directly versus
  Rsensor's controlled value, should tell you exactly
  what you want to know.
  
  Youi can also ALTERPARAM Rsensor in a looped .TRAN
  script and only look at a given list of Rsensor
  values.
  
  -marcel
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Holger Vogt - 2018-11-04

status: open --> closed

assigned_to: Holger Vogt
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Resistor Parameter Sweep

Group

Searches

Help

#32 Resistor Parameter Sweep

Related

Discussion

Related

Index res-sweep v(12) v(14)