QUESTIONS & ANSWERS
■ FIGURE 7
automotive audio amplifiers. I would
like to know if you have any literature
at hand that you could send to me
detailing the different classes of
amplifiers (A/B/D/Push Pull/etc.).
— Tyre Bernard Daniely
inductor, L3, is needed to reduce the
DC current through the speaker. This is
not efficient or high fidelity, but a lot of
engineering went into making it cheap.
Figure 6 is a push-pull, class A
power stage. The advantage of push-
AAn audio amplifier consists
of: a linear gain stage,
power stage, and transformer.
The linear gain stage may
contain a volume control and bandwidth (tone) controls. Some amplifiers
for band instruments have (yuck) “fuzz”
and reverberation controls; I won’t
consider those here. If the signal source
is low level — like a turntable with
magnetic pickup — a preamp is needed; I am not going to cover that either.
Consider the classes of amplifier:
A, B, A/B, C, and D.
Class A could be a single transistor
or two in push-pull (push-pull will
be explained below). The defining
parameter is that the current flows all
the time in a more or less linear
manner shown in Figure 5. This circuit
was used for many years in Delco auto
radios. The driver circuit was a module
that was laser trimmed to provide a
stable 0.7 volts DC to the DS501
power output transistor. If you calculate the emitter current of the DS501,
it is about one amp average DC. The
In the November ‘07 issue of Nuts & Volts under “Parallel Transistors,”
you stated in your response to Daniel’s question that, in your experience,
MOSFETs cannot be paralleled, while it’s one of their great benefits in
omparison with other (bipolar) semiconductors.
All others have negative resistance/junction temperature curve
(meaning when you heat NPN/PNP/IGBT, “resistance” goes down; the “hot”
device draws even more current, and it goes on quickly till its soul (smoke)
leaves its body. By the way, it proves that semiconductors work on smoke
and when you let it out they stop working. MOSFETs are quite the opposite;
junction temperature rises causing Rdson to go up; effectively compensating
power distribution with paralleled devices.
What is commonly overlooked in such attempts is that when
you parallel two (or more) MOSFETs, the input capacitance multiplies
effectively, extending the on/off switching time.
Even if your driver circuit has been declared as 5,000 pF and one
MOSFET as 1,500 pF, it does NOT mean that you can assume that, with no
changes in driver circuit, you can drive two MOSFETs with 3,000 pF, as the
on/off slope (time to switch) will significantly go up (will double). As the
most dissipation comes from the time the unfortunate device is trying to
move from one state to another, doubling the time can have disastrous
effects; not because the MOSFETs cannot handle the current, but because
you asked them to spend too much time in the linear region.
If you get time (and parts to spare), try to parallel MOSFETs while
beefing up your driver circuit.
Response: What you say makes a lot of sense, but I had separate
three amp drivers for each power MOSFET and still the transistors were
exploding. I am still looking for an explanation of what I was doing wrong.
March 2008 29