The chassis must always be earthed - and especially so in any DIY stuff since DIY'ers (and that's me with my DIY hat on also) don't have the money or the wherewithal to undertake the testing and certification a double insulated product demands.
The 0V of the amplifier PSU must either go directly to the chassis earth point (and there is only ONE chassis earth point in order to avoid ground loops) OR it can go through a suitable Ground Lifter.
If the mains is diverted in any way from the live neutral as would be the case if the live or neutral touched the chassis or some part of the amplifier PCB, the GFI (Ground Fault Interrupter) will trip. You only need a few mA of imbalance for this to happen and the GFI usually responds to this type of fault within 2-3 mains cycles.
The GFI is designed to protect against electrocution.
On the mains side, the first thing the live feed has to encounter is a fuse, then a switch and from there to the transformer primary and any mains filtering etc.
The fuse is there for fire protection arising from a potential overload or short within the equipment.
re the ground lifter:- do not use a Ground Lifter unless you are absolutely sure the equipment will be used in tandem with an electrical system that specifies a GFI. These are mandated by law in most advanced economies, but this may not be the case in some other areas - check your local regulations.
Further, I recommend a 35A bridge rectifier for the Ground Lifter job. The surge rating is typically 5-10x the average current rating for a specified number of mains cycles.
re AndrewT's original test that caused the mains fuse to blow:- If the live for example touched the amplifier 0V on the secondary side or to the chassis, two things would be going on. Firstly, the current flowing to Earth (either through a Ground Lifter or directly if there was none) would shoot up into the many 10's of amps, and on a system with low loop resistance on the mains side, potentially 100's of amps. A 'T' type fuse would limit the amount of time this condition could persist to a few mains cycles as it appeared happened in his test. Secondly, the GFI would be detecting a massive imbalance in the live<>neutral currents and it would trip. GFI's typically trip in 40-60ms and would probably go before the mains fuse. I don't recall the details of AndrewT's original test - I don't know if he mentioned if the GFI had tripped, but its likely it would have as well as the fuse popping.
So, this is why the Ground Lifter Diodes/Bridge should be rated to carry the full mains trip current for a few hundred ms - if it popped, the user would reset the mains and now have an appliance where the mains could touch something on the secondary without it tripping the GFI until someone touched it of course, and then the GFI would trip - but, that is still not considered safe or acceptable.
The 0V of the amplifier PSU must either go directly to the chassis earth point (and there is only ONE chassis earth point in order to avoid ground loops) OR it can go through a suitable Ground Lifter.
If the mains is diverted in any way from the live neutral as would be the case if the live or neutral touched the chassis or some part of the amplifier PCB, the GFI (Ground Fault Interrupter) will trip. You only need a few mA of imbalance for this to happen and the GFI usually responds to this type of fault within 2-3 mains cycles.
The GFI is designed to protect against electrocution.
On the mains side, the first thing the live feed has to encounter is a fuse, then a switch and from there to the transformer primary and any mains filtering etc.
The fuse is there for fire protection arising from a potential overload or short within the equipment.
re the ground lifter:- do not use a Ground Lifter unless you are absolutely sure the equipment will be used in tandem with an electrical system that specifies a GFI. These are mandated by law in most advanced economies, but this may not be the case in some other areas - check your local regulations.
Further, I recommend a 35A bridge rectifier for the Ground Lifter job. The surge rating is typically 5-10x the average current rating for a specified number of mains cycles.
re AndrewT's original test that caused the mains fuse to blow:- If the live for example touched the amplifier 0V on the secondary side or to the chassis, two things would be going on. Firstly, the current flowing to Earth (either through a Ground Lifter or directly if there was none) would shoot up into the many 10's of amps, and on a system with low loop resistance on the mains side, potentially 100's of amps. A 'T' type fuse would limit the amount of time this condition could persist to a few mains cycles as it appeared happened in his test. Secondly, the GFI would be detecting a massive imbalance in the live<>neutral currents and it would trip. GFI's typically trip in 40-60ms and would probably go before the mains fuse. I don't recall the details of AndrewT's original test - I don't know if he mentioned if the GFI had tripped, but its likely it would have as well as the fuse popping.
