Dot convention with magnetically-coupled coils

As I slowly move my way through all of my EE classes, I keep running into things that seem to come out of left field. The first one was the unit "nepers per second", which apparently is just another way of saying "radians per second." I swear someone makes this stuff up just to torture undergrads.

The latest, though, is the dot convention on magnetically-coupled coils, aka transformers (including autotransformers). Can someone please explain to me what in the physical construction of a transformer creates this weird kind of independent polarity on each coil? I don't remember ever encountering this before.

> "nepers per second", which apparently is just another way of saying "radians per second."

No. Nepers are logarithmic like db.

> explain to me what in the physical construction of a transformer creates this weird kind of independent polarity on each coil?

Huh? What to explain? Coils have two ends. You can wire two coils on the same core in-phase or out-phase. The raw leads all look the same. The schematic may have been more conveniently drawn mis-phased. Placing dots is a necessary clue to getting the polarity correct.

You are too young to have built a "tickler" regenerative radio. A coil in the plate lead is put near the coil in the grid lead. One way is negative feedback and just clobbers gain; the other way is positive feedback and increases gain and selectivity. You can build it and try, or you can follow the dots.

PRR said:

> No. Nepers are logarithmic like db.

Click to expand...

I knew it! The other students DID have it wrong! I spent several hours of searching around trying to figure out what this unit is and what it means in terms of what's happening physically inside the circuit.

That lab basically ended with me just writing down what the TA told me to, with his being unable to explain to me what this unit actually is. The further I get in this degree, the more useless I think it is. I'm figuring more out on my own.

Huh? What to explain? Coils have two ends. You can wire two coils on the same core in-phase or out-phase. The raw leads all look the same. The schematic may have been more conveniently drawn mis-phased. Placing dots is a necessary clue to getting the polarity correct.

Click to expand...

I guess I'm getting the idea now, but I'd still like to see some physical examples of how this works on real transformers.

dmlandrum said:

PRR said:

> No. Nepers are logarithmic like db.

Click to expand...

I knew it! The other students DID have it wrong! I spent several hours of searching around trying to figure out what this unit is and what it means in terms of what's happening physically inside the circuit.

That lab basically ended with me just writing down what the TA told me to, with his being unable to explain to me what this unit actually is. The further I get in this degree, the more useless I think it is. I'm figuring more out on my own.

Click to expand...

As PRR says, the Neper is aloarithmic unit, based on the number e
http://en.wikipedia.org/wiki/E_(mathematical_constant)
instead of powers of 10.
Although mathematically unavoidable, it is less intuitive than powers of 10; I think that why it's gone into obsolescence.
Neper-per-second is even more obscure because it defines a rate of gain or attenuation per time unit; useful only in dynamic control and even then most people don't fully grab the concept. 

Huh? What to explain? Coils have two ends. You can wire two coils on the same core in-phase or out-phase. The raw leads all look the same. The schematic may have been more conveniently drawn mis-phased. Placing dots is a necessary clue to getting the polarity correct.

Click to expand...

I guess I'm getting the idea now, but I'd still like to see some physical examples of how this works on real transformers.

Click to expand...

Depends what culture you were born in. If you're British, the dot is the start of the winding (closest to the core); if you're a human being ;D, the dot is the "hot" point, the end of winding.
And that's all there is. The physical construction of the windings defines some differences between one end and the other, mainly in regard to stray capacitance.
If you take a standard construction, the end of the primary is close to the start of the secondary, so the capacitance between hot point of primary and cold point of secondary is much higher than the capacitance between cold of primary and hot of secondary. This unbalance results in decreased HF CMRR and response. A good audio transformer designer takes steps to minimize these problems (bi-filar construction, segmented or sandwiched layers).

abbey road d enfer said:

As PRR says, the Neper is aloarithmic unit, based on the number e
http://en.wikipedia.org/wiki/E_(mathematical_constant)
instead of powers of 10.

Although mathematically unavoidable, it is less intuitive than powers of 10; I think that why it's gone into obsolescence.

Neper-per-second is even more obscure because it defines a rate of gain or attenuation per time unit; useful only in dynamic control and even then most people don't fully grab the concept. 

Click to expand...

What you're saying makes perfect sense, except for one thing: our book associates nepers per second with omega (usually used for angular frequency), implying that it's a unit of frequency, rather than a change in amplitude over time. Is that where I'm going wrong?

dmlandrum said:

What you're saying makes perfect sense, except for one thing: our book associates nepers per second with omega (usually used for angular frequency), implying that it's a unit of frequency, rather than a change in amplitude over time. Is that where I'm going wrong?

Click to expand...

What book is it? can you post a fragment?

dmlandrum said:

As I slowly move my way through all of my EE classes, I keep running into things that seem to come out of left field. The first one was the unit "nepers per second", which apparently is just another way of saying "radians per second." I swear someone makes this stuff up just to torture undergrads.

Click to expand...

I've done some digging and I've come up with an explanation of the Np/s, which is the unit for the damping attenuation alpha in a resonating system. A much more useful parameter is the damping factor =alpha/omega zero, which unit is the Neper.
I must admit that, in 40 years working with resonant circuits, I don't remeber having had the need to put a unit on a damping factor!

> dot is the start of the winding (closest to the core); if you're a human being , the dot is the "hot" point, the end of winding.

Whichever, but when used to indicate _relative_ phase, start/end is meaningfull only if you also specify that all windings were wound in the same direction.

On a one-bobbin E-I core, they are wound the same because that's easiest for the winder.

Multiple bobbins or other core geometry can have same-direction come out opposte polarity.

Okay, I think I get the idea now.

I think part of the issue I have with understanding stuff that should, quite frankly, be easy for me is the fact that our teacher this semester is actually a TA who got thrown in the deep end after the regular professor got injured and ended up with blood clots in his leg. It's better than getting the whole class cancelled, but all she does is just read out of the book, and when you challenge her with a question, she has a lot trouble providing good explanations extemporaneously.

She's still better than my DSP professor, though. I might tell you all that story sometime.

Anyway, thanks a lot for the help!

> our teacher this semester is actually a TA

You should have a Study Group. Different students "get" different parts of the lesson. Pooling insights helps.