# 95 VAC to 120 VAC Treshold Detector

Discussion in 'Electronic Components' started by Dr. Phibes, Feb 19, 2007.

1. ### Dr. PhibesGuest

I want to trigger a relay so that when the input voltage is at 95 VAC (134 V
peak), the relay is off. When the input voltage is at a full 120 VAC (170 V
peak), or it goes beyond the 95 VAC threshold, then the relay is energized.

This circuit needs to be discrete enough to where I can put it on a small PC
board inside of an enclosure. So I can't be using hulky 120 VAC relays or
anything like that. The only source voltage is the input voltage that
switches between 95 and 120 VAC. The relay will be switching on a seperate
circuit that uses less than an amp of current (a series of LEDs). Thus,
another reason why the relay doesn't have to be hulking.

I'm a newbie to this board, so I hope someone out there can help.

2. ### LostgallifreyanGuest

I've not thought through the control system, but two ways you can detect
that voltage (or something close) are a chain of three diacs, or a small
neon lamp. In either case you'll need a gain stage, probably a triac, to do
the switching. As either the diacs or the neon will conduct at beyond
threshold, no other comapator method is needed. You'll need to limit the
current through either the diacs or the neon.

I might have missed something, so don't rely on this post.

3. ### LostgallifreyanGuest

Comparator...

4. ### EeyoreGuest

Does the switched load have to be via a mechanical relay ? Is the load even
connected to the same ac supply ?

Graham

5. ### Mark AitchisonGuest

1. If you are new to high voltage wiring get some help from a local
expert, if only to check over the wiring before plugging it in.

2. The approach you need to take is to rectify the AC with a
current-limiting resistor in series, smooth it, shove it through a 150
volt zener, and drive something like the base of a high voltage
transistor via a safe variable resistor (you really need expert help in
knowing what variable resistors are safe in this context), then use that
transistor to trigger an opto-isolator or sensitive relay.

6. ### Dr. PhibesGuest

Does the switched load have to be via a mechanical relay ? Is the load
even
No, it doesn't have to use a mechanical switch like a relay. Yes, the load
is connected to the same power supply. Here's the jyst of it

I'm replacing regular incendescent stage lights with home made LED lamps. I
have a light controller that controls the various color lamps. Since the
LEDs are much more sensitive to voltage and current than regular lights are,
I've found out that the controller actually puts out 90-something volts when
the lamp should be "off". This works fine with regular lights. But with
LED lamps, the lamps go dim rather than switch completely off. Thus, I'm
trying to come up with some sort of a gate circuit that will detect the
undervoltage and electrically (or mechanically) cut the LED string off.

The LED string simply consists of a string of LEDs connected in series with
a resistor in front and in back of the circuit, then connected directly to
the light controller that puts out 120 VAC when "on". And anywhere from
75-90VAC when it's supposed to be "off".

I decided to put a switching transistor in the circuit, that would switch
the LEDs on and off. But the problem I ran into with that was I couldn't
figure out how to bias the base of the transistor in relation to the voltage
change between 90 VAC and 120 VAC.

I really appreciate everyone's help! I love this kind of stuff, and I know
I can learn a lot too!

7. ### EeyoreGuest

That's not tricky ! Can do that fine.

It seems to me that your controller is the problem though since it doesn't
really switch off. Is there any way to modify that ?

Graham

8. ### Dr. PhibesGuest

It seems to me that your controller is the problem though since it doesn't
Yes, it's definitely the problem. It's a DMX light controller, very commonly
used for stage light controlling. Definitely not made to drive LED lights.

9. ### Dr. PhibesGuest

Sorry, I didn't fully answer your question. No, there's no way I know of to
modify the DMX controller.

10. ### Peter BennettGuest

I wonder if the light controller is getting confused by the
characteristics of the LED light. Have you tried putting a normal
incandescent lamp in parallel with the LED lamp? If things work that
way, you could then use a suitable resistor in place of the
incandescent lamp (although this would increase the power consumption
somewhat.)

--
Peter Bennett, VE7CEI Vancouver BC, Canada
peterbb4 (at) interchange.ubc.ca
new newsgroup users info : http://vancouver-webpages.com/nnq
GPS and NMEA info: http://vancouver-webpages.com/peter

11. ### Dr. PhibesGuest

I wonder if the light controller is getting confused by the
You have a VERY good point. I'm going to try that. Hopefully, it's really
that simple.

12. ### ehsjrGuest

I made safety the top priority for this circuit, so it uses
a very small transformer and an opto triac driver. This will
do it for you if you have room on your PC board for the parts.
The transformer needs a bit less than 1" sq and 1" high. The
opto is a 6 pin dip, the TL431 is a TO-92. The trimpot is 5K.