So, this is why the Ground Lifter Diodes/Bridge should be rated to carry the full mains trip current for a few hundred ms - if it popped, the user would reset the mains and now have an appliance where the mains could touch something on the secondary without it tripping the GFI until someone touched it of course, and then the GFI would trip - but, that is still not considered safe or acceptable.
Hi Bonsai!
Thanks for that excellent explanation,and especially the article you wrote. Being a novice amp builder I find these things very helpful. Being lazy, My solution would be to always make monoblocks, but there is still grounding issues to resolve.
I have a question about the drawing "Example #1 - Split secondary rectification with GND lifter".
The issue is that the basic rules #2 seems to differ from what you've drawn. In particular, the speaker return going to the PCB appears to contradict the rule of keeping high current wires away from small signal wires. Because, no doubt, there can be high currents on the speaker return.
Did I miss something or am I making something out of nothing?
In my amplifier, an F5T, the intention is to bring the speaker return to the 0 volt line, which should handle the bulk of the current. The amplifier circuit grounds will be attached a bit further from the PS common than the speaker return, as they will have far less current flow. In the F5T these are the FE board and the output boards.
The Front end board ground goes next, as the LTspice sim shows significant current across the HBR from the feedback circuit.
Finally, the decoupling caps on the output boards, are attached to the zero volt line.
However, I have an odd split... my zero volt line is really long. it streches from the capacitor bank at the front of the amp, to the ground lifters at the rear. Since this is dual mono, their are two ground lifters, one per channel.
Does that make sense to do that way? I'm about to wire it up....
Thanks for that excellent explanation,and especially the article you wrote. Being a novice amp builder I find these things very helpful. Being lazy, My solution would be to always make monoblocks, but there is still grounding issues to resolve.
I have a question about the drawing "Example #1 - Split secondary rectification with GND lifter".
The issue is that the basic rules #2 seems to differ from what you've drawn. In particular, the speaker return going to the PCB appears to contradict the rule of keeping high current wires away from small signal wires. Because, no doubt, there can be high currents on the speaker return.
Did I miss something or am I making something out of nothing?
In my amplifier, an F5T, the intention is to bring the speaker return to the 0 volt line, which should handle the bulk of the current. The amplifier circuit grounds will be attached a bit further from the PS common than the speaker return, as they will have far less current flow. In the F5T these are the FE board and the output boards.
The Front end board ground goes next, as the LTspice sim shows significant current across the HBR from the feedback circuit.
Finally, the decoupling caps on the output boards, are attached to the zero volt line.
However, I have an odd split... my zero volt line is really long. it streches from the capacitor bank at the front of the amp, to the ground lifters at the rear. Since this is dual mono, their are two ground lifters, one per channel.
Does that make sense to do that way? I'm about to wire it up....
Hi Bonsai,
I'm thinking some issue in your paper, more explains are welcome.
1. Thimios has measured speaker return to 0V PSU and THD result more better than your way (to 0V amp PCB). So if this thing is correct, it would be 2 loops: one for HF (including local decouplings) and load loop. Is it right?!
2. Affect of HBR to THD. Someone measured to report that HBR increase THD.
The one of highlight contents in your paper are HBRs which are applied to stereo amp to reduce cross channel ground loop. So i'm thinking a way to layout to minimize cross channel ground loop: both channels on a PCB and ground plane.
However that is a challenge and it maybe require 4 layers PCBs to keep ground plane solid.
Another way better is dual mono config. There is no cross channel loop, therefore no need HBR. Just doubling GLB and connect chassis earth at center.
Regards,
I'm thinking some issue in your paper, more explains are welcome.
1. Thimios has measured speaker return to 0V PSU and THD result more better than your way (to 0V amp PCB). So if this thing is correct, it would be 2 loops: one for HF (including local decouplings) and load loop. Is it right?!
2. Affect of HBR to THD. Someone measured to report that HBR increase THD.
The one of highlight contents in your paper are HBRs which are applied to stereo amp to reduce cross channel ground loop. So i'm thinking a way to layout to minimize cross channel ground loop: both channels on a PCB and ground plane.
However that is a challenge and it maybe require 4 layers PCBs to keep ground plane solid.
Another way better is dual mono config. There is no cross channel loop, therefore no need HBR. Just doubling GLB and connect chassis earth at center.
Regards,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.