Chef
In Memoriam
Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
Low-cost DIY practice/coffeehouse tweeterless 110 [long]
- Thread starter DukeLeJeune
- Start date
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Thank you to MX21 and greenboy for image hosting.
This response-shaping circuitry is OPTIONAL. It adds a lot of cost and complexity, and my not be a net improvement, depending on the tone you're looking for.
If you look at the [Invalid or Expired Link Removed] of the S2010, you'll see a wide lop-sided peak roughly centered on 1.5 kHz, and a tall narrow peak roughly centered on 2.5 kHz. These peaks emphasize the upper midrange region, and some people prefer a smoother sound.
In order to get a smoother sound we don't need to eliminate those peaks entirely, but we do want to bring them down enough so they aren't shouty. A slightly rising response trend is okay to partially offset the narrowing radiation pattern as we go up in frequency.
So what we're going to do is use a medium-width parallel notch filter (so called because the components are wired in parallel) to take down the output in the 1.5 kHz region by about 4 dB, and a narrow parallel notch filter to take down the output in the 2.5 kHz region by about 4 dB. That will leave some bumpage, but the overall variation will be a lot less. Perfectly smooth is perfectly unaffordable; and this is still going to be pricey.
Here's the response-shaping filter; the labelling that my program uses is a little confusing so I will explain:
V1 is the amplifier, and the little triangle represents the "ground" or "negative" wire. Over on the other end you see the speaker itself, labelled K3, with one wire going to ground. That wire actually connects back to the negative terminal on the speaker cab, which in turn connects to the negative terminal of the amplifier.
The biggest "stack" of parts is the 1.5 kHz notch filter. From top to bottom, we have a 5 microfarad capacitor (5.1 microfarads is close enough); a 50 microfarad capacitor; a .2 (point two) millihenry 20 gauge aircore inductor (the program converts to microhenries); and then three 20-ohm, 20-watt sandcast resistors.
The two capacitors in parallel in effect form a single 55 microfarad capacitor, and the three 20-ohm resistors form a single 6.7 ohm resistor. So in effect each notch filter consists of a capacitor, an inductor, and a resistor. We'll come back to this later.
Okay the next stack of components consists of two 40 microfarad capacitors; a .05 (point zero five) millihenry 18 gauge aircore inductor; and an 8-ohm, 20-watt sandcast resistor. This is the 2.5 kHz notch filter.
The reason the .05 mH inductor is heavier gauge is, Parts Express doesn't sell a .05 mH inducor in 18 gauge. A 20-gauge inductor saturates at 200 watts, while an 18 gauge inductor saturates at 300 watts. Since the speaker's thermal and mechanical power handling are in the 150 watt balllpark, the 20 gauge inductor is okay for the first notch filter. However for about a dollar more you can move up to the 18 gauge inductor.
The reason the capacitor values are broken down like they are is to minimize cost. There are other combinations of capacitor values that will give the same total value, but they are more expensive... at least if you get 'em from Parts Express.
Each of these clusters of parts are wired in parallel; that is, all of the wire leads on one side are soldered together, and all of the wire leads on the other side are soldered together. The signal comes from the + input terminal through one clump of parts, meets on the other side and then goes through another clump of parts, meets on the other side, and goes to the + tab on the speaker.
The final, in this case optional, part is the switch. This is a heavy-duty single pole-single throw switch wired to bypass both notch filters. So if you want a more aggressive tone for some songs but not for others, you have the option of bypassing the notch filters and just hearing the native frequency response of the speaker. If you leave out the switch, also leave out the wires that connect it to either side of the notch filters.
None of these components have a particular polarity; that is, they can each be connected with either end being the input or the output side.
The threaded shaft of the switch will have a grooved side, which accepts a little tab on the metal on/off label. This groove is always on the "OFF" side. So with the switch "OFF", the signal will actually be passing through the notch filters. With the switch "ON" the signal will be bypassing them, and the upper mids will be louder.
Here is the list of parts; you can substitute equivalents from other sources:
75 uF polypropylene capacitor (Dayton 5%)
2 x 40 uF polypropylene capacitors (Dayton 5%)
5.1 uF polypropylene capacitor (Dayton 5%)
.20 mH inductor, 20 gauge or 18 gauge aircore (Jantzen)
.05 mH inductor, 20 gauge aircore (Jantzen)
3 x 20 ohm, 20 watt sandcast resistors
8 ohm, 20 watt sandcast resistor
SPST toggle switch, heavy duty
toggle switch boot
dish for toggle switch
gasket for dish
And, here are the pages where you'll find them:
[Invalid or Expired Link Removed]
Invalid Link Removed (scroll down)
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
The hole in the dish is smaller than the 1/2" diameter needed to mount the toggle switch, so you'll have to drill it out. Mount the dish from the inside of the cabinet; if you mount it from the outside, the toggle switch will stick out too far and it can be broken of easily. So, you want to place the gasket on what would have been the upper surface of the "lip" around the dish, so that it's between the dish and the MDF.
