@Mister Audio
I'm pretty sure it is Post #1 (without the 4 ohm resistor in series with 0.33uf cap) that is at least theoretically the short at some very high frequency.
I'm pretty sure it is Post #1 (without the 4 ohm resistor in series with 0.33uf cap) that is at least theoretically the short at some very high frequency.
That branch is not part of the short.
It depends on whether you're trying to understand the issue or just looking for smoke. Every amplifier is different. Amps tend to have mitigations against parasitics and while this is a different thing, the reality has something to say. However a trained individual may see this as a red flag for at least a few reasons.
This is veering of the original topic of the thread.
AllenB has said
Yes, I am sure there are quite a few designs out there that use this idea. In the past I have used the combination of just a 0.22 to 0.5uF capacitor to do something in the several kHz region. The addition of the resistor obviously lowers the effect.
Last night in my response I was almost tempted to say does the use of C6 imply the need of a higher order filter to kill the breakup region?
I admit to having used a parallel capacitor on the bass coil and I have made a few speaker for friends over the years where I have used it.
I haven't had any issues with blown amps. But that is not to say it wont happen at some point in the future. I usually use a cheap ebay second hand Japanese amp for test purposes as during tweaking and listening to crossovers it is easy to have an accident even with the speaker leads themselves. Once i am happy they can go onto an ACA, or old quad II valve amp,
I think the topic of crossover and amp interactions outside of the audio band needs a complete discussion thread looking at the many different amp technologies available to us, output stages and how much feedback is used. I guess for some crossover circuits and loudspeaker combinations past the 100kHz range they will be doing interesting things.
AllenB has said
I think we should take this point onboard as several other people have commented about it.
Yes, I am sure there are quite a few designs out there that use this idea. In the past I have used the combination of just a 0.22 to 0.5uF capacitor to do something in the several kHz region. The addition of the resistor obviously lowers the effect.
Last night in my response I was almost tempted to say does the use of C6 imply the need of a higher order filter to kill the breakup region?
I admit to having used a parallel capacitor on the bass coil and I have made a few speaker for friends over the years where I have used it.
I haven't had any issues with blown amps. But that is not to say it wont happen at some point in the future. I usually use a cheap ebay second hand Japanese amp for test purposes as during tweaking and listening to crossovers it is easy to have an accident even with the speaker leads themselves. Once i am happy they can go onto an ACA, or old quad II valve amp,
I think the topic of crossover and amp interactions outside of the audio band needs a complete discussion thread looking at the many different amp technologies available to us, output stages and how much feedback is used. I guess for some crossover circuits and loudspeaker combinations past the 100kHz range they will be doing interesting things.
Just put a thorough HF blocking network at the output of your amp, thus optimizing the Zobel that already is present. A well-designed one has it already. Read Self, pp227. About 4µH does fine, add a 1W resistor of 10Ω in parallel. Make the coils yourself, wrapping two layers around the resistor.
I have offered the question over at the Solid State amplifier forum so we might get a wider perspective.
https://www.diyaudio.com/community/threads/problems-associated-with-a-capacitive-load.413083/
https://www.diyaudio.com/community/threads/problems-associated-with-a-capacitive-load.413083/
Wooww I didn’t expect to open such a can full of worms! Now I am ever more confused 
Let me share the difference C6 makes regarding the break up region:
I actually copied that from a well known design as well.
Let me share the difference C6 makes regarding the break up region:
I thought so, but one can’t deny that it looks much nicer when you knock it down
3rd order didn’t work for some reason but i can give it another go tonightAnybody else have a remark on that being barely noticable?
@hautparlurker
Sorry if I dragged this out too long, but I had tried to get a good answer to this issue a few months ago and never did. I have used this approach in 2 to 4 builds and have never had a problem but the comment came up a few months ago that it causes a short. I tried to understand the alternatives but the thread pretty much ended with no comments.
One of my designs uses your exact woofer, so I understand exactly why you ended up with your crossover. Here are my responses without and with a small 0.10 uf capacitor. It makes a significant difference to the cone breakup just as you had noticed in your design. To AllenB's point, there very well could be better alternatives, but this (at least with a resistor in series after the capacitor) is a very common filter.
Sorry if I dragged this out too long, but I had tried to get a good answer to this issue a few months ago and never did. I have used this approach in 2 to 4 builds and have never had a problem but the comment came up a few months ago that it causes a short. I tried to understand the alternatives but the thread pretty much ended with no comments.
One of my designs uses your exact woofer, so I understand exactly why you ended up with your crossover. Here are my responses without and with a small 0.10 uf capacitor. It makes a significant difference to the cone breakup just as you had noticed in your design. To AllenB's point, there very well could be better alternatives, but this (at least with a resistor in series after the capacitor) is a very common filter.
Oh please, I appreciate your contribution on the contrary, I am very confused right now.
That nasty break up combined with the baffle loss (also the fact that I put my tweeter in a waveguide) makes it very hard to get a good response and seriously limits your options. I really don't understand how so many people opposes this given the fact that so many designs incorporate it.
(I actually saw some guy made some physical modifications to the SB15nrx to dampen the breakup)
Last edited:
Yeah something like that. I made a quick search but I couldn't find the link, sorry
I think they have similar behavior at the break up frequencies, but aluminum's break up is more audible, harsher etc. At least from what I understand...
Below are the responses with and without the capacitor in question. This is my first proper build so I don't know if there will be an audible difference.
Yup, that is the post indeed.
Wow, that is quite a difference. This much improvement would solve all my issues but I don't have the confidence for such an undertaking right now
I modelled your driver and crossover in Basta, with a 200mm by 350mm baffle. This doesn't use the measured response (and I'm guessing your baffle size), but it helps me to see what your crossover is doing.
With your original crossover, it would be easy to add a resistor in series with the 0.33uF cap, maybe even as much as 12 ohms. You'd still get a decent depth of notch. But such a circuit always gives a pretty narrow notch, no matter the resistor value, and you're trying to reduce a broad breakup area. So it's probably best to omit the 0.33 capacitor altogether, as you've done.
I'm surprised by your LCR trap - it's giving a broad reduction centred around 1kHz. (And the revised version is maybe centred around 800 Hz?) But the driver datasheet shows only a minor surround resonance (or similar) peak in that region. Are you using the LCR trap to reduce baffle step? If so, why not use a more conventional LR baffle step circuit? (Then, if you do want to reduce the breakup region, maybe use an LCR trap? But you're probably OK without that.)
With your original crossover, it would be easy to add a resistor in series with the 0.33uF cap, maybe even as much as 12 ohms. You'd still get a decent depth of notch. But such a circuit always gives a pretty narrow notch, no matter the resistor value, and you're trying to reduce a broad breakup area. So it's probably best to omit the 0.33 capacitor altogether, as you've done.
I'm surprised by your LCR trap - it's giving a broad reduction centred around 1kHz. (And the revised version is maybe centred around 800 Hz?) But the driver datasheet shows only a minor surround resonance (or similar) peak in that region. Are you using the LCR trap to reduce baffle step? If so, why not use a more conventional LR baffle step circuit? (Then, if you do want to reduce the breakup region, maybe use an LCR trap? But you're probably OK without that.)
Last edited: