NTC inrush current limiter

Discussion in 'General Electronics Chat' started by Jesse Spiessens, Jun 4, 2016.

  1. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    So I've been thinking on using an NTC to not blow up my TRIAC.
    The thing I'm building is a programmable multi-socket/powerstrip. It would have 5 sockets and the load will most likely be some kind of christmas lights (incandescent). Each of the sockets should be able to handle a 16 amp load@240V (more as a safety measure as opposed to real use, you know, 'cause of breakers and stuff).
    I'm planning on using 22 ohm NTCs ( http://www.produktinfo.conrad.com/datenblaetter/450000-474999/468053-da-01-en-Heissleiter_NTC.pdf ), but I have no idea if this value is sufficient etc. I've used PTCs etc, but NTCs are pretty new to me (but I know how they work).
     
  2. MrAl

    Well-Known Member

    Jun 17, 2014
    2,443
    492
    Hi,

    The 22 ohm'ers you are looking at only go up to 2.8 amps.
    In fact, none of them go up to 16 amps.

    Also keep in mind that the devices get hot when in use, and if you size them so they dont get hot then they dont work :)
     
    Last edited: Jun 4, 2016
    Jesse Spiessens likes this.
  3. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Hi MrAl

    Thank you for your response.
    The Imax is the reason I wasn't sure about it.
    I can't even find one remotely close to 16 amp ratings.

    And I know, I will keep that in mind :)
     
  4. #12

    Expert

    Nov 30, 2010
    16,355
    6,852
    Save your triacs from what?
    Over current?
    Excessive dV/dT?
     
  5. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Overcurrent, that's the reason I'd like an inrush current limiter ;)
    (More specific the inrush current caused by cold filaments)
     
  6. dl324

    Distinguished Member

    Mar 30, 2015
    3,250
    626
    One disadvantage using an NTC thermistor is they don't offer protection for "hot restart".

    Above a few hundred watts, you should use an active solution.
     
    Jesse Spiessens likes this.
  7. MrAl

    Well-Known Member

    Jun 17, 2014
    2,443
    492
    Hi,

    Some very good points being brought up here.

    First, triacs are fairly hardy, although if you have a specific app that always draws high current, that could be a problem unless you over size the triac, which is entirely possible. Triacs go up pretty high as do SCR's.

    Also brought up was the thermal time constant of the NTC type, which could take 60 seconds to cool down. During that time they only provide partial protection.

    I second the "active solution" suggestion, as that will force it to work right without any hot parts.
    If the triacs can be oversized, then that would be a good idea too. Rate them based on the max current you end up with rather then the normal running current. Of course that does not prevent line dips due to high current surges.
     
    Jesse Spiessens likes this.
  8. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Thanks to all of you who already replied.
    The TRIACs I was planning to use can handle 25A.
    The active solution would indeed be nice, although then I'd have a zero crossing TRIAC driving other TRIACs, or are their other solutions?
    Or should I just trust the makers of the loads in that they don't exceed 25 amps for a period so long that it would damage the TRIACs?
     
  9. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    That indeed is something I've thought about. Although I'm not expecting the swap of loads, it's always possible that it happens.
     
  10. MrAl

    Well-Known Member

    Jun 17, 2014
    2,443
    492

    Hi,

    If this is important enough you may want to measure that. There are current clamp on meters that can catch the current surge when a device is turned on. For example my AC unit normally draws only maybe 5 amps, but when first turned on it draws a whopping 20 amps for a second. That 20 amps gets caught on the peak catching clamp on ammeter. Really any way you can come up with to measure it would be good though.
    A 25 amp triac can take a lot of abuse, but measuring the actual current surge would allow us to figure out just how good it might work long term. Anything else is just guessing.
     
    Jesse Spiessens likes this.
  11. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Alright. I'll try to give some more information as soon as possible :)
    If it exceeds the 25A though, I will propably go for an active solution.
    Maybe something worth mentioning too. I have a zero crossing optocoupler which gives a microcontroller a pulse everytime the sinewave goes through the zero point. Could this measurement be used to soft start the load? For instance make it so that the first times a load is switched on, it only makes the TRIAC trigger right near the end of the wave? (Sorry if my explenation isn't clear)
     
  12. MrAl

    Well-Known Member

    Jun 17, 2014
    2,443
    492
    Hi,

    Sounds good.

    I suppose it could make a difference if the surge ends fast, but if it takes 20 cycles then it probably wont do much. If you had a scope that would be nice, to measure the surge.

    Yeah 25 amps sounds high. 10 amps would be much better.
     
    Jesse Spiessens likes this.
  13. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    How about measuring the current with using the voltage across a shuntresistor, using an opamp to substracht the voltage before and after the resistor (using a voltage divider because opamps won't like 240V) and then going to an adc of the microcontroller? Maybe not the easiest configuration, but it is one which would give me a lot of good information. And then say if current is greater than 16A, limit it by phase cutting.
     
  14. MrAl

    Well-Known Member

    Jun 17, 2014
    2,443
    492
    Hi,

    Probably doable. Either rectify or add a DC offset to the current, which will be AC also.
    Test for accuracy first with resistive load.
    Watch for excessive power dissipation in the shunt resistor or shunt.
     
    Jesse Spiessens likes this.
  15. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Will do. Only thing to do now is to use optocouplers between microcontroller on the board and the external microcontroller which will be connected to a PC because well, it doesn't have a galvanic isolation.
     
  16. Jesse Spiessens

    Thread Starter New Member

    Jun 4, 2016
    9
    0
    Thank you for all your feedback. You've been an amazing help!
     
Loading...