I see buzz in spectrum even when PSU is turned off(!) but is connected to power outlet. Buzz disappears when PSU is unplugged from wall outlet.
Outlet is ~120V
PSU is for tube preamp HT +300V, built on toroidal transformer.
HT is rectified by diode bridge with snubbers and smoothed with 5 RC stages 270uF, 1K.
Heaters are feed from separate PSU (switch mode). It is also turned off.
Power switch disconnects both wires, the hot and the neutral, but does not disconnect the Earth.
Earth is connected to PSU chassis (alum box) and further is passed through umbilical cord to preamp chassis (alum box).
Earth is connected with the ground only in one point at input XLR of the preamp to avoid loops.
Where the buzz comes from? Is it from outlet Earth? How to eliminate it? I would not want to remove connection to Earth because of safety reasons.
Outlet is ~120V
PSU is for tube preamp HT +300V, built on toroidal transformer.
HT is rectified by diode bridge with snubbers and smoothed with 5 RC stages 270uF, 1K.
Heaters are feed from separate PSU (switch mode). It is also turned off.
Power switch disconnects both wires, the hot and the neutral, but does not disconnect the Earth.
Earth is connected to PSU chassis (alum box) and further is passed through umbilical cord to preamp chassis (alum box).
Earth is connected with the ground only in one point at input XLR of the preamp to avoid loops.
Where the buzz comes from? Is it from outlet Earth? How to eliminate it? I would not want to remove connection to Earth because of safety reasons.
Attachments
probably SMPS breaking through after power Off.
The SMPS runs from the charge on the very large input capacitor. Usually 400V or a pair of series connected 200V.
These/this capacitor holds an enormous amount of charge that keeps the SMPS running after shut down.
The SMPS runs from the charge on the very large input capacitor. Usually 400V or a pair of series connected 200V.
These/this capacitor holds an enormous amount of charge that keeps the SMPS running after shut down.
I think that you have either a test setup problem or are seeing what electricians call phantom voltage. (that's when a high impedance DMM picks up a voltage reading on an unconnected wire).
for reference here is schematics of PSU and grounding layout schematics of the preamp.
Also, output goes to computer that is connected to UPS that is connected to one outlet. Preamp is connected to another outlet.
I am not sure how these outlets are related.
I tried to plug PSU into UPS, result is even worse. Noise floor is 20dB higher, and lots of 60x harmonics.
Also, output goes to computer that is connected to UPS that is connected to one outlet. Preamp is connected to another outlet.
I am not sure how these outlets are related.
I tried to plug PSU into UPS, result is even worse. Noise floor is 20dB higher, and lots of 60x harmonics.
Attachments
Disconnect all the wires from XLR pin #1. Now connect pin # 1 to the chassis near the jack. Create a Main Audio Ground point near the TRS jack. Connect all those other wires to the MAG, then connect the MAG to the chassis near the TRS.
My guess is some sort of ground loop. You're measuring this with a computer of some sort. If it has a ground connection then you can easily create a loop that way.
Install a ground loop breaker and see if that solves it.
Earthing (Grounding) Your Hi-Fi - Tricks and Techniques
Without those installed in my equipment I cannot measure a thing with the PC otherwise ground loops just destroy everything.
Install a ground loop breaker and see if that solves it.
Earthing (Grounding) Your Hi-Fi - Tricks and Techniques
Without those installed in my equipment I cannot measure a thing with the PC otherwise ground loops just destroy everything.
I read somewhere that MAG should be located at most sensitive input. I did not mentioned, it is ribbon Mic preamp, very sensitive. That is why I've placed MAG at input XLR.
Why MAG needs to be near the TRS jack?
Thanks, very helpful article. I currently do not have 35A bridge. Will 8A one be sufficient for proof of concept? There is 2.5A 250V slow fuse on hot mains wire. I suppose it should open first.
Also, there is an error on 1st diagram. The hot mains Fuse is before the Switch1.
I tested the doubled diode version of the Disconnecting Network, i.e. shorting diagonally across both ways, ~ to ~ and + to -
This gives the 35A bridge rectifier an effective 70A capability.
The peak transient capability is possibly approaching a kA.
The diodes and the other small components survived the test across the Mains.
The fuse holder did not survive the test !
I have not tested any other version.
This gives the 35A bridge rectifier an effective 70A capability.
The peak transient capability is possibly approaching a kA.
The diodes and the other small components survived the test across the Mains.
The fuse holder did not survive the test !
I have not tested any other version.
Forgive me my ignorance, but I do not understand how the Earth Loop Breaker works.
There are two "positive" diodes conduct positive pulses of 60Hz sinusoidal noise, and two other "negative" diodes conduct negative pulses.
From my perspective this connection of diodes is equivalent of a short, just a unidirectional conductor with resistance that is less than R1 10 Ohm. Current avoids R1 and C1 in my SPICE simulations, and diodes conduct ground noise.
