Active cross-over, solid state amps connect directly to tweeter?

[Drew Eckhardt]

I finished my Linkwitz Orions last weekend, finished populating
my cross-over boards yesterday, and went to work with the plan of coming home
early and mixing the two once I've stuck op-amps in sockets and tested the
boards.

While idle at work, I became paranoid and feared DC/turn-on/turn-off problems
frying a tweeter. Do people actually have problems like that or are my worries
groundless?

If it matters, the tweeters in question are Seas T25CF-002 millenmium tweeters
and amplifiers Adcom 545iis. My 545iis don't have audible power on/off
artifacts; although the 555ii on my sub does and I don't know if it started
life that way or got cranky over time.

Do I want to install some capacitors between amp and tweeters? Cross-over
is at 1440Hz 4th order, presumably I want to roll-off a couple octaves below,
and impedance is 6 Ohms nominal so a pair of 39uF polypropylene Solens
producing an F3 point arround 340 Hz would do the trick? Is that likely
to be audible?

 

[Konrad]

While idle at work, I became paranoid and feared DC/turn-on/turn-off problems
frying a tweeter. Do people actually have problems like that or are my worries
groundless?

I had the same thought's abt series cap, using 55u. But realy the resonans f to that tweeter is abt 600Hz maybee abt 20uF will do better?
Also the Q value needs to bee increased, normaly by a series resistor. this is also linearising the impedance and the freq response.

 

[leadbelly]

Do I want to install some capacitors between amp and tweeters?

Yes, you do. 78uF is overkill. Your money is better spent using more and smaller caps to wind up with 20uF total C and lower DCR.

 

[MBK]

I have a biamped dipole setup with chip amps. Fullrange driver doing the job of mid and tweeter, and no cap. The chip amps are great in not having a thump, but my (Linkwitz style) crossover does produce one.

What happens however is that the HP to the fullrange (313 Hz in my case) almost completely gets rid of that, so that the preamp or EQ thump only affects the woofers (who have handled it for 3 years now). I see them moving in and out 1/2 inch at on or off ... but the fullrange hardly moves, and with chip power amp inputs disconnected, no thump whatsoever.

So, under the condition that the power amp doesn't have a massive thump, it can work w/o a cap. And your tweeters have a high power handling as far as I know.

 

[thylantyr]

I've never experienced any blown tweeters on a direct
connect to the amplifier, but I do insert an inline fuse
to protect the speaker and the fuse value is determined
by trial and error to find the breaking point as desired.

Removing the capacitor removes another variable in
the signal path that could taint the sonics depending
on how picky you are.

 

[Davey]

Siegfried addresses this concern on one of his FAQ questions. (middle paragraph.)

http://www.linkwitzlab.com/faq.htm#Q26

You probably don't need to worry, but it doesn't hurt to check the turn on/off thumps of your amplifier with an oscilloscope.

Cheers,

Davey.

 

[Konrad]

Thylantyr: please let's know your fuse size.

Mbk : I am talking aktive x-over.

 

[MBK]

Mbk : I am talking aktive x-over.

Me too - but for some reason the LP of the 24dB/oct L-R cum dipole EQ I have does send close to DC turn-on and off transients to the power amps while the mid only gets a brief "squeak". If I disconnect the EQ, no thump or squeak at all, so it's not the power amp.

In addition, the most dangerous thumps are from line level devices (pre, CD player) because they get amplified - and this, an active line level HP will almost completely get rid off (if it is not itself causing one).

 

[Drew Eckhardt]

I don't have a power switch on my XO, so it's always on.

I scoped the turn-on output of the 545ii I'm using for my tweeter amp, found that the brief DC offset was less than the ~1.5V of the battery I used to check polarity, and figured it should hurt even less because it's a more gradual change that isn't even audible - you just see a slow retract/relax on the tweeter.

Still works fine after over a hundred cycles.

 

[Drew Eckhardt]

Yes, you do. 78uF is overkill. Your money is better spent using more and smaller caps to
wind up with 20uF total C and lower DCR.

My cross-over point is 1440Hz.

Using the nominal 6Ohm impedance, 20uF of capacitance yields a 1327Hz F3 point. With the actual impedance at XO of arround 5.5 Ohms I end up with a 1448Hz F3.

IOW, I'd be missing 1/4 of my power at the XO point for a 1.2dB dip, and have a phase shift there which negates the effects of my all-pass filter used to align tweeter+mid output.

Doesn't seem like a good idea.

As far as price, 4 39uF Solen polypropylene caps would total $42 (and 2 82uF parts would be $36). I had been thinking that on a $2500 project that was not an interesting number.

In practice, no one has reported any problems with the direct connection, the scoped turn-on transient doesn't look bad, I haven't had any scary sounds on startup/shut down, and a pair of replacement dome/coil assemblies are just $52 so I I'm going to pass on the capacitors until I see something bad happen.

 

[thylantyr]

Thylantyr: please let's know your fuse size.

Mbk : I am talking aktive x-over.

Fuse size is dependent on your driver and power. You have
to experiment with your setup.

First you need to turn on your stereo system to the maximum
level you feel comfortable that the drivers won't blow up.

Second, you insert inline fuses starting with low amperage
ratings and progressively work up until you find the correct
value.

For tweeters start with 1/2 amp and work your way up
to 1 1/2 amp until you find the sweet spot.

At 1/2 amp, the fuse may blow too early, at 1 1/2 amp,
the fuse may never blow, instead your tweeter may
blow.

This is a one-time calibration, then remember which fuse
rating works for you.

3/4 amp or 1 amp fast blow AGC is average for the common system.

 

[Centauri]

In my main domain of pro audio, where HF driver diaphrams are hundreds of dollars, and solid state amplifiers can & DO develop faults where DC voltage can appear on the outputs, we DO use such DC blocking capacitors. If the value is high enough, then there is no discernable degredation to the system.

In your case, a lot will depend on your perceived likelyhood of anything going amiss.

Cheers

 

[Drew Eckhardt]

So what's a high enough value? Placing the pole an octave below XO shouldn't affect the frequency response, but what about phase shift? Two octaves below seems consistant with other aspects of XO design?

 

[zygibajt]

Wouldn't it be easier to put some relay with some delay at the output of the power amp feeding the tweeters?
Bartek

 

[Konrad]

Relay at the output? (as DC switch of/fuse ) The responsetime is normaly set minimum 1/2SEC or more. More important is the amp's DC drift, There should not bee any! anny value abow ca50mv is not tolerable. Then designing a dc sensor with this sensitivity, and not a to loong responsetime ( in case of failure ) is quite a task. I prefer DC blocking Cap's And DC sensor to switch of the output of amp.