Simple Current Limiting Add-On

Discussion in 'General Electronics Chat' started by Xplode, Sep 1, 2012.

  1. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    Hi guys,
    I'm still fairly new to electronic design and I'm trying to work out a simple current limiting circuit for a project I'm working on. I will have variable loads in the 1.3 - 5ohm range that will be attached through spring terminals, and is wired loose in the field, so the chance of a short circuit is fairly high. Also, the load itself (depending on Vdrop across the field wiring) may try to draw more than my intended limits but each load only requires ~1amp to function...

    I have many parallel loads that will be turned ON/OFF via a set of FETs powered by a common 24VDC rail that I would like to limit to 4 or 5Amps... (Ideally so 4 parallel sets of load at ~1amp required each, will function simultaneously).
    These FETs will be triggered ON for only a short burst of time (100-200ms max) but if there is a wiring error in the field, I don't want the circuit to run uncontrolled and allow 20+AMPs to flow, burning up my FETs and melting wires.

    I found a simple Current limiting circuit through google (http://www.vidisonic.com/2008/07/10/current-limiting-circuit/) but I don't fully understand the instructions on how to select the Transistors and bias resistors. Plus all the circuits I've seen seem to be geared towards smaller currents...

    Help? :)

    My List of Requirements:
    * 24VDC Battery Powered overall
    * Minimal Drop across limiter under normal operation
    * Will be on the Positive side of the load (V+ to Limiter to Load to GND)
    * Must auto-reset back to normal operation the instant the fault is cleared (ie the Load's controlling FET is turned off)

    I think the circuit I found will work - so maybe just help with parts?

    Thanks in advance!

    P.S. - I tried searching but the search seems to be broken? I just get a blank white page....
     
  2. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    The problem with a simple current limiter is that, in the event of a shorted load, the limiter will have to dissipate 24V*5A=120 Watts. This is difficult to deal with.
     
  3. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    Thanks for the reply. I really appreciate the help.

    There are some rather large devices that are rated for 120W and 15A or higher. I had my eye on a 8A/80W unit since I was only going to use 4A as my Imax, but is there a design/implementation issue with me using a much larger rated unit?

    I poked onto Digikey and found this - http://www.digikey.com/product-detail/en/2N3773G/2N3773GOS-ND/918208
    I'm still new to this so I think this is an ok choice (minimal Vce drop @ Saturation and 150W power handling with a reasonably low Ibase)

    Could I get away with using something like this?

    Or is there another method of controlling the current so I don't lose parts to a short circuit? I can't fuse/breaker it because 200ms later I need the system to keep functioning after a short.
     
  4. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    I would prefer a circuit that turns off completely the load if the current is above the limit. If you have 200ms time the circuit could recheck every 150ms if the short / overcurrent condition is still present.

    This needs something to limit the rate of rise of load current, like an inductor in series with the output. If you turn on the output into a shorted load this will still give you enough time to cut the output voltage with overloading your protection circuit switch.
     
  5. thavinator

    New Member

    Jul 4, 2011
    20
    3
    If one of the outputs is shorted, it's not going to work 200ms later or 20 minutes later until someone comes by and un-shorts it. So why not a fuse/breaker on each output? That's the simplest solution. A hard short might still kill an output transistor, so a series inductor like praondevou suggests might still be in order.
     
  6. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    The LM723 chip can solve this. Look particularly at the foldback current limiting circuit. Minimal dissipation during a short, analog speed of sensing, and nearly instant recovery.

    Yes, it's an antique chip but, I specialized in "self protecting" power supplies when I was a pup. I have another way to do this but the 723 chip uses less parts and doesn't even need a "time to recovery" calculation.
     
  7. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    thavinator - I tried to explain that it was a current control for a power rail that fed multiple outputs that would be triggered for a short period of time.

    The project is to control firing of pyrotechnic effects. The electric igniters (ematches) only need to be powered for 100-200ms to fire.

    So I have multiple FETS, possibly setup through mutliplexer chips, to trigger the output cues for a fireworks/pyro show. I just need the limiter circuit to clamp the current to 4amps, or kill the output completely during the short circuit (which would be over when the associated FET opned and turned off)


    #12 - I'll look into that chip - although I am curious about your other options as well. what ideas do you have for me?
     
