Activity for ngspicehttps://sourceforge.net/p/ngspice/activity/Recent activity for ngspiceenWed, 13 Dec 2017 06:04:21 -0000marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=250#927b/d053Nice, compact, overview of the main concepts! This could go in the manual somewhere. -marcelmarcel hendrixWed, 13 Dec 2017 06:04:21 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=250#927b/d053Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 17:10:02 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 13:19:47 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 13:17:42 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 13:15:32 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 13:12:25 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bHolger Vogt posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bWould this work? Yes, with some more effort: Define a vector (let ...) long enough to hold all the spliced pieces. Re-format the vector using the 17.5.53 Reshape command into a 2D vector. All 'let' commands generate a vector in their actual plot. There is no need for 'save all' because this is the default. Each time a linearize, dc, run , tran etc. command is given, a new plot is generated, where the new simulation result is stored. Read about plots in the manual, chapt. 17.3. 'Display' shows all...Holger VogtTue, 12 Dec 2017 13:11:41 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#927bmarcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=250#a547/1d4b(not tested) You can setplot a new plot, then modify the LET commands to reference vectors in the previous plot(s) to compose blahblah. Save the new plot with set filetype=ascii; write xyz.raw. The LET commands need more typing now, but it may be possible to use a small script or the readline facilities to paste "prevplotname." and vector names together. -marcelmarcel hendrixTue, 12 Dec 2017 09:29:54 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=250#a547/1d4bjstor modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547Say we have a linearized vector called v(out) with 100 elements and you want to trim this vector to a specific time interval i.e. let x = linearize(v(out)) let trimmedvector[0] = x[20,40] %Does this save data from sub-index 20 "to" 40? let trimmedvector[1] = x[60,80] Would this work? Also, if have multiple vectors created using let command, and I would like to write them all using wrdata command, where should I include save all to save not only signal vectors generated at the end of analyses but...jstorTue, 12 Dec 2017 05:15:20 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547jstor modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547Say we have a linearized vector called v(out) with 100 elements and you want to trim this vector to a specific time interval i.e. let x = linearize(v(out)) let trimmedvector[0] = x[20:40] let trimmedvector[1] = x[60:80] Would this work? Also, if have multiple vectors created using let command, and I would like to write them all using wrdata command, where should I include save all to save not only signal vectors generated at the end of analyses but also the new vectors created using let command?...jstorTue, 12 Dec 2017 04:44:30 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547jstor modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547Say we have a linearized vector called v(out) with 100 elements and you want to trim this vector to a specific time interval i.e. let x = linearize(v(out)) let trimmedvector[0] = x[20:40] let trimmedvector[1] = x[60:80] Would this work? Also, if have multiple vectors created using let command, and I would like to write them all using wrdata command, where should I include save all to save not only signal vectors generated at the end of analyses but also the new vectors created using let command?...jstorTue, 12 Dec 2017 04:42:52 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547jstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547Say we have a linearized vector called v(out) with 100 elements and you want to trim this vector to a specific time interval i.e. let x = linearize(v(out)) let trimmedvector = x[60:80] Would this work? Also, if have multiple vectors created using let command, and I would like to write them all using wrdata command, where should I include save all to save not only signal vectors generated at the end of analyses but also the new vectors created using let command? i.e. save all dc ............ let blah...jstorTue, 12 Dec 2017 04:42:02 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/f5553e91/?limit=25#a547jstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=25#9c76Thanks for the pointer!jstorTue, 12 Dec 2017 04:07:06 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=25#9c76Miguel Gomes posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#3413Hi. I tried Holger’s suggestion and it worked (ngspice stopped hanging when simulating those circuits). Thank you all.Miguel GomesMon, 11 Dec 2017 20:00:36 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#3413marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=250#5167/bbedWith a console, use "shell grep <instance> thisfile" E.g. D:\spice\mhx\transverter50MEg>spicew trans50MEG.cir Processing 'd:\spice\share\ngspice\scripts\spinit' ****** ** ngspice-27 : Circuit level simulation program ** Compiled with KLU Direct Linear Solver ** (NDEV and URC models not supported yet) ** The U. C. Berkeley CAD Group ** Copyright 1985-1994, Regents of the University of California. ** Please get your ngspice manual from http://ngspice.sourceforge.net/docs.html ** Please file your bug-reports...marcel hendrixSun, 10 Dec 2017 18:07:09 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=250#5167/bbedjstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=25#5167Hello, Is there a way to get instance node names from ngspice i.e. let's say one would like to know the nodes for subcircuit instance called "nmos6" - what command or set of commands can be