installing a temp indicator on the voice coil

[scampo77]

it seems there is very little that can be done to accuratly figure out how mych power a speaker can handle. and even if i can figure out how much it can handle it seems to change due to the type of music or duration etc. so what i would like to do is install a couple of wires directly to the aluminum former of the voice coil and hook up a guage for each sub.

can anyone off up any ideas on how to do something like this. if i can get any ideas i will just be using 2 wires soldered to the inside of the former and using my multimeter with e temp monitor to watch the temperature. i would think that a little readout or an actual gauge would work much better.

[trd1587]

You can try and install a small thermocouple.

[scampo77]

what is that? i was thinkinking of just running 2 wires?

[davygrvy]

You don't need a thermo couple. You can use the copper of the voice coil as is. Let me explain.. Copper has a positive thermal coefficient and increases in resistance as it gets hotter.

If one was to apply a constant bias current across the coil (ac coupled to the amp across a bank of capacitors), you could monitor the voltage change and apply a little math and viola! It doesn't matter what the current is, as long as it is measurable. If it's too high, you'll push the driver forward. A simple constant current source is a depletion mode mosfet like a DN2540 in a T0-220 package run source voltage at least the positive rail of the amplifier used.

The math is this shown in Tcl from a test app I wrote.

proc thermalCoef {T0 {material "cu"}} {
    switch -- $material {
	"cu" {set C 0.00385}
	"al" {set C 0.00401}
    }
    return [expr {1.0/((1.0/$C)+($T0-25))}]
}

proc nowTemp {VT0 T0 VT1} {
    return [expr {((1/[thermalCoef $T0])*(($VT1/$VT0)-1))+$T0}]
}

proc tempRise {VT0 T0 VT1} {
    return [expr {[nowTemp $VT0 $T0 $VT1] - $T0}]
}

VT0 = rest voltage
T0 = rest temperature
VT1 = coil voltage to request temp for

So basically, you start from a known temp and voltage, and read either the rise or absolute in Celsius. Be aware that you'll need some strong filtering if you use that with a sub. Maybe as much as an 8-pole bessel on the test signal and maybe as much as 20,000 uF for C1 depending upon your load.

[davygrvy]

Here's a paper on the technique, but equation #7 can be simplified to use a constant current source as resistance is directly proportional to the voltage the current produces.

[Magura]

The easy solution is a small ntc. As this is for a big brutal car sub, 0.5 gram won't make any difference to the unit.


Magura

[davygrvy]

I didn't see it before, but those procedures can be a little more simple.

proc nowTemp {VT0 T0 VT1} {
    return [expr {[tempRise $VT0 $T0 $VT1] + $T0}]
}

proc tempRise {VT0 T0 VT1} {
    return [expr {(1/[thermalCoef $T0])*(($VT1/$VT0)-1)}]
}

[davygrvy]

Originally Posted by Magura

The easy solution is a small ntc.

I disagree. I say don't modify your driver at all. Just push a little DC bias current through it and monitor the DC voltage and apply a little math.

[Andrew Eckhardt]

Originally Posted by davygrvy

I disagree. I say don't modify your driver at all. Just push a little DC bias current through it and monitor the DC voltage and apply a little math.

How will you force a current through the driver and not the output of a voltage controlled power amplifier unless it is temporarily disconnected?

Maybe you could sense average current and voltage in the power amplifier and apply some processing to that to find changes in DCR.

[Magura]

He forgot to tell you the rest of the story. He is to replace the VC.


Magura