Choke AC signal - inductor?
afeu...@hotmail.com
We have a switchable signal source. With the switch in position 1
there is an AC signal that fluctuates from 0V to 12V at about 60Hz (it
does not reverse polarity). If the switch is in position 2 the signal
is 12VDC constant.
The signal is connected to a standard electromagnetic relay. We would
like for the relay to turn on ONLY when the 12VDC signal is applied
and NOT when the AC signal is applied. Currently, the relay reacts
fast enough that when the 60Hz signal is applied, it will switch on
and off at 60Hz.
Will this work with a suitable sized inductor in series with the relay
coil? If so, how to calculate the proper size? If not, what is the
best method?
Thank you in advance...
Tom Bruhns
The series inductor may or may not work. The AC signal you describe
has some DC component; if the waveform is sinusoidal, for example, or
if it alternates between 0 and +12 with equal "on" and "off" times,
the DC component is 6 volts, and if that is enough to pull in the
relay, an inductance alone won't do the job. I can imagine a circuit
that should be able to do it, if the inductor alone is not enough: an
inductor in series between the switched source and the relay coil, a
capacitor across the relay coil, and a diode in parallel with the
inductor, cathode to the switched source side. That way, the
capacitor across the relay coil will be discharged each time the input
voltage goes to zero, and depending on the part values, you can keep
the relay coil voltage from ever getting very large. Instead of an
expensive inductor, you could probably use just a series resistor
instead, assuming it's OK to put a somewhat heavier load on the input
AC signal and that the relay will pull in well below 12 volts.
Assuming a relay coil of 1000 ohms and a maximum (high temperature is
usually worst-case) pull-in voltage of 9V, I'd suggest a 100 ohm
resistor, and something like a 470uF 16V electrolytic capacitor, and a
1N4001 (or 4002 or 4004 or...) diode. The 42 millisecond time
constant of the resistor+coil and the capacitor should do a decent job
of killing the 60Hz on the coil. If the 60Hz source is very "stiff,"
it may be advisable to put a small resistance in series with the diode
to control peak currents.
Cheers,
Tom
afeu...@hotmail.com
Hi Tom,
Wow! Thanks for the great ideas! I think your method is way better
than the inductor alone...
A few questions though:
1) If the source never goes to zero, but instead went from say 4V to
12VDC would this still work?
2) The relay coil (automotive relay) impedance is about 100 ohms,
what formula did you use to calculate R and C?
Thanks again!
Tam/WB2TT
<afeu...@hotmail.com> wrote in message
news:281200f2-01b3-40c3...@m34g2000hsb.googlegroups.com...
Connect a retriggerable monopulser to the source. The time constant should
be longer than the cycle time of your AC. So long as there is AC present the
MP output will be a constant 1. When the source switches to DC the MP output
will revert to a 0 within a time constant. Operate the relay when the MP
output is a 0.
You don't even need a "real " MP; just a diode, capacitor, resistor and FET.
Tam
Fred Bloggs
There is no best way because most people don't work themselves into a
situation like this. Re-select the switch to be a DPDT:
View in a fixed-width font such as Courier.
.
.
.
. DPDT SW
. --------------
. | |
. | 1a |
. Fluct------------- o |
. | \ coma |
. | o -------> some other crap
. | |
. 12VDC-+----------- o |
. | | 2a |
. | | |
. | | |
. | | 1b |
. | | o | .-|<|----.
. | | \ comb | | |
. | | o -------+-relay--+
. | | | |
. '----------- o | |
. | 2b | ----
. | | ///
. --------------
.
.
Tom Bruhns
Well, to do things right, you need to characterize the relay a bit
better. At what voltage does it pull in, minimum? (This will set the
max you can allow across the relay coil to insure it stays in the de-
energized state.) At what voltage is it guaranteed to pull in? (This
will set the maximum series resistance you can allow.) The R-C time
constant should be long compared with the cycle time of the pulsating
AC. R is the parallel combination of the relay coil and the resistor,
and C is just whatever size capacitor you use. R*C ought to be at
least a couple times the cycle time of the slowest wave-form. The
value of the resistor itself must be small enough so that it provides
enough current to let the relay pull in reliably: for example, if the
relay pulls in reliably at 8 volts, and you can guarantee that you
supply 12V DC minimum in the DC state, and the relay coil is max 100
ohms (when hot: worst case), the resistor must be no more than 50
ohms: 8V = 12V* 100/(100+50). 47 ohms is a common value that should
work in such a case. Power dissipation in the resistor would be R*I^2
= R*(Vin/(R+CoilR))^2 = 50*(12/150)^2 = 0.32 watts; in this case, I'd
recommend using at very least a 1/2 watt resistor, but 1 watt would be
better, to allow for higher excitation voltage and lower relay coil
resistance than I used there. Then 47 ohms in parallel with 100 ohms
is about 30 ohms, and to get an R*C time constant of, say, 50
milliseconds (about 3 cycles of your ~60Hz waveform), you'd need
1500uF or so--2200uF would be a common value, easy to find, but
starting to get physically big.
Use Tam's way by adding an FET for a more efficient circuit and one
that can handle the 4-12V pulsating signal more easily. Here's a
variation on that theme: N-channel FET, source to common, drain to
"bottom" of relay coil. Other side of relay coil to DC/pulsating AC
source. Capacitor from FET gate to common. Resistor from FET gate to
common. Another resistor from FET gate to the DC/pulsating AC
source. Adjust the ratio of the resistors so that with the DC input,
the FET gate is positive enough to turn it on, and with the pulsating
AC, it's not. R*C, as above, should be longer than the cycle time of
the slowest pulsating wave form. You could also get fancy and rectify
the pulsating AC part of the waveform to get a negative bias to
_really_ keep the FET off, but probably don't need to go to that
trouble. Big advantages: using the FET, you can get by with a much
smaller capacitor value, since the gate current is practically zero
and you can then use high value resistances; the FET will turn on with
very low resistance, so the relay coil gets very nearly the full
effect of the 12V DC, and with the pulsating AC, it's very "off" so
the relay coil won't have any significant residual voltage across it.
Also, this way doesn't significantly load the pulsating AC source, as
the diode could in my initial suggested circuit. The relatively
stable and narrow range of gate voltage between "on" and "off" for a
given part is another plus; I assume this is one-off and you can
select the resistors to match the particular FET's gate turn-on
voltage. It's a simple circuit, using inexpensive small parts that
should be quite reliable. One thing to beware about: any possibility
of excess voltage on the FET gate.
Cheers,
Tom