Twisted Pair (EMI Cancellation)

Thread Starter

AutoNub

Joined Oct 14, 2011
44
Hello,

I am putting together a wiring harness schematic for outsourcing to reduce company labor costs, and we adhere to strict electromagnetic radiation standards. One of the techniques we use to reduce EMI is twisted pairing for cancellation.

I'm curious, though, is there a standard for the ratio of twists per some measure of length? For example, 2 twists (two complete loops) per inch? How much of a difference does the frequency of twisting have on the EMI? Can you twist too much? Any details would be appreciated.

Thank you!
 

t06afre

Joined May 11, 2009
5,936
Unless the signal is truly diferential. I doubt you will see any improvments in radiation from the cable.
 

Thread Starter

AutoNub

Joined Oct 14, 2011
44
All of our products have to pass MIL-STD-461 and 462 testing. We know twisted pairs improve our test results by reducing EMIs, but for the sake of outsourcing the construction of our wiring harnesses, I'm wondering about the amount of twisting necessary to achieve optimum results.
 

#12

Joined Nov 30, 2010
18,076
I think I saw, "5 turns per foot" in reference to Cat5 cable.
 

Thread Starter

AutoNub

Joined Oct 14, 2011
44
Thanks for the information #12. One thing I'm wondering is if there is a linear relationship between the number of twists per unit of length. For example, if you have twice as many twists per inch, would the EMI be reduced by twice as much? I doubt it is exactly linear, but I'd be surprised if someone didn't actually put together some calculations and/or measurements and maybe a graph of some kind showing these relationships.
 

#12

Joined Nov 30, 2010
18,076
I am not equipped to measure EMI, but apparently you are. Set up a test and do the work. The results will probably be about the law of diminishing returns and best measured in decibels.
 

Thread Starter

AutoNub

Joined Oct 14, 2011
44
We do have a non-certified anechoic chamber for our preliminary testing prior to sending out for official testing, which I could use to conduct some experiments and compile some data to establish some relationships specifically addressing twisted pairs. I was simply wondering what data was already available on the subject. Posting on forums is usually an effective strategy to supplement my learning.

Thanks for your time.
 

wmodavis

Joined Oct 23, 2010
739
I just googled "twisted pair emi reduction" and found lots of references. You'll have to sort through to find what if any are helpful to you.

Remember - GOOGLE is your friend!
 

Thread Starter

AutoNub

Joined Oct 14, 2011
44
I just googled "twisted pair emi reduction" and found lots of references. You'll have to sort through to find what if any are helpful to you.

Remember - GOOGLE is your friend!

I've been google surfing since about an hour prior to my creation of this thread. I have found lots of interesting material, but nothing quite satisfactory yet. I'm thinking it is time to close this thread ;).

Thanks again.
 
Top