used to return output as "net@16 net@7 net@12 net@0" i.e. just the names of the nodes of this instance? Example of instance card: xnmos6 net@16 net@7 net@12 net@0 .........jstorSun, 10 Dec 2017 17:12:40 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/5bfff21f/?limit=25#5167Clyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5/0f30As I expected, it prints pi just fine, but it does not use the symbol in const.Clyde R. ShappeeSun, 10 Dec 2017 13:32:37 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5/0f30Clyde R. Shappee modified a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5Yes, thank you. This was what I was after. Although I did not try your edited netlist, print pi produced the desired result. Perhaps it will print but not work in the computations. Thanks again.Clyde R. ShappeeSun, 10 Dec 2017 13:28:05 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5Clyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5Yes, thank you. This was what I was after. Although I did try your edited netlist, print pi produced the desired result. Perhaps it will print but not work in the computations. Thanks again.Clyde R. ShappeeSun, 10 Dec 2017 13:21:52 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227/6c73/01c5marcel hendrix modified a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#f81b/efd2Hi Robert, Sorry that I called it a bug, "strangeness", then? Proposed standard parameters. HSPICE has PI (can't find any others). LTspice has this (different names): pi = 3.141593 boltz = 1.380650e-23 echarge = 1.602177e-19 planck = 6.626070e-34 i = 0,1 e = 2.718282 NGSPICE itself has more. The name "c" may cause errors because it is so short and common. kelvin = -2.73150e+02 c = 2.997925e+08 When these four are added .if() etc. become easier (probably .if is already aware of these parameters)....marcel hendrixSun, 10 Dec 2017 12:40:41 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#f81b/efd2marcel hendrix posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#f81b/efd2Hi Robert, Sorry that I called it a bug, "strangeness", then? Proposed standard parameters. HSPICE has PI (can't find any others). LTspice has this (different names): pi = 3.141593 boltz = 1.380650e-23 echarge = 1.602177e-19 planck = 6.626070e-34 i = 0,1 e = 2.718282 NGSPICE itself has more. The name "c" may cause errors because it is so short and common. kelvin = -2.73150e+02 c = 2.997925e+08 When these four are added .if() etc. become easier (probably .if is already aware of these parameters)....marcel hendrixSun, 10 Dec 2017 12:36:22 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#f81b/efd2Robert Larice posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#f81bHello Marcel (There is a very old bug in ngspice in that it doesn't know pi although that is in the constant plot). The namespace of the "constant" plot has nothing to do with the .param namespace. So this is not really a bug, though one can argue that some .params should be predefined. Perhaps you find a list of predefined .params of some other well known spice which we could reuse. Regards,Robert LariceSun, 10 Dec 2017 10:42:37 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#f81bmarcel hendrix modified a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#af6bThe transfer function one. Here is my netlist. high pass transfer function test v1 1 0 ac 1 *.param rad2by2 = {sqrt(2.0) / 2.0 } *.param freq = '2.0*pi*20.0' .model filter s_xfer(gain=1 + num_coeff=[1.0 0.0 0.0] + den_coeff=[1.0 0.707 1] + int_ic=[0 0] denormalized_freq = 125.66) a1 1 2 filter .control destroy all ac dec 1000 1 1000 let resp = db (v(2)/ v(1)) settype decibel resp plot resp .endc .end Yes, I do know what poles and zeros are, this was a simplied example.marcel hendrixSun, 10 Dec 2017 00:27:58 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#af6bmarcel hendrix posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#3227/6c73When I uploaded the netlist, somehow the * before the .param got taken out. Ok, I edited that for you. Put exactly three '~' characters above and below a netlist. Maybe you wanted something like the below? (There is a very old bug in ngspice in that it doesn't know pi although that is in the constant plot). .TITLE high pass transfer function test .param pi = 3.1415926535897932384626433832795028841971693993751058 .param Q = sqrt(2)/2 .param freq = 20 .param omega = 2*pi*freq .param omega1 = 1 v1 in...marcel hendrixSun, 10 Dec 2017 00:27:05 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#3227/6c73Clyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227No worries, Marcel. I found my way with gnu octave 4.2.1 with the math. I was looking for a way to confirm that I had input the model correctly. Where I am going with this is a circuit design where the desired transfer function is well known. With the s_xfer model, I can create a precise vector of the desired response and then, analyze my design and compare frequency respsonse vectors to see how it is doing. If there is interest, I will post the simulation when I finish. Where I could use some help...Clyde R. ShappeeSat, 09 Dec 2017 19:12:52 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#3227Holger Vogt posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#f0ffIn standard ngspice we still use a very old X11 (and Windows) plot interface. The pointplot plotting characters are directly linked to the font selection and size used for labeling the axes. In git branch utf-8+2 I have started coding some improvements. You may select a different font (together for both points and axes) by something like set xfont="-adobe-helvetica-bold-r-normal--17-120-100-100-p-92-iso8859-1" Fonts like the one selected above are listed in /usr/share/fonts, e.g. in /usr/share/fonts/100dpi/fonts.dir....Holger VogtSat, 09 Dec 2017 18:06:59 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#f0ffHolger Vogt committed [d450e0]https://sourceforge.net/p/ngspice/ngspice/ci/d450e0566818e77f6c19e30d8c10c91c1dd55fe0/variable gridwidth, not gridlinewidthHolger VogtSat, 09 Dec 2017 11:52:32 -0000https://sourceforge.net/p/ngspice/ngspice/ci/d450e0566818e77f6c19e30d8c10c91c1dd55fe0/Holger Vogt committed [45a14b]https://sourceforge.net/p/ngspice/ngspice-manuals/ci/45a14b41a66be619f2fcd961498b4d356228b8d1/Chapt. 