  8. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,227
    If you only need to protect the circuit for a 200ms pulse, then the circuit you referenced in you initial post would work, but be aware that its voltage drop at the current limit point will be about 1.4V. Is that acceptable?

    If you want a lower drop you could use a comparator, such as an LM339, to sense the voltage across a small shunt resistor. That would allow a much small voltage drop of less than 100mV for the current limit trip point. The comparator will likely oscillate between its on and off trip points, as determined by its hysteresis setting, but that just makes it work as a switching limiter to minimize the MOSFET dissipation under limit conditions.

    The comparator output would remove the drive to the MOSFET switches, shutting them off during the limit time.

    Edit: An added RC delay network in the positive feedback of the comparator could be used to keep the output off for a few tenths of a second, as desired. This would avoid any oscillations.
     
    Last edited: Sep 2, 2012
  9. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    My "other" circuit is made of discrete transistors, resistors, and capacitors. Not an improvement on the 723 chip. The one I have in mind measures the peak current and shuts off the drive to a triac for half a second, then resets and checks to see if the current is still too high. It would be better if it checked to see if the load resistance was still too low.

    You don't even need to use the whole 723 chip. You can steal the idea from that chip and make a foldback current limiter with transistors or maybe a chip with several npn transistors in it. When faced with an over current condition, it simply lowers the output voltage until it is useless. As soon as the overload disappears, it bounces back to normal operation in milliseconds. It doesn't check the current, measure time, or digitize anything. It simply responds like an analog amplifier.

    This is the basic shape of a foldback circuit with a few parts researched and calculated for your purpose.
    And then there is the "Stop&Check" circuit.
     
    Last edited: Sep 2, 2012
  10. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,227
    Below is my simulation of a limit circuit using an LM339. The limit trips at about 5A and stays tripped for about 0.3s.

    The comparator output is the signal that would inhibit the drive voltage to the control FETs.

    The current is sensed in the return line from the pyro's to the battery, which should not be a problem.

    Edit: R5 and V2 are for simulation purposes and not part of the circuit.

    Current Limit.gif
     
    Last edited: Sep 3, 2012
  11. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    Much cleaner than that stop&check I did in 1979!

    Then again, that circuit I did was designed for AC and built with what I had on hand.
     
    Last edited: Sep 3, 2012
  12. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    I'm going to go back and change my last post a bit.
    Also, perhaps I should clear up what the circuit is doing overall.

    This is an electronic pyrotechnic trigger system, and the device that is field wired will be 1.3-2ohms at most. The idea is that it is purposely overloaded with 12-24volts and 1amp to produce the necessary heat (as fast as possible) to heat and ignite an explosive compound which in turn lights the fuse of the effect being used.

    So at 1.3Ohms, and lets assume 5ohms for the scab wires its connected with (but could be up to ~10 or 12 ohms) each effect that is firing will be trying to draw more than the 4amps I wish to keep the limit at... They will reliably ignite and work at 1 amp, so 4amps was a buffer and allowed the idea of parallel effects being used simultaneously.

    Ok. the 200ms time wasn't a dead set (I forget i'm talking to engineering minds, sorry) I'll need to test fire some effects and get the time dialed in, but I expect 100-150ms pulse ON, and then 20-30ms minimum (but generally muuuuuch larger... in the 10s of seconds) before the next effect is triggered ON for the next 100-150ms. I would be happiest if the circuit was capable of being operated for several seconds at max limiting without damage. Just so I've got a super good buffer of reliability.

    The only goal is to keep the current flow below 5amps (so I picked 4) so I can use lower cost/rated components elsewhere in the circuit (50 fets rated for 5A is much cheaper than 50 fets rated for 20A) plus the board traces wouldn't handle 20A pulses for long before burning up.

    Accuracy of the limit isn't essential so long as it stays below the 5A hard limit. And voltage drop across the limiter isn't terribly important to me either (within reason). so a 1-3V drop is probably just fine...

    I'm super tired so I'm going to look at the attachments in more detail in the morning.

    I REALLY appreciate the help on this one - this forum is fantastic.
     
  13. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    This suggestion may sound stupid, and it's probably the result of not understanding your setup.
    Have you considered using a humongous 5Ω resistor as the current limiter?
    You would get ≈5A short circuit limiting. You would get more than one Amp with any load less than 19Ω.
     
