Hi,Think of the curve as the top of your bent leg. The knee is clearly where the red line is.
Bob
Take a look at the second plot and see if you change your mind.
Hi,Think of the curve as the top of your bent leg. The knee is clearly where the red line is.
Bob
Hi,Is there any formal definition of a knee voltage? Because I don't know any.
Hi,It's a rather arbitrary judgement call but in most cases I'd probably opt for something in the vicinity of the red line, roughly where the increase in conductivity per applied volt is greatest. But there are situations where I'd be most interested in what the Zener is doing around the green line or even below, and situations where I'd want to know what it's doing around the violet line or even higher.
So what I'd call the "knee" kinda depends.
Hi,Take your pick... Lots of "knee point" opinions.
View attachment 177638 View attachment 177639 View attachment 177640 View attachment 177641 View attachment 177642 View attachment 177643
Hi,Well, I'm colorblind but the low line looks like the 45° point for one definition and the top line where the "Knee" flattens out and becomes a constant slope for another definition.
I think that's true.It also looks like the knee may not be as important anyway
Hi,I think that's true.
The exact location of the Zener "knee" is generally not of interest for normal circuit design.
The typical parameters of interest are the Zener voltage at a desired operating current, and the Zener dynamic resistance (a measure of it's regulation accuracy).
In my very limited experience with Zeners, it seems like the available nominal values are spaced out far enough that the small differences you're pointing out would be negligible.Hi,
Well the other view is that the knee should indicated something comparative about two different zeners. But if everybody defines it differently this wont be possible. Compare this idea to the cutoff frequency of a low pass filter.
There would still have to be other specs too of course.
Hi,In my very limited experience with Zeners, it seems like the available nominal values are spaced out far enough that the small differences you're pointing out would be negligible.
In other words, if I want a 5V clamping action, it's not like I get to choose between a variety of "5V" Zeners and find myself wishing they all used a standard methodology with greater precision, listing 4.97 vs 5.05, etc. In my experience, I end up choosing between 4.7 and 5.1V options. For any more detail, I have to read the datasheet for the parameters that actually matter in my given situation. I'm not sure a more precise or consistent Zener knee description would help with this process.
Hi
Maybe a different approach is needed...
Repectfully MrAl....What problem are you trying to solve that isn't already solved using current spice modelling methods? Or maybe your just experimenting?
Just curious...
Edit: Oh...you posted just before I did...
eT
What is your metric for "better"? For instance, given the statistical nature of real devices, a marginal improvement of fitting a particular, single diode's characteristic may not be viewed as "better" if it comes in significantly increased computational costs.Hi,
Yes just looking at different ways to do this. The model i have created looks better than the standard spice model.