25.1 Original XSPICE manual is no longer availableHolger VogtSat, 09 Dec 2017 11:46:06 -0000https://sourceforge.net/p/ngspice/ngspice-manuals/ci/45a14b41a66be619f2fcd961498b4d356228b8d1/Holger Vogt committed [df537c]https://sourceforge.net/p/ngspice/ngspice/ci/df537ca281004a725362f5ddd77d5e3387a8c361/Variables xbrushwidth and gridlinewidth (to be set e.g. inHolger VogtSat, 09 Dec 2017 11:08:37 -0000https://sourceforge.net/p/ngspice/ngspice/ci/df537ca281004a725362f5ddd77d5e3387a8c361/marcel hendrix posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#af6b/0c6b[My previous comments reported what I found in the NGSPICE sources, I was not trying to be pedantic.] Unfortunately, and maybe unexpectedly, the PZ analysis algorithm does not handle the s_xfer model.marcel hendrixFri, 08 Dec 2017 22:54:43 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#af6b/0c6bClyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#af6bThe transfer function one. Here is my netlist. ~~~~~ high pass transfer function test v1 1 0 ac 1 .param rad2by2 = {sqrt(2.0) / 2.0 } .param freq = '2.0pi20.0' .model filter s_xfer(gain=1 + num_coeff=[1.0 0.0 0.0] + den_coeff=[1.0 0.707 1] + int_ic=[0 0] denormalized_freq = 125.66) a1 1 2 filter .control destroy all ac dec 1000 1 1000 let resp = db (v(2)/ v(1)) settype decibel resp plot resp .endc .end ~~~~~~ Yes, I do know what poles and zeros are, this was a simplied example.Clyde R. ShappeeFri, 08 Dec 2017 16:21:43 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#af6bmarcel hendrix posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#529a/7972It is certainly not possible with every XSPICE model, as sometimes there is simply no MIF_AC section (see, e.g. the multi_input_pwl definition). Sometimes (oneshot) it is stated that AC behavior is not modeled. Digital logic models normally only look at DC and TRAN states. Which mixed signal circuits are you interested in? -marcelmarcel hendrixFri, 08 Dec 2017 16:01:17 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=250#529a/7972Clyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#529aIs it possible to do a pole zero analysis on a circuit with a mixed signal behavioral model? I tried this and get an error saying my input is shorted. I tried a series R and C between the input generator, but it really upset the simulation. ClydeClyde R. ShappeeFri, 08 Dec 2017 15:28:19 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/16fc503e/?limit=25#529aClyde R. Shappee posted a comment on discussion Helphttps://sourceforge.net/p/ngspice/discussion/120973/thread/987accdd/?limit=25#e728Is it possible to do a pole zero analysis on a circuit with a mixed signal behavioral model? I tried this and get an error saying my input is shorted. I tried a series R and C between the input generator, but it really upset the simulation. ClydeClyde R. ShappeeFri, 08 Dec 2017 15:28:17 -0000https://sourceforge.net/p/ngspice/discussion/120973/thread/987accdd/?limit=25#e728Amal Banerjee posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44/1de7/c13f/c262/3a43Many, many thanks. Guanella and Ruthoff are transmission line transformers, which are truly broadband devices, and so are used in all RF/microwave signal processing equipment for impedance matching. Initailly fabricated from coaxial cables, these are fabricated easily and cheaply using the same methods used for making printed circuit boards.Amal BanerjeeThu, 07 Dec 2017 04:53:26 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44/1de7/c13f/c262/3a43marcel hendrix modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=250#ec90/4639Do you have gnuplot installed? It should look better than that. -marcel Test file .TITLE Test the AGAUSS function .csparam isnom=1 .csparam isavar=0.6 .csparam issigma=3 $ mean = 10, sigma = 2 => range <4,16> Itest 0 resistance 1A R1 resistance 0 10 .control let mc_runs = 16 let run = 0 set curplot=new $ create a new plot set scratch=$curplot $ store its name to 'scratch' setplot $scratch $ make 'scratch' the active plot let data=unitvec(mc_runs) $ create a vector in plot 'scratch' to store the data...marcel hendrixWed, 06 Dec 2017 18:43:59 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=250#ec90/4639marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=250#ec90/4639Do you have gnuplot installed? It should look better than that. -marcel Test file .TITLE Test the AGAUSS function .csparam isnom=1 .csparam isavar=0.6 .csparam issigma=3 $ mean = 10, sigma = 2 => range <4,16> Itest 0 resistance 1A R1 resistance 0 10 .control let mc_runs = 16 let run = 0 set curplot=new $ create a new plot set scratch=$curplot $ store its name to 'scratch' setplot $scratch $ make 'scratch' the active plot let data=unitvec(mc_runs) $ create a vector in plot 'scratch' to store the data...marcel hendrixWed, 06 Dec 2017 18:41:33 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=250#ec90/4639Mateusz modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#ec90Hello, I have problem with size of plotted points using pointplot. Plot example in attachments. These dots are very small. In manual i can't find instruction to change it. There is any possibility to do this?MateuszWed, 06 Dec 2017 12:27:24 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#ec90Mateusz posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#ec90Hello, I have problem with size of plotted points using pointplot. Here is example: These dots are very small. In manual i can't find instruction to change it. There is any possibility to do this?MateuszWed, 06 Dec 2017 12:26:21 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/0e2ddea0/?limit=25#ec90Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Unfortunately there is no list. There