  14. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,227
    Below is the simulation of a simple two transistor, two resistor current limit circuit which has a drop of less then 1V before it limits. The limit current is approximately 0.7V/R2. The P-MOSFET can be any device that has an ON resistance of less than 50mΩ, a current rating of >5A, and a 40V or greater Vds rating.

    Edit: The MOSFET shouldn't need any heat sink if the pulses are no more than a couple hundred milliseconds. For longer than that, you will.

    Current Limit.gif
     
    Last edited: Sep 3, 2012
    Xplode likes this.
  15. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    I was working on that, but our OP mentioned possible short circuiting of the load, which I assume could bypass the switching transistor, resulting in 120 Watts being dissipated by the MOSFET for an indefinite period of time. 120 Watts in a MOSFET is hard to get rid of.
    I have a foldback-enhanced version that I'm working on, but it isn't finished, and I'm not sure he even needs it.
    What's wrong with the seres resistor idea? It is big, and probably costs more that a semiconductor-based solution, but it doesn't have time constraints on the short circuit.
     
  16. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    @ RonH - The giant resistor idea has crossed my mind but a 120W resistor (or I guess something even bigger so its oversized) would probably cost me waaaaay more than the semi-conductor version, and I'm assuming be muuuuch larger and heavier. The system is going into a SeaHorse waterproof case and needs to be ultra portable. In the hopes of keeping the weight down, i'd be happier with a circuit designed to limit.

    I will be fusing the overall input to the circuit as well, so if the limiter ever failed in some fashion a 5/6A fuse will burn up and hopefully protect my boards/etc...

    The output pulse is controlled by an arduino so it won't ever be longer than what is deemed necessary to reliably activate the ignition system. (so I'm positive it won't be more than 200ms) BUUUT multiple effects may be fired in close succession, so for example a finale may have 6-10 groups of effects to be activated in a 3 second period.

    The circuit HAS to be able to handle a short circuit... When the product ignites, its possible the wiring will be damage from the resulting explosion and so if the igniter triggers the effect 50ms into a 150ms (theoretical) pulse, there may be 70-100ms of short circuit...

    @crutschow - that looks very similar to the circuit I posted in the OP, but with calculated values (thanks) and using a MOSFET... I DID find Mosfets rated for 150-200Watts so I could probably get away with a FET producing 120W of heat in short circuit condition no? Unless something fails in the programming of the controller (hopefully worked out in the controlled testing) then it shouldnt be longer than the 200ms Cap I'm expecting the FET to be on for...


    Think of the circuit as a giant rotary selector just made digital. This would be the power in limiter, and then i quickly rotate the knob through 50 contacts where each on each contact a load of 1.3-10Ohms exists. but the knob only stays on the contact for 200ms before being rotated to the next one. (except in practicality, the knob goes to an off position between the loads and waits for the signal that the next effect is needed, then rotates to the next ON position)


    I hope this makes sense - these ideas and discussion is exactly what I was hoping for. Much thanks for everyone's help.
     
  17. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,227
    Why do you think a short would bypass the switching transistor? :confused: Are you assuming a low-side switch? He stated he was worried about a short zapping his switching transistors, which would imply a high-side switch.

    Certainly the resistor is a simple solution if the 5Ω series resistance is not a problem. Of course 120W resistors are not cheap. ;)
     
  18. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    I'm not up to speed today. Be back when brain works.
     
  19. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,227
    I went to the MOSFET circuit to reduce the normal voltage drop about a volt when it is operating in the not-limiting mode.

    I would think a 150-200W MOSFET would tolerate 120W for 200ms without a problem, as long as it has a few seconds between pulses.
     
  20. Xplode

    Thread Starter New Member

    Sep 1, 2012
    19
    0
    Ok - I did up a quick n dirty block diagram for how I envision this working.
    As you can see there are parallel output drawing from this limited output.

    There are times when only one output FET will be on, but there are times when 2 or 3 might be on simultaneously. and this is to be used in a live, real time, show environment. So using a simple fuse/breaker isn't practical if there is a wiring error.

    The system needs to clamp the output to 5A and automatically recover from a fault rapidly. hope this helps clear up what i'm aiming for.
     
Loading...