Opto = Sharp PC3SD12NTZAF Mouser # 852-PC3SD12NTZAF
TL431 - Mouser #512-LM431ACZ
Transformer Mouser # 823-BV020-5384

+------------------------------------------------+
| |
| ----- D1+ |
AC-+-|Trans|-->|--+----+------+---[1.5K]---+ |
| | |+ |+ | | 1 | 4
| 12V | [D2][470uF] | [Opto] [OptoOut]
| | | | | | 2 | 6
AC-+-| |------+----+ P --- |
| ----- | 5K O<--[1.5K]--/_\TL431 |
| | T | |
| | | | [LEDCircuit]
| +------+------------+ |
| |
+------------------------------------------------+

Here's a drawing of the opto:
----
1| |6 Pin 1 is + input
2|Opto|. Pin 2 is - input
3| |4 Pins 4 & 6 is nonpolarized output
---- There is no pin 5

The opto specified normally drives a triac, but in
this circuit driving LEDs the current is assumed to
be low enough that it can be used directly. It will
handle 100 mA on the output side, and your LEDs should
be configured to run at less than 30 mA. If your LEDs
draw appreciably more, a different output circuit is
needed.

The TL431 acts like an adjustable zener. When it "sees"
adjust it to turn off at ~ 95 volts. You said you had
a control signal that could be either 95V or 120V, so
no hysterisis is needed and you can make the adjustment
using your control signal. Set it ~half way between the
point where the LEDs go out at 95 volts and turn on at
120 volts.

How it works:
The voltage across the 10 uF cap will be ~ 16.3V when the
line voltage is 120. When line voltage drops to 95, the
voltage across the cap will be about 12.8V. The small
5K trimpot is a voltage divider which sets the voltage
at the TL431 adj pin. Say it happens to be set to provide
2.77V to the adj pin when the control signal is 120.
That will turn the TL431 on, which in turn makes the
opto output conduct. When the control signal drops to
95 volts, the voltage at the adj pin will be 2.17 volts,
and the TL431 will turn off, which turns the opto off.

Ed

13. ### Dr. PhibesGuest

THIS IS AWESOME! I can't tell you how much I appreciate this! Especially
in that you took the time to explain it.

Again, I appreciate EVERYONE'S input. I've already learned a lot!

I definitely have my work cut out for me now.

14. ### Homer J SimpsonGuest

Are these back to back LEDs or do you use a diode in series?

Your suggested solution seems very ass backwards - there must be a better
way.

BTW, why are you using LED lamps for this? Seems odd.

--
..

--
..
..
..
..
..
..
..
..

--

15. ### Dr. PhibesGuest

Are these back to back LEDs or do you use a diode in series?

They're LEDs wired in series with each other.

--|<---|<---|<---... and so on.
I'm trying to replace my conventional stage lights with the much lighter,
much more power efficient LED lamps.

16. ### LostgallifreyanGuest

Not to me. I remember when I was eight, that there was a big similarity
between stage lighting, and traffic lights. Anything that makes an
efficient narrow projection is useful, and LED's are particularly good at
that, as their use in traffic lights shows.

17. ### Homer J SimpsonGuest

But are they? ISTM that they still produce fewer lumens/watt than
incandescents - or so I've been told.

Do they really dim as well? And if so, isn't the curve very different?

--
..

--
..
..
..
..
..
..
..
..

--

18. ### Mark AitchisonGuest

I see. The answer may be a lot simpler than I suggested before. The
problem seems to be that the controller puts out a bit of leakage
current when it is supposed to be off - enough to get a glow from the
sensitive (and very non-linear current/voltage curve) LEDs yet not
enough to light a normal bulb.

There are two solutions (a mixture of the two might be the best for
efficiency). One is to put some load ACROSS the LEDs to soak up some of
that extra current, enough to bring the voltage down to less than the
threshold for light emission (about 2.something per LED in series);
ideally a load resistor that has a lower resistance at low voltages -
such as a PTC thermistor of the correct value (expensive/rare) or a
light bulb (cheap, common, and somebody else suggested that).

The bulb in parallel will drop the efficiency somewhat. Another
solution is to put a zener diode in series, and I suggested this before
with a whole lot of other circuitry that would have been needed if the
load wasn't LEDs. As it is, you will probably need two zeners in
series, back-to-back, because I presume your controller is putting out
AC and your rectification for the LEDs is taking place within the LED
unit? (If we are talking about DC then you only need one zener). So
long as the current to your LEDs is under about 0.2A RMS I'd start by
trying a pair of 3 Watt 15 Volt zeners back-to-back in series with the
LEDs, and add a low power 120V bulb (say a 5 Watt pilot bulb) in
parallel if it still doesn't switch off enough. (Just the bulb by
itself might work, and would be cheaper, but less energy-efficient).

Mark A

19. ### Franc ZabkarGuest

http://en.wikipedia.org/wiki/Digital_Multiplex

It seems that your controller uses RS-485 to send a digital value to
an intelligent dimmer. If the output of this dimmer is merely chopping
the AC by means of a TRIAC, then the idea of a 95VAC threshold makes
no sense.

- Franc Zabkar

20. ### Franc ZabkarGuest

You need to wire the LEDs in anti-parallel, otherwise they will only
light during each alternate half cycle.

|--|>|--| |--|>|--|
+--| |--- / ---| |--- R ---+
| |--|<|--| |--|<|--| |
| |
AC o--| +-- C --o AC
| |
| |--|>|--| |--|>|--| |
+--| |--- / ---| |--- R ---+
|--|<|--| |--|<|--|

- Franc Zabkar