has been a recent discussion about this issue: https://sourceforge.net/p/ngspice/mailman/message/36060875/. Some ideas are there, but no solution is available on short notice. In shared ngspice the actual procedure is simply to catch all output strings with ng_getchar by prepending stderr to all message strings sent by C functions like fprintf(stderr, ...) or prepending stdout for fprintf(stdout, ...). So it might be o.k. to check for "stderr" only. HolgerHolger VogtTue, 05 Dec 2017 22:12:12 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Unfortunately there is no list. There has been a recent discussion about this issue: https://sourceforge.net/p/ngspice/mailman/message/36060875/. Some ideas are there, but no solution is available on short notice. In shared ngspice the actaul procedure is simply to catch all output strings with ng_getchar by prepending stderr to all message strings sent by C functions like fprintf(stderr, ...) or prepending stdout for fprintf(stdout, ...). So it might be o.k. to check for "stderr" only. HolgerHolger VogtTue, 05 Dec 2017 22:11:39 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Holger Vogt posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Unfortunately there is not list. There has been a recent discussion about this issue: https://sourceforge.net/p/ngspice/mailman/message/36060875/. Some ideas are there, but no solution is available on short notice. In shared ngspice the actaul procedure is simply to catch all output strings with ng_getchar by prepending stderr to all message strings sent by C functions like fprintf(stderr, ...) or prepending stdout for fprintf(stdout, ...). So it might be o.k. to check for "stderr" only. HolgerHolger VogtTue, 05 Dec 2017 22:11:12 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889/a6c1Miguel Gomes posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889Hello. When using ngspice as a shared library, I’m parsing ngspice output through ng_getchar function in order to detect if there were any errors in the simulation (for the time being I’m using only transient simulations). It seems to me that this is what is suggested in some ngspice example files: in some of these files, function ng_getchar reads like this: int ng_getchar(char* outputreturn, int ident, void* userdata) { printf("%s\n", outputreturn); /* setting a flag if an error message occurred...Miguel GomesTue, 05 Dec 2017 21:02:20 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/86eec9da/?limit=25#e889marcel hendrix modified a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=250#b667/79ee/122b/9f06Dear Miguel, Holger is 100% right that you have found a nasty ng-spice bug that should be killed forthwith. However, you may try your circuit with the BSIM4 and a (then) more appropriate Vdd = 1.8V and drive level of 1.2V. (Search the manual for examples on BSIM4, e.g. 17.9 p337.) I found it instructive to experiment with levels 1 - 54 and try different supply voltages -- your circuit clearly shows how important it is to have a good model with good parameters. -marcel PS: Shortened the simulation...marcel hendrixTue, 05 Dec 2017 20:27:47 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=250#b667/79ee/122b/9f06marcel hendrix posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=250#b667/79ee/122b/9f06Dear Miguel, Holger is 100% right that you have found a nasty ng-spice bug that should be killed forthwith. However, you may try your circuit with the BSIM4 and a (then) more appropriate Vdd = 1.8V and drive level of 1.2V. (Search the manual for examples on BSIM4, e.g. 17.9 p337.) I found it instructive to experiment with levels 1 - 54 and a try different supply voltages -- your circuit clearly shows how important it is to have a good model with good parameters. -marcel PS: Shortened the simulation...marcel hendrixTue, 05 Dec 2017 20:23:35 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=250#b667/79ee/122b/9f06Holger Vogt posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79ee/122b/829cThat's right. HolgerHolger VogtTue, 05 Dec 2017 20:17:43 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79ee/122b/829cMiguel Gomes posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79ee/122bHi. Thank you all for your feedback on this topic. Holger, I’ll try recompiling ngspice (still in v26) with your suggestion: updating cktop.c so that line 186 reads factor = max(sqrt(factor), 1.00005); /* has to be larger than 1 */ The thing is that in my machine, line 186 of cktop.c is empty (and I’m pretty sure I never edit this file before) but it may just be a matter of different file’s version. In my version of cktop.c, some lines around line 190 read like this: (...) 182 for (n = ckt->CKTnodes;...Miguel GomesTue, 05 Dec 2017 17:18:42 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79ee/122bHolger Vogt posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79eecktop.c:186 should read factor = max(sqrt(factor), 1.00005); /* has to be larger than 1 */ HolgerHolger VogtTue, 05 Dec 2017 10:09:46 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667/79eeFrancesco Lannutti posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#8affHi Holger, it should exit from dynamic_gmin at a certain point, if I remember correctly. The problem is that it may take a very long time, depending on the circuit. I will wait for your findings. FraFrancesco LannuttiTue, 05 Dec 2017 09:53:23 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#8affHolger Vogt modified ticket #351https://sourceforge.net/p/ngspice/bugs/351/ngspice hangs in transient analysisHolger VogtTue, 05 Dec 2017 08:44:34 -0000https://sourceforge.net/p/ngspice/bugs/351/Holger Vogt posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667The bug is that ngspice enters a loop and does not leave it anymore. This occurs in function dynamic_gmin() in cktop.c. This should not happen, despite of what customer input we have. If there is no convergence, ngspice should simply stop with an error message. I will look into