The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Skin effect on DC transient?
https://designers-guide.org/forum/YaBB.pl?num=1234001690

Message started by 7-7 on Feb 7^th, 2009, 2:14am

Title: Skin effect on DC transient?
Post by 7-7 on Feb 7^th, 2009, 2:14am

Hello everyone.

I would be grateful if anyone could answer my following query.

Skin effect limits a conductor's cross-sectional area through which an AC current can flow. This has an effect on the resistance of that conductor to the AC current.

Is a DC transient also restricted to flow into a partial region of the cross-sectional area of a conductor, or does a DC transient, no matter how steep, flow through all the area of a conductor, like a steady DC current?

Many thanks.

Title: Re: Skin effect on DC transient?
Post by pancho_hideboo on Feb 7^th, 2009, 4:03am

7-7 wrote on Feb 7^th, 2009, 2:14am:

Is a DC transient also restricted to flow into a partial region of the cross-sectional area of a conductor,
or does a DC transient, no matter how steep, flow through all the area of a conductor, like a steady DC current ?

I can't underestand what you mean by "DC transient".
Also I can't understand what you mean by "like a steady DC current".

"Skin effect" is a real physical phenomena.
If frequency components of current are high, "Skin effect" is appeared for these frequency components.
But low frequency components don't undergo "Skin effect" as far as a conductor is linear.

Title: Re: Skin effect on DC transient?
Post by 7-7 on Feb 7^th, 2009, 7:24am

Pancho_hideboo thank you for replying.

By DC transient I mean a steep increase in current flow, say from 2 A to 1000 A in say 10^-3 sec, but with the current flowing always in the same direction.

By steady DC current I mean that the rate of change of current is zero i.e. there is no current change with respect to time.

I am aware of the skin effect for a current whose direction of flow changes alternatively i.e. an AC current. I would like to know if a similar effect applies to a DC transient, which was described above.

Thank you.

Title: Re: Skin effect on DC transient?
Post by pancho_hideboo on Feb 7^th, 2009, 8:56am

7-7 wrote on Feb 7^th, 2009, 7:24am:

By DC transient I mean a steep increase in current flow, say from 2 A to 1000 A in say 10^-3 sec, but with the current flowing always in the same direction.

Response is similar to LPF output driven by step input signal, since frequency characteristics of resistor is proportional to sqrt(freq).

Title: Re: Skin effect on DC transient?
Post by 7-7 on Feb 8^th, 2009, 1:43am

I take your answer to mean that yes, skin effect conditions apply to a DC transient. I take it also to mean that I would have to do the harmonic analysis of the step slope. Could you please confirm if this is correct.

In that case, how do I do the harmonic analysis of the step of my previous example, from 2 A to 1000 A in 10^-3 sec ? Could you give me the full procedure and result?

Many thanks for your help.

Title: Re: Skin effect on DC transient?
Post by pancho_hideboo on Feb 8^th, 2009, 1:56am

7-7 wrote on Feb 8^th, 2009, 1:43am:

I take it also to mean that I would have to do the harmonic analysis of the step slope.

What do you mean by "harmonic analysis" ?

7-7 wrote on Feb 8^th, 2009, 1:43am:

In that case, how do I do the harmonic analysis of the step of my previous example, from 2 A to 1000 A in 10^-3 sec ? Could you give me the full procedure and result ?

Generally use Transient Analysis. But about details of procedure, it depends on what simulator engine you have.
For example, if you approximate step signal as periodical pulse signal with long period, you can use Harmonic Balance Analysis with pure frequency domain description of skin effect for model of line.

See TLINP (2-Terminal Physical Transmission Line) of chap-7 in the following.
http://cp.literature.agilent.com/litweb/pdf/ads2005a/pdf/ccdist.pdf
http://www.designers-guide.org/Forum/YaBB.pl?num=1227194040

If you use Agilent ADS, you can use various line models such as "TLINP" in Transient Analysis.
If you use Synopsys HSPICE, you can use "W-Element" in Transient Analysis.
If you use Cadence Spectre, you can use "mtline" in Transient Analysis.

Also you can make original line model with skin effect.
http://www.designers-guide.org/Modeling/ind.pdf
http://www.designers-guide.org/Modeling/suite.pdf

Title: Re: Skin effect on DC transient?
Post by 7-7 on Feb 8^th, 2009, 4:38am

I'll rephrase it:

Am I correct in saying that I should find the constituent sine waves of the transient ramp by Fourier transform, then calculate the skin effect on each sine wave, sum the sine waves together again, and the net result, as you say, is a low pass filter on the original ramp?

If I am not correct please stop me here.

Title: Re: Skin effect on DC transient?
Post by pancho_hideboo on Feb 8^th, 2009, 5:39am

7-7 wrote on Feb 8^th, 2009, 4:38am:

I should find the constituent sine waves of the transient ramp by Fourier transform, then calculate the skin effect on each sine wave, sum the sine waves together again, and the net result, as you say, is a low pass filter on the original ramp?

Conceptually correct.
But you can do this task by using AC Analysis or Harmonic Balance Analysis.

What was your question after all ?

If you want to know response of current flow output for line having skin effect, just do Transient Analysis by using line model having skin effect and step signal drive.

I can't understand what you are confusing so much.

http://sigcon.com/Pubs/edn/MixSkinandDielectric.htm

Title: Re: Skin effect on DC transient?
Post by 7-7 on Feb 8^th, 2009, 6:24am

pancho_hideboo wrote on Feb 8^th, 2009, 5:39am:

[quote author=icodar link=1234001690/0#6 date=1234096724]
What was your question after all ?

You just answered it!

Thanks.

Title: Re: Skin effect on DC transient?
Post by pancho_hideboo on Feb 8^th, 2009, 7:16am

7-7 wrote on Feb 8^th, 2009, 6:24am:

You just answered it!

After all your question was how to calculate a response of linear system having transfer function H(freq) when it is driven by step signal. Very easy issue.

http://www.designers-guide.org/Forum/YaBB.pl?num=1234001690/0#0
From first your explanation, I couldn't understand what was your question at all.