Active crossover folks: do you use DC blocking caps on your tweeters?

[Pallas]

I've never actually done an active project for the home. My car system is active, though that might not count, as the Aura Whisper widebanders I use from ~800Hz up are quite a bit more hardy than the typical dome tweeter.

So, questions for folks who use active crossovers (no "why bother" snarking from the passive set, please)

• Do you use DC blocking caps between amp and tweeter?
• If so, how low below the tweeter's highpass do you design them to work?
• Leaving cost completely out of it, is there any performance reason to choose a 'lytic over a massive poly Coke-can? (The latent audiophool idiot snob in me does not like the idea of a 'lytic in the tweeter's signal path, and considering the scope of the project one giant poly cap per speaker is reasonably priced. Especially since replacement tweeter diaphragms will cost $140ish shipped, while they're available.)

I ask because I'm going to be taking my mains active in the next month and change - setting up system in new home, may as well change things up fairly radically, right? I'll be using six channels (LCR 2-ways) of an 8x8 miniDSP board, and six channels of a multichannel amp. My amp sends no extraneous pops through its output terminals upon turn on/off. (All one hears is the click of the relays.) Hopefully, once I get the miniDSP's delayed 12V trigger I/O sorted, I will also be able to avoid its turn on/off thump. But even then, I suspect there's a chance of turn off/on thump if for some reason the miniDSP board temporarily loses power and then comes back online.

PS: Yes, I searched. But I didn't find anything. If I missed a previous discussion with which you're familiar, please link it here.

[Sydney]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

Do you use DC blocking caps between amp and tweeter?

No

[Jonathan Anspach]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I don't.

[ryanbouma]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I don't bother. I've even had some DC thumpt with dome tweeters without issue. However, if you're talking very expensive tweeters, it may be worth it. For DC protection, you can use a really big cap, but the price goes up. I'd use an NPE personally, but if your tweeters are expensive and the price doesn't bother you, then go for the poly. Where is your cross over point?

I've thought about adding a cap, and then having a jumper once I know everything is playing nice. But the extra work wouldn't be worth it. I'd probably just leave the cap there and listen. And like you said, if the mini turned off then on abruptly... pop.

[dcgold2000]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I do. The tweeters were expensive and the idea that a failed amp could take them out is unacceptable. Caps won't really protect against on/off thumps but they will protect (hopefully) if an amp fails and the output goes to one or the other supply rails. 63 volts in my case. Tube amps are safe I believe.

[charlielaub]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I don't use capacitors in series with the tweeter in my active designs.

One thing that many active crossover folks do not realize is that a cap introduces phase rotation within several octaves of its corner frequency because the phase of a 1st order RC circuit changes quite slowly with frequency. This can mean that the cap will change the phase angle well into the tweeter's passband unless the value of the cap is very large. This is one reason why such HUGE cap values are a good idea if you want to ignore that they are in circuit. If you are planning on doing some nice crossover work, use something well above 100 uF or model the cap as a passive crossover.

-Charlie

[Pallas]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I don't bother. I've even had some DC thumpt with dome tweeters without issue. However, if you're talking very expensive tweeters, it may be worth it. For DC protection, you can use a really big cap, but the price goes up. I'd use an NPE personally, but if your tweeters are expensive and the price doesn't bother you, then go for the poly. Where is your cross over point?

Crossover point to be determined experimentally by polar measurements (IMO, the single most important thing in choosing a crossover frequency is getting a good directivity match between woofer and tweeter) followed by listening for strain. I expect it to be between 1.1 and 1.5 kHz. (The factory passive is 1.4kHz, with low-order electrical slopes. The drive unit is a 12" concentric, Tannoy 3133GG.) The tweeters do have easily-replaceable diaphragms. But said diaphragms are currently $132 plus shipping each,

One thing that many active crossover folks do not realize is that a cap introduces phase rotation within several octaves of its corner frequency because the phase of a 1st order RC circuit changes quite slowly with frequency. This can mean that the cap will change the phase angle well into the tweeter's passband unless the value of the cap is very large. This is one reason why such HUGE cap values are a good idea if you want to ignore that they are in circuit.

