Creating a reliable transistor switch / alternative ideas?

Discussion in 'The Projects Forum' started by rokz_2005, Oct 29, 2013.

  1. rokz_2005

    Thread Starter Member

    May 14, 2011
    68
    0
    Hi All,

    I want to create a simple switch. Wondering if someone can help me out.

    The input voltage varies between 0v to 14v DC (Automated)
    Output - Light bulb which requires 12v app to light up (as required) anything less than the required voltage still lights up the bulb but dimly.
    What I want to do is to restrict the on off state according to voltage supplied.
    Basically I want the light bulb to turn on if the voltage supplied is = or > 11v

    I was thinking of using a transistor with a resistor so that the transistor is active only when 11v or more is applied. This will collect the voltage from the same source and pass it through to the light bulb. And if the voltage drops below 11v then the transistor will switch off?

    Is there a better design? I was thinking of using a relay but I want it to be reliable and longer lasting.

    And what type of components can I use to run this cool or would it better to build it in a heat sink casing?
     
  2. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    1.What is the supply voltage? (12.6V???)

    2.What is the variation of the supply voltage? (12V to 14V????)

    3.How accurate does the trip point have to be? (11.00 to 11.01V???)

    4.If the supply voltage varies, how does that effect the trip point?

    5.How much current is being switched?
     
  3. cornishlad

    Member

    Jul 31, 2013
    196
    25
    I would think the simplest way to control the turn on of a switching transistor is to put a zener diode in series with the voltage feed to the base, In your case a 10.3v zener would be required. You can fiddle with other nearby zener values by adding single small signal silicon diodes. Each one will add 0.7v to the zener voltage. Or locate two lower value zeners such as 2 x 5.1v
    You can use LEDs as well which have different turn on voltages depending on the colour if you have some to hand. You will still need a current limiting resistance in series as well.
    If you need precision though, a single op amp wired as a comparator would be a better bet - and easy to adjust with a pot.
     
  4. rokz_2005

    Thread Starter Member

    May 14, 2011
    68
    0
    The supplied voltage varies between 0v - 14v - I cannot control this.

    The bulb has 2 state.
    1. around 1.3v
    2. around 12v
    As state 1 has a really low voltage input the trip can can be anywhere between 5v to 14v. I said 11 so that if the supplied voltage does go up to 10v then the bulb won't light up. I guess this bit is a flexible.

    All I'm trying to do is get rid of the 1st state so that bulb only comes on when the input is > 11v

    I'm not exactly sure about the current drawn. |It's a 12v 21w bulb calculated to 1.71A
     
    Last edited: Oct 29, 2013
  5. cornishlad

    Member

    Jul 31, 2013
    196
    25
    21/12 = 1.75 amps. I would use a darlington transistor or a darlington pair. then allow for 1.4v base voltage. It doesn't sound like a precision requirement so a zener voltage around 9.6v should suit.
     
  6. #12

    Expert

    Nov 30, 2010
    16,306
    6,815
    Here's the basic concept. Change values to account for the proper current needed and enough amplifier transistors to get that current.
     
  7. BobTPH

    Active Member

    Jun 5, 2013
    783
    114
    I got the impression that the OP wanted to the bulb to snap on fully at 11V and not ever go on dim. The Zener circuit will not accomplish that, when the voltage is a little over 11 it will light diimly. The best way to do what I describe is use a comparator and a voltage reference.

    Bob
     
  8. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Three circuits: Not so good(red trace), Better (blue trace), and Best (green trace).

    I plot the power in the three lamps (as a proxy for brightness) as a function of the input voltage. The Best circuitry dissipates little power, so no heat sinking is required. Not so on the other two, which would require the 3055 to be on a big heatsink.

    If you have a lot of wire resistance between the source and the lamp circuit, it may chatter as the input voltage approaches 11V. If that is a problem, we can add some hysteresis.
     
    absf, #12 and mcgyvr like this.
  9. #12

    Expert

    Nov 30, 2010
    16,306
    6,815
    That's why I called it a basic concept instead of a finished circuit. It shows the concept of using a zener to create a switching point better than any of the previously posted schematics.

    @MikeML
    Very slick use of a 431. :cool:
     
  10. wayneh

    Expert

    Sep 9, 2010
    12,137
    3,054
    +1 Very nice.

    I'm too lazy to read the fine print on the data sheet: How would this TL431-as-comparator work compared to a genuine comparator such as LM339? I mean, in terms of transition sharpness and hysteresis.
     
  11. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Open loop voltage gain of a comparitor= ~200K
    Open loop voltage gain of a TL431= ~1000
    Gain of TL431 and PFET combined = 18000

    No intrinsic hysteresis, but it is easy to add some...
     
    Last edited: Oct 29, 2013
    absf likes this.
  12. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Rather than use a specialty component, you can add schmidt/hysteresis effect with another transistor and two resistors.

    I have no idea if the 1k value is right, it will need to be adjusted to set the amount of hysteresis. Partly depends on the amount of current the lamp draws (low beta of the 2N3055).

    This circuit will also snap switch well even with no zener, so it could save a specialty part there too although it would be a little bit more temperature dependent.

    A small cap across the 2k2 resistor will improve the snap switching, but the cap may not be needed.
    :)
     
    absf likes this.
  13. cornishlad

    Member

    Jul 31, 2013
    196
    25
    he wanted a simple switch using a transistor. some thoughts....

    a 2n3055 has low current gain and unnecessarily high current rating.

    a tip120 darlington has a gain of around 1000 and handles 5 amps. It would need a small heatsink. And very cheap on eBay.

    can't see much difference between the "better" and "best" circuits in MikeML's circuits..maybe that's just me. TIP120 could be better again ?

    Can't see any need for hysteresis in that circumstance.
     
  14. rokz_2005

    Thread Starter Member

    May 14, 2011
    68
    0
    OK So I've decided to try out MIKEML,s best method. going to buy the components asap. Will let you know how it goes :)

    1/4 watt resistors ok?

    Thanks all for your valued input :)
     
    Last edited: Nov 4, 2013
  15. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Yep. You might as well order five or ten TL/LM431. I use them for lots of things. Very handy little IC chip...
     
  16. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I have an idea that rokz_2005 isn't giving us the whole story...

    Is this for a taillight on a vehicle, like a 1156 bulb? Because reasonably modern vehicles use PWM on the lamps to not only control brightness, but to prolong the filament life AND to provide lamp failure information to the vehicle's computer system. If you monkey around with the lighting circuit, you're not only jeopardizing your own safety, but the lives of your passengers and those around you on the highways - along with the likely possibility of frying some pretty expensive modules in the vehicle.

    If you ARE trying to modify a vehicle's lighting, I strongly suggest that you abandon that effort, and restore/have it restored to the same condition as when delivered by the dealer.
     
  17. rokz_2005

    Thread Starter Member

    May 14, 2011
    68
    0

    You're kind of right but nothing to do with 'something I'm not telling you guys'

    I bough a upgraded set of rear cluster for my car. Previously the bulbs (That I'm trying to apply the circuit to) came on dim as it was that design. Since upgrading the bulb still lights up dim but does not suit the new set up. The bulbs etc are compatible with the old and new cluster. the difference is the design and I'm trying to change it to so that it looks the same as the newer model.
    It does not throw any error etc

    Is it still a problem?
     
Loading...