STEORN DEMO LIVE & STREAM in Dublin, December 15th, 10 AM

[Omnibus]

@teslaalset,

That's strange because the least squares method which I suppose is the method for calculating the slope and intercept in Excel should yield the accurate result and not just by finding the slope of the line between the first and the last point. Will have to try it with Excel 2010 when it arrives (USPS shows on their website that it's in Springfield, MA at this moment; wonder if it's going to arrive today, being a holiday),

[Omnibus]

@teslaalset,

That's very interesting. Think about it, the method from the theoretical writeup also uses in effect the last and the first point, that is, uses a brute force approach. However, when you get into the nitty-gritty the result changes even in the theoretical calculations. How can this be?

[teslaalset]

@teslaalset,

That's strange because the least squares method which I suppose is the method for calculating the slope and intercept in Excel should yield the accurate result and not just by finding the slope of the line between the first and the last point. Will have to try it with Excel 2010 when it arrives (USPS shows on their website that it's in Springfield, MA at this moment; wonder if it's going to arrive today, being a holiday),

The difference is that the 'slope' function gives you the average slope of the sinus, while the formula I used instead uses the slope of the start and end value (the linear line between two points). If the sinus would be a sawtooth, they would give the same results.

[Omnibus]

The difference is that the 'slope' function gives you the average slope of the sinus, while the formula I used instead uses the slope of the start and end value (the linear line between two points). If the sinus would be a sawtooth, they would give the same results.

That shouldn't be, though. Least squares method, using all the points, not just two, should be the really accurate method. Could it be there's a problem in Excel? Alternatively, could it be that the type of function we're dealing with isn't prone to the kind of integration (I'm talking about the theoretical, not the numerical) we usually do.  You know how there are requirements for a function to be smooth (this one is smooth, though), no poles, no special points etc., etc. There must be something subtle which is escaping me.

[tagor]

 
do a google search with :
"Synthesis of Passive RC Networks with Gains Greater than Unity"

there is so much overunity !