Thermal protection with PTC device?

Discussion in 'The Projects Forum' started by ProZac, Dec 5, 2010.

  1. ProZac

    Thread Starter New Member

    Nov 23, 2010
    8
    0
    Hello.

    I'm working on a performance automotive ignition project for uni, that will be situated in the engine bay, so will be subject to some heat. The IC's the product uses have built in thermal protection that kicks in when they reach 94°C.

    The initial design called for this thermal protection to be used as the main protection for the device, but its effect is so immediate that the IC's cool to below 94°C very quickly, and the device re-activates. The protection scheme manifests itself during real world testing as an erratic misfire in the engine... Misfires can be a real PITA to track down, so we really dont want to make a device that could potentially introduce one!

    After doing some googling and what-not, I'm wondering if it might be possible to create a thermal protection circuit from a PTC resetable fuse in thermal contact with the case of the device?

    The IC's we're using have an enable pin, than lets the device run with this pin grounded, but turns the IC's off if it is tied high, or left floating.
    The idea i had was to run a known current through the PTC fuse, which has been chosen to trip at this known current, and a specific temperature. When this trips, the enable pin will be floating the and device will turn off.

    The case of the device is basically one large heatsink, so it would take some time for this to cool down and the fuse to reset.

    The worry i have is that all the PTC fuses I've looked at have a fairly large section between normal current handling mode, where their resistance is known and fairly low, and the tripped mode, where their resistance is high. During this mode, they can apparently be un-tripped, or tripped...

    Any of you smart chaps out there have and ideas? Whats the 'typical' way to go about thermal protection like this? The ideal result would be for the device to shut down for a good 30 seconds or so, as that would make tracking the problem down very easy for the end user.

    Cheers :).
     
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Use a thermistor (PTC or NTC) and a comparator IC to enable/disable the chips.
     
  3. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Along with what mik3 said, use hysteresis feedback to make the enable/disable more positive; ie: if the temp reaches 94°C, then disable until the temp falls below 90°C, then enable again.
     
  4. ProZac

    Thread Starter New Member

    Nov 23, 2010
    8
    0
  5. wayneh

    Expert

    Sep 9, 2010
    12,151
    3,058
    This approach will work well, and I've recently built something similar.

    But your time might be even better spent solving the root problem of keeping the circuits cooler in the first place. It's your call, but it seems like that would be a far better long-run solution, to make a thermal shut-off a rare event.
     
  6. ProZac

    Thread Starter New Member

    Nov 23, 2010
    8
    0
    Ahhh, I dont think i explained myself well enough to begin with on that one. The module itself will function at full noise for aslong as you would like to run it aslong as it is subjected to ambient temperatures of 65°C or less. The problem strikes if it is installed somewhere silly in the engine bay (near exhaust or turbo's, etc) and the ambient temperature is far to high. I'm trying to make it as 'idiot proof' as possible you see.

    However, having a bit of an idiot designing it in the first place might not be the wisest course of action ;).
     
Loading...