it. HolgerHolger VogtTue, 05 Dec 2017 08:44:34 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=25#b667marcel hendrix modified a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655/fd2fConverges (still difficult) when putting a 10 Ohms resistance in series with the PWL sources: Va a1 0 DC 0 PWL (0 0 .999999 0 1 5 2.999999 5 3 0) Ra1 a1 1 10 Vb a2 0 DC 0 PWL (0 0 1.999999 0 2 5) Ra2 a2 2 10 Vtgt a3 0 DC 0 PWL (0 0 .999999 0 1 5 1.999999 5 2 0 2.999999 0 3 5) Ra3 a3 200 10 Made minor adjustment here ".tran 1ms 4 0 1ms" and here ".option method=gear reltol=1m minbreak=200ps", but it also worked without these changes. The problem seems to be bad convergence on node 4 (or 5). Output:...marcel hendrixTue, 05 Dec 2017 06:31:21 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655/fd2fmarcel hendrix posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655/fd2fConverges (still difficult) when putting a 10 Ohms resistance in series with the PWL sources: Va a1 0 DC 0 PWL (0 0 .999999 0 1 5 2.999999 5 3 0) Ra1 a1 1 10 Vb a2 0 DC 0 PWL (0 0 1.999999 0 2 5) Ra2 a2 2 10 Vtgt a3 0 DC 0 PWL (0 0 .999999 0 1 5 1.999999 5 2 0 2.999999 0 3 5) Ra3 a3 200 10 Made minor adjustment here ".tran 1ms 4 0 1ms" and here ".option method=gear reltol=1m minbreak=200ps", but it also worked without these changes. Output: D:\spice\mhx\bug_problematic>spice64w problematic_circuit_after_unsuccessful_bg_halt_0_MOD.cir...marcel hendrixTue, 05 Dec 2017 06:27:52 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655/fd2fmarcel hendrix posted a comment on ticket #351https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655Executing a transient analysis over any of the attached circuits leads to ngspice hanging: in interactive mode it becomes unresponsive to ctrl-c; With ngspice-27, in console mode, three Ctrl-Cs break the .OP pass (this is a new feature). I only tried problematic_circuit_after_unsuccessful_bg_halt_0.cir . ... Trying gmin = 9.7119E-12 Note: One successful gmin step Trying gmin = 9.7119E-12 Note: One successful gmin step Trying gmin = 9.7119E-12 Interrupted again (ouch) Killing, since 3 interrupts have...marcel hendrixTue, 05 Dec 2017 06:07:55 -0000https://sourceforge.net/p/ngspice/bugs/351/?limit=250#7655Miguel Gomes created ticket #351https://sourceforge.net/p/ngspice/bugs/351/ngspice hangs in transient analysisMiguel GomesMon, 04 Dec 2017 22:40:57 -0000https://sourceforge.net/p/ngspice/bugs/351/Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743The following is an automatic cutting, scaling and re-plotting of a vector. Please check if you can create an eye diagram. If you provide an example simulation file, we may generate a new ngspice tip. Holger * test for automatically splitting vectors .param tstep = 1m .param tstop = 1000m * time intervall to be cut out .csparam period = 600m * some data into the .control loop, available in plot 'const' .csparam tint = 'tstep/tstop' .csparam tstop = 'tstop' * create a transient simulated vector v(1)...Holger VogtMon, 04 Dec 2017 21:59:29 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743The following is an automatic cutting, scaling and re-plotting of a vector. Please check if you can create an eye diagram. If you provide an example simulation file, we may generate a new ngspice tip. Holger * test for automatically splitting vectors .param tstep = 1m .param tstop = 1000m * time intervall to be cut out .csparam period = 600m * some data into the .control loop, available in plot 'const' .csparam tint = 'tstep/tstop' .csparam tstop = 'tstop' * create a transient simulated vector v(1)...Holger VogtMon, 04 Dec 2017 16:39:56 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7/c13f/c262I know nothing of Guanella. With your hint about R0 and R1 the circuit indeed doesn't run. Inspection of the circuit shows that because of the inherent XFMR behavior of transmission lines node 6 has no DC path to ground. I grounded node 6 with 1G and then the simulation runs. Maybe R0 should be 160 and R1 40 Ohms, you could have been more explicit about that. .tran 10ps 0.2us 0 10ps What is the extra 0 after 0.2us That is the TSTART parameter: before that time no data is written to the output. Without...marcel hendrixSun, 03 Dec 2017 08:12:31 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7/c13f/c262Amal Banerjee posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44/1de7/c13fI am afraid there is a small mustake in the sample code that you have provided. You have r0 1 2 10 r1 4 5 40 This is valid for a Guanella 1:4 transmission line transformer, where the source/load impedance ratio is 1:4. For a Guamella 4:1 transmission line transformer, the source/load impedance ratio is 4:1, i.e., the impedance matching is to be performed between a source impedance that is 4x the load impedance. I am using Ngspice 26 myself. Also, I am not sure about the syntax: .tran 10ps 0.2us 0...Amal BanerjeeSun, 03 Dec 2017 04:25:44 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44/1de7/c13fHolger Vogt modified a comment on discussion ngspice tips and exampleshttps://sourceforge.net/p/ngspice/discussion/ngspice-tips/thread/6a5b9dd2/?limit=25#7029Isotel has devised a very nice example on how to combine analog and digital simulation using ngspice, Yosys, and GTKWave and for comparison Icarus Verilog. All tools used are open-source. The demo circuit is a digital sine generator. The digital circuit is described in Verilog, synthesized and translated to XSPICE by Yosys and is finally combined with an analog circuit in an ngspice netlist for simulation. Please check the Yosys and Isotel web pages and the ngspice manual for further information...Holger VogtSat, 02 Dec 2017 22:50:23 -0000https://sourceforge.net/p/ngspice/discussion/ngspice-tips/thread/6a5b9dd2/?limit=25#7029Holger Vogt posted a comment on discussion ngspice tips and exampleshttps://sourceforge.net/p/ngspice/discussion/ngspice-tips/thread/6a5b9dd2/?limit=25#7029Isotel (http://www.isotel.eu/) has devised a very nice example on how to combine