[Amps] Direct rectification of AC mains to derive the amp VDD, supply

[Manfred Mornhinweg]

Leigh,

> I can in principle concur with your 3 salient points here Manfred.

Yes, we concur on the points, and only weigh them differently. No
problem with that. Given the choice, I also find an isolated power
supply and grounded amplifier block to be a cleaner layout - but the
economy and greater efficiency of a direct line supply is attractive
too, and for this I'm willing to use a floating amplifier!

> directly rectifying the AC power mains would never pass regulatory
> approval for commercial equipment sale in most jurisdictions as the
> rules have tightened up in recent years, particularly in respect of a
> DC and harmonic components going back into the mains, and power
> factor requirements mandated by CISPR, CENELEC, EN standards, IEC,
> etc.

I'm aware of the new power factor standards in some countries, but not
of any of the others you mention. Can you point me at some, to get up to
date? Since I live pretty much at the end of the world, where there is
essentially no electronics industry and no enforcement of any standards
whatsoever in regard to RFI/EMC, I'm a bit out of touch with such
standards. If I should design and build an amp which I also want to
publish, I would like to make it legal in most jurisdictions, even if
here at home I can use anything I like.

> For these reasons we're unlikely to see such a dubious direct AC
> mains rectification scheme deployed in the QRO amp commercial
> marketplace.

I don't know. But for a simple, inexpensive homebuild amp, I still think
direct line powering is attractive.

The issues one has to address when doing this, as I see it, are the
following:

- Safety. This is easily handled by using good enough insulation in the
RF transformers, along with grounding the amplifier cabinet (of course
NOT the neutral!!!), and using a ground fault interrupter, which could
be internal or external. In my country GFIs are legally required in all
homes for all outlets, so it's not necessary to add an internal one.

- Since you mention it, I will pick it up: DC fed into the power line.
But this is a non-issue, when using a bridge rectifier across the line.
Of course it would be unacceptable to use half wave rectification!
Apparently some people think about half wave rectification when reading
about direct line rectification. Antique radios did that, but modern
high power devices don't!

- Power factor. This is indeed an important factor, not only for legal
reasons, but also because legal limit amps draw so much power that if
they have bad power factor, they require special dedicated feed lines.
In my area any outlet can provide 2200 watts continuously - and that
means 2200 voltamperes, or unity power factor. It's usually possible to
run a legal limit amp from them, as long as its power factor is decent,
and perhaps only in SSB mode. With bad power factor and in CW or RTTY, a
special circuit is needed.
But there are important things to consider: First: All the usual, widely
used tube type amps have very poor power factor. Second: A direct
off-line rectifying power supply that has the same ripple as a simple
transformer-based one, also has the same power factor, as long as a
series resistor is used that has the same amount of loss as the windings
of a transformer would have. Note that the efficiency of this supply is
still better than that of the transformer based supply, because the iron
loss of the transformer is avoided.
And third: If a very good power factor is required by law, such as above
0.98, then anyway an active power factor correction circuit is needed.
This needs to be added to ANY kind of power supply! So, a simple, direct
rectification supply turns into a minimal active power correction
circuit, and the amp stage might be powered off a roughly regulated
350VDC line coming from this PFC circuit. An isolated supply instead
would have to start with this PFC, and then add a complete basic
isolated supply! So it's still simpler to use direct off-line power,
even when requiring excellent power factor.

In my amplifier design I go a middle route: My power supply is
non-isolated, but extremely simple, yet it delivers a well regulated,
current-limited DC output, and has an input power factor that is much
better than that of the average tube type amplifier, although it is not
close to 1.

- RF fed into the line. Some people worry about this. I don't. Using
basic RFI filters at the AC input, the amount of RF fed into the AC line
by an amp is at least 50 to 60dB below the power output, even with a
very crude, direct power supply. This amount of RF on the AC line might
matter, if the amp is only used into a dummy load. But if it is used
into an antenna installed in the backyard or on the roof, the power
lines will pick up far more RF out of the air, than directly from the
amplifier!

> I would also not wish sacrifice the immense superiority of TLT xfmrs
> for conventional RF xfmrs to provide the necessary robust electrical
> isolation from the potentially lethal AC mains supply.

I see your point, but I wonder why almost all manufacturers of HF radios
and solid state amps use conventional transformers...

There is a common misconception. Many people think that any transformer
wound with a coaxial conductor or with a twisted pair is a transmission
line transformer. But in truth only some of these are. It depends on how
these lines are connected.

> Facilitating robust AC mains electrical isolation in a traditional
> flux coupled RF transformer covering the entire 1.8 to 30 MHz HF
> range is nontrivial and somewhat problematic.

I have done it, and found no problems at all with the insulation. Simple
teflon tubing is good enough, as long as it is kept from being cut by
some metal edge. Instead the problem is achieving flat enough, low loss
response over that frequency range, when working with higher voltage
supplies. There is a problem with the total wire length required,
compared to the shortest wavelength, to push the flux density low
enough. I developed a transformer that handles legal limit, goes from
1.8 to 30MHz (but barely - at 35MHz it turns resonant!), provides a true
center point, and provides safe line voltage insulation, while being
compact and quite inexpensive. It took me some tinkering, though.

