EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: bd139 on November 16, 2017, 01:44:30 pm

Title: X2 capacitor across line - what defines suitable capacitance?
Post by: bd139 on November 16, 2017, 01:44:30 pm
I've got an HP E3630 here with a 100nF X2 cap across the line. A good old RIFA of death. So it's being replaced. No one has any stock of exactly the replacement part.

Ergo, I'm looking at substitutions. What defines a suitable capacitance in these circumstances?

Quick analysis of what's in there: 100nF, I assume that as these are across line that the reactance at 50Hz is ~31K. Yet reactance at say 20KHz is 80R so it'll just sink any HF interference across the line?

But if I stick a 150nF in, reactance at 50Hz is ~21K. Yet reactance at 20KHz is 53R. Is that 50Hz reactance too low? What's the limit or suitable range here?

Edit: I looked in AoE but there's no discussion of suitable selection.
Title: Re: X2 capacitor across line - what defines suitable capacitance?
Post by: macboy on November 16, 2017, 03:09:39 pm
It is just an interference suppression capacitor, so don't worry too much about the exact value. I've seen anywhere from 47 nF to 2.2 uF (which is a rather large box). Make sure it is in fact X2 rated, and the appropriate voltage as well. Find something with the right lead pitch so that it mounts to the board neatly and safely.

The little bit of capacitive reactance at 50 Hz is not a bad thing, it helps counteract the inductive reactance of the transformer, usually giving a slightly better overall power factor.
Title: Re: X2 capacitor across line - what defines suitable capacitance?
Post by: bd139 on November 16, 2017, 03:16:25 pm
Cool thanks for info. Priority was X2 rated and 20.3mm pitch so  :-+
Title: Re: X2 capacitor across line - what defines suitable capacitance?
Post by: T3sl4co1l on November 16, 2017, 08:24:22 pm
I'd say, within a factor of 2.

Line filters are not usually very carefully constructed, so I would be surprised if this level of error would cause problems.

More change, and you're really starting to fundamentally change the filter.  Large values can have other consequences, like inrush current and reactive power.

Tim