Min Voltage to turn on curcuit

9.1 volt zener from radio shaft used to feed a Resistor of lets say 680 ohms to the base of a common 2n2222 or some larger type transistor. one end of the R goes to the base, the other end goes to the anode of the Zener Diode, the cathode of the diode comes from your 6..12 volt source. the collector will connect to one side of the relay coil, the emitter will go to common/ground. the other side of the relay coil will connect to the same location as the Cathode side of the zener to obtain power for the relay unless you have a constant voltage some where else? in any case, the Zener will turn off any biasing to the transistor that will release the coil of the relay and not allow any noticeable amount of current to flow. also you need a diode across the relay coil to keep flyback effects from damaging the transistor and other things in your voltage source. so what you have is a a circuit with 1 zener, 1 common low power silicone diode, 1 resistor and 1 transistor.. that should do it.

The device normally used to make such decisions is called a comparator. They have two voltage inputs labeled + and -. When the one labeled - is more positive than the one labeled +, the output turns on and pulls down to the negative supply rail. However, to make a comparator based decision like you need, you also need some sort of voltage reference to compare the supply voltage to, and a switching device able to boost the few milliampere output current to a current capable of driving the relay coil (65 ohms / 10 V =154mA in this case).

Here is the kind of thing I am talking about: (View with font set to something fixed width per character, like Courier)

+-----+-------+------+--------+---+--- + | | | / | | | | 22k | .-. o o | | | .-. | | | |( - 1N4148 z | | | | | |( ^ A | | | '-' |(_ | | '-' | |10k +--+ | | | | | | |\\|LM393 | | +-----|-----|-\\ | |< | | | >-----+-------| 2N4401 | +--+--|+/ | |\\ | | .-. |/| | | | | | | | | | .-. .-.| | | | | | | | |'-' | | | | | | | |220k| | | '-' '-' +----|------+ | |10k |10k | | +-----+-------+----------------+------ -

(created by AACircuit v1.28.4 beta 13/12/04

For all voltages lower than 10, the zener diode will drop more voltage than the 22k resistor, but for voltages higher than 11, the 22k will drop more. So for voltages higher than 11, the - input of the comparator will be more positive than the + input and the output of the comparator (actually, the LM393 contains 3 comparators, so you van make 2 of these with one chip) will pull down to the - supply rail, turning the 2N4401 transistor on, and energizing the relay. The diode across the coil prevents a large voltage spike when the comparator turns off. The 200k feedback resistor lowers the switching threshold voltage once the relay is on, to make sure the decision doesn't chatter.

Hello Ducky,

Look at the Texas TL431. It comes in SOT23, is quite precise and cheap. Unless it's a really big relay you need two resistors to set the trip voltage, plus a flyback diode and maybe other stuff to protect it from the relay.

Regards, Joerg

Hello Ducky,

You might also need some hysteresis to prevent the relay from chattering when the voltage is near the cut-off. Unless it's a rapid enough rise and fall.

Regards, Joerg

Exactly. You need a relay driver after the TL431. Here's a diagram:

• Vin ------------+----------+-------------+ | | | R1 33K R3 220 e | | / | +----R4-----b| 2N4403 | | 3.3K \ | ----- c +---------/ \ TL431 | | --- +------+ | | | | | | ----- R | | / \ L R2 4.7K | --- Y | | | | Gnd -------------+----------+-------------+------+

This will turn the transistor on at ~ 10 V. Whether the relay will energize is another issue, as we don't know its pull in current. Darn thing draws ~ 200 mA at 12 volts. Ed

Looks like you have the values of R3/R4 interchanged.

Hello Ed,

Watch out for excess base current as you don't limit it anywhere. The TL431 can draw quite a bit.

Also, you'd need a little hysteresis in there.

Regards, Joerg

I bet it has to be simple...

