Constant K Crossover Mod?

[geoff.]

Should the last capacitor before the tweeter in a constant K crossover be changed to three times the value of the first? 

 

For instance, the AA crossover would have a 2uF capacitor and the second 2uF capacitor would be replaced with a 6uF (or better).

 

I have seen several posts over the years, the oldest by ALK, stating it is prudent to do this but it doesn’t seem to be “driven home”.

 

Is that because it is common knowledge to most?

 

I am in the process of recapping/cloning some crossovers and this just ran through my head.

[mboxler]

9 hours ago, geoff. said:

Should the last capacitor before the tweeter in a constant K crossover be changed to three times the value of the first? 

 

For instance, the AA crossover would have a 2uF capacitor and the second 2uF capacitor would be replaced with a 6uF (or better).

 

I have seen several posts over the years, the oldest by ALK, stating it is prudent to do this but it doesn’t seem to be “driven home”.

 

Is that because it is common knowledge to most?

 

I am in the process of recapping/cloning some crossovers and this just ran through my head.

 

Sounds like you are trying to turn the tweeter filter to a 3rd Order Butterworth, like the Universal's filter?  Then yes, the second capacitor is 3 times larger than the first.  

 

If you want to get the the Universal's 6000hz crossover point, simply changing the second capacitor in an AA to 6uf won't work.  An ideal 6000hz 3rd Order Butterworth filter into an 8 ohm load would be 2.2uf, 160uh, and 6.6uf values.  Since the first 2uf capacitor on the AA is in series with the 13uf capacitor, it effectively becomes a 1.75uf cap.  A 1.75uf, 245uh, 6uf combination ends up crossing around 8800hz.

 

In order to get a 6000hz filter, you would need to replace the first cap with a 2.7uf, replace the inductor with a 160uh, and replace the third cap with a 6.6uf.  A 2.7uf cap in series with a 13uf cap is around 2.2uf.

 

Keep in mind, the tweeter filter attached to tap 5 in a stock AA drops the voltage to the squawker (tap 3) by an extra 1.5db or so.  Looks like changing the tweeter filter to the 2.7uf, 160uh, 6.6uf values will drop it another .5db.

 

I think I got my simulations correct.

 

Mike

[geoff.]

Wow, thanks Mike.

 

I was not aware of the effect the tweeter filter has on the squawker as a voltage divider when it comes off of the same capacitor, until now.

 

 

 

 

 

 

[mboxler]

1 hour ago, geoff. said:

Wow, thanks Mike.

 

I was not aware of the effect the tweeter filter has on the squawker as a voltage divider when it comes off of the same capacitor, until now.

 

 

 

 

 

 

 

I think it's actually more interesting than that.  

 

If the tweeter filter is connected before the 13uf capacitor, the load on the 13uf capacitor stays consistent, and the voltage drop across the capacitor will approach zero as frequency increases.  Taps 0-5 will eventually get the full voltage.

 

If the tweeter filter is connected to tap 5, the load on the 13uf capacitor becomes a bell curve, peaking out around 500-700hz.  After that, the load drops quickly.  At the rate the load is dropping, the voltage drop across the 13uf capacitor actually levels off, even though the frequency is increasing.  That extra voltage drop never gets to taps 0-5.  

 

I'm guessing that's why Klipsch changed autoformers when they went with the elliptical tweeter filter.  Since that filter is attached before the 13uf capacitor, 
it doesn't mess with the load on the capacitor, and more autoformer attenuation was needed.