Where is a place where loop breaks?
Or the bridge should be connected differently (back-to-back / head-to-head)?
There are two "positive" diodes conduct positive pulses of 60Hz sinusoidal noise, and two other "negative" diodes conduct negative pulses.
From my perspective this connection of diodes is equivalent of a short, just a unidirectional conductor with resistance that is less than R1 10 Ohm. Current avoids R1 and C1 in my SPICE simulations, and diodes conduct ground noise.
Where is a place where loop breaks?
Or the bridge should be connected differently (back-to-back / head-to-head)?
Attachments
Think further, is the above schematic of the input transformer to op-amp input complete? Or is it a simplified version of the op-amp circuit?
It is a schematic of my tube preamp. all tubes and passives are combined into "opamp" triangle for simplicity. Heater's supply enters there as well. However the input transformer is outside.
The input transformer primary does not have the chassis reference any way, wherever the MAG is. I think it is correct design because it makes mic input balanced (though without a center).
Well, I've read about diodes and have understood, that they don't conduct until some "Forward Voltage Vf" is reached.
For some it is 0.7V for others it is 1.2V...
Also, for this Earth Loop Breaker circuit, what would be a recommended bridge?
If voltage in my case 300V RMS. And it may be higher, what diode voltage will be ok? Will be 600V OK? Or it can be less, anyway high voltage will be dropped by another "positive" pair of diodes. So reverse voltage of "negative" diodes will not go too high. Am I right?
the diode bridge starts conducting at around 400mVdc.
If it is asked to pass 50Apk then the forward voltage will rise to maybe around 1000mVdc.
If the Fault current hits around 1kApk then the forward voltage will hopefully not exceed 5000mVdc
you DO NOT need a 600V bridge rectifier. a A 50V PIV will do perfectly well. It's the current rating that is important.
I have used 25A and 35A bridge rectifiers wired in the criss cross shorting as you have shown in post10. this arranges the diodes in parallel to roughly double the current rating.
You have correctly shown the IN and OUT on adjacent terminals of the bridge. Make sure that is the way you actually wire it in.
Now to what the bridge rectifier does.
It replaces the direct wire link between Chassis and Main Audio Ground (MAG).
A direct wire connection is the normal way to do this. It has NOTHING at all to do with Audio. Forget about noise on the Chassis. You are not interested in the Audio. This is about SAFETY and ONLY safety.
This is the Disconnecting Network (DN) and it can be supplemented with some optional components. Try adding a low value resistor and/or a low value VHF capacitor and/or a power Thermistor and/or a switch.
These reduce the impedance of the DN to signals that want to return to Earth.
If it is asked to pass 50Apk then the forward voltage will rise to maybe around 1000mVdc.
If the Fault current hits around 1kApk then the forward voltage will hopefully not exceed 5000mVdc
you DO NOT need a 600V bridge rectifier. a A 50V PIV will do perfectly well. It's the current rating that is important.
I have used 25A and 35A bridge rectifiers wired in the criss cross shorting as you have shown in post10. this arranges the diodes in parallel to roughly double the current rating.
You have correctly shown the IN and OUT on adjacent terminals of the bridge. Make sure that is the way you actually wire it in.
Now to what the bridge rectifier does.
It replaces the direct wire link between Chassis and Main Audio Ground (MAG).
A direct wire connection is the normal way to do this. It has NOTHING at all to do with Audio. Forget about noise on the Chassis. You are not interested in the Audio. This is about SAFETY and ONLY safety.
This is the Disconnecting Network (DN) and it can be supplemented with some optional components. Try adding a low value resistor and/or a low value VHF capacitor and/or a power Thermistor and/or a switch.
These reduce the impedance of the DN to signals that want to return to Earth.
Thanks Andrew, it is very clear.
I am reading http://www.diyaudio.com/forums/power-supplies/115698-understanding-star-grounding-10.html
and there is an idea to disconnect MAG from Earth completely, not even via diodes and a resistor. This should be safe. Right?
I am reading http://www.diyaudio.com/forums/power-supplies/115698-understanding-star-grounding-10.html
and there is an idea to disconnect MAG from Earth completely, not even via diodes and a resistor. This should be safe. Right?
if your issues are just at shutdown and none during normal listening,
then a simple muting switch to disconnect output or short the input is
probably all that you need...
how is the hum and noise at normal listening? do you have to
put an ear an inch away from your speakers to hear hum and noise?
then a simple muting switch to disconnect output or short the input is
probably all that you need...
how is the hum and noise at normal listening? do you have to
put an ear an inch away from your speakers to hear hum and noise?
The issue exits during normal work as well. 60Hz and harmonics stick up above noise floor and reach up to -95dB.
- Status
- Not open for further replies.