analog and digital simulation using ngspice, Yosys, and gtkwave and for comparison Icarus Verilog (http://iverilog.icarus.com/). All tools used are open-source. The demo circuit is a digital sine generator (http://www.isotel.eu/mixedsim/intro/prssine.html). The digital circuit is described in Verilog, synthesized and translated to XSPICE by Yosys (http://www.clifford.at/yosys/) and finally combined with an analog circuit...Holger VogtSat, 02 Dec 2017 22:14:39 -0000https://sourceforge.net/p/ngspice/discussion/ngspice-tips/thread/6a5b9dd2/?limit=25#7029Holger Vogt posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743The following is an automatic cutting, scaling and re-plotting of a vector. Please check if you can create an eye diagram. If you provide an example simulation file, we may generate a new ngspice tip. Holger * test for automatically splitting vectors .param tstep = 1m .param tstop = 1000m * time intervall to be cut out .csparam period = 159m * some data into the .control loop, available in plot 'const' .csparam tint = 'tstep/tstop' .csparam tstop = 'tstop' * create a transient simulated vector v(1)...Holger VogtSat, 02 Dec 2017 20:49:31 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3d/0743Holger Vogt committed [4d988c]https://sourceforge.net/p/ngspice/ngspice/ci/4d988c95ae0ef8e3b13f9704d1ba857e06d4501e/clarify the licenses, add a table of files licensedHolger VogtSat, 02 Dec 2017 16:18:06 -0000https://sourceforge.net/p/ngspice/ngspice/ci/4d988c95ae0ef8e3b13f9704d1ba857e06d4501e/marcel hendrix modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7This does not happen for ngspice-26 and ngspice-27 when using the code below. Upgrade? -marcel D:\spice\mhx\guanella>spice64w guanella2.cir Using d:/spice64/bin/spinit ****** ** ngspice-27 : Circuit level simulation program ** The U. C. Berkeley CAD Group ** Copyright 1985-1994, Regents of the University of California. ** Please get your ngspice manual from http://ngspice.sourceforge.net/docs.html ** Please file your bug-reports at http://ngspice.sourceforge.net/bugrep.html ** Creation Date: Dec...marcel hendrixSat, 02 Dec 2017 15:57:38 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7It makes no difference for ngspice-26 and ngspice-27. Upgrade? -marcel D:\spice\mhx\guanella>spice64w guanella2.cir Using d:/spice64/bin/spinit ****** ** ngspice-27 : Circuit level simulation program ** The U. C. Berkeley CAD Group ** Copyright 1985-1994, Regents of the University of California. ** Please get your ngspice manual from http://ngspice.sourceforge.net/docs.html ** Please file your bug-reports at http://ngspice.sourceforge.net/bugrep.html ** Creation Date: Dec 2 2017 16:41:31 ****** Warning:...marcel hendrixSat, 02 Dec 2017 15:45:13 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#3b44/1de7Holger Vogt committed [56c6e4]https://sourceforge.net/p/ngspice/ngspice-manuals/ci/56c6e41f4cc061f0b0231bf394a67e753595c38b/Update on Chapter 33 Copyrights and licensesHolger VogtSat, 02 Dec 2017 15:05:23 -0000https://sourceforge.net/p/ngspice/ngspice-manuals/ci/56c6e41f4cc061f0b0231bf394a67e753595c38b/Holger Vogt committed [80f804]https://sourceforge.net/p/ngspice/ngspice-manuals/ci/80f8041fb061044734ca31aa70603eb35fa6e5a8/Declare as deprecated using &( ... ) instead of .param ...Holger VogtSat, 02 Dec 2017 13:56:07 -0000https://sourceforge.net/p/ngspice/ngspice-manuals/ci/80f8041fb061044734ca31aa70603eb35fa6e5a8/Holger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dI can only imagine the complexity of loops to get the job done. Yes, but this is trivial programming The last I recall, there's nothing that can return index number of a time vector for a specific unique time point. Don't give up so quickly! Our .control language is not too comfortable, but you can do a lot of things with it. Below you will find a simple tran simulation and a way to cut out some pieces from the plot and re-plot it onto a new x scale. What is still missing is a loop to do the cutting...Holger VogtSat, 02 Dec 2017 12:53:19 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dI can only imagine the complexity of loops to get the job done. Yes, but this is trivial programming The last I recall, there's nothing that can return index number of a time vector for a specific unique time point. Don't give up so quickly! Our .control language is not too comfortable, but you can do a lot of things with it. Below you will find a simple tran simulation and a way to cut out some pieces from the plot and re-plot it onto a new x scale. What is still missing is a loop to do the cutting...Holger VogtSat, 02 Dec 2017 12:52:20 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dHolger Vogt modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dI can only imagine the complexity of loops to get the job done. Yes, but this is trivial programming The last I recall, there's nothing that can return index number of a time vector for a specific unique time point. Don't give up so quickly! Our .control language is not too comfortable, but you can do a lot of things with it. Below you will find a simple tran simulation and a way to cut out some oices from the plot and reoplot it onto a new x scale. What is still missing is a loop to do the cutting...Holger VogtSat, 02 Dec 2017 12:50:39 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dHolger Vogt posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dI can only imagine the complexity of loops to get the job done. Yes, but this is trivial programming The last I recall, there's nothing that can return index number of a time vector for a specific unique time point. Don't give up so quickly! Our .control language is not too comfortable, but you can do a lot of things with it. Below you will find a simple tran simulation and a way to cut out some oices from the plot and reoplot it onto a new x scale. What is still missing is a loop to do the cutting...Holger VogtSat, 02 Dec 2017 12:49:36 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88c/ec3dAmal