With some success for triode models, I thought I'd take the information from Koren, Leach, Rydel etc. and expand the Microsoft Excel Solver routine:
Here's a first approximation using the minima of log squashed error and all the data points:
When the anode voltage data points Vak<=50 are removed, the SPICE parameters become more useful:
I'm all ears to anyone who can improve this.
Here's a first approximation using the minima of log squashed error and all the data points:
An externally hosted image should be here but it was not working when we last tested it.
When the anode voltage data points Vak<=50 are removed, the SPICE parameters become more useful:
An externally hosted image should be here but it was not working when we last tested it.
I'm all ears to anyone who can improve this.
Yes. I was going to do them myself, but my laptop died a horrible death this morning
Since the pentode is most likely to be used as a power amplifier, wouldn't the mismatch of "knees" cause significant error for output power and distortion calculations?
Here's how the curves fit, with the parameters used to get best fit in the previous chart:
Optimizing using solver can make it appear better:
But the new parameters result in this mess:
An externally hosted image should be here but it was not working when we last tested it.
Optimizing using solver can make it appear better:
An externally hosted image should be here but it was not working when we last tested it.
But the new parameters result in this mess:
An externally hosted image should be here but it was not working when we last tested it.
I took Rydel's parameters and compared them with the "limited" solver solutions:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
6LU8 with Ayumi's pctube
Here is Ayumi's 6LU8 model vs. the datasheet. Ayumi's model building process is described here. The key is not to use the Koren parameters...
Here is Ayumi's 6LU8 model vs. the datasheet. Ayumi's model building process is described here. The key is not to use the Koren parameters...
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Here is the Ayumi 6AC7 model vs. the datasheet, while the knees do not match up well, the transfer characteristic is spot on, so it should still be good for voltage amplifier applications.
An externally hosted image should be here but it was not working when we last tested it.
I used Google Translate to get through it, a lot of the text are math anyway, so you should have no difficultly understand them. But basically, Ayumi's model building process relies on electrical equivalent circuits, KCL and Child-Langmuir law unlike Koren's phenomenological approach. Ayumi's pctube program also uses a curve fitting technique similar to what you are doing with the Excel Solver routine.
Hi jazbo8,
I always use the Bing Translator
This translator give good results but it is for the math I need a translator ...
This is the kind of accurate curves I like to have for every pentode or "beam power" tubes. I am not a math crack and I never try making a tube spice model by myself but I think this site will help me to understand how to do it.
This page explain the equivalent circuit of a triode and a pentode :
ayumi.cava.jp/audio/pctube/node47.html#SECTION00952000000000000000 - Translator
This is nice but I wonder if it is possible to get all the necessary data for the formulas only from the datasheet specifications and curves ?
I think the first step for a beginner like me to understand how the tubes models works is experimenting with fictive models and test tube data, just to see what kind of curves they produce. I also need to know more about the "URAMP" function because it is used almost everywhere ...
I think the pentode curves should be very accurate in the upper left where is the 0V knee and the lower right because this is where the tube is really operating in a power amplifier circuit, where pass the AC load line.
Thank you for this link !
The data usually comes from the datasheet, but sometimes, the data comes from traced curves from actual tubes, the formulas then try to describe in mathematical terms how to best describe the tubes' properties (not always successful though).
You are being too modest, you are not a beginner
uramp(x) = x if x > 0, else 0. I'm not sure if SIMetrix uses it though, in any case, the Ayumi pctube library contains all the models in SIMetrix .MOD format, so you can use them directly.That's actually the hardest part - around the knees, as shown by Robert in this post.
Great to see you back on the forum!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.