best Spice models for tubes?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

jhaible

Well-known member
Joined
Jun 24, 2004
Messages
530
Location
Germany
I know that Spice models for tubes are not on the level of semiconductor models, but everything is relative, and I'd like to know a source of some (relatively) good Tube spice models.

What would you recommend?

Thanks in advance,

JH.
 
Hi Jh,
I have been looking for something like that myself, especially a Pspice software that allows you to input the tube specs and replicate its curves as on the data sheet mainly looking to test the biasing range.

Here is the only simulator that is dedicated on tubes which I have not heard of any feedback from any where but am interested to tested out if there was a demo trial version, check it out....

http://www.glass-ware.com/Audio_Software_Descriptions.html
 
I know I got some at Duncan Munroe's site http://www.duncanamps.com/spicevalves.html, There are a few other spots that have some basic models I can't recal. Maybe a Google search.

I use a 3F4 Spice program I got inexpensively (can not afford Pspice) and I made a few models myself. They are simple, hybread models derived from Koren,Munroe and data sheets as well as trial and error.

Most valve models give a reasonable approximation until you get to the extremes of the operating curves.
 
[quote author="Learner"]Hi Jh,
I have been looking for something like that myself, especially a Pspice software that allows you to input the tube specs and replicate its curves as on the data sheet mainly looking to test the biasing range.

Here is the only simulator that is dedicated on tubes which I have not heard of any feedback from any where but am interested to tested out if there was a demo trial version, check it out....

http://www.glass-ware.com/Audio_Software_Descriptions.html[/quote]


Thanks for the link.
I have seen these, but that's a set of specialized programs, not compatible with Spice, is it?

JH.
 
[quote author="thomasholley"]I know I got some at Duncan Munroe's site http://www.duncanamps.com/spicevalves.html, There are a few other spots that have some basic models I can't recal. Maybe a Google search.

I use a 3F4 Spice program I got inexpensively (can not afford Pspice) and I made a few models myself. They are simple, hybread models derived from Koren,Munroe and data sheets as well as trial and error.

Most valve models give a reasonable approximation until you get to the extremes of the operating curves.[/quote]


Thanks, I'll try these!

JH.
 
Thanks for the link.
I have seen these, but that's a set of specialized programs, not compatible with Spice, is it?

JH.

Yea, I they are an independent program that assist you to design/calculate the biasing value and predict the outcome. However, I don't think it will display transient analysis..... :sad:

I donno why no Pspice program allows you to customise your own active device models by inputing the data sheet parameters and curves... :shock:

Or is there!? :?:
 
[quote author="Learner"]
Thanks for the link.
I have seen these, but that's a set of specialized programs, not compatible with Spice, is it?

JH.

Yea, I they are an independent program that assist you to design/calculate the biasing value and predict the outcome. However, I don't think it will display transient analysis..... :sad:

I donno why no Pspice program allows you to customise your own active device models by inputing the data sheet parameters and curves... :shock:

Or is there!? :?:[/quote]


I think PSpice has such features for diodes, not sure about transistors, certainly not for tubes.

A quick check of the 12AX7 and 12BH7 models from the duncanamps site shows their severe limitations. Very simple modeling, lower part of plate courves just cut off ...

Ok to find bias points, at least in the quick circuit I've tried.

No way to find a sweet spot for certain distortion behaviour, though ...

I found a much "better" model for the 12AX7 (must look up the link), plate courves look really nice, but it has tremendous problems with convergence. So I could plot the data sheet courves easily, but a simple grounded cathode amp with bypassed Rk already runs into errors ...

Sigh.

JH.
 
Here is the link:

http://www.lynx.bc.ca/~jc/Koren_Maillet.html

Looks extremely nice if you plot ip vs. vp, but I'm running into problems with real life applications. Maybe I've not tried hard enough (just a very quick test). Anybody else wanna try it ?

JH.
 
Here's more about this:

http://www.normankoren.com/Audio/Tubemodspice_article.html

Needs closer study.

JH.
 
[quote author="jhaible"]Here's more about this:

http://www.normankoren.com/Audio/Tubemodspice_article.html

Needs closer study.

JH.[/quote]

Whoo....yes, we shall definitely take a closer look into that.... if I can just upgrade the simm ram in my brain to rambus, along with the new P4 800FSB MB and a 3Ghz CPU plus some 15K rpm scsi HD :shock: :oops:
 
I think if you use PSPICE, the Norman Koren models work pretty well. The best I have run into at least. I use this method when I make my own models but I usually end up having to simplify something because of convergence problems with my 3F4 program. the very simple 3/2 law subcircuit using diode models works OK for rudimentary work.

Mostly I have found these models only get you in the ball park. From there it is breadboard work. U certainly would like to find a way to make better models.
 
I use a 3F4-type program. The small-signal triode models I've tried have given very misleading results. I'm back to drawing load lines on paper copies of plate curves and old-fashioned breadboarding and measurement. For active circuits, anyway, I just don't trust SPICE.
 
yeah, if you have a transformer in the circuit with the tubes, then you probably won't see subtle stuff, like the effect of varying impedance of the transformer messing with the loadlines, changes of impedance with level, and IM distortion caused by the tramsformer/tube interaction.
Great for passive filters, though.
 
[quote author="thomasholley"]I use a 3F4 Spice program I got inexpensively (can not afford Pspice)[/quote]
Why not just use a demo version? I still use the PSpice "student edition" (or whatever it's called) I got at the university. It's version 6.1 for DOS as far as I remember. I know it has limitations, but I haven't reached the limits yet...

Best regards,

Mikkel C. Simonsen
 
Back
Top