Hi there. I wanted to start developing an amplifier for audio using JFET's, and my question is, is LTSpice reliable enough for that purpose. Can I trust that if it works on LTSpice I can at least buy the same transistors and I will be able to adjust the resistances and make it work satisfactorily? Cause here is the thing. Just to test things out, I tried to simulate the last stage of the Philips rh749 receiver's amplifier, from the capacitor at the base of the BC557B(differential pair) until the output darlingtons. It did not work. The output has almost no level and it's clipping most of the top of the wave, ramaining below the 0V line. That happens on the emitters of the differential pair as well. Then I went to multisim, drew the same circuit, even replacing transistors that weren't in the library for similars, and it worked fine, level, waveform, everything... After many days trying to understand, I got the spice models from the components I used on the multisim schematic, and used them on LTSpice's. Then it worked fine for one time on LTSpice(waveforms OK, levels I forgot to check). Then, I changed it back to LTSpice's models and began changing on by one to multisim's to spot what was the component that was causing the problem. It turns out that I kept changing until it was all multisim's, exactly like the time it worked, and it didn't work again. I just got the same clipping as before. It doesn't make any sense. I tried to delete .log and .raw files... nothing. Hasn't worked again. Does anyone know why is that? I just can't understand. Link to the schematics below, although you'll have to delete the first stage of pre-amp and tone control, reconnect the transistors(or use the symbols I provided that have the pins in a slightly different position) and add a 4ohm resistor at the output to simulate a speaker to mimic what I'm doing. This amplifier sounds very good by the way. For audiophiles it's worth doing.

https://drive.google.com/file/d/1xWSZcXFM6UVka6wHchOJcwxETWPWsRyF/view?usp=drivesdk

I tried to copy the models that I originally used on LTSpice to use in multisim, but multisim doesn't run them because it doesn't recognize many of the variables. Here are the models from multisim that I ran on LTSpice:

.model bc557bmultisim PNP (
+ Level=1 IS=3.44897e-011 BF=10000 NF=1.40366 VAF=540.707 IKF=0.728255
+ ISE=9.39256e-013 NE=1.68285 BR=2.18737 NR=1.5 VAR=1.08697 IKR=3.78894
+ ISC=1.00018e-016 NC=3.99992 RB=18.2485 IRB=0.01 RBM=0.1 RE=0.403156
+ RC=2.01578 CJE=1.05101e-011 VJE=0.99 MJE=0.238496 TF=4.23692e-010
+ XTF=16.763 VTF=859.777 ITF=3.7096 PTF=0 CJC=1.73742e-011 VJC=0.04
+ MJC=0.348553 XCJC=0.91739 TR=1e-007 CJS=0 VJS=0.75 MJS=0.5 XTB=0.1
+ EG=1.206 XTI=4 FC=0.1 KF=0 AF=1
+ )

.model bc546bmultisim NPN (IS =1.8E-14 ISE=5.0E-14 NF =.9955 NE =1.46 BF =400 BR =35.5
+IKF=.14 IKR=.03 ISC=1.72E-13 NC =1.27 NR =1.005 RB =.56 RE =.6 RC =.25 VAF=80
+VAR=12.5 CJE=13E-12 TF =.64E-9 CJC=4E-12 TR =50.72E-9 VJC=.54 MJC=.33)

.model bc548amultisim NPN (
+ Level=1 IS=1.89314e-014 BF=1536.18 NF=0.942746 VAF=16.0274 IKF=0.095245
+ ISE=4.13018e-014 NE=1.2599 BR=0.1 NR=1.13089 VAR=1.05111 IKR=0.622681
+ ISC=1.79252e-014 NC=2.7038 RB=3.11982 IRB=0.1 RBM=0.1 RE=0.0001
+ RC=0.001 CJE=7.68434e-012 VJE=0.4 MJE=0.349069 TF=5.80579e-010
+ XTF=4.00934 VTF=28.4173 ITF=0.176065 PTF=0 CJC=3.9003e-012 VJC=0.821072
+ MJC=0.329281 XCJC=0.8 TR=1e-007 CJS=0 VJS=0.75 MJS=0.5 XTB=0.1
+ EG=1.05 XTI=1 FC=0.8 KF=0 AF=1
+ )