> Achieving the requisite isolation with coupling / blocking capacitors
> is a risk as they have a propensity to fail...and not in a failsafe
> manner :-(

Yes. I wouldn't go that route.

> In summary such schemes whilst seemingly attractive and tempting to
> do on the surface of things have many ugly hairs growing on them that
> make them unattractive.

I'm almost fully determined to go through it and turn my experiments
into a complete amplifier. I just don't know when... since I spend my
time posting stuff on this forum! ;-)  If I ever finish my amp, I will
of course report on it.

> BTW Manfred, I really like your succinct piece titled "Output
> architectures of conventional class AB push-pull amplifiers" at:
> 
> http://ludens.cl/Electron/mosfetamps/amps.html
> 
> I concur with what you say here about how circuit design
> misinformation is perpetuated.

Good. One more member in the club! What I would like to hear is the
reasons anyone who did this "the wrong way" had! I still wonder if all
who use the wrong configuration do so simply because they copy older
wrong designs without understanding them, or if they see a good reason
to use that configuration!

Peter,

> In some countries, connecting the neutral to an outside earth could
> be extremely dangerous, leading to fires or even electrocution.

In Chile, connecting the neutral to earth inside any equipment would 
instantly trip the ground fault interrupter. The neutral is grounded at 
a single point, that being at the source of the circuit: The step down 
transformer, or the generator, or the inverter, or whatever is powering 
the circuit. That's before the GFI.

> Where
> a Protective Multiple Earth system is used, a ruptured neutral could
> pass many amps down any earth lead, or lead the coax outer to float
> up to full mains voltage. This is why it is not permitted to bring
> any conductor within the unipotential zone unless it is bonded to the
> system earth, and  earth leads need to be able to carry large fault
> currents.

Chile has a double-safe standard there: It requires GFIs, so that the 
fault current can never get large, but at the same time it is required 
that protective earth conductors be as thick as the main conductors.

In olden times, GFIs didn't even exist, and earth ground in power 
outlets was unknown here. Outlets were two prong, non-polarized, most 
tube radios had one side of the line connected to the chassis, the back 
covers usually fell off and got lost, and if a knob pulled off, the 
shaft poking out had a 50% chance of being live at 220V, depending which 
way the power plug had been connected. Wire insulation was mostly 
rubber, which crumbled and fell off after ten or twenty years. In 
retrospect, it's surprising that so few people were electrocuted. 
Compared to that, a teflon-insulated device, with its metal case 
grounded, and powered through a GFI, is extremely safe!

> 
> Where PME is not used and the neutral is earthed somewhere else, you
> can still get quite a large potential between neutral and earth. When
> I lived on a housing estate in Swindon where the whole estate was fed
> from a substation where the neutral was earthed, on a Sunday morning
> with washing machines, electric cookers, immersion heaters etc all
> blasting away, I could get 25 volts at 500mA - free - between a
> decent ground and the mains neutral. No RCD (GFI in the US) there -
> if there was one fitted, earthing the neutral to a decent external
> earth would be very safe - you would have no electricity supply
> because the RCD would trip!! At this QTH, at the end of the line
> feeding other houses and a half mile of wire from here to the
> transformer, the neutral is usually around 1 volt above earth
> potential.

So that's much the same as here. When I still lived in the city and 
bought my electricity, I typically got about 4V between ground and 
neutral. Now I make my own power, and the grounding spot of the neutral 
is right under my house, so I get only a fraction of a volt between 
ground and neutral - but If I join them at an outlet, the GFI still trips!

> So any system earthing  the neutral is not a good idea and may even
> be illegal in some places.

Agreed. But just to make sure all readers get it: Powering an amp 
directly from the line does not require earthing the neutral. Instead 
the entire amp module floats at line potential.

> If one used TLT instead of conventional transformers, the blocking
> capacitors would need to be Class X, and for 160m, at least 10nF. But
> Class X capacitors are not designed for handling large amounts of RF
> current (ca. 3 amps at 500 watts) , and are not as low inductance as
> one would want for 10 metre operation. 

Exactly.

 > Cascaded RF transformers is, to my mind, asking for trouble.

Some amps use an unbalanced to unbalanced transformer to step up the 
impedance, and then cascade a balun to provide coax output. That works 
perfectly. I see no problem with cascading RF transformers, other than 
the added cost and loss. Usually it is possible to find a single 
transformer solution, which is more cost-effective.

 > From memory, an isolating transformer
> to meet the regulations needs to withstand 2kV, although I could be
> wrong on that.

I have seen a 4kV rating in many things, such as the required insulation 
between primary and secondary sides for cellphone chargers, or the specs 
of safety-rated optocouplers. So I design my own stuff for 4kV 
insulation. My legal limit HF output transformer comfortably handles 
4kV. This is probably overkill, as it is the requirement for equipment 
that has no grounding on the secondary side. When there is grounding, 
probably insulating for 2kV is plenty. Even 1kV might be enough!

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================