From: Winfield Hill
Subject: Re: Watt Meter Project....If your interested.
Date: 21 Dec 2002 06:49:25 -0800
Organization: Rowland Institute
X-Newsreader: Direct Read News 4.20
Gary Lecomte wrote...
> I have just put a good Watt-meter design on my site. It is also
> capable of measuring wattage into complex loads! And its quite
> simple to make.
> In the Projects Area at: http://www3.telus.net/chemelec
Gary, the LM394 has NPN rather than PNP transistors. And one of
your diodes is backwards, damaging to the LM394's beta. :>) I
could see this at a glance, but it also squares with NSC's AN222
fig 12, one place where the circuit was published many years ago.
It's a clever circuit. Bob Pease credits it to Carl Nelson. Hmm,
I have a dim memory of saving an ED article long ago. Also Bob had
a few ED columns earlier this year on making current shunts and on
making expanded versions of Carl's wattmeter circuit.
. | |
. V ~ Iac c c
. ,------ b b --,
. | e--+--e |
. ---+--/\/\----- | -------+----
. AC I ~ Vac AC
. LINE | LOAD
Many engineers know that BJT transistor long-tail-pair differential
amplifiers suffer from serious gain-dropping distortion for inputs
above 25mV or so. This is implied by the familiar Ebers-Moll
equation, I = Is e^Vbe/VT, and results in a tanh() gain function.
Bob and Carl solve this problem by keeping the signal voltage low,
15mV full scale in Bob's case. But this results in a weak output
signal, hence the specified 50uA meter.
As an alternate solution, NSC offers the LM13700, which features a
pair of linearizing diodes that allows the use of higher signals,
such as from a conventional 50 or 100mV current shunt.