Electrical Engineering - Power fail indicator circuit?

Author

Power fail indicator circuit?

John E.

2007-07-16, 3:25 am

Two 120vac sump pumps do main and backup duty. Each is powered via dedicated
circuit breaker from different phases of 3-phase 208 supply. If either
circuit fails the other pump will continue to function.

I need to flash an LED, or such, if one circuit fails, and to sound an
audible alarm if both fail. A simple relay circuit, or such, will do.
Simplicity is king.

Ideas?

Thanks,
--
John English

John Tserkezis

2007-07-16, 3:25 am

John E. wrote:
> Two 120vac sump pumps do main and backup duty. Each is powered via dedicated
> circuit breaker from different phases of 3-phase 208 supply. If either
> circuit fails the other pump will continue to function.
>
> I need to flash an LED, or such, if one circuit fails, and to sound an
> audible alarm if both fail. A simple relay circuit, or such, will do.
> Simplicity is king.
>
> Ideas?

A neon across the breaker/fuse? There should be more than enough leakage
through the motor to allow for a neon indicator.

If you need more than that, perhaps being a little more creative with an LDR
or some such near the neon, giving isolation, and giving you all sorts of
possibilities to alarms, indication, network, whatever.

--
Linux Registered User # 302622
<http://counter.li.org>

MooseFET

2007-07-16, 3:25 am

On Jul 16, 12:03 am, John Tserkezis
<j...@techniciansyndrome.org.invalid> wrote:
> John E. wrote:
>
>
>
> A neon across the breaker/fuse? There should be more than enough leakage
> through the motor to allow for a neon indicator.

That will only tell you that the breaker is tripped not if the mains
connection drops a phase.

A simple "normally closed" relay running on each pumps will do the
trick. The horn and LED system can be low voltage and battery backed
up to be sure it always works. ASCII art:

+9V------+---------------------------------
! !
\ \ SW2
/ O
\ !
! !
V LED1 !
--- !
! !
+----------O HORN O---------------+
! !
! /
! \
! /
! !
O V LED2
/ SW1 ---
! !
GND GND

If either switch closes, one or the other LED lights. If both do, the
horn blows. The relays would normally be energized by the mains
holding both switches open.

>
> If you need more than that, perhaps being a little more creative with an LDR
> or some such near the neon, giving isolation, and giving you all sorts of
> possibilities to alarms, indication, network, whatever.
>
> --
> Linux Registered User # 302622
> <http://counter.li.org>

JohnR66

2007-07-16, 9:25 am

"John Tserkezis" <jt@techniciansyndrome.org.invalid> wrote in message
news:469b185f$0$10708$afc38c87@news.optusnet.com.au...
> John E. wrote:
>
> A neon across the breaker/fuse? There should be more than enough leakage
> through the motor to allow for a neon indicator.
>
> If you need more than that, perhaps being a little more creative with an
> LDR or some such near the neon, giving isolation, and giving you all sorts
> of possibilities to alarms, indication, network, whatever.
>
> --
> Linux Registered User # 302622
> <http://counter.li.org>

If a circuilt fails you can't assume voltage is available to operate the
neon lamp.

A simple relay that closes when power is lost on that circuit would do. A
battery could operate the buzzer and LED.
John

MooseFET

2007-07-16, 9:25 am

On Jul 16, 5:07 am, "JohnR66" <nos...@att.net> wrote:
> "John Tserkezis" <j...@techniciansyndrome.org.invalid> wrote in message
>
> news:469b185f$0$10708$afc38c87@news.optusnet.com.au...
>
>
>
>
>
>
>
> If a circuilt fails you can't assume voltage is available to operate the
> neon lamp.
>
> A simple relay that closes when power is lost on that circuit would do. A
> battery could operate the buzzer and LED.
> John

If you go the route of the relay, (see my reply), you need two of
them. I just though of perhaps a lower cost way.

Relays can be expensive compared to cheap "wall wart" power supplies.
Assuming you have some of those consider this circuit:

1N400X
DC IN ------+---->!-------+---------------
! ! !
! |/e ---
+---/\/\----! PNP --- HUGE
! R2 !\ Q1 !
\ ----OUT !
/ R1 !
\ !
! !
------+-----------------------------

Assuming the "wall wart" has some leakage, R1 may not be needed.

