my question is whether adding a series protection cap on an active crossover system (lm3886 based) is necessary or not.
what's the common thinking, here?
crossover is behringer dcx2496 and I was going to run its analog outs into a stepped atten and then into a chipamp directly feeding the speaker driver.
the running dc offset is not what I care about (it will be low enough to not matter) but I wonder if turn on/off thumps, however small they are, would best be avoided for a direct wired tweeter?
I'm very open to having relays do turn on/off delays. maybe that's my solution right there (?)
have not built the speaker system yet, but just at the planning stage right now. I do want to use that crossover and a group of 2 or 3 chipamps to drive each speaker element.
go capless or not?
what's the common thinking, here?
crossover is behringer dcx2496 and I was going to run its analog outs into a stepped atten and then into a chipamp directly feeding the speaker driver.
the running dc offset is not what I care about (it will be low enough to not matter) but I wonder if turn on/off thumps, however small they are, would best be avoided for a direct wired tweeter?
I'm very open to having relays do turn on/off delays. maybe that's my solution right there (?)
have not built the speaker system yet, but just at the planning stage right now. I do want to use that crossover and a group of 2 or 3 chipamps to drive each speaker element.
go capless or not?
Try the amps on a speaker first to see if they have any turn on/off movement or noises. I'd be tempted to use a muting relay rather than cap as high value polypropylene caps are both large and expensive.
If you're using the behringer, you want to avoid and filter components - thats the whole point of the thing. The behringer itself has a little thump when powering up or down itself. A good idea might be to use the muting circuit on the lm3886 chip, that way you're not adding anything to the signal path.
forgot about the muting circuit! I wired it in (using the chipamp.com boards for now) and forgot about it.
that might be the thing to do. if you can avoid the relays AND the caps its a double win.
still, part of me still likes the ease and insurance of the relay method. and its also amp-agnostic; if I switch to an amp module that does not have muting, I'll have to do my own thing then anyway.
would an opto coupler work for the muting switch (using the 10k Rm typically)? I will have a cpu nearby (arduino thingie) and I don't want that digital supply near the analog stuff. I don't mind having the arduino logically turn on/off an opto isolator and that would mute/unmute the 10k on the lm3886. does that make sense?
that might be the thing to do. if you can avoid the relays AND the caps its a double win.
still, part of me still likes the ease and insurance of the relay method. and its also amp-agnostic; if I switch to an amp module that does not have muting, I'll have to do my own thing then anyway.
would an opto coupler work for the muting switch (using the 10k Rm typically)? I will have a cpu nearby (arduino thingie) and I don't want that digital supply near the analog stuff. I don't mind having the arduino logically turn on/off an opto isolator and that would mute/unmute the 10k on the lm3886. does that make sense?
How much are your tweeters worth? How difficult will it be to replace them if they fry?
I think even a stepped volume control might generate DC spikes or nasty AC noise if the contacts get dirty.
A large blocking cap before the tweeter is not the end of the world, no matter how "pure" you think the signal will be without one. You need to buy them only once.
Peace,
Tom E
I think even a stepped volume control might generate DC spikes or nasty AC noise if the contacts get dirty.
A large blocking cap before the tweeter is not the end of the world, no matter how "pure" you think the signal will be without one. You need to buy them only once.
Peace,
Tom E
In my experience with LM3886s, not. But if my tweeters were expensive ones, I'd probably change my mind.
Turn on/off thumps will be passed through the cap anyway. I'd say the reason for employing a cap is primarily fault protection. Keep coupling caps small in the tweeter circuit so turn on thumps will be shorter. Or run bridged so as to practically eliminate them.
not overly expensive, in the $50 range or less.
Im using sealed telco quality relays (its a relay based atten) and so the atten system is pretty clean. I also make sure to mute between selections, so it simulates a make before break style when you run up or down the vol 'knob'.
the purist stuff is what I'm trying to wrestle with. I sure do like the insurance of an 'install and forget' cap.
Hi,
an 8ohm tweeter crossed over at 3kHz needs a 6u6F (6u8Fn should be available or two //3u3F) for a single pole passive 3kHz filter.
If you adopt this as a protection cap then you must reduce the DCX treble filter by one pole to maintain the same crossover characteristic.
Alternatively, you can increase the cap value by a decade to 66uF to move the filter of the protection cap down to 300Hz. This will have virtually no audible effect on the crossover characteristic set in the DCX.
I suspect you can move the protection filter up quite a bit from 300Hz towards 1300Hz (15uF) with increasing audible effect. Experimentation will prove what you find acceptable.
You can adopt the same with a Mid driver using a passive single pole filter as your protection cap. Don't go a decade lower as it forces adoption of an electrolytic.
an 8ohm tweeter crossed over at 3kHz needs a 6u6F (6u8Fn should be available or two //3u3F) for a single pole passive 3kHz filter.
If you adopt this as a protection cap then you must reduce the DCX treble filter by one pole to maintain the same crossover characteristic.
Alternatively, you can increase the cap value by a decade to 66uF to move the filter of the protection cap down to 300Hz. This will have virtually no audible effect on the crossover characteristic set in the DCX.
I suspect you can move the protection filter up quite a bit from 300Hz towards 1300Hz (15uF) with increasing audible effect. Experimentation will prove what you find acceptable.
You can adopt the same with a Mid driver using a passive single pole filter as your protection cap. Don't go a decade lower as it forces adoption of an electrolytic.
Last edited:
thanks andrew. I'll try for a 15uf or larger as a general target.
the turn on/off's on the 3886 really seem mild, to me. maybe I can get by with no cap if I don't use these speakers as 'testers' and only connect known good amps to them, or diy projects that seem known/good 😉
the turn on/off's on the 3886 really seem mild, to me. maybe I can get by with no cap if I don't use these speakers as 'testers' and only connect known good amps to them, or diy projects that seem known/good 😉
ever thought of using the kind of capacitors that ceiling fans have?
They are cheap, rated for high voltage, fully solid construction, durable and i guess they're available in 4uf and thereabouts. Parallel two of those and you're done.
They are cheap, rated for high voltage, fully solid construction, durable and i guess they're available in 4uf and thereabouts. Parallel two of those and you're done.
ceiling fan caps?
nice high frequency response AND a cool breeze at the same time! sounds good to me, lol.
nice high frequency response AND a cool breeze at the same time! sounds good to me, lol.
hey linux, you do have speaker protection just incase the amp blows right? there are circuits out there that detect voltages over a certain amount, once that amount is detected, it enables a relay which turns off the speakers to keep from frying them. now, the reason why they put cap on speakers is to limit that speaker to a specific tonal degree, lows, mids, highs, and if you'll notice that use different caps for all 3 bands. now, we can only hear so low, and we can only hear so high, so the caps also double as a cut off point beyond what we can hear, and they go bye bye.
Yes, a capacitor will still pass thumps through it. But it gives the tweeter protection from any DC on the outputs. If DC offset is not an issue, you need not consider that.
If you're trying to use the mute function in LM3886, try this:
Building a buffered Gainclone chip amp.
scroll down to the very bottom.
you can probably substitute the relay for a optoisolator in the circuit.
That gives you a simple, yet elegant solution for your thump problem.
If you're trying to use the mute function in LM3886, try this:
Building a buffered Gainclone chip amp.
scroll down to the very bottom.
you can probably substitute the relay for a optoisolator in the circuit.
That gives you a simple, yet elegant solution for your thump problem.
Oversizing the series capacitor means the tweeter receives a higher proportion of LF energy during transients that are not part of a musical signal, such as turning on and off and in fault conditions.
Reliably protecting a tweeter is extremely difficult. I've been struggling with a new design that needs to dump 50W or so into a compression driver, and have either a fried tweeter or complex filter calculations to deal with. Particularly since my crossover is at about 600-900Hz. Older Bose speakers used a lightbulb in series with the tweeter, which is extremely reliable and sounds pathetic.
The DCX2496 has caps at the output, but still pops quite a bit when switching on and off if you are not running balanced. It is also very intrusive sonically, though that may be considered a necessary evil.
OTOH, if you build the 3886 well you can eliminate any clicks and pops (mainly ground errors or improper sizing of input and feedback resistors). In which case it becomes a simple task to switch the amplifiers on last and off first...
Reliably protecting a tweeter is extremely difficult. I've been struggling with a new design that needs to dump 50W or so into a compression driver, and have either a fried tweeter or complex filter calculations to deal with. Particularly since my crossover is at about 600-900Hz. Older Bose speakers used a lightbulb in series with the tweeter, which is extremely reliable and sounds pathetic.
The DCX2496 has caps at the output, but still pops quite a bit when switching on and off if you are not running balanced. It is also very intrusive sonically, though that may be considered a necessary evil.
OTOH, if you build the 3886 well you can eliminate any clicks and pops (mainly ground errors or improper sizing of input and feedback resistors). In which case it becomes a simple task to switch the amplifiers on last and off first...
- Status
- Not open for further replies.