Y'all have some serious mains voltages over there. But, the 10% variation makes me wonder how much of that is from load variations on the power line, and how those loads effect the purity of mains delivered to you, and how much of that mains noise is coupled through the various amplifier power supplies to amplifier signal path. Sub-100Hz junk seems well within the passband at first blush. To the extent that this might matter, the amplifier's mains transformers are designed for 50Hz, and power supply filtering is designed for 100Hz and up (in your time zone). Many mains noise sources have components down to and including DC.
A definitive test would be (at your friend's house) a comparison between "with vs. without" a resonant transformer upstream. If you have access to one, this is easy; if not..... well, probably a red herring. Maybe interesting if you don't find anything else easier.
All good fortune,
Chris
Thanks Chris. Most of the variation here is caused by roof-top solar. Solar here is now the largest generator in the state...when the sun is shining.
I keep a power metre on the mains so I can monitor the voltage whenever I'm doing measurements. On a cloudy day you can watch the voltage going up and down!
I keep a power metre on the mains so I can monitor the voltage whenever I'm doing measurements. On a cloudy day you can watch the voltage going up and down!
You live in an enlightened place. Let's hope we all get there before it's too late.
Along this same tangent, is it even possible for a small-ish (less than 10,000?) number of binary (on or off) uncorrelated noise sources (my imaginary model of the relevant solar generators in parallel) to look like 1/f ? I'm not smart enough to know, but at least one of the posters here might venture an opinion. I see it as a problem in maths, rather than something requiring knowledge of 1/f, which is in the category of dark energy, or a systemic medical symptom: a description of an observation, mechanism currently unknown.
All good fortune,
Chris
Along this same tangent, is it even possible for a small-ish (less than 10,000?) number of binary (on or off) uncorrelated noise sources (my imaginary model of the relevant solar generators in parallel) to look like 1/f ? I'm not smart enough to know, but at least one of the posters here might venture an opinion. I see it as a problem in maths, rather than something requiring knowledge of 1/f, which is in the category of dark energy, or a systemic medical symptom: a description of an observation, mechanism currently unknown.
All good fortune,
Chris
The people are enlightened. The government not so much. They were told 15 years ago that they'd need to plan for this day and they did nothing. They've just implemented a new requirement that inverters need to include a facility to be shut off remotely when there's too much solar!
I don't know about the noise profile of the mains but my fellow retirees who spent they're careers in the power industry scoff at any suggestion that the mains is particularly sinusoidal.
I don't know about the noise profile of the mains but my fellow retirees who spent they're careers in the power industry scoff at any suggestion that the mains is particularly sinusoidal.
Hear, hear!
I don't know how to calculate it either, but I vaguely remember having read something about popcorn noise and a 1/f spectrum. What you describe is essentially a lot of sources of popcorn noise added together, I think that resulted in 1/f somehow.
To get into the amplifier, it either has to go via the heaters and get filtered by the heaters' thermal time constant, or via the supply with all its LC and RC filters. The spectrum at the amplifier output would probably roll off faster than -10 dB/decade then.
Last edited: