Elements between PCB-GND and Mains Earth in SolidState Power Amps - How to Calculate?
Well, the best way it is, to avoid the use of the mains earth and thus this additional components.
In case of dual monaural constructions the RCA input connector is now mounted without isolation to the chassis envelope.
Unfortunately most transformers require mandatory the use of the mains earth for safety reasons.
This means at the same time, I have to introduce components between the metal chassis envelope and the GND from main board/star GND from power supply (to avoid double grounding).
I observe a wide range of different ways in commercial amp products - each manufacturer chooses other components resp. values resp. topologies.
Sometimes only a resistor (10 ohms until 1000 ohms) is present, somtimes only a MKT cap (10 nF until 1 uF), and sometimes both of them (the attached gainclone example with LM3875 e. g. uses a 100 ohms resistor and a cap with 0,22uF).
Also the wattage and the kind of the cap are highly variable (very small until very large).
The choosen position at the chassis and at the PCB are also different. Sometimes near the input GND, sometimes near the output (loudspeaker-terminals) and sometimes at the power caps (star ground).
Also, the thickness and the length of the connecting cables between the main-PCB GND and the chassis (over the mentioned components) are very different. In any cases I did observe an extremely thin wire.
Follow question rises up:
How did I calculate this devices between chassis/mains earth and secundary GND for lowest possible hum (highest possible signal to noise ratio) so as highest possible reliability and where are to find papers and test reports in this matter?
Rod Elliot has an article about this. Basically it's like this:
The power resistor should be large enough to prevent currents flowing that would cause a ground loop hum, but small enough that if there is a primary-secondary short, enough current will flow to mains earth to blow fuse/trip breaker. It should be of sufficient wattage that it will survive a direct 240v to earth short for long enough to ensure the fuse/breaker trips.
The capacitor is seen as a short at high frequencies, for RFI purposes.
Generally, 10 Ohm/5W resistor and 100nF 275V X2 rated capacitor do the job. If the transformer is an EI type, or has an internal electrostatic shield, it should be connected directly to mains earth.
There's a few variants on this. Rod shows the use of a 35A rectifier connected as back-to-back diodes between mains earth and 0V. Nelson Pass has used a 10 ohm NTC thermistor instead of the 5W power resistor. A lot of amps dont even have a loop breaker - they just connect 0V to mains earth.
mean you this article?
Earthing (Grounding) Your Hi-Fi - Tricks and Techniques
Yep, that and his "Power supply wiring" article pretty much covers it :)
The pic shows the Audio Ground connected to the Safety Earth via a Disconnecting Network of 100r//0u22F.
TEST that arrangement.
1.) the Disconnecting Network blows up.
2.) that the mains fuse may remain intact.
3.) that if the mains fuse remains intact that the potentially fatal result will be MAINS voltage on the Chassis.
Tief, you have made a grave mistake placing that dangerous schematic on the Forum where every Member can see it.
Please ask to have the dangerously wired pic removed before someone gets killed.
If a Disconnecting Network is substituted for that direct wire connection then That DN must survive up to a few kA of fault current.
If you are not sure then TEST it.
I have and I know that the DN in the ESP site which is basically copied from this Forum works. Beware there is an ambiguity in the ESP schematic.
He shows the PE connected to the DN and then going to Chassis.
That pic is wrong !!!!
The PE goes direct to Chassis.
The Audio Ground goes to Chassis either direct or via a DN. This Chassis connection can be many inches, cms away from the Safety Earth where the PE connected to Chassis.
The DN must use inverse parallel diodes to pass the Fault current from Audio Ground to Chassis. The other components in parallel to the high power diodes are all optional.
These can be
a.) a switch, of low current capability
b.) a resistor, of low power dissipation.
c.) a capacitor, of low power/voltage
d.) an NTC, of low power.
I have not tested a high power NTC as the main fault current carrying element.
It may be that a High Power NTC can be used instead of the inverse parallel High power Diodes.
Would some Member/s like to carry out thorough safety testing of this NTC substitution for 110/120Vac and for 220/240Vac mains supply systems?
"Please ask to have the dangerously wired pic removed before someone gets killed. "
What's your message DJ?
|All times are GMT. The time now is 09:51 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio