Please consider below drawing - will these two alternatives be equal or is any one preferred when it comes to the two consumers experience of clean DC feed? Note that Consumer 1 and Consumer 2 share ground on one and the same circuit board. If it matters the board is an XMOS board and C1 is board main power and C2 is clock section.
I would assume A is better but will it really be in reality? And if so, why? Is there in fact a ground loop in B?
//
I would assume A is better but will it really be in reality? And if so, why? Is there in fact a ground loop in B?
//
Yes I think so, and it would be more of a constant rather than a changing quantity.
If I squint, can see a loop possibility in B which doesn't seem to be able to occur in A - at least not galvanically - is it dangerous?
//
//
It shouldn't be, but as with everything there are numerous possibilities for things to be less than ideal. In post #1 you say the board shares a common ground and so that immediately puts that point as our 'ground reference'. What happens elsewhere should not matter. The twin SMPS can conceivably cause this non correlated ground current to flow through the ground tracks in the board and because they are of finite resistance they will cause an uncorrelated and changing volt drop to develop across the ground tracks.
How much of an issue it is remains an unknown.
How much of an issue it is remains an unknown.
Ground loops can work as antennas for RF, a properly grounded metal chassis can help to avoid this. You should make some sketches to find out where there might be connections between signal ground, PSU ground and wall-power ground. Ideally, these should be separate or connected via a high-ohm (ground-lift) resistor. Often an unbalanced part of the circuitry demand some connection between PSU-ground and signal ground, then it is good practice to have such connection at one place and one place only. PCB's should have a separate layer for grounding with as much copper as possible to help shielding and to have less resistance. Sometimes nothing seems to help (TV-transmitter close by, 60dB amplification for DC-1Ghz). Lots of small-value C's close to the chips can make wonders. From Your sketch it seems You need some filtration between the regulators and the consumers. Manufacturers application notes will often provide hints on what to do. One or two SMPS? - best is none (SMPS produce high-frequency noise and spurious) but proper grounding and filtration can make this (mostly) disappear. After all, even batteries produce noise... Hope this can be of some help.
Do you mean the 3 first and as opposed for all of them and "wall-power ground" or between basically between all of them?
I plan to place the whole system into a dubbel shielded / class 2 enclosure that has no metallic surfaces (not outwards, nor inwards) but where all of the different parts will reside within an own compartment that are a faraday cage. A metallic net will be placed between isolated outer sheets - as a sandwich (plastic-metal-plastic). So no mains earth involved.
OK, I might want to place also filters between the regulators and the load. One compartment will hold smps and a filter - so that will be essentially dirty and the load will be in an other compartment - I will place yet a filter before the load in a "cleaner" department.
How does that sound?
//
Between all of them.
I think Your way of RF-shielding (Faraday...magnetic eddy currents...) should work fine as long as the mesh-size of the metallic net is small.
Best is to have a separate shield-grounding connected to wall socket ground or to Mother Earth to get rid of the voltage produced by the eddy currents. Might be useful to make a sketch for shield-grounding of Your system. If I remember correctly, Your wall power in Sweden has 0V(N) and 230Vac from power company, so it would be nice to know if/where this N is connected to Mother Earth (can be in a local transformer, central transformer, the generator...).
If You make one sketch for each (shielding, wall power, PSU, signal) You will see if You have any loops.
The main goal is to have the total system look like a tree-structure, with one central starting point (which can be Mother Earth).
Unsure/can't find out about the power company's P-E-N, or if Your neighbor has a bad boiler? Make Your own central point (if legal....) with a dedicated 1:1 transformer between the wall socket and Your system (search for "Site Transformer").
Your filtering scheme should work, try to avoid tantal-C's wherever possible.
I think Your way of RF-shielding (Faraday...magnetic eddy currents...) should work fine as long as the mesh-size of the metallic net is small.
Best is to have a separate shield-grounding connected to wall socket ground or to Mother Earth to get rid of the voltage produced by the eddy currents. Might be useful to make a sketch for shield-grounding of Your system. If I remember correctly, Your wall power in Sweden has 0V(N) and 230Vac from power company, so it would be nice to know if/where this N is connected to Mother Earth (can be in a local transformer, central transformer, the generator...).
If You make one sketch for each (shielding, wall power, PSU, signal) You will see if You have any loops.
The main goal is to have the total system look like a tree-structure, with one central starting point (which can be Mother Earth).
Unsure/can't find out about the power company's P-E-N, or if Your neighbor has a bad boiler? Make Your own central point (if legal....) with a dedicated 1:1 transformer between the wall socket and Your system (search for "Site Transformer").
Your filtering scheme should work, try to avoid tantal-C's wherever possible.
Thanks! I do own 1:! 5kVA isolationstransformer with produces 115/115V - it has an electrostatic shield between the windings. A drawing is a good advice. I will also try ground the faraday cage to mother earth i.e. something that is dug into the ground. If I dare - there are quite a lot of thunderstorms here and I don't want to lead lightning into the house.... But I will probably test it at least.
//
//
5kVA should be ample 🙂 Sometimes the circuit breakers don't like the inrush-current of big transformers, so don't use the same circuit as the freezer/fridge... IF you choose to make Your own central point (connected to as many square meters of soil/sea as You can afford, copper is expensive) please be sure to avoid using this (central point) to anything connected directly to mains outlet, or You may start a fire(in the night, when everybody is asleep, from potential differences) or get an explosion(from lightning hitting the power grid).
You will not guide lightning into the house as long as You avoid exposed parts over the ground. I have seen lightning hit the ground, missing grounded metal mast by a few meters...a channel of ions are formed from earth to the sky before the lightning strikes, would be nice to know where it is formed 🙂
You will not guide lightning into the house as long as You avoid exposed parts over the ground. I have seen lightning hit the ground, missing grounded metal mast by a few meters...a channel of ions are formed from earth to the sky before the lightning strikes, would be nice to know where it is formed 🙂
The trafo has a soft start using a AC relay and a resistor - simple and effective - it lets that ENK build up for a few cycles via the resistor and then as the voltage is built up, the relay shorts the resistor out.
It's a lot of rock around here so I don't know if it's possible. But if so, I have no intention of connecting any mains lead of any sort to this mother ground. To lead away energy from the cage to a common potential is a good idea never the less.
//
It's a lot of rock around here so I don't know if it's possible. But if so, I have no intention of connecting any mains lead of any sort to this mother ground. To lead away energy from the cage to a common potential is a good idea never the less.
//
For rocky ares (and mountains) a deep hole in the ground, copper wire to the bottom, fill with (elektrodemasse https://efobasen.no/productresources/646/fdv/GEM datablad.pdf).
A crow-foot is the best, a central hole with 2+ holes around it, drilled as far down as possible, but now we are talking of extremes...
Q: Did You experiment with multi-entry horns? I wonder what spec's can be somewhat relaxed; distance from throat, horn area at the ports or distance between the ports. Any tip?
A crow-foot is the best, a central hole with 2+ holes around it, drilled as far down as possible, but now we are talking of extremes...
Q: Did You experiment with multi-entry horns? I wonder what spec's can be somewhat relaxed; distance from throat, horn area at the ports or distance between the ports. Any tip?
You're welcome. Forgot to mention - it would be very beneficial if You can get Your earthing electrode to or below ground water level.
OK - that's 25 meter down ;-) it's where I get my drinking water. But perhaps that is indeed a solution..... it has a pump sunk to the bottom of the hole....
//
//