Testing Shielding Effectiveness of Coaxial Cables: Using a comb generator

[@kavinduratnayake]

Thanks for the video Can you also teach the theory behind the test please

[@MrThedatastream]

You've got a good strong Digital TV signal where you live Min!

[@MachOneDesignEMC]

haha James, indeed, just got myself a nice radio receiver, I will see if I can listen to all the ambient signals in my house.

[@RogierYou]

Can you test various shielded Cat 6 and 7 cables?

[@MachOneDesignEMC]

you can certainly use this method, we can look into testing CAT 6 & 7 cables in the future.

[@reinerfranke5436]

If i not misunderstud the effectiveness means against magneto static field as the wavelenght is clearly longer than the coupling structure. In the beginning you refer to an electric current. Do you have a clou why there is a difference? Is that because of the different impedance transitions at the end? I assume that that they are either open or terminated. Do not remember if you state this condition clear.

[@MachOneDesignEMC]

Hi Reine, See my reply to Pitti, you got me think hard!

[@pitti2602]

I was with you on the test with the ESD gun, but I think here you are wrong and I think you are actually measuring the differences in wave impedance either along the cable or in relation to the termination - and not the shielding effectiveness. If you were to put the cables on a well grounded (to the shield of the comb gen NOT the termination) metal plane then I might believe you.

[@MachOneDesignEMC]

Hi Pitti, again, thanks for your constructive feedback. This really got me scratch my head. The way I see it, first, shielding effectiveness is a wrong word to use in this case, okay, I might have used the word simply to attract more viewers, but I think a better wording for this should be the quality of the co-axial cable. Now moving to your question, it got me thinking hard. Besides shielding quality, it is also cable impedance/mismatch that would influence the result. I could try squeezing the coaxial cable with a clamp or vise, thereby creating a discontinuity of the cable impedance (perhaps somewhere halfway of the cable, if we see a significant difference in the measured spectrum amplitude, mismatch would be important. if not, the result would be determined by the shielding quality. What do you think? Could make a follow up video on it, but I doubt many people would care....

[@pitti2602]

​@@MachOneDesignEMC Sorry did not notice you replied. I'd go with quality, but for me the question is, why do you actually measure current there. From my perspective this can only be CM current and then the question is where does it come from and does the length of the cable affect it? So in my opinion there might be tiny openings in the shield which provide leakage for a displacement current - and while that would be an indicator for the shielding performance one can also amplify this effect by providing a metal plate (good return path) in close proximity like I mentioned. As for the reflections or impedance mismatches you could also experiment with different cable lenghts and clamp positions. Will do some further thinking 😅

[@pitti2602]

Ok .. no new conclusion yet but more ideas: -your comb gen generates a 600ps..1ns long pulse every 1/20MHz (50ns) -in 1ns the pulse travels 20cm (~2/3 c) -it looks like you short cables are 10cm long so parts of the pulse might cancel each other, but there is atleast only one pulse on a perfect cable -if there is impedance mismatch the height of pulse will drop upon reflection, so sth should be visible at the base frequency of 20MHz - however you probably use peak detector so this will not matter IF you measure at a position where the fwd and bwd part of the pulse cannot be at the same time -however you also do not know the exact output inpedance of the comb gen and how well it matches the cable - but probably its ~47Ohm and should not create a big difference So my suggestion would be: -at least 30cm cables -measure at 5..10cm distance from gen output -reflection should not matter due to use of pk detector (here i'd currently have to correct myself, but verify at 20MHz) -still use a metal reference plane at also be more careful on the distances between cables and clamb and positions of items in the setup 😅😅😅