Lessee- could it be that the 6V "idle voltage" means you have an identical relay in series? Whatever- be sure you bypass all relays with supressor diodes. One way to do it with garden variety discretes: View in a fixed-width font such as Courier. . . . input >--+------+-------------+ . voltage | | | . | | | . [1k] | 2n4403 | . | | e . | | |/ . +------|---[560]---| . | | |\\ . | | c . | | | . | [560] +---------+--------------------+ . | | | | | . | | | | | . | 10v - | | | . | zener^ [22k] - 9v | . | | | ^ zener + | . c | | | +-----+ . \\| | | | 1n4001 V |relay| . |----+---+--[1k]---+---|

The Vref for the TL431 is 2.5V, and not 1.24V, so that you have R3=R2*(Vtrip/Vref-1), neglecting leakage thru relay due to feedback resistor, for Vtrip=10V, and R2=3.3k, R3= 3.3*(10/2.5-1)=10k. Then if you want this condition to sustain itself to 9.5V or so, you would want R3=3.3*(9.5/2.5-1)=9.25k so that an R5 feedback such that R5||R2=9.25k would be R5~100k. Also, at 12V, you want Ic/Ib=10 for low loss and clean switching. This makes R4=(12-0.8-2.5)/20m=470 ohms. The circuit should look more like this:

View in a fixed-width font such as Courier.

. . . + Vin ------------+----------+-------------+ . | | | . R1 10k R3 3.3k e . | | / . | +----R4-----b| 2N4403 . | | 470 \\ . | | c . | | | . | | | . +----------|----R5-------+ . | | 100k | . | | | . | | | . | ----- +------+ . +---------/ \\ TL431 | | . | --- ----- R . | | / \\ L . R2 3.3k | --- Y . | | | | . Gnd -------------+----------+-------------+------+ . .

Obviously the R3 in the calcs refers to R1. Make the real R3 anything,

3.3k makes it the same as R2.

This one works well enough with fewer components. Multiplied Vbe of T2 is used for cutoff threshold:

View in a fixed-width font such as Courier.

. . . . input >-+-----+------+-------------+ . voltage | | | | . | | | | . | [1k] [560] 2n4403 | . | | | e . | | | |/ . | +------|---[560]---| T1 . | | | |\\ . | | | c . | | | | . | | | +---------+----------+ . | | | | | | . === | | | | | . 0.47u | 10v - | | + | . | | zener^ [15k] | +-----+ . | | | | - V |relay| . | c | | 1n4001 ^ R | | . | \\| | | | +-----+ . | T2 |----+---+ | - | . | /| | | --- . | e | | /// . | | 2n4401 [1k] | . | | | | . +-----+----------+---------+ . | . --- . /// .

Hello Ed,

Sorry, I missed R4 so you do have base current limiting. However, 3.3K might be a bit much, the beta of the transistor may not be sufficient to drive a large relay.

Regards, Joerg

Hysteresis is an excellent point - I didn't think of it at all. But current is limited - cathode-anode current is limited to ~ 54ma by the 220 R3. The 2N4403 base has a 3.3k limiter about 3 mA. And the input pin on the TL431 is limited to 300 uA

But see Fred's post below yours. He spotted a major error - I computed with Vref of ~1.25, but it is 2.5. That changes the values of the R1/R2 divider - 10K would work for R1 with 3.3K for R2.

Ed

Thanks! Yup - I was thinking LM317 Vref 1.25. It is 2.5 for the TL431, as you said. So R1 should be 10K, R2 3.3K.

And the hysteresis was something I completely overlooked.

Regarding driving the transistor into saturation, you are right once again. It will work without that, but it is better with your design. Bottom line, I *like* your much better design. Thanks again.

Ed

Then if

Hi Joerg,

That's a good point. Fred has a better solution that drives the transistor into saturation using a much smaller base R. I computed the base drive 10/3300 at 3 mA, and the hfe for the transistor is 100 min at 150 mA IC, so it will work, but not as crisply as a harder drive, and it wastes more power (Ic*Vce) than when it is saturated with a lower vce. Fred's version includes the hysteresis you mentioned, too.

Ed