I'm confused by your reply. If one is using anything in the signal path, why would one base a design off measurements with with that component excluded in the signal path? That is to say, I'd work on the active xo transfer function based on measurements taken with any blocking cap in the tweeter's signal path. Also, given that it's bloody easy to send a large LF sinewave to a tweeter during measurements, I would think that a person worried about the DC blocking generally would use blocking caps on tweeters during measurement.

[scottq]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

If one is using anything in the signal path, wouldn't one base a design off measurements with with that component not in the signal path? That is to say, I'd work on the active xo transfer function based on measurements taken with any blocking cap in the tweeter's signal path.

Provided you're doing the modeling/design with that in the measurements/simulation, you should be in good shape in terms of compensating for the phase shift. I think Charlie is trying to prevent someone from using a 2.7uF cap just to save costs since the f3 doesn't need to be anywhere near DC for a tweeter, which could very likely induce some phase shifts. On the other hand, if you're accounting for this in your design, it can be compensated for.

There is a performance reason not to use electrolytic; they produce significant distortion near their corner frequency. Mr. Douglas Self seems to be the most well respected person that has documented and published this distortion. In general, film caps have lower distortion. If you're of the "they sound different" camp, then use your favorite flavor (and may I also suggest bypassing with higher voltage/quality, smaller caps).

[charlielaub]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I'm confused by your reply. If one is using anything in the signal path, wouldn't one base a design off measurements with with that component not in the signal path? That is to say, I'd work on the active xo transfer function based on measurements taken with any blocking cap in the tweeter's signal path.

That's correct, you should take measurements that include the cap in series with the tweeter. Essentially it's a one component high pass filter. You could also model it in PCD and then export the result (I think PCD can do that) as the driver's frequency response. You then go ahead an design the active crossover based on that data (measured or modeled).

If, on the other hand, you make the value of the cap so large that the corner frequency is several (e.g. 3+) octaves below where you will be using the tweeter (e.g. below the lower edge of its passband) then the effect of the cap can more or less be neglected. When the cap corner frequency is closer to the tweeter passband, you will want to include it in your model and measurements.

I'm not sure that everyone gets this point, so I thought I would mention it.

-Charlie

[dcgold2000]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

One should also be aware that a capacitor in series with voice coil inductance forms a resonant circuit that can cause peaking. A discussion here: http://www.pispeakers.com/Speaker_Crossover.pdf page 27. Also a good reason to make the cap large.

[Pallas]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

I think Charlie is trying to prevent someone from using a 2.7uF cap just to save costs since the f3 doesn't need to be anywhere near DC for a tweeter, which could very likely induce some phase shifts. On the other hand, if you're accounting for this in your design, it can be compensated for.

Makes sense. I assumed that everyone understood by "blocking cap" a capacitor large enough to be at least one octave below the tweeter's Fc.

There is a performance reason not to use electrolytic; they produce significant distortion near their corner frequency. Mr. Douglas Self seems to be the most well respected person that has documented and published this distortion. In general, film caps have lower distortion. If you're of the "they sound different" camp, then use your favorite flavor (and may I also suggest bypassing with higher voltage/quality, smaller caps).

[billfitzmaurice]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

Do you use DC blocking caps between amp and tweeter?

Yes, not so much for blocking DC as for taming turn-on transients. Size the cap as a 1st order high-pass with the corner 1 octave down from the crossover high-pass frequency.

[charlielaub]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

Here's a web page that shows a plot of the phase response of a first order high pass filter:

http://www.play-hookey.com/ac_theory/hi_pass_filters.html

Scroll down to "phase response".

As you can see, the phase rotation takes place over more than two decades (that's well over 6 octaves!). In the plot, the filter corner frequency has been normalized to 1. So, if you select a capacitor that has a corner frequency one octave below the passband edge, make sure to include it in your model! The phase change from the cap will still be present in the crossover region...

-Charlie

[johnk...]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

No.

[scottmc]

Re: Active crossover folks: do you use DC blocking caps on your tweeters?

Yes, Yes.

I certainly have hooked all those wires up wrong before. Tweeter on the woofer channel? Poof.
Using a Windows PC to drive it? Never trust it. Any software can screw up. You can accidently change something.
You shouldn't have a thump on any reasonable equipment today.

Just include the cap when taking measurements. If not, just add it as a passive component to Jeff's crossover simulator and use the active configuration at the same time.

Time correction will account for most of the phase issue. 2 octaves from crossover seems reasonable.