Banerjee posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44A million thanks for your initial suggestion yesterday. However, if insteaf of a 1:4 Guanella TLT, I have a 4:1 TLT, like: .SUBCKT GUANELLA4_1 1 2 3 4 1 IN 1 2 IN 2 3 OUT 1 4 OUT 2 T0 1 5 3 4 Z0={Z0} T1 5 2 3 4 Z0={Z0} .ENDS Here, Z0 = 20 ohm Then I keep getting a "timestep too small" runtime error, as: doAnalyses: TRAN: Timestep too small; initial timepoint: cause unrecorded. I have adjusted the time step and tried several times, but the problem never goes away. Any suggestions ?Amal BanerjeeSat, 02 Dec 2017 11:15:03 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#3b44marcel hendrix modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#3080/6e02This syntax is not valid: .tran 1us 100us 101us 1us TSTART is invalid, must be less than TSTOP. I don't see what you are getting at? Ask Francesco Lannutti about NGSPICE's PSS branch. It needs a bit of work, but ISTR it has the hooks to experiment with / implement vector extrapolation. Speaking of acceleration, do we have APS like feature in Ngspice to accelerate simulations (besides OpenML models and/or Cuspice)? APS == 'Spectre Accelerated Parallel Simulator' ? It might be possible to accelerate...marcel hendrixSat, 02 Dec 2017 09:13:20 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#3080/6e02marcel hendrix modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#a88c/a607Did you look at 17.5.78, 'Stop: Set a breakpoint' and 17.5.55 'Resume: Continue a simulation after a stop' ? -marcelmarcel hendrixSat, 02 Dec 2017 09:10:17 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#a88c/a607marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#a88c/a607Did you look at 17.5.78, 'Stop: Set a breakpoint' and 17.5.55 Resume: 'Continue a simulation after a stop' ? -marcelmarcel hendrixFri, 01 Dec 2017 22:35:30 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#a88c/a607marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#3080/6e02This syntax is not valid: .tran 1us 100us 101us 1us TSTART is invalid, must be less than TSTOP. I don't see what you are getting at? Ask Francesco Lannutti about NGSPICE's PSS branch. It needs a bit of work, but ISTR it has the hooks to experiment with / implement vector extrapolation. Speaking of acceleration, do we have APS like feature in Ngspice to accelerate simulations (besides OpenML models and/or Cuspice)? APS == 'Spectre Accelerated Parallel Simulator' ? It might be possible to accelerate...marcel hendrixFri, 01 Dec 2017 22:29:43 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#3080/6e02Robert Larice committed [efcf84]https://sourceforge.net/p/ngspice/ngspice/ci/efcf8447adbb42a66abe74902050e7da9e141cca/test-subckt-m.cir, expose a .subckt m= issueRobert LariceFri, 01 Dec 2017 21:03:39 -0000https://sourceforge.net/p/ngspice/ngspice/ci/efcf8447adbb42a66abe74902050e7da9e141cca/Robert Larice committed [c646d7]https://sourceforge.net/p/ngspice/ngspice/ci/c646d7790f189b3cd8c51e156c22c9ad47c65f4b/fix the m= issueRobert LariceFri, 01 Dec 2017 21:03:39 -0000https://sourceforge.net/p/ngspice/ngspice/ci/c646d7790f189b3cd8c51e156c22c9ad47c65f4b/marcel hendrix modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#f910.TITLE guanella .subckt guanella1_4 1 2 3 4 * 1 in 1 * 2 in 2 * 3 out 1 * 4 out 2 L0 3 4 5.0nH T0 1 2 3 5 z0=20 T1 1 2 5 4 z0=20 .ends guanella1_4 r0 1 2 10 r1 4 5 40 vd 1 0 dc=5mV sin(0,10,500MEG) vtst0 2 3 0 vtst1 5 6 0 xb 3 0 4 6 guanella1_4 .options method=gear reltol=1m minbreak=5ps .tran 0 0.2us 0 10ps uic * Input/source voltage/current .PRINT TRAN V(3) I(VTST0) * Output/load voltage/current *.PRINT TRAN V(4,6) I(VTST1) .meas tran p1 AVG par('v(4,6)*i(vtst1)') .meas tran p2 AVG par('v(3)*i(vtst0)')...marcel hendrixFri, 01 Dec 2017 20:41:41 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#f910Robert Larice committed [89ca99]https://sourceforge.net/p/ngspice/ngspice/ci/89ca99e0c86fba488eadc256fc52ef02738439a4/numparam, drop "old style" '&' syntaxRobert LariceFri, 01 Dec 2017 17:18:00 -0000https://sourceforge.net/p/ngspice/ngspice/ci/89ca99e0c86fba488eadc256fc52ef02738439a4/jstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88cThe idea here is to splice this vector into smaller set of vectors so that one can overlay these individual traces to plot an eye-diagram. I thought (did this long time ago when I was working on a steady state solver using control block) to find vector index coresponding to let's say 5us and 10us and then fetch all index points between timepoints 5us and 10us. I can only imagine the complexity of loops to get the job done. The last I recall, there's nothing that can return index number of a time...jstorFri, 01 Dec 2017 16:53:01 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#a88cmarcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#f910.TRAN 100.0us 30.0s 50.0us UIC This is never going to work with a 500MEG sine input and the desire to have precise power. (Also, the output samples will not be equidistant.) And 30s in total is a bit optimistic. Suggested: .TRAN 10ps 30us 10ps UIC -marcelmarcel hendrixFri, 01 Dec 2017 10:01:26 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=250#f910marcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#64d2I'll have more to say later on, unless somebody else does, but is it acceptable to use (time) interpolation? Then subindexing the vector becomes trivial. A simple loop over the vector with a 'time > marker then remember index' test will also do. -marcelmarcel hendrixFri, 01 Dec 2017 07:35:54 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=250#64d2jstor modified a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#70d6Okay, I shouldn't have accidently hit the return key while typing the thread title. Had a question about slicing a signal vector from transient simulations. Say you have a vector called v(x) from a transient run of 50 us. Is it possible to slice this signal into 10 vectors, each segmenting 5 us worth of signal? I though of