Sanity check please.
Mains transformer with centre-tapped secondary ... so ...
Primary = 0 - 240 (actually 2 120s in series, but for simplicity)
Secondary = 400 - 0 - 400
The 0v secondary should be grounded, but should it go to mains neutral or mains earth ?
Objectives of grounding that I can see would be;
1) provide a 0v ground reference
2) provide a low-impedance drain for any parasitic capacitance
3) provide a safety earth in case of a short in the secondary winding
For reason 3 I believe it should ground to chassis earth, which connects to mains earth. Any current will trip. Welcome comments confirming I'm right or putting me back on the straight and narrow.
The transformer is actually hidden inside my amp (wooden case) but for complete safety I also connect the transformer case to chassis earth, along with all of the choke cases which are also hidden.
Mains transformer with centre-tapped secondary ... so ...
Primary = 0 - 240 (actually 2 120s in series, but for simplicity)
Secondary = 400 - 0 - 400
The 0v secondary should be grounded, but should it go to mains neutral or mains earth ?
Objectives of grounding that I can see would be;
1) provide a 0v ground reference
2) provide a low-impedance drain for any parasitic capacitance
3) provide a safety earth in case of a short in the secondary winding
For reason 3 I believe it should ground to chassis earth, which connects to mains earth. Any current will trip. Welcome comments confirming I'm right or putting me back on the straight and narrow.
The transformer is actually hidden inside my amp (wooden case) but for complete safety I also connect the transformer case to chassis earth, along with all of the choke cases which are also hidden.
The 400V secondary CT should go to the negative tag on the reservoir capacitor. This point will usually eventually have a connection to the chassis via the PSU circuitry. The one thing you should never do (unless you enjoy buzz) is to connect the CT directly to the chassis; this technique was used in the 1950s, which is why most 1950s equipment buzzes.
Heater circuits should have a DC reference voltage. This voltage may be zero. The reference voltage may be applied to one side of the heater wiring or to a CT (either real or artificially made from resistors). Different options may give different levels of hum.
Mains neutral is not ground. Have you made and do you use a lamp limiter when debugging new PSUs? The sad thing about lamp limiters is that the people who most need to use them are likely to be the people who are least likely to use them.
Heater circuits should have a DC reference voltage. This voltage may be zero. The reference voltage may be applied to one side of the heater wiring or to a CT (either real or artificially made from resistors). Different options may give different levels of hum.
Mains neutral is not ground. Have you made and do you use a lamp limiter when debugging new PSUs? The sad thing about lamp limiters is that the people who most need to use them are likely to be the people who are least likely to use them.
Thanks - yes I have made and tested a lamp limiter and always use it. Any changes made, the lamp goes into circuit. When I'm confident I switch to bypass. One of the great tips I've picked up in these forums (it may have been you ?).
Interesting about taking the CT to chassis earth via the negative terminal of the first reservoir cap (I'm using LCLC smoothing). I use that terminal as star ground with separate grounds from each valve cathode, each valve grid input resistor, signal input, etc. Had you not mentioned that comment I would indeed have gone CAP- to SEC 0v to chassis earth. Thanks. I'll reference from CAP- direct to chassis earth as you suggest.
I understand that Mains Neutral and Chassis Earth aren't the same. Both are (should be) at 0v and my question was checking my assumption that earth is the correct 0v reference to use for the SEC CT. Grounding is an area I've struggled to get my head around but with a lot of reading I think I'm getting it. Some articles on the web use those terms interchangeably and I stop reading as soon as I see that !
Don't connect the reservoir negative directly to chassis, and don't use it as a star point. Instead, connect reservoir negative to smoothing cap negative, then connect smoothing cap negative to chassis. You could use smoothing cap negative as a star point, although I personally prefer a bus ground. You must keep PSU charging pulses well away from the audio circuits.RhythMick said:
Mains Neutral could be anything. Don't assume that it is even near 0V. Chassis could be anything; depends on what it is connected to. Normally it will be connected to Safety Ground - which may or may not be near local earth ground in your garden. You choose what to call 0V; the universe does not care, as it only knows about potential differences and current loops.