When the wpoer fails Q1 turns on. This could replace one relay
contact in my suggested circuit. If you flip the 1N400X and use an
NPN, you can replace the other switch.

John E.

2007-07-16, 1:25 pm

> A simple "normally closed" relay running on each pumps will do the
> trick. The horn and LED system can be low voltage and battery backed
> up to be sure it always works.

Looks very simple and reliable.

I'm having trouble locating small 120vac coil relays. Since all these 2
relays will do most of the time is remain open contact, and when closed,
power an LED, I'd like them to be miniature. But all the 120vac relays are
(relatively) huge.

Any suggestions for small line-voltage relays?

Or should I just half-wave rectify the AC and use a dropping resistor (and
fuse!) with a 12vdc relay?

Thanks,
--
John English

John E.

2007-07-16, 5:25 pm

> Use a 12V wall wart and a 12VDC relay. Measure the voltage on the coil
> and if it's really high (more than 10-15%) add a series resistor.

The circuit, as proposed by MoosFET uses 2 relays held open by 120vac.

How do you suggest doing this?

Thanks,
--
John English

Palindrome

2007-07-16, 5:25 pm

John E. wrote:
>
>
> The circuit, as proposed by MoosFET uses 2 relays held open by 120vac.
>
> How do you suggest doing this?

What substitute a plug-in dc supply + dc relay for a supply-voltage ac
relay? That's simple enough.

Personally, I'd go for a "working ok" lamp in preference to a "fault"
lamp, almost every time. In which case a couple of mains small wattage
lamps or neons would do fine.

Similarly, only differently, I'd want to sense the state of the sump, in
preference to sensing whether the pumps have power - for operating the
alarm. Pumps can still fail or block with all the power needed available.

--
Sue

2007-07-16, 5:25 pm

On Mon, 16 Jul 2007 19:55:30 GMT, John E. <incognito@yahoo.com> wrote:

>
>The circuit, as proposed by MoosFET uses 2 relays held open by 120vac.
>
>How do you suggest doing this?
>
>Thanks,

Substitute one wall wart + one low voltage DC relay for one 120VAC
relay (plus R if required)

Multiply each of the above by the number of AC relays.

MooseFET

2007-07-16, 9:25 pm

On Jul 16, 10:24 am, John E. <incogn...@yahoo.com> wrote:
>
> Looks very simple and reliable.
>
> I'm having trouble locating small 120vac coil relays. Since all these 2
> relays will do most of the time is remain open contact, and when closed,
> power an LED, I'd like them to be miniature. But all the 120vac relays are
> (relatively) huge.
>
> Any suggestions for small line-voltage relays?
>
> Or should I just half-wave rectify the AC and use a dropping resistor (and
> fuse!) with a 12vdc relay?

Many DC relays will chatter when you try that.

D1
AC ---/\/\--->!-----+----
! )
X )
! )
---------------------

X could be several uF of capacitor.

For some relays another diode will work for the X

Others have suggested the "wall wart + relay" option. You could also
use the wall wart to charge a huge capacitor so that no batteries
would be needed.

You could also use opto-isolators to keep some transistors turned off.

John E.

2007-07-17, 1:25 pm

> What substitute a plug-in dc supply + dc relay for a supply-voltage ac
> relay? That's simple enough.

I see. Too many mains receptacles required for the limited space. I'm using a
GFCI receptacle for each pump due to their -- by definition -- proximity to
water. By taking a wire from the "load" side connector of the GFCI i can
monitor if these trip.

> Personally, I'd go for a "working ok" lamp in preference to a "fault"
> lamp, almost every time. In which case a couple of mains small wattage
> lamps or neons would do fine.

.... so as to reduce the possibility of a failed indicator not being noticed?
I commend your approach. In my experience, a regularly lit indicator, if
failed or turned off, draws less attention than does one that appears. Maybe
add a "failed indicator" indicator? (c: But if that fails... Oh, the tragedy!

> Similarly, only differently, I'd want to sense the state of the sump, in
> preference to sensing whether the pumps have power - for operating the
> alarm. Pumps can still fail or block with all the power needed available.

Good idea. I'll be adding a level sensor to the sump to trigger the alarm if
the level goes above norms.

