What the ammeter shows for the line current on a particular transformer
will depend on the ammeter, the transformer core, the load, and the line
voltage.
In a plate transformer application the load current is distorted, just
short pulses of current at the line voltage peaks to replenish the
charge in the filter capacitors that was taken out by the continuos
(e.g. DC) load current during that half cycle. The width of the pulses
is narrowed by a larger filter capacitor bank which lowers the load
ripple voltage, but the peak current gets bigger. This effect is purely
load dependent. The winding impedance of the transformer will tend to
broaden the width of the pulse and lower the amplitude of the peak. I'll
mention that more down a ways.
The core material causes the line current to be distorted, with a
stronger peak in the current than the voltage. That's because the core
is approaching saturation. This is a function of core material, core
construction, and the winding design, as well as the applied line
voltage. Classical transformers with laminated E-I cores can't make full
use of the grain oriented steel of the Silectron E tape wound cores and
so run at a lower flux density, more remote from saturation and so
demonstrate a softer saturation. The Silectron E cores can be run up to
twice the flux density of the ordinary E-I cores, but if the line
voltage is a little high and they are wound too close to the limit, a
little rise in line voltage will cause a leap in that peak current.
I've designed transformers both ways, the silectron E core transformer
is likely to end up requiring half the core weight and half the copper
in the windings for the same load rating because it can use twice the
peak flux density. This is good and it is bad. Its good for compactness
and with a little conservatism in design can lead to a transformer that
operates cooler at rated load. But the fewer turns and lower resistance
of the windings means the peak currents caused by the capacitive
rectifier filter load will be greater. The tape wound cores cost more,
but the copper savings generally will cover that extra core cost.
When a core is run closer to saturation, it often tends to remember that
last peak when turned off. By Murphy's law, when energized the applied
voltage will be opposite to the last cycle and the magnetized core will
need a massive switch in polarity which will cause an extra large spike
of current in the first half cycle and an audible "TUNG" sound. The
lower flux density cores generally don't show these nearly as much.
Line current wave form does affect the meter reading, depending on how
the meter is built. The typical clamp on type ammeter uses a step up
transformer, then a rectifier and a resistor to the meter. That meter
actually reads PEAK current but is calibrated for RMS assuming a
sinewave. The input current to a transformer with rectifier load and
capacitor input filter (or all capacitor as is common in solid state
supplies) is a long ways from a sinewave. If the ammeter is a moving
vane type (often detected by not being damped much) it will give a
different reading than the rectifier type meter. And if the digital
meter used is actually made with a "true RMS" circuit, the reading will
different yet again, though the true RMS meter and the moving vane
movement should agree reasonably close.
A year or so ago, I built an AC current regulator circuit that had flat
topped sine waves as the output current. I checked it with my true rms
digital meter, a moving vane meter, an electronic true RMS instrument, a
Simpson 260 (rectifier), and a wide band oscilloscope. It was like the
man with two watches, never knowing the time... The input current to a
transformer with a rectifier load is less well behaved than the output
of my AC current regulator.
A class B linear should run about 60% DC to RF efficiency. A portion of
the output in a grounded grid linear comes from the driver, probably
about 80 watts of the output on a pair of 3-500Z is from the driver.
That adds to the apparent final efficiency. But for 1500 watts out, the
final is adding about 1420 watts, at 60% means 2367 watts into the
plates. Add 50 or 100 for filaments and neglecting transformer and
rectifier efficiency you have about 2450 watts total input. The
transformer will loose probably 2% for a silectron E core, maybe 4% for
a regular core, so you could easily show over 2500 watts. If the line
current has a peak to average ratio to mislead the meter by a factor of
two (which is easy) a clamp on ammeter might show 5000 watts input. If
you make the breaker too close to the true RMS load current it will trip
from the peak currents.
You can use the watthour meter on the house wall or out on the pole to
read true watts ignoring the waveform distortion. On the meter face
there should be a note Kh = 7. That means that 7 watt hours will turn
the disk one turn. Meters with different ratings will have a different
constant. An electronic meter might not have that marking. So you time
the rotation of the disk, then watts are equal to Kh * 3600 / rotation
time in seconds. You can time multiple revolutions and then divide the
total time by the number of revolutions before feeding the formula for
watts. I keep a couple watthour meters around, mounted on boards for
just such measurements. They are cheaper than pointer type wattmeters
and more rugged.
73, Jerry, K0CQ
--
FAQ on WWW: http://www.contesting.com/tentecfaq.htm
Submissions: tentec@contesting.com
Administrative requests: tentec-REQUEST@contesting.com
Problems: owner-tentec@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|