The CT does not need a '0V reference' and if connected incorrectly it will pollute any reference. What the CT needs is a circuit loop so the charging pulses have somewhere to go. The 0V reference if you do it correctly is the negative terminal of the smoothing capacitor. If so, the CT will not be 0V.
OK I need to check terminology. I'm using LCLC so my PSU is..
MAINS-SEC -> TUBE RECT 5R4WGB -> CHOKE -> CAP1 -> CHOKE -> CAP2 -> B+
which would you refer to as the reservoir vs smoothing caps ?
ok - so that flies in the face of what I thought I understood. I understand that the whole of the power supply could be left floating, but that it was bad practise to do so - better to tie it down to a known voltage reference.
Similar would apply to the OPT secondary - speaker terminal should be grounded to prevent build up of parasitic capacitance ?
Am I mixing up different concepts or are there genuinely different views ?
Cap1 is the reservoir cap. Cap2 is the smoothing cap. If there was a Cap3 then this would be the second smoothing cap, and Cap2 would then be called the first smoothing cap.
Any circuit can be left floating, but for safety reasons it should not be. Connections to safety ground need to be made at the correct point. The correct point in a PSU is the clean smooth DC end, not the noisy dirty AC end. Even more important, the correct point to connect a PSU to the rest of the circuit is the clean end.
The OPT secondary should be grounded for safety reasons. In many cases it also needs to be grounded so feedback can work. If the latter, then it needs to be grounded at whatever is the same reference point as the input signal. Parasitic capacitance is not the reason; insulation breakdown is the safety issue, while the definition of voltage is the circuit issue.
As on all subjects, there are genuinely different views; some of these views are wrong.
Any circuit can be left floating, but for safety reasons it should not be. Connections to safety ground need to be made at the correct point. The correct point in a PSU is the clean smooth DC end, not the noisy dirty AC end. Even more important, the correct point to connect a PSU to the rest of the circuit is the clean end.
The OPT secondary should be grounded for safety reasons. In many cases it also needs to be grounded so feedback can work. If the latter, then it needs to be grounded at whatever is the same reference point as the input signal. Parasitic capacitance is not the reason; insulation breakdown is the safety issue, while the definition of voltage is the circuit issue.
As on all subjects, there are genuinely different views; some of these views are wrong.
I've connected all exposed metal parts to safety earth and as mentioned above have also connected the cases of the chokes and transformers, even though they are inside the (wooden) case.
One exception to this - the metal bolts connecting things like the valve bases and cap holders. Were a live wire to come loose and touch one of those metal bolts at the same time as someone touched the outside that would be a problem. The metal bolts holding the transformers aren't a problem, because the case of the transformers are earthed.
What's the best way to solve that problem ?
- nylon bolts ? wary of their strength over time
- fix to underside of wood without metal protruding through the case (eg threaded inserts or wood screws)
- insulating covers over internal metal bolts and nuts
One exception to this - the metal bolts connecting things like the valve bases and cap holders. Were a live wire to come loose and touch one of those metal bolts at the same time as someone touched the outside that would be a problem. The metal bolts holding the transformers aren't a problem, because the case of the transformers are earthed.
What's the best way to solve that problem ?
- nylon bolts ? wary of their strength over time
- fix to underside of wood without metal protruding through the case (eg threaded inserts or wood screws)
- insulating covers over internal metal bolts and nuts
Good point I hadn't thought of taking the B+ from there. I don't have a 5VCT yet, but I will have when I get the proper mains transformers.
That is what we do. But that is NOT what is required
1.) the metal enclosure must be permanently and mechanically connected to the Protective Earth (PE) wire in the three core mains cable.
2.) Any exposed conductive parts should be connected to the protected Chassis/Enclosure.
It's this second SAFETY requirement that confuses many Members.
If your enclosure has no exposed conductive parts, then you do not need an extra SAFETY connection to comply with requirement 2.
But you MUST always comply with requirment 1.
PE is what blows the mains Fuse if a mains wire comes loose and touches something else.
You are now entering audio performance. This is completely separate from the two SAFETY requirements listed earlier
Now back to SAFETY. Separate the two issues from each other.
Decide on how you make the equipment safe to operate. Implement this.
Decide quite separately what is required to allow the equipment to operate as good audio gear.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.