John Curl's Blowtorch preamplifier

[Sigurd Ruschkow]

Sure:

************************************************************************************

******************* 2SK170 *********************************************************

*** N-Channel JFET
.MODEL 2SK170GR_4m1 NJF (AF=492.527m BETA=17.3669M CGD=19.8997p CGS=24p IS=10f
+ KF=.0007956085f LAMBDA=1m RS=5.51589 VTO=-0.50)

*** N-Channel JFET
.MODEL 2SK170V_10m NJF (AF=499.843m BETA=29.5953M CGD=19.8997p CGS=24p IS=10f
+ KF=.001681153f LAMBDA=1m RS=7.19294 VTO=-0.65)

*** N-Channel JFET
.MODEL 2SK170Y_2m NJF (AF=499.84582m BETA=13.6348M CGD=19.8997p CGS=24p IS=10f
+ KF=.002599428f LAMBDA=1m RS=1.74483 VTO=-0.39)

*** N-Channel JFET
.MODEL 2SK170BL_8m NJF (AF=500.504m BETA=27.7612M CGD=19.8997p CGS=24p IS=10f
+ KF=.002229965f LAMBDA=1m RS=8.03465 VTO=-0.60)

*** N-Channel JFET
.MODEL 2SK170V_12m NJF (AF=499.843m BETA=32.0953M CGD=19.8997p CGS=24p IS=10f
+ KF=.001681153f LAMBDA=1m RS=7.19294 VTO=-0.70)

*** N-Channel JFET
.MODEL 2SK170V_18m NJF (AF=499.843m BETA=40.0953M CGD=19.8997p CGS=24p IS=10f
+ KF=.001681153f LAMBDA=1m RS=7.19294 VTO=-0.80)



************************************************************************************
************************************************************************************

******************* 2SJ74 *********************************************************

*** P-Channel JFET
.MODEL 2SJ74BL_7m8 PJF (AF=500.257m BETA=21.351M CGD=106.132p CGS=73p IS=10f
+ KF=.0005386405f LAMBDA=1m VTO=-0.6)

*** P-Channel JFET
.MODEL 2SJ74GR_4m PJF (AF=499.959m BETA=17.9759M CGD=106.132p CGS=73p IS=10f
+ KF=.00133578f LAMBDA=1m VTO=-0.47)

*** P-Channel JFET
.MODEL 2SJ74V_16m3 PJF (AF=500.034m BETA=35.5713M CGD=106.132p CGS=73p IS=10f
+ KF=.00163621f LAMBDA=1m RS=4.75307 VTO=-0.75)

*** P-Channel JFET
.MODEL 2SJ74Y_2m1 PJF (AF=500.053m BETA=15.441M CGD=106.132p CGS=73p IS=10f
+ KF=.001223835f LAMBDA=1m RS=448.083m VTO=-0.37)

*** P-Channel JFET
.MODEL 2SJ74BL_10m PJF (AF=500.257m BETA=23.351M CGD=106.132p CGS=73p IS=10f
+ KF=.0005386405f LAMBDA=1m VTO=-0.65)

*** P-Channel JFET
.MODEL 2SJ74BL_12m PJF (AF=500.257m BETA=24.351M CGD=106.132p CGS=73p IS=10f
+ KF=.0005386405f LAMBDA=1m VTO=-0.70)


************************************************************************************
************************************************************************************


******************* 2SK246 *********************************************************

*** N-Channel JFET
.MODEL 2SK246BL_8m NJF (BETA=0.95111M CGD=8.29156p CGS=6.5p IS=10f RD=260 LAMBDA=1m VTO=-2.9)

*** N-Channel JFET
.MODEL 2SK246GR_4m NJF (BETA=0.983573M CGD=8.29156p CGS=6.5p IS=10f RD=390 LAMBDA=1m VTO=-2.0)


************************************************************************************
************************************************************************************


******************* 2SJ103 *********************************************************

*** P-Channel JFET
.MODEL 2SJ103BL_8M PJF (BETA=1015.438U CGD=11.9398p CGS=14.4p IS=10f RD=200 LAMBDA=1m VTO=-2.8)

*** P Channel JFET
.MODEL 2SJ103GR_4M PJF (AF=500m BETA=3.01875M CGD=13.2918p CGS=14.3694p IS=10f
+ KF=.001f LAMBDA=1m PB=806.232m RS=186.826 VTO=-1.9)

************************************************************************************
************************************************************************************



Sigurd

Hi Sigurd ,
do you mind posting them?
Rüdiger

 

[syn08]

Sure:

<snip>

Sigurd

May I suggest some improved models?

http://www.diyaudio.com/forums/showthread.php?postid=1420730#post1420730

 

[GRollins]

Isn't it fun watching simulation people? Pop some popcorn, sit back and watch the posts roll by...see how long before one tells another that his software and/or models aren't up to snuff.
Gimme reality. I spent twenty minutes with real parts and found out what I wanted to know. No muss, no fuss, just dependable answers.

Grey

 

[scott wurcer]

Isn't it fun watching simulation people? Pop some popcorn, sit back and watch the posts roll by...see how long before one tells another that his software and/or models aren't up to snuff.
Gimme reality. I spent twenty minutes with real parts and found out what I wanted to know. No muss, no fuss, just dependable answers.

Grey

Virtually no one working in any state of the art process breadboards any circuit these days. It just is not possible, so this line of reasoning has run its course.

 

[GK]

Virtually no one working in any state of the art process breadboards any circuit these days. It just is not possible, so this line of reasoning has run its course.

And most designers don't spend 20 minutes breadboarding for answers to something that 2 minutes with the calculator would solve....................

 

[uli]

Isn't there a difference between designing very complex circuits which are built in an IC and designing fairly simple things, where the only target is sound quality?

Maybe I missed something but IMO it is not possible to spice intrinsic sound quality.

So let's talk real audio electronics instead of modelling.

Uli

 

[jacco vermeulen]

Or a bit of both ?
I don't see how any engineering type can disqualify (decent) simulation tools, nor how someone who is after sonic performance can do without prototyping.
Spicing saves time, real modelling saves from errors, sauve the virtual exception.

(not very polite, but the body part nomenclature was rather funny)

 

[PMA]

I knew you would like to comment, Jacco 😀

 

[john curl]

Simulation is OK, much of the time. However, most audio circuits can be designed without simulation and still work well. The simulation models still leave a lot to be desired, compared to direct measurement.
However, 'what if' questions can be easily and quickly answered with simulation. Older engineers, people like me, who used a slide rule for the first five years of their professional design life, can often do more with simple tools than people who often get lost in their computer models and crashes.

 

[uli]

This is a problem of blindness on one eye. Spicing everything is like playing golf on a decent simulator. There IS a difference between reality and computing. Using tools without having critical distance to the tool's abilities is simply foolish.

Uli

 

[jacco vermeulen]

Pavel,

proletarians may be disguised, their true origin always surfaces.
(fortunately humor has not boundaries, i guess my slide rule makes me a geezer too)

 

[Sigurd Ruschkow]

Uli,

I don't think that anyone will say that there is a one-to-one
correlation between simulations and the real world.

I also do aeordynamic simulations. Even the best simulators that we use, are maybe 50-70 % correct only. It does give us a hint where we should focus our attention, and also quickly evaluate new ideas.

Another thing that SPICE does for me is to quickly check how component tolerances (e.g Idss and Hfe spreads) will affect a circuit. Monte Carlo analysis is a very neat tool.

The more people work on models and simulation algorithms, the more exact simulations will become.

I am using the professional version 9 of Spectrum Software's MicroCAP and am very happy with it.




Sigurd

 

[syn08]

Or a bit of both ?
I don't see how any engineering type can disqualify (decent) simulation tools

My problem with simulations start when simulation results attempt to replace knowledge and understanding of a circuit functionality. There is a simulation empiricism ("because my simulator tells me so") that I dislike as much as plugging a pair of JFETs in the breadboard for determining if a differential amp works properly. The simulator can always answer your questions (how correct? - that's a different story), but will never tell you what to ask.

 

[scott wurcer]

My comment was just food for though with reference to where the larger engineering community has gone. Plenty of people here are enjoying the availability of new chip amps and other IC's and the hard reality is they were probably completely designed by simulation. I can't comment on the hit rate, but there are plenty of examples of very complex chips that have gone right from the computer to the customer meeting every spec.

Modeling for discrete devices seems way behind, I think that's where the problem is. With enough effort and a simulator that has temperature as a third independent variable I think you could get results with "numbers" so close it wouldn't matter. OTOH if you included everything like power transformer magnetics, etc. the problem seems insurmountable. Of course no amount of simulation will tell a difference between Vishay's and Dale
RN55C's.

 

[syn08]

My comment was just food for though with reference to where the larger engineering community has gone. Plenty of people here are enjoying the availability of new chip amps and other IC's and the hard reality is they were probably completely designed by simulation. I can't comment on the hit rate, but there are plenty of examples of very complex chips that have gone right from the computer to the customer meeting every spec.

Modeling for discrete devices seems way behind, I think that's where the problem is. With enough effort and a simulator that has temperature as a third independent variable I think you could get results with "numbers" so close it wouldn't matter. OTOH if you included everything like power transformer magnetics, etc. the problem seems insurmountable. Of course no amount of simulation will tell a difference between Vishay's and Dale
RN55C's.

The problem is the gap between device modeling and macro modelling.

IC simulators do not work at device level. Instead, they use a set of "building blocks" (usually proprietary to the IC design house) that are fully characterized and a relevant set of parameters are already measured and integrated in a macro model. Take, for example, noise. A macro model may only describe the output noise (perhaps as a function of frequency) without any details on the origins of this noise. The circuit designer doesn't care if that noise is of a shot, thermal or 1/f origin.

Discrete device modelling goes to the root of the device physics and it's indeed well behind, as much as discrete devices are about to get extinct.

Edit: Scott, my comments are not targeted to you - obviously you are aware of all these.

 

[PMA]

You cannot mix DIYers playing around 4 transistors (supported by audio designers, who are not friends with modern simulation SW) and professionals, designing complex circuits. For rather complex circuits, there is no way than to use simulation. For primitive circuits, experience counts, but even in this case simulation helps to optimize the design.

 

[john curl]

It should be pointed out that none of my often successful designs have ever been fully characterized by simulation.
I am not an adversary of computer simulation, by any means!
In fact, I was doing professional circuit simulation as early as 1966, and was originally hired, partially because I had so much enthusiasm for using computer simulation. However, as I matured in understanding, I started to see its flaws, and when a simple slide rule demonstration PROVED in one case, that the computer simulation gave the wrong results, we were very disappointed, and realized that it was not infallable.
I was there when SPICE was first developed at UC Berkeley. My favorate engineering professor, Dr. Pederson, is considered the father of SPICE. However, even then, I saw examples of REVERSE ANALYSIS, such of the 741 made, long after the design was released to the public, rather than anything really new and special being shown. This was around 1971-1973.
When the personal computer became available to me, in 1978, I found it useful for limited analysis, such as AC filter design, but that was about all. By 1980, or so, I purchased the first transient analysis program made by Microcap. It was VERY limited.
However, Microcap continued to develop and became very useful for AC, DC and Transient Analysis, BUT only with simple models. The transistor models were not that good. However, I used this program for almost 25 years, and made great headway with it, especially with RLC interactions and filter design.
Berkeley Spice was adapted for home computers as well, but it was VERY difficult to make work, at first. It appeared to be very 'unforgiving' so, even though I have had it installed in my computer(s) for the last 15 years, I did not use it much.
Still, I could develop successful circuit designs without the use of SPICE. What a concept!
This is why I gave Bob Cordell an implied 'raspberry' when I said " I don't need no stinkin' Spice simulations!" that launched a new thread here.
To be sure SPICE can be valuable, but often just a simple solderless breadboard will work as well. Sometimes, it takes less effort as well.

 

[GRollins]

I don't mind the false positives...those would be weeded out the moment someone actually tried to build something. What bothers me most is the problem of false negatives. It's fun to watch proponents of simulation suddenly discover that they forgot to defrost the cat or paint the roof when I bring this up. To date, not one single person has admitted the possibility that a nice idea might be lost forever if their simulation program tells them it won't work. Not one. Rare is the simulation person who trusts their gut enough to build something their simulator tells them is impossible.
Don't believe it can happen? Look at the first ten or twenty pages of my Aleph-X thread, wherein a veritable army of people told me it would never work, based on their simulations. I wasn't encumbered by the 'impossibility' of it...in fact, I knew better because I already had a functioning prototype. Impossible? Obviously not. But they believed it was impossible because their simulations told them so. Would any of them have ever built the circuit if they had had the idea independently? Hell, no! The truth of the matter is that if Nelson hadn't backed me up that, yes, the concept was viable (given that he had a commercial product of that nature coming out), there would still be a sizable number of doubters to this day. That thread was the final nail in the coffin as far as I was concerned.
Until such day as simulation software manages to discriminate between possible and impossible with more finesse I refuse to waste my time on such things. I believe Scott Wurcer has said that AD has in-house software that is much better than the stuff available to the masses. This doesn't invalidate my point for the simple reason that, by definition, that software is effectively nonexistent as far as we're concerned. For all I know NASA, IBM, and Intel have effective internal software. So what? If it isn't available, or not affordable, it's nonexistent.
People sneer because I use real parts. Too bad. The difference is that I know and they merely believe. This is science as religion. Your beliefs are only as solid as their underpinnings, and given that the (publicly available) software and models are demonstrably flawed (as evidence, I offer every single thread here at DIY wherein people get conflicting simulation results [in other words, every single thread where simulation plays a non-trivial part]), your beliefs are flawed.
It's painful to think of how many bright ideas fell by the wayside because someone was clever enough to come up with something new--a genuinely novel, useful idea--but was not strong enough to tell his simulation program to take a flying leap.
Me? I'm just sitting here munching popcorn, watching people defend a weak belief system. What's needed is what's called an "intervention" in today's vernacular. Someone needs to snatch some of the worst offenders by the scruff of their necks and rub their noses in the differences between computer games and reality. It won't happen--I know that. I also know that they'll be offended that I'm not bowled over by the strength of their beliefs. It's always that way with religious zealots. So I'll continue to munch popcorn and build real circuits and they'll stay up late at night congratulating themselves over getting simulated distortion down to .000000001% on their computer games. Might as well spend their time playing World of Warcraft for all the good it does them.
But as has been said, audio is part of the entertainment industry, and if some find computer fantasies entertaining, then I suppose it has fulfilled its purpose.

Grey

 

[forr]

Hi,

The following has been written thirty years ago in Wireless World, march 1978 :

p41
"Audio power amplifer design - 2
.Negative feddback concepts"
.The best result of mathematics is to be able to do without it - Oliver Heaviside"

by Peter Baxandall


p43
"Some people may say that arriving at optimum values for some components by trial and error does not constitute a respectable modern design technique but I cannot agree with this outlook. One way to regard such a trial-and-error approach is to say that one is using the actual amplifier itself as an analogue computer - changes are made to the circuit and the values are displayed in an analogue form on an oscilloscope"

 

[syn08]

It should be pointed out that none of my often successful designs have ever been fully characterized by simulation.

John,

Simulation software did huge advances in the last thirty years. In despite of these advances, it is almost impossible today (as much as it was impossible thirty years ago) to persuade a device simulator to precisely predict the amplitude of a Hartley RF oscillator.

It's nothing wrong with the computers and the simulation software, but with the models. In the last 30 years very little was invested in developing new device models or extending the old ones. Even less money went to extracting good quality device models from device characterisation/measurements.

With all regrets I have to admit that these facts are entirely supported by the today's business requirements and practice, and this is told by someone which started with ECAP on IBM mainframes in the 70's, and ended up the simulation adventures by working for OrCad/Cadence sometimes during this decade, in between doing a PhD in electronic devices characterisation and modelling.