"linear"power supply "roundtable"

[quote author="Brian Roth"][quote author="VacuumVoodoo"]And to be on the safe side: double fused double switched with an X1 cap across the mains.[/quote]

If you mean a fuse in the "neutral" line (USA usage), that is NOT permitted. Only the "hot" should be fused.

Bri[/quote]

I should've mentioned that this is a requirement IF the mains transformer has a grounded electrostatic shield between primary and secondary winding. At least according to EU norm EN-60065.
I'm currently in the process of having my amps fully CE certified and in ongoing discussions with an accredited lab who are doing all the testing etc. Had to fuse all secondaries too.....shorting tube heater windings is not enough to blow primary fuse, that's why.

The grapevine has it that US has began to recognize CE-certification as proof of safety allowing a product on domestic market.

Why power Schottkys for rectification in place of good'ol bridge rectifiers?? What does that get you (besides being 3 times the cost)??

I often see RC snubber circuits in parallel with each schottky as well. What does that get you that a choke clamped around the power cord does not?

Snubbers on the rectifier diodes are used to bypass large transient spikes that occur on the power line that are passed through the power transformer. A typical spike may be as much as 3-4 times (or more) the peak value of the power line waveform. If you are supplying 36VAC output from your power transformer for a +- 15VDC supply, the transient could be as much as 100 volts or more which will fry a 60VRMS diode. Along with snubbers, a MOV is usually placed across the transformer input (or output) to clip off the transient. MOVs should have series resistors of some value to limit the peak power to something less than the MOVs rating (typically 1500 - 6000 joules - and if you want to get fancy, a fuse and indicator since the MOV can punch through if it gets hit by a really big spike like a lightening strike). Obviously, the problem is enhanced when the transformer represents a step-up in voltage.

I am also a big fan of power line RF filters.

Okay,

An integrated Corcom RFI plug connector and a good pair of MOVs make sense to me, but why fancy Shottky diodes instead of a good'ol molded bridge?

Actually the primary purpose of caps across typical power diodes is to push the energy in the ringing of the stray inductance/capacitance, induced by the reverse recovery spike of the diodes, down to lower frequencies where it doesn't radiate as efficiently---that is, it is done out of emission concerns. Putting a little damping R and/or ferrite bead in series with the cap can help as well.

The reason to use schottkies is that they look more like a simple capacitor in reverse bias. As previously remarked, silicon metal-semiconductor schottkies have a small region that behaves like a regular P-N diode and thus has some reverse-recovery activity and associated noise generation capability. SiC appears to be perfect by comparison.

Considerable work has given us "soft-recovery" P-N diodes wherein the reverse-recovery transient is made as smooth as possible (note that this conflicts to some extent with the desire to make the recovery interval as short as possible for high-speed rectification efficiency).

Note that it is not just an issue of passing EMC. Once this energy from the spike radiates it gets into everything in the vicinity---emitter-base junctions and other nonlinear devices then detect it and we get a mains-harmonic buzz in the audio, or a periodic glitch in our data acquisition. The spectrum of the radiated energy can reach up into the GHz region.

The silicon carbide diodes have a higher forward drop so are most suitable for higher voltage supplies. They are a boon to power-factor-correction circuits which operate at hundreds of volts, and are being enthusiastically embraced for this application.

As far as high-voltage audio: there's a suspicion among some bottleheads that the reverse recovery junk from conventional silicon rectifiers gets into the signal path and accounts to some extent for the degradation heard compared to hollow state rectifiers. Of course the higher impedance of a GZ34 may have as much or more to do with the difference...

If you mean a fuse in the "neutral" line (USA usage), that is NOT permitted. Only the "hot" should be fused.

Click to expand...

Could someone give me a hint why this is not permitted? I thought it were considered good practice here in Switzerland and I do it most of the time.

Samuel

IIRC it has to do with nuetral being at ground in some cases. If the neutral line fuse blows it will disrupt ground connection while HOT would remain live. Correct me if my i got it wrong out of my protein-based database...

you are correct. most residential wiring(USA) has neutral and ground bonded at the breaker panels. if you fuse neutral your box will shut off but the interior will still be "hot" even though the circuit is broken.

[quote author="Svart"]you are correct. most residential wiring(USA) has neutral and ground bonded at the breaker panels. if you fuse neutral your box will shut off but the interior will still be "hot" even though the circuit is broken.[/quote]

I guess the problem is you can't assure that the hot fuse has blown if the neutral fuse blows. So you (as the hapless tech, or consumer disregarding the warning label) go in to service and get potentially zapped.

I don't know if it would be to code, but having a spring-loaded double circuit breaker that interrupted both side of a single-phase system, regardless of which side tripped, would be best. This would take care of situations in which the plant wiring was reversed by mistake, or the neat trick I just dealt with in a client's samples from China, where they had wired the polarized line cord backwards :twisted:

In my "pro" designs i always try to have a breaker on the Line side and then polyfuse certain areas of the board. Having a safety mechanism on neutral is of course a good idea but rarely used due to parts cost etc.. However, in residential wiring it would seem that crossing a wire's output at either end would cause a grounding fault, but as you mentioned and I have also experienced, sometimes you get cheap cables that have a color backwards and can give you nightmares in your box.. :shock: so sometimes it does pay to purchase something a little more costly to ensure a less costly outcome..

> double fused double switched

We must draw a distinction between Service Equipment (the building fusebox) and Appliances (our boxes).

In the USA, we don't fuse or switch the White (groundED, "neutral") in the fusebox. If we did, and it went open, a "dead" circuit would still be 120V to the nearest radiator, which could surprise someone.

If it appears to be necessary to break both sides of a 120V-0V circuit, we must use a ganged fuse (bet you never see one) or a ganged circuit breaker. However that is so unusual that I doubt it is ever done for a 120V-0V circuit.

US-type "240V" (120V each side of Ground) circuits are complex. We generally use a 2-gang circuit breaker, though individual fuses used to be legal IF they were mounted so you had to take both out at once. Since US-culture 240V boxes are generally above DIY scale, I'll ignore them.

General power in the US is 100V-120V because we electrified before good insulators, and 100V is less lethal than the higher voltages we would prefer for reduced copper-cost. Most of Europe was electrified later, when decent insulation was available, and general outlets are 230V, apparently single-ended (one side grounded in fusebox or power-pole). Looking at where you fuse and switch, a EU 230 system works the same as a US 120V system, just check your voltage taps. Since one side is supposed to be earthy and "safe", we can just switch/fuse the other "live" side and have a safe "dead" mode.

It is presumed that the Service Entrance will be wired correctly 99% of the time.

Now when you get to the Appliance, 60% of the outlets in my kitchen were wired with Black and White reversed (and all the Green wires cut off). This is almost typical in US wiring: we had millions of 2-pin unpolarized outlets installed 1910-1960, and even if someone has put polarized or 3-pin outlets in their place, they are often wired wrong. Inspectors do pay attention to the Service Entrance, and most power companies won't connect until the panel is Inspected, but no inspector can check ALL outlets and many outlets are maintained without inspection. And then there are miswired power cords.

Inside the Appliance, you must assume that either wire may be HOT (or even both, as in bcarso's odd tale of +/-60VAC). However you can also assume that only Qualified Persons will be inside the box while it is plugged in (see big warning on all commercial boxes). If your wiring is done perfectly, especially if any exposed metal is properly Grounded, it "should" be adequate to fuse one side, and Black is the preferred side.

Two fuses are usually cost-prohibited in US-market commercial production (however I do have a Cisco router with a warning that it has a "fused Neutral"). They may be a good idea, and VV says they are required under his regulations in some cases. (This may reflect the reality of the New Eurpose: while UK, West Germany, and many other areas have generally good wiring, some backwaters and ex-USSR areas may be wired as bad as much of the US, and at more-lethal voltage and/or more-hot circuit power limits.)

In DIY, where wiring may be imperfect, and fuse-cost is negligible compared to your labor and your life, I see no reason you can't fuse both sides of the line, so long as you remember that "does not light up" is NEVER the same as "dead inside".

Always ASSUME the inside of a box is LIVE and DEADLY until proven otherwise. If it is plugged in, it can kill you. If it was plugged in, and has energy storage (capacitors), it can kill you. When you open the box, always assume it is full of angry rattlesnakes until you check it every which way. I generally work with one voltmeter on the main power cap all the time: I need to see voltage to be sure the meter works, and I need to see low/no voltage before I put my flesh on conductors. Often I just work without letting my flesh contact any conductor.

Yeah, yeah: keep one hand in pocket, stand on a slab of glass.... my shocking experience is that these methods fail as soon as you get excited or careless. And despite incredible luck (I'm still alive), I have permanent damage from a one-hand shock (600V through my little finger).

good grounding link

http://en.wikipedia.org/wiki/Earthing_system


check the other blue links for more info.