using meas tran slicedsignal FROM=5us TO=10us for let's say one of the segments (and then wrapping it in a loop) but I suspect this will work. Thanks.jstorFri, 01 Dec 2017 06:03:46 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/8ff1ff2f/?limit=25#70d6jstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=25#3080Hello Marcel, I was concerned about the speed. Thanks for the vector optmiziation pointers - I was reading some material on it (Univeristy of Victoria, CA). Now here's the thing -> .tran 1us 100us 101us 1us vs .tran 1us 100us 1ns 1us. The fact that tmax is greater than tstop => time points would be even finer than time steps scaled due to tmax of 1ns, right? Let me know if that makes sense. I wonder, now that I'm without the Cadence armor of PSS+PNoise (driven and autonomous circuits), transient...jstorFri, 01 Dec 2017 05:56:39 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=25#3080Amal Banerjee posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#d8deCould some Ngspice guru please help ? I have a simple Guanella 1:4 source/load impedance transmission line transformer. The input/output RMS power are orders of magnitude different. To obtain the RMS input/ output power is obtained by processing the raw transient anaysis output current/voltage at both the input/output ends with a C language program. The netlist is below: .SUBCKT GUANELLA1_4 1 2 3 4 1 IN 1 2 IN 2 3 OUT 1 4 OUT 2 L0 3 4 5.0nH T0 1 2 3 5 Z0=20 T1 1 2 5 4 Z0=20 .ENDS R0 1 2 10 R1 4 5...Amal BanerjeeFri, 01 Dec 2017 04:26:07 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/04e62075/?limit=25#d8demarcel hendrix posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#e99e/cdbaSay a PLL is destined to settle at 100 us. Do I need to run transient simulation for first 100 us or I have the option to jump to 48 us using tstart option and then store simulation results thereon? It is unclear what you are asking. Are you concerned about speed or about size? As you apparently have no problem with jumping 48us ahead, you must be only interested in the steady-state. With .tran 1us 101us 100us 1u the steady-state is found slowly, but with very low memory footprint. Accelerating the...marcel hendrixTue, 28 Nov 2017 19:00:56 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=250#e99e/cdbajstor posted a comment on discussion ngspice-usershttps://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=25#e99eHello, Say a PLL is destined to settle at 100 us. Do I need to run transient simulation for first 100 us or I have the option to jump to 48 us using tstart option and then store simulation results thereon? Any "tips" for settling time simulations? Running transients on PLLs can be time expensive. Thanks.jstorTue, 28 Nov 2017 17:10:20 -0000https://sourceforge.net/p/ngspice/discussion/133842/thread/fe199254/?limit=25#e99eRobert Larice committed [7fa808]https://sourceforge.net/p/ngspice/ngspice/ci/7fa808e0b376199821799a3b032907566961bb9f/struct card, #1/6, minor cleanupRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/7fa808e0b376199821799a3b032907566961bb9f/Robert Larice committed [5d00bd]https://sourceforge.net/p/ngspice/ngspice/ci/5d00bdc6bb720d184f433fdb1e788b957a76575c/struct card, #2/6, drop struct lineRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/5d00bdc6bb720d184f433fdb1e788b957a76575c/Robert Larice committed [959c14]https://sourceforge.net/p/ngspice/ngspice/ci/959c14e26d580d475743136cd603941f8544e720/struct card, #4/6, drop typedef cardRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/959c14e26d580d475743136cd603941f8544e720/Robert Larice committed [1288a7]https://sourceforge.net/p/ngspice/ngspice/ci/1288a71535596a2e49c6d03f7cde18f1b606287f/struct card, #6/6, drop some redundant castsRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/1288a71535596a2e49c6d03f7cde18f1b606287f/Robert Larice committed [580768]https://sourceforge.net/p/ngspice/ngspice/ci/580768424ef6a846e952352516cf6d6d386e949c/numparam, #1/2, nupa_eval(), use struct card * argumentRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/580768424ef6a846e952352516cf6d6d386e949c/Robert Larice committed [91e58b]https://sourceforge.net/p/ngspice/ngspice/ci/91e58b5cc5ae8184680a6a9028c7b5a865a486bf/numparam, #2/2, nupa_copy(), use struct card * argumentRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/91e58b5cc5ae8184680a6a9028c7b5a865a486bf/Robert Larice committed [80585a]https://sourceforge.net/p/ngspice/ngspice/ci/80585a6838718c051ce6a7d088fb9d43343e7993/numparam, #3/4, nupa_scan(), use struct card * argumentRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/80585a6838718c051ce6a7d088fb9d43343e7993/Robert Larice committed [887ede]https://sourceforge.net/p/ngspice/ngspice/ci/887ede1ae1cc8313fcaa71fa69b33082cad30412/numparam, #4/4, defsubckt(), use struct card * argumentRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/887ede1ae1cc8313fcaa71fa69b33082cad30412/Robert Larice committed [dcf507]https://sourceforge.net/p/ngspice/ngspice/ci/dcf50710f56252b63aca505787c7c64b9a5f4df0/numparam/xpressn.c, #2/5, cleanup, drop 'err' and return earlyRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/dcf50710f56252b63aca505787c7c64b9a5f4df0/Robert Larice committed [951e2c]https://sourceforge.net/p/ngspice/ngspice/ci/951e2c188f7d8a1b31715ab57cdef4fc49de9b90/numparam/xpressn.c, #3/5, whitespaceRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/951e2c188f7d8a1b31715ab57cdef4fc49de9b90/Robert Larice committed [bf33c8]https://sourceforge.net/p/ngspice/ngspice/ci/bf33c841fea5cd95ecad3aa9f8b1a8ba82a81a6b/numparam/xpressn.c, #4/5, use a local 'c' instead of 'dt'Robert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/bf33c841fea5cd95ecad3aa9f8b1a8ba82a81a6b/Robert Larice committed [22cc54]https://sourceforge.net/p/ngspice/ngspice/ci/22cc545d9bcd5846b6281f2bea7f62ed53665f80/numparam/xpressn.c, #5/5, drop unused assignmentRobert LariceSun, 26 Nov 2017 12:38:24 -0000https://sourceforge.net/p/ngspice/ngspice/ci/22cc545d9bcd5846b6281f2bea7f62ed53665f80/