[AMPS] Re: Surge voltage, HV connectors.

[Tom Rauch]

Hi Wilbert,

Hopefully we can move away from this parasitic nonsense, and get 
on to useful things that really are common problems.

> > >Microwave oven transformers have horrible ESR, and make very 
> > >poor linear amplifier power supplies.
> 
> Measured:
> 2093 VRMS (no load)
> 2038 VRMS @ 400 mARMS (projected amplifier load)
> I can't see what's wrong with this.
> 
> The fact the supply under test sags, is not due to poor regulation. It is
> simply a result of the fact the capacitor will float to Vpeak under no
> load. I note that all HV supplies in the 1990 ARRL handbook are of this
> type. All will exhibit voltage surging under light loads.

I suspect you are measuring RMS voltage, and not peak voltage. 

Excessive ESR in the transformer does not show up on an RMS 
meter.

That's because the power supply depends on the peak voltage, and 
being able to have many amperes of peak current with minimal 
voltage drop right at the peak. A very large change in peak voltage, 
because the peak is such a small fraction of the cycle, barely 
shows when the measuring device is a RMS indicating device.

A good capacitor input supply on a normal power main in the USA 
will have better than 10% regulation from zero load to full load. 
You'll have a bit less because of the 50 Hz mains and the system 
used in NZ where the power transformers are a long distance from 
the house in many cases.

When I visited NZ, I was amazed that often times transformers 
were several hundred feet from the last house on the drop! But the 
bulk of the regulation problem you have is in the oven transformer.

A good 2000 volt transformer 500 mA transformer for capacitor 
input supplies should have under 15-20 ohms ESR, which is about 
ten ohms secondary dc resistance in a typical transformer design.

A choke input will help, because the supply will draw current over 
virtually all of the cycle with the correct size choke. Peak current 
can be reduced by a factor of 10 or more.
 
> Voltage surging can be reduced by using an inductor, and by increasing the
> stand-by load (more bleeder current). Inductor-input filters don't seem to
> be used in the commercial designs I have seen (FL2100, AL811). I would
> prefer not to increase bleeder current (a crude form of shunt regulation)
> but will do so if I have to.

That's because it costs much less to use a better transformer than 
to add a choke, to say nothing about size and weight. 
 
> What I wanted to know is: is this voltage surging really a problem? When I
> put the key down, the supply will 'collapse' to its working voltage of 2.1
> kV. If I tune up the amplifier for steady-state clean output (no
> flat-topping) at 2.1 kV, will the initial reduction from 2.8 kV at the
> start of the transmission cause non-linear distortion?

There will be two types of distortion, one caused by short term 300-
4000 Hz variations and one caused by syllabic variations at a few 
Hz. The higher frequency variation is easy to handle with capacitor 
size, the low frequency one is a problem.

The low frequency problem does not show up in conventional two-
tone IMD tests or noise-notch IMD tests, so people often miss it. 
To find that problem you need a three tone test with the envelope 
modulated at a few Hz rate.

Some PA's will handle the LF supply variations without severe 
problems, others won't. What you will have is a PA that has 
leading edge overshoot, where peak power on leading edges can 
be up to the square of voltage increase under light load conditions 
above the steady state condition.

IMD is a funny thing, sometimes we can get away with 
unbelievable amounts of distortion with no real IMD and other times 
it is the opposite. It's the shape of the non-linearity that effects the 
stuff we dislike to hear.
      
> If the answer is 'yes', what should I limit the surge voltage to?

Everyone will have an opinion, and they will all be guesses 
because there are too many unknown variables from your speech 
habits to the radio to the tube behavior. But as a general rule 10-
15% is at the limit.

The largest problem you will notice is the PA will go way overpower 
until the voltage drops. If you combine that effect with a radio with 
overshoot (and that is common) you can easily have 5 or more 
times the leading edge peak as the continuous peak power.

Poor regulation increase the odds of bandswitch or tank arcs, tube 
arcs, and occasional reports of splatter dramatically. But no one 
can answer "how much", because there are to many "it depends 
on".
  
> > >> Ideally I would like an HV connector with a set of extra contacts for
> > >> carrying mains to the amplifier filament transformer as well.
> > >
> > >Never ever do that, unless you have a death wish for you or your 
> > >family. Especially in NZ, where the power mains are not grounded.
> 
> The mains here is of the Multiple Earthed Neutral type.

It's my understanding the mains are not ground referenced, that 
both sides of the line float with no CT reference to earth, although 
you have a safety ground. I had a specific conversation about that 
with two ZL's. Was I misinformed?
 
> Mains and HV are mixed inside the amplifier anyway. I can't see how 
> combining mains, interlock and HV in one properly rated cable would make
> things more unsafe. In the case of a short from B+ to mains neutral, the
> HV supply would be unable to lift the mains supply. The resulting short
> would cause it to blow a fuse.

What about leakage from HV to one side of the mains? Or a direct 
fault that discharges the HV into the floating mains leads? Is it safe 
to tie a charged bank of capacitors across the mains or from the 
mains leads to earth? 

If it isn't, you need to isolate the HV lead from the mains with some 
form of guard area that is always well grounded, so there is no 
possible path. I prefer not having it anywhere around the other stuff, 
if nothing else to prevent blowing up extra parts if there is a 
insulation or connection failure.

Sometimes we all get a little "fussy" and this is a hobby, but its 
nice to try and keep each other alive a bit longer. In 1970, I learned 
all about resonant choke filter and Millen connectors when I was 
almost killed by following a Bill Orr resonant choke design and 
using Millen connectors INSIDE a rack cabinet. The resonant 
choke surged to several kV when the power switch was closed 
because I wasn't warned that exact resonance can produce 
extreme voltages. The voltage shattered a Millen plug and the lead 
came flying out of the rack through an open door and attached 
itself to my wrist, while my fingers were on a metal toggle switch. I 
drew a few inch arc off my fingers while I waited for the mains to 
open.

The shack smelled like burnt Ham for a month.
73, Tom W8JI
w8ji@contesting.com

--
FAQ on WWW:               http://www.contesting.com/ampsfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm