Power Resistor for Audio Measurement
Question:
this is Elizma from South Africa. I
need your advice on some audio measurement aspects.
What type of resistors do you use when
making measurements acroos 1ohm ,2ohm ,4ohm and 8ohm constant loads?.
Do you always measure across speaker loads or do you use high power
resistors? Please help me on this subject.
How many Amperes travel through a
speaker. Can one use a constant resistor to substitute a speaker?
perhaps you can give me the part number or company name for resistors
that you know, works well.
Answer:
the quality of resistors for
audio power measurements, it is not so very important. Problems occurs
with many types of wire wounded resistors, they change their complex
impedance with increasing frequencies. My experience, for DC up to
about 10 kHz they remain like an pure ohm's resistor, above they tend
to be more and more inductive, much higher they reach their resonance
frequency and the capacitive part dominates.
When the resistors changes the U/I phase
for only some degrees to be an inductor - no problem - the amp see's a
higher impedance and the output current decrease a little. The question
is: "how precise you want to measure?" Doing a device data sheet
specification you'll should use good quality. For test purpose only not
so important. If
the amps works now with 4 ohms or 4.6 ohms at e.g. 55 kHz and 100
Watts, it doesn't matter. The amp wont reach that power level in
practice under that frequency.
I use for my 4 ohm measurement many times
this
power resistor. I guess and hope this type use a bifilar
winding (not sure) resulting in being ohmic up to almost 100 kHz.
Mounted on a big aluminium block. If you are not sure with your own
wire wounded resistors, check them easily with a sine generator and a
two channel oscilloscope, connect an additional series resistor and the
power resistor as a voltage divider and observe the phase shift with
increasing frequency. Do a second test with 1/4 watts film resistors to
verify the measurement before, there should be almost no phase shift.
Don't use to high ohmic resistors for this test, because - consider the
parallel capacitance of the scope probe could also create a lowpass
with the series resistor resulting in a small phase shift error at
higher frequencies. Sorry I can't recommend a specific manufactorer for
a wire wounded power resistor, I haven't test enough different types of
them. The one I got was good enough for me and better than my old ones,
so I stopped searching.
What would be also a cheap but very good solution?
Build your own resistor!
do a parallel
connection of a lot of e.g. 1 Watt or 2 Watt film resistors
until the wished maximum power level reaches. The quality will be very
excellent with a much better high frequency behaviour than any wire
wounded resistor. No matter if carbon, metal film or metaloxid
resistors, to do power measurement it's very ok. Use a two plate
construction, a thin copper or brass metal sheet, one for each pole. Do
drills in every sheet and solder the resistors close as possible on the
metal sheet. The advantage of this construction, the metal sheet acts
as very low inductance conductor, like a ground plane. Second, the
metal sheets acting as a heat sink. Third, the resitors having only
short remaining leads - the heat in the resistors can flow easily in
the metal sheets. If you are tired of paralleling so many resistors,
why not use a fan to keep the less one at life. Test the frequency
response. Things like temperature cofficient not so important
for these tests. I prefer the method of paralleling film resistors to
build a cheap but very good load resistor. use an old electric cooker.
The electric
plates are good loads too (their wires are not so extremly long
inside). The load can be easily varied with the rotary swith. And the
plates - believe -the amplifier really must be very big to destroy the
plates. I used them also for my experiments, but never measured their
high frequency characteristics.
What would be a expensive
solution?
build uoyour own power resistor e.g. with
paralleled expensive bulk foil resistors in a TO-220 package, but for
power measurements - nonsense not necessary - OK!, if you have the
bugdet and the need for a commercial data sheet specification - why
not.
Do I measure many times
with a real speaker?
No first,
I hate the loud noise of a sine - OK, a wobbling sine sounds a little
more interested than a single frequency. Specially at 3 to 4 kHz
somebody could get mad about the loud sound, in this frequency range
the human ear is very sensitive. Or at high frequencies may be all cats
and dogs running away - I don't know.Yes, I
could do sine sweeps with a lower amplitude to prevent noise, but in
most cases the amplifier copes very good with real speakers -
boring.the
measurements are not comparable, only with the same type of speaker.the
speaker can be damaged, special at high amplitudes - not all speakers
are designed for high continous levels of power. You don't
know where are the real limits of your speaker pair. You have only two
trials - haha. You don't know the increasing temperature on the coil
without checking removing from the chassis.Yes, a
nice characteristic of a real speaker is the effect of the inducted
voltage by it's own movement, to see the effects on the amplifier. But
there are many other tests to see how the amplifier will behaviour with
a difficult speaker.
Which
other loads are good to test an amplifier?
using
a parallel capacitance with a ohmic load. A capacitance on an
amplifiers output - that's poison for the amplifier, they don't like
that. The phase shift and the increasing currents at high frquencies is
a demanding load for the amplifiers control loop. But be careful - the
capacitance can cause the amp to oscillate, some "real" nanofarads
could be already enough for a audio amplifier. My opinion, a poor amp
has also problems to drive a capacitance. Use for that test high
quality capacitors (no electroly, tantals or poor foils) good are
ceramics, styroflex or small foils in parallel. General, better many
small ones parallel than a big one. Start with low amplitude and
frequency and increase slowly and always check the heat dissipation in
the capacitor - a power amp can kill the capacitor. Special bad
capacitors getting hot very fast due their higher ESR.
inductive loads are not a big problem as capacitors.
if an
amplifier can cope with the inducted voltage of an speaker? - check the
amplifiers bandwidth, slew rate and the dynamic output resistance
(damping factor). After measuring them - the results tells you if the
inducted voltage will be a problem or not.
test
one or two channels at the same time, more stress for the power supply
and increased heat.
Speaker
Impedance and currents?
hard
to say in general. Check different speaker manufactorer datasheets for
the impedance vs. frequency curves. Adjust a load current same as the
low impedance under the specified frequency and check the amplifier for
errors and distortion. Get the scenario more worse by increasing power
and adding a capacitance and/or inductance. Do such tests with
different frequencies and levels (also low amplitudes). It's in your
fantasy which additional stress applying to the amp. But please keep in
mind, such stress can easily damage the amplifier. You never know the
real limit. You don't know if the designer has protected the amp
against all bad conditions e.g. short circuit currents, clipping,
oscillation, a high applied induced voltage, or over temperature? These
tests I hate to do with stranger amplifiers, you never know what
happens - not enough stress, the limits can't be seen easily - Too much
stress - much work finding and repairing the errors of the property
others, destroyed by yourself - no thank you. With your own electronic
- oh yeah - it's fun to reach the limits, either easy to repair or
knowing no damage occurs because of save circiut desing.
