https://www.diyaudio.com/forums/solid-state/2080-dc-filter-6.html#post331815https://www.diyaudio.com/forums/solid-state/2080-dc-filter-6.html#post331815
22nd February 2004, 08:09 PM #57 Eva diyAudio Member
Join Date: Oct 2003 Location: Near the sea
I've done some measurements that show what's really hapening when you connect a transformer to mains
Using a lowpass filter [100k + 47uF] I have measured about 50mV average over time of DC on my mains supply
I also have an old electric heater that in half-power mode uses a diode in series with the heat element to pass only half of the mains waveform. When I plug this heater in half-power mode I get an additional 1V of offset on mains supply
To test the need and the efficiency of DC filtering, I've done some measuremens of the current through the primary of a 750VA toroidal transformer
This oscillogram shows what happens when I connect the transformer to mains and let it deal with the 50mV DC offset
An externally hosted image should be here but it no longer works. Please upload images instead of linking to them to prevent this.
Blue trace is mains waveform at 100V/div [230V AC], it looks more like a clipped triangle wave instead a sine wave due to the line inductance limiting the slew rate and all the rectifiying applications consuming all the current only during the peaks [30% of total time]
Red trace is the current through the primary at 200mA/div. Transformer saturation towards the negative side is evident, reaching 350mA peak of leakage current. The transformer buzzs slightly due to the saturation
The noise present in the current waveform is common mode and was suppressed in further measurements adding a common mode filter between mains and measurement point
The second oscillogram shows what happens when I plug the electric heater in half-power mode
An externally hosted image should be here but it no longer works. Please upload images instead of linking to them to prevent this.
Red trace this time is in 2A/div so the leakage peak current exceeds 6A. The transformer is heavily saturated towards the upper side and buzzs loudly.
The third oscillogram shows what happened when I placed a DC filter consisting of two 1000uF 16V and some diodes in series with the primary
An externally hosted image should be here but it no longer works. Please upload images instead of linking to them to prevent this.
This time, red trace is 20mA/div and shows the small leakage current due to both the magnetizing inductance and the parasitistic capacitance between adjacent turns. The transformer is no longer saturated and performs silently
the fourth oscillogram shows the induced voltage in a loop of wire of 10cm diameter placed vertically, paralell and 1cm away from the transformer [placed horizontally] obtained when the transformer was saturating with more than 6A peak [with the electric heater plugged]
An externally hosted image should be here but it no longer works. Please upload images instead of linking to them to prevent this.
The red trace is 2mV div and shows the induced voltage in the loop of wire
That measurement demonstrates that when a 50-60Hz transformer is saturating, it produces electro-magnetic-interferences that induce noise voltages on everything in the nearhood
Actually, I think that +-5mV of low frequencies induced in a loop of wire of 10cm diameter near the transformer is a serious thing since this EMI is at audio frequencies, it's not RF so it's 100% audible and it may be happening in all your transformers
In the other hand, all the tests were performed with open secondaries but if we add load so that primary current has peaks of 6A, then the EMI radiated would be the same or higher
With load, the EMI is produced due to flux in the leakage inductance, resonances due to parasitistic inter-turn capacitance, RF ringing due to diode turn-off characteristics and the fact that the peak current through the transformer is 3 times or more the average DC current after rectification
In conclusion : Rectifiying the output of 50-60Hz transformers produces EMI as any SMPS does and this phenomena gets aggravated when the transformer is saturating due to direct connection to mains without a DC filter
50-60Hz transformers are nothing but big, bulky and crappy antennas
.
22nd May 2004, 05:20 PM #71 Eva diyAudio Member
Join Date: Oct 2003 Location: Near the sea
Bricolo :
DC levels on mains cannot be measured by directly connecting the multimeter to the line. You have to place a low-pass filter between the line and the DMM. I use a first order filter made of a 100K 1W resistor and a 22uF non-polar capacitor for that purpose
Non-gapped transformers must be operated on 'pure AC'. Even 50mV of DC may be enough to see signs of saturation on toroids
.