Description 
There are plenty of Passive crossover kits and PCBs around but this one also addresses a problem that many do not touch, Tweeter protection.
Many of you will have noticed that tweeter Failures are not uncommon particularly in Sound reinforcement systems.  
I have heard many theories attempting to explain this involving clipping and distortion, but really the reason is very simple:
Tweeters, particularly horn loaded compression drivers are very sensitive, often 6 to 12dB more so than the lower frequency drivers.
High frequency power of 'normal music' expected from the Tweeter is often as low as 25% of the total acoustic power.
What all this means is that a tweeter only needs to have a fraction of the power handling of the speaker system as a whole and 50 Watt tweeters in 500W speakers are not entirely unheard of.
Now when everything is working as it should this is fine, but when something starts feeding back, and your Amplifier starts trying to deliver full power to your Tweeter, then its future life expectancy is often measurable in fractions of a second....

Having had a number of Tweeter failures I decided it was time to investigate protection.
The speakers I had on hand with blown tweeters had been protected with poly switches yet had still failed.
I had heard that selecting the best poly switch was always a balance between adequate protection and nuisance tripping so I ran some tests on three Poly switches: RXE110 (Rated at 2.2A or 38.72W @ 8 Ohms) RXE090 (Rated at 1.8A or 25.9W @ 8 Ohms) and RXE075 (Rated at 1.5A or 18W @ 8 ohms).
To test the response times of the poly switches I used the following test rig:

The RXE090 took almost 3 Seconds to open and even the RXE075 took 2.5 Seconds to open.

Having tested several Tweeters to Destruction (by no means an exhaustive test as Tweeters are expensive, but more to get an indication of how long a tweeter could last) I had found that most nominal 50Watt tweeters would be lucky indeed to last 2 seconds at 128 Watts.
I also found that many of the cheaper '50 Watt' tweeters were in fact lucky to survive several seconds at full power, although to give credit where credit is due the two more expensive tweeters I tested did handle rated power for several hours - despite being mounted inside an insulated wooden box (the test got rather loud otherwise).

The conclusion I reached was, that while Poly switches are better than nothing, they are barely adequate as Speaker protection.

Lightbulbs likewise proved to be problematic - either getting consistent lightbulbs or lightbulbs that provided adequate protection without false tripping or failure.

What I needed, was protection that tripped at a preset point, allowed a slight (but not significant delay) before tripping (to avoid tripping on otherwise harmless signals), preferably a circuit that reset itself when the danger had passed and just to add to the challenge a circuit that required no external power....
This is the circuit I came up with:

C9 and L3 act as a high pass filter, while R9 provides a constant/predictable load for the filter.
D1 to D4 are UF4001 high speed diodes acting as a full wave rectifier and C6 - 1uF (May be varied to suit application) acts as a Filter/Delay.
R6 (Set as required) and R7 (560 Ohm) act as a voltage divider to trigger for the SCR SC1, and determine the voltage at which the protector trips. R6 may be calculated by R6= (Tv-0.56)*1000 where Tv is the trip Voltage.
R1 allows the Input filter to be bypassed if/when the protection is triggered this would mean that protection remains active until it is reset by a compete absence of signal - I have not used this option yet but I would suggest R1 of around 39 Ohms (if used).
C8 (10uF) ensures that the SCR is reset when the Relay drops out.
Finally Q1(MJE340), R4 (85 Ohm), ZD1 (3V3) and R5 act as a current limit to stop the Relay burning out when protection is active.
RLY1 is a Relay with a 400 Ohm coil and R9 is set at 150 Ohms, This sort of load should be insgnificant when compared to typical 4, 8 or even 16 ohm speaker systems.
While I did experiment with alternative methods of resetting the SCR that did not involve an extra pair of Relay contacts, the Relay I ended up going with only cost $1 more than the SPST Relay and has 5 Amp contacts instead of 3 Amps.
The reason for having a high pass filter in front of this circuit rather than installing this circuit after the existing high pass filter in the crossover is that if the tweeter alone is disconnected, then in the case of a second order crossover, the high pass filter for the tweeter will act as a series resonant circuit and this may cause stability issues for some Amplifiers.
Of course this is a Crossover, so by the time I add all the other components for a second order crossover we get: