Anybody using a SFX40 filter?

[MindBender]

If 'yes', how is it working for you?

I'm having my fuse box replaced next week, and getting in that SFX40 filter that's been laying around here the past 7 years may be a bit harder than expected; All my extra wishes are making it a real tight fit. I've been doing fine without it for 7 years, but I would have felt better during these years if it were installed. Now to determine if it actually makes a difference, I have attached it to my signal analyzer and I came to the rather shocking measurement below:

Xanura SFX40 filter spectral diagram

SFX40 @120kHz.jpg (124.83 KiB) Viewed 6118 times

This diagram shows a the spectrum of the SFX40 filter under test. The left-hand side represents 12kHz, the right-hand side 1.2MHz, everything in between is scaled logarithmically. The green trace represents the input signal to the filter, the yellow trace the output signal from the filter. The red vertical line indicates 120kHz, where the X10 carrier signal is. The difference between the green and yellow line show the actual attenuation of the filter at the given frequency. I have put markers on both traces and the difference in printed in yellow: At 120kHz the SFX40 filter attenuates an underwhelming -0.24dB. That's waaaay too little (I feel the need to add more a's here, but I won't) to be of any use.

But at 2kHz and further, the output signal takes a dive, so I figured I'd see where it goes:

Xanura SFX40 filter dip

SFX40 @5MHz.jpg (122.51 KiB) Viewed 6118 times

I have to give it to them: It goes deep. Impressively deep even. But at 5.030MHz a whopping attenuation of -22.08dB is completely useless to block X10 signals at 120kHz.

I wonder how much engineering and testing went into this filter. Perhaps it's not possible to design a filter capable of handling 40Ampere that blocks 120kHz at a decent rate without introducing huge amounts of 'blind current'. But it is useless the way it is now, and a simple test would have proven that during development already. So I really wonder why it's on the market.

[rances]

You didn't have any problems in your network at all for 7 years? I've installed 3 of them in my fuse box (3 phase system), but honestly I have no clue if it was needed at all. I did it because the manual said so.

I don't know if it makes any sense, but a simple guess might be that they wanted to block other noise, not only x10 signals from outside your house, keeping your network clean.

I wonder how FD10 is doing it compared to sfx40.

[MindBender]

You didn't have any problems in your network at all for 7 years? I've installed 3 of them in my fuse box (3 phase system), but honestly I have no clue if it was needed at all. I did it because the manual said so.

Same here; It's recommended, it's paid for, so it's going in

I don't know if it makes any sense, but a simple guess might be that they wanted to block other noise, not only x10 signals from outside your house, keeping your network clean.

Perhaps, but I doubt that signals over 1MHz carry so far they need to be filtered out.

I wonder how FD10 is doing it compared to sfx40.

Bring one over and I can show you.

[MindBender]

I have just pulled an AFX2 filter and measured it. It's not doing much better.

My AFX2 filters are modified, mainly because they didn't do anything at all. Internally they seem pretty standard common mode filters. But both windings have been put in parallel, and together they are in series with the load. My modification is nothing more than putting both windings in series. I doubt if it can still bear a 16A load with this modification, but at least it's improving X10 reliability near specific equipment, so it must be doing something.

Unfortunately I don't have any standard AFX2 filters laying around. And after spending a whole saturday figuring out how to connect my signal analyzer to my network, I felt too lazy to un-modify it for these measurements.

Xanura AFX2 filter spectral diagram

Modifed AFX2 @120kHz.png (28.84 KiB) Viewed 6077 times

Again the spectrum of a modified AFX2 filter under test. The left-hand side represents 12kHz, the right-hand side 1.2MHz, everything in between is scaled logarithmically. The green trace represents the input signal to the filter, the yellow trace the output signal from the filter. The red vertical line indicates 120kHz, where the X10 carrier signal is. The difference between the green and yellow line show the actual attenuation of the filter at the given frequency. I have put markers on both traces and the difference in printed in yellow: At 120kHz the modified AFX2 filter attenuates -1.66dB. That's not much, but enough to make a difference. I'm assumed here that an unmodified AFX2 does about half of this, around -0.85dB. Not a lot, but it could be enough in some case. In reality it's a bit more complicated than dividing it in two, but I think it's a good estimate.

And again a sweep to see how deep the rabbit hole goes:

Xanura AFX2 filter dip

Modified AFX2 @5MHz.png (27.2 KiB) Viewed 6077 times

Again the filter dips around 5MHz: -26,56dB @5,029MHz.