Thanks,
--
John English

Palindrome

2007-07-17, 1:25 pm

John E. wrote:
>
>
> I see. Too many mains receptacles required for the limited space. I'm using a
> GFCI receptacle for each pump due to their -- by definition -- proximity to
> water. By taking a wire from the "load" side connector of the GFCI i can
> monitor if these trip.
>
>
>
>
> ... so as to reduce the possibility of a failed indicator not being noticed?
> I commend your approach. In my experience, a regularly lit indicator, if
> failed or turned off, draws less attention than does one that appears. Maybe
> add a "failed indicator" indicator? (c: But if that fails... Oh, the tragedy!

I would agree, however:

One solution is to make the "ok" lamp indispensible. eg it provides
illumination for something that has to be done on a regular basis. Like
read a meter or provide (most of) the room light(s).

Another is to use the light to illuminate a semi-silvered mirror with a
warning notice behind it. You won't see the warning sign whilst the
light is on... it can be very effective..

"If you can read this notice, start praying"... ;)

--
Sue

John E.

2007-07-17, 1:25 pm

> "If you can read this notice, start praying"... ;)

Ha!

I think I have the critical modes covered. The lights are only lit for the
non-critical modes: if one pump's electrics fail, one or the other light goes
on; If both fail, the piezo alarm sounds; if the sump level rises above
normal max, another alarm screams bloody murder. Battery backed-up.

Equipment is to be used only during business hours (and sump is used only
during these hours), so no "dial the manager in the middle of the night" mode
is required.

Thanks,
--
John English

Rich Grise

2007-07-17, 5:25 pm

On Tue, 17 Jul 2007 16:24:37 +0000, John E. wrote:
[someone wrote]

>
> ... so as to reduce the possibility of a failed indicator not being
> noticed? I commend your approach. In my experience, a regularly lit
> indicator, if failed or turned off, draws less attention than does one
> that appears. Maybe add a "failed indicator" indicator? (c: But if that
> fails... Oh, the tragedy!

Just use a green LED for "working", and also power a reed relay or
transistor switch, to turn on a red LED when the green LED goes out,
albeit you might need a teensy bit more logic so you don't get a fault
light when it's simply off.

Have Fun!
Rich

Don Kelly

2007-07-20, 8:25 pm

"JohnR66" <nospam@att.net> wrote in message
news:jcJmi.171641$Sa4.13656@bgtnsc05-news.ops.worldnet.att.net...
> "John Tserkezis" <jt@techniciansyndrome.org.invalid> wrote in message
> news:469b185f$0$10708$afc38c87@news.optusnet.com.au...
>
> If a circuilt fails you can't assume voltage is available to operate the
> neon lamp.
>
> A simple relay that closes when power is lost on that circuit would do. A
> battery could operate the buzzer and LED.
> John

Two relays with two sets of normally closed contacts. one set of contacts of
each relay supplies a lamp (independent power source for lamps-possibly
battery). The other sets are in parallel supplying the hooter (also
independently energised) so that when both are down you get an alarm.
Now it may be cheaper to take this switch logic and do it electronically.
The main thing is that the alarm and light sources be independent of the
motor supply as there is no use shutting down the alarm system when you need
it. The wall wart scheme suggested by MooseFET appears at first glance to be
violating this necessity -one can't turn Q1 on or off if there is no DC
supply from the wall wart and, if it is independent- then how does one sense
the failure of a circuit? Possibly I am missing something?
--

Don Kelly dhky@shawcross.ca
remove the X to answer
----------------------------

MooseFET

2007-07-20, 8:25 pm

On Jul 20, 4:54 pm, "Don Kelly" <d...@shaw.ca> wrote:
> "JohnR66" <nos...@att.net> wrote in message
>
> news:jcJmi.171641$Sa4.13656@bgtnsc05-news.ops.worldnet.att.net...
>
>
>
>
>
>
>
>
>
> Two relays with two sets of normally closed contacts. one set of contacts of
> each relay supplies a lamp (independent power source for lamps-possibly
> battery). The other sets are in parallel supplying the hooter (also
> independently energised) so that when both are down you get an alarm.
> Now it may be cheaper to take this switch logic and do it electronically.
> The main thing is that the alarm and light sources be independent of the
> motor supply as there is no use shutting down the alarm system when you need
> it. The wall wart scheme suggested by MooseFET appears at first glance to be
> violating this necessity -one can't turn Q1 on or off if there is no DC
> supply from the wall wart and, if it is independent- then how does one sense
> the failure of a circuit? Possibly I am missing something?

You only need one set of contacts per relay.

LinkBot