.model baw62multisim D (
+ IS = 3.126E-9
+ N = 1.925
+ BV = 85
+ IBV = 0.0005
+ RS = 0.8109
+ CJO = 7.27E-13
+ VJ = 0.7226
+ M = 0.01318
+ FC = 0.5
+ TT = 4.04E-9)

.model bzx79c5v6multisim D (IS= 2.524e-016 RS= 0.5164 CJO= 1e-012 VJ= 0.75
+ TT= 5e-009 M= 0.3333 BV= 5.694 IBV= 0.2113 N= 1
+ EG= 1.11 XTI= 3 KF= 0 AF= 1 FC= 0.5
+ TNOM= 27)

 

Hi there. I wanted to start developing an amplifier for audio using JFET's, and my question is, is LTSpice reliable enough for that purpose. Can I trust that if it works on LTSpice I can at least buy the same transistors and I will be able to adjust the resistances and make it work satisfactorily? Cause here is the thing. Just to test things out, I tried to simulate the last stage of the Philips rh749 receiver's amplifier, from the capacitor at the base of the BC557B(differential pair) until the output darlingtons. It did not work. The output has almost no level and it's clipping most of the top of the wave, ramaining below the 0V line. That happens on the emitters of the differential pair as well. Then I went to multisim, drew the same circuit, even replacing transistors that weren't in the library for similars, and it worked fine, level, waveform, everything... After many days trying to understand, I got the spice models from the components I used on the multisim schematic, and used them on LTSpice's. Then it worked fine for one time on LTSpice(waveforms OK, levels I forgot to check). Then, I changed it back to LTSpice's models and began changing on by one to multisim's to spot what was the component that was causing the problem. It turns out that I kept changing until it was all multisim's, exactly like the time it worked, and it didn't work again. I just got the same clipping as before. It doesn't make any sense. I tried to delete .log and .raw files... nothing. Hasn't worked again. Does anyone know why is that? I just can't understand. Link to the schematics below, although you'll have to delete the first stage of pre-amp and tone control, reconnect the transistors(or use the symbols I provided that have the pins in a slightly different position) and add a 4ohm resistor at the output to simulate a speaker to mimic what I'm doing. This amplifier sounds very good by the way. For audiophiles it's worth doing.

https://drive.google.com/file/d/1xWSZcXFM6UVka6wHchOJcwxETWPWsRyF/view?usp=drivesdk

I tried to copy the models that I originally used on LTSpice to use in multisim, but multisim doesn't run them because it doesn't recognize many of the variables. Here are the models from multisim that I ran on LTSpice:

.model bc557bmultisim PNP (
+ Level=1 IS=3.44897e-011 BF=10000 NF=1.40366 VAF=540.707 IKF=0.728255
+ ISE=9.39256e-013 NE=1.68285 BR=2.18737 NR=1.5 VAR=1.08697 IKR=3.78894
+ ISC=1.00018e-016 NC=3.99992 RB=18.2485 IRB=0.01 RBM=0.1 RE=0.403156
+ RC=2.01578 CJE=1.05101e-011 VJE=0.99 MJE=0.238496 TF=4.23692e-010
+ XTF=16.763 VTF=859.777 ITF=3.7096 PTF=0 CJC=1.73742e-011 VJC=0.04
+ MJC=0.348553 XCJC=0.91739 TR=1e-007 CJS=0 VJS=0.75 MJS=0.5 XTB=0.1
+ EG=1.206 XTI=4 FC=0.1 KF=0 AF=1
+ )

.model bc546bmultisim NPN (IS =1.8E-14 ISE=5.0E-14 NF =.9955 NE =1.46 BF =400 BR =35.5
+IKF=.14 IKR=.03 ISC=1.72E-13 NC =1.27 NR =1.005 RB =.56 RE =.6 RC =.25 VAF=80
+VAR=12.5 CJE=13E-12 TF =.64E-9 CJC=4E-12 TR =50.72E-9 VJC=.54 MJC=.33)

.model bc548amultisim NPN (
+ Level=1 IS=1.89314e-014 BF=1536.18 NF=0.942746 VAF=16.0274 IKF=0.095245
+ ISE=4.13018e-014 NE=1.2599 BR=0.1 NR=1.13089 VAR=1.05111 IKR=0.622681
+ ISC=1.79252e-014 NC=2.7038 RB=3.11982 IRB=0.1 RBM=0.1 RE=0.0001
+ RC=0.001 CJE=7.68434e-012 VJE=0.4 MJE=0.349069 TF=5.80579e-010
+ XTF=4.00934 VTF=28.4173 ITF=0.176065 PTF=0 CJC=3.9003e-012 VJC=0.821072
+ MJC=0.329281 XCJC=0.8 TR=1e-007 CJS=0 VJS=0.75 MJS=0.5 XTB=0.1
+ EG=1.05 XTI=1 FC=0.8 KF=0 AF=1
+ )

.model baw62multisim D (
+ IS = 3.126E-9
+ N = 1.925
+ BV = 85
+ IBV = 0.0005
+ RS = 0.8109
+ CJO = 7.27E-13
+ VJ = 0.7226
+ M = 0.01318
+ FC = 0.5
+ TT = 4.04E-9)

.model bzx79c5v6multisim D (IS= 2.524e-016 RS= 0.5164 CJO= 1e-012 VJ= 0.75
+ TT= 5e-009 M= 0.3333 BV= 5.694 IBV= 0.2113 N= 1
+ EG= 1.11 XTI= 3 KF= 0 AF= 1 FC= 0.5
+ TNOM= 27)

You asked:

I wanted to start developing an amplifier for audio using JFET's, and my question is, is LTSpice reliable enough for that purpose?​

LTspice is reliable enough, but your ability to model actual circuits accurately is not. The purpose of simulation is understanding. It is manifestly not for the purpose of design or design verification. The models of components use typical values. Most designer do not use typical values, they make designs that do not depend on particular device parameters.

You then asked:

Can I trust that if it works on LTSpice I can at least buy the same transistors and I will be able to adjust the resistances and make it work satisfactorily?​

I would not be comfortable doing that without conducting some tests. As it stands there is no central authority that will vouch for any particular model. Some of the extant models are 2 or more decades old. For example, I have in my collection about two dozen models for the popular TIP31C power transistor. I think four of them are close to the behavior of the actual part. The rest are garbage. How will you know going in if a particular model is good or bad -- you won't you'll have to verify things.

I'm unable to follow what you are doing switching back and forth between LTspice and Multisim. It's nice that you are bilingual, but nobody anywhere would believe or expect that models developed for one would necessarily work on the other. It is often the case that models developed for Pspice will work in LTspice, but there are some that won't.

I'm not sure that simulation will be much help until you become more familiar with what it does and does not tell you. This comes with experience and unfortunately it can't be rushed. If you are only making one of these it might be to you benefit to ignore simulation for now and just concentrate on making forward progress.

The good news is that there are plenty of folks here, including some audiophiles, who can help you with simulation problems and issues. The best approach is to attach you simulation files along with the symbols and subcircuits to your post so we can run the same simulations and help you get them the way you want them.

ETA: Just in case you imagine that you can create your own models from a datasheet, this has proven in most cases to be very difficult for any but the most basic models for diodes or the Gummel-Poon model for a BJT. The models for MOSFETs are more complicated than you can imagine.

 

an amplifier for audio using JFET's

The problem isn't SPICE, it's JFETs. Their parameters vary wildly. If you are using long-tailed pairs, then you will not get satisfactory results unless you pay silly money for dual matched JFETs.
Erno Borberly is the expert on JFETs for audio, but most of his papers are behind paywalls.
JFETs have their uses in audio for high-impedance sources, but for power amplifiers, bipolars will give better results.