I've been reading about the fact that a bit of DC on the mains can cause toroidal transformers to hum more than normal. So I decided to try to measure it I've got no toroidal transformers I'm just interested to try to measure it. Any how I ran a 150 watt bulb via a half way rectifier and got around 480 millivolts DC from neutral to earth and around 10 millivolts live to earth and live to neutral I then reversed the diode but still the voltage was only from neutral to earth why would this be? Incidentally with the diode the other way round the earth was 480 millivolts POSITIVE to neutral. The way I measured it was with my best quality meter ( the true RMS 6000 count one) switched to DC Volts range. My other meters couldn't make sense of anything when connected to the test terminals on DC range with AC mains present. Why could I only measure the offset on the neutral side? Surely it shoul be present on both sides.
The simplest low pass filter is basically is a series R and a C to ground, such that the RC time constant is long compared ot the period of the mains, and then the cap charges to the mean DC value, with a ripple that is attenuated. Clearly the resistor and capacitor need to be suitable for the votages (both AC and DC ) that you expect.
If your DC meter is a digital one that has say a megohm input impedance, then you can afford perhaps 100K in series, as this will push the DC reading off by ~10%
If we take that 100K as an example the capacitor value needed to reduce 720V p-p (==230V RMS) to say 1`% of its original value (7.2Vp-p or 2.3V RMS) would be 1000j ohms, or at 50Hz, a touch over 3uF. (the reactance of the capacitor Xc is 1/(2pi FC) )
You would get greater suppression but more phase shift splitting the 100k in half into a pair of 47K, and using a paiir of 1uf capacitors to create an R-C-R-C network, to get two lots of 47K against 3k (15 squared is more than 100....)
Take care how you do this - if some how you end up with rectified mains on the caps, the energy stored may be heart stopping.
As a useful rule of thumb about when to start to worry, in the lab we treat anything where (CV2)/2 is more than about 1 joule as potentially deadly, and require resistors for automatic discharge. Above 10 joules we require multiple automatic discharge paths, so several things go wrong before we reach a dangerous situation.
The simplest low pass filter is basically is a series R and a C to ground, such that the RC time constant is long compared ot the period of the mains, and then the cap charges to the mean DC value, with a ripple that is attenuated. Clearly the resistor and capacitor need to be suitable for the votages (both AC and DC ) that you expect.
If your DC meter is a digital one that has say a megohm input impedance, then you can afford perhaps 100K in series, as this will push the DC reading off by ~10%
If we take that 100K as an example the capacitor value needed to reduce 720V p-p (==230V RMS) to say 1`% of its original value (7.2Vp-p or 2.3V RMS) would be 1000j ohms, or at 50Hz, a touch over 3uF. (the reactance of the capacitor Xc is 1/(2pi FC) )
You would get greater suppression but more phase shift splitting the 100k in half into a pair of 47K, and using a paiir of 1uf capacitors to create an R-C-R-C network, to get two lots of 47K against 3k (15 squared is more than 100....)
Take care how you do this - if some how you end up with rectified mains on the caps, the energy stored may be heart stopping.
As a useful rule of thumb about when to start to worry, in the lab we treat anything where (CV2)/2 is more than about 1 joule as potentially deadly, and require resistors for automatic discharge. Above 10 joules we require multiple automatic discharge paths, so several things go wrong before we reach a dangerous situation.
