Problems in a transistor switching circuit with a high pitch sound

Discussion in 'The Projects Forum' started by c1rcu1ts, Dec 11, 2014.

  1. c1rcu1ts

    Thread Starter Member

    Oct 19, 2013
    62
    0
    Hello all,

    The attached pdf file shows the circuitry i used.

    (Please keep in mind i am still sort of new to transistors)

    Goal:
    The goal of the project is to create a voltage regulator that can regulate 9V and at the same time, disconnect a load from the circuit at conditions of 40 degrees celcius and 1A.

    Basic Operation:
    The regulator would be designed such that it outputs a constant 9V which would be the voltage drop across the load. It is desired that the load be separated from the circuit for temperatures greater than 40 degrees celcius and current exceeding 1A. Therefore, sensing circuits are implemented such that it would detect these parameters and control the load accordingly. The load is controlled by a switching circuit, comprising of MOSFETs. The MOSFETs are arranged such that it produces an AND gate (Tmf1 and Tmf2) and receives signals from the temperature and current sensing circuit so it would behave accordingly when the conditions are not met. This means that a current would only allow a current to flow through the load when the temperatures is less than 40 degrees celcius and current is less than 1A. The alternative path for the current to flow is a short circuit to ground (probably a poor decision but i was limited to time), so that when the 1A flows, the load would not be going on and off repeatedly very quickly.

    Problem:
    There are multiple problems however these are the main ones.
    -The problem i am having with my circuit is that when the current passes the 1A, the current is sent into the alternative path with the mosfet Tmf4 which is fine but the load keeps going on and off repeatedly irregardless (even though i took it into consideration) and also making a sort of high pitch sound. I know the load is on and off repeatedly because there is a small glow in the LED.

    -When currents greater than 1A flows, the voltage across the Zener diode changes greatly which messes up the regulated output which causes a problem because my references (potentiometers) are set using it.

    - The current limiting circuit also messes up the regulated voltage a bit but i guess it is because the MOSFET is not ideal and draws little current from the base even when it is off.

    -Even my LM35 temperature transducer is affected by the high currents. It gives very wrong output values at the higher currents.
    -------------------
    Please ask questions about my design for any clarification or other purposes. I am open to any such criticism. I have considered letting the load disconnect at 1A and coming back on at around say 0.95A. In addition to that, i am considering how to keep the current through my zener constant. I more than likely have more problems but these are my main ones.

    ~Thanks :) , all responses are much appreciated
     
  2. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    Your circuit is much too complex! A constant voltage power supply with a current limiter set to 0.95A (or whatever) would accomplish most of what you need. You just need to add over-temperature cut-out, which is a common feature. Personally, I would use a LM339 (quad) to control a single MOSFET to shut everything off if the temperature limit is exceeded. But there are probably other techniques.

    I think your circuit oscillates by design. The current through the shunt is not controlled at a steady level, it's cycling on and off. Things may improve by fixing that, but you'll always need to consider the paths for higher currents, especially the grounds. A shifting ground wreaks havoc.
     
  3. JWHassler

    Member

    Sep 25, 2013
    201
    33
    What is 'Tmf4' doing? It looks to be connected across the supply. (in series with a tiny resistor: still kinda shorted.)
    Your oscillation may be the supply being crowbarred by a sudden load..
     
  4. c1rcu1ts

    Thread Starter Member

    Oct 19, 2013
    62
    0
    The tiny resistor is another current sensing resistor used to detect when the current reaches about 1.3A. If it reaches this then the transistor Tmf5 draws current away from the base of the NPN transistor (TIP31C) which reduces the current in the current(less current being amplifier). It essentially operates as current limiting circuit. The short to ground was an alternative pass for the current when the load was off. Consider it as if the branch was not there and the current flowed through the load and Tmf3. The current sensing resistor ( 1ohm 25W ) would read the 1A (a diferential amplifier is used to read the voltage of the two terminals to determine when the voltage across it is 1V) and stop current from flowing into the load (by controlling the mosfet, Tmf1). This would inhibit the current from flowing in the circuit so therefore the current sensing resistor would no longer have a voltage drop which means that the load would turn back on. then the process is repeated continuously and rapidly. Aaand i also do not know what you mean by "crowbarred".

    Wayneh, the voltage regulator IC was not an option as i had to use a series pass voltage regulator. (sorry about that detail). And no, the current limiter is not an option.
     
  5. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    ?? So when the circuit cuts out due to over-current, what will keep it from trying to turn on again? Do you want to build in a manual reset?
     
  6. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    I don't see why you need the 'alternative path' for current. That merely wastes power as heat. Just provide the power source with a current-limiter as Wayneh suggested (post #2).
    If you are relying on a zener for regulation then the zener supply needs to come from the regulator output.
     
  7. c1rcu1ts

    Thread Starter Member

    Oct 19, 2013
    62
    0
    I don't see why you need the 'alternative path' for current. That merely wastes power as heat. Just provide the power source with a current-limiter as Wayneh suggested (post #2).
    If you are relying on a zener for regulation then the zener supply needs to come from the regulator output.


    I used the alternative path because i was limited to time to think of another way. I only realised recently however that i can let the circuit cut-off at 1A and let it turn back on at say 0.95A. Unfortunately i do not know how to do it.

    ?? So when the circuit cuts out due to over-current, what will keep it from trying to turn on again? Do you want to build in a manual reset?

    I was hoping to go with an automatic turn on. The alternative path would remain on as long as currents greater than 1A flows through the circuit. This is because the output of the differential amplifier used to measure the voltage across the 1V,5W resistor would read something higher than the reference voltage (1V). This would keep he MOSFET Tmf1 off and hence the load off and the alternative path on. When it drops to a current below 1A, then the MOSFET Tmf1 would turn back on.
     
  8. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    Would this be a suitable alternative? The LM723 provides both voltage regulation and current-limiting.
    When a set temperature is exceeded the LM723's reference voltage is pulled down by Q1 so the IC switches off the supply to the load.
    CurrentLimitedVreg1a.gif
    CurrentLimitedVreg1b.gif
     
    Last edited: Dec 12, 2014
    c1rcu1ts likes this.
  9. c1rcu1ts

    Thread Starter Member

    Oct 19, 2013
    62
    0
    Thank you, that is a pretty cool circuit and i would look into it (although i do not this i have access to the LM723 part). Do you know of any current regulator circuits. A circuit that outputs a constant 50mA for a range of input voltages.

    By the way i agree that the "alternative path" was not a good option but it was a "quick fix". If i had more time then i would have definitely looked into the band for the circuit to operate (cut off at 1A and come on again at 0.95A, cut off at 40 degrees celcius and come on at 35 degrees celcius) or maybe even a more viable solution. In order to demonstrate cut-off, the current cannot be limited at the 1A (a little higher will do) and also the voltage regulator IC would not be permitted. I would definitely still be working on it after it is due (next thurdsay).
     
    Last edited: Dec 13, 2014
  10. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    A quick google should give you lots of constant-current schematics.
     
  11. c1rcu1ts

    Thread Starter Member

    Oct 19, 2013
    62
    0
    I did try the google search and found some but i do not think it could have been used with my circuit. I shall try again.

    Whenever 0.9A flow through my circuit, all the references change. This would be inclusive of the output of the LM35. I am not sure why but i know that it occurs at higher currents. I tried a buffer but it did not work. I suspect that it is the voltage across the zener diode that changes (reference of regulator) which is why i am going to create a current regulator which sends a constant 50mA through the branch to keep that zener voltage constant.
     
  12. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,773
    1,103
    You won't need as much as 50mA through the zener. 10mA should be enough. A zener is not a particularly good reference. A better reference would be a TL431A for example, which is programmable with two resistors.
     
  13. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,281
    1,232
    Maybe this would work for a Thursday project.

    I'm not sure what you want to happen with high current and high temperature, but this one holds the current at 1 amp max and shuts off the supply if the temperature goes above 40C.
    It will then turn back on if the temperature goes back down.
     
Loading...