The parts cost for the whole shebang, assuming that .20 mH inductor is 20 gauge, is about $51.56 if my arithmetic is correct. Ouch!
Well, we can omit stuff. We can omit the switch, and that brings us down to $44.09.
That may still be on the high side. Okay, the lower peak is going to be the more audible one because it's over twice as wide. Narrow peaks are generally not as audible as gentler, wider ones. So if we just do the 1.5 kHz filter (and no switch), that would be about $19.93.
Incidentally, if I showed these filters following convention, each would look like one capacitor, one inductor, and one resistor, for reasons described earlier. In other words, this circuit would be said to consist of seven elements (including the switch), even though physically there are eleven elements. The crossover for my 1x12+horn cab consists of twenty-three elements. My point is not that more = better, but that in some cases the crossover is eating up a bigger share of the budget and design effort than you'd expect.
This response-shaping circuitry is OPTIONAL. It adds a lot of cost and complexity, and my not be a net improvement, depending on the tone you're looking for.
If you look at the [Invalid or Expired Link Removed] of the S2010, you'll see a wide lop-sided peak roughly centered on 1.5 kHz, and a tall narrow peak roughly centered on 2.5 kHz. These peaks emphasize the upper midrange region, and some people prefer a smoother sound.
In order to get a smoother sound we don't need to eliminate those peaks entirely, but we do want to bring them down enough so they aren't shouty. A slightly rising response trend is okay to partially offset the narrowing radiation pattern as we go up in frequency.
So what we're going to do is use a medium-width parallel notch filter (so called because the components are wired in parallel) to take down the output in the 1.5 kHz region by about 4 dB, and a narrow parallel notch filter to take down the output in the 2.5 kHz region by about 4 dB. That will leave some bumpage, but the overall variation will be a lot less. Perfectly smooth is perfectly unaffordable; and this is still going to be pricey.
Here's the response-shaping filter; the labelling that my program uses is a little confusing so I will explain:
V1 is the amplifier, and the little triangle represents the "ground" or "negative" wire. Over on the other end you see the speaker itself, labelled K3, with one wire going to ground. That wire actually connects back to the negative terminal on the speaker cab, which in turn connects to the negative terminal of the amplifier.
The biggest "stack" of parts is the 1.5 kHz notch filter. From top to bottom, we have a 5 microfarad capacitor (5.1 microfarads is close enough); a 50 microfarad capacitor; a .2 (point two) millihenry 20 gauge aircore inductor (the program converts to microhenries); and then three 20-ohm, 20-watt sandcast resistors.
The two capacitors in parallel in effect form a single 55 microfarad capacitor, and the three 20-ohm resistors form a single 6.7 ohm resistor. So in effect each notch filter consists of a capacitor, an inductor, and a resistor. We'll come back to this later.
Okay the next stack of components consists of two 40 microfarad capacitors; a .05 (point zero five) millihenry 18 gauge aircore inductor; and an 8-ohm, 20-watt sandcast resistor. This is the 2.5 kHz notch filter.
The reason the .05 mH inductor is heavier gauge is, Parts Express doesn't sell a .05 mH inducor in 18 gauge. A 20-gauge inductor saturates at 200 watts, while an 18 gauge inductor saturates at 300 watts. Since the speaker's thermal and mechanical power handling are in the 150 watt balllpark, the 20 gauge inductor is okay for the first notch filter. However for about a dollar more you can move up to the 18 gauge inductor.
The reason the capacitor values are broken down like they are is to minimize cost. There are other combinations of capacitor values that will give the same total value, but they are more expensive... at least if you get 'em from Parts Express.
Each of these clusters of parts are wired in parallel; that is, all of the wire leads on one side are soldered together, and all of the wire leads on the other side are soldered together. The signal comes from the + input terminal through one clump of parts, meets on the other side and then goes through another clump of parts, meets on the other side, and goes to the + tab on the speaker.
The final, in this case optional, part is the switch. This is a heavy-duty single pole-single throw switch wired to bypass both notch filters. So if you want a more aggressive tone for some songs but not for others, you have the option of bypassing the notch filters and just hearing the native frequency response of the speaker. If you leave out the switch, also leave out the wires that connect it to either side of the notch filters.
None of these components have a particular polarity; that is, they can each be connected with either end being the input or the output side.
The threaded shaft of the switch will have a grooved side, which accepts a little tab on the metal on/off label. This groove is always on the "OFF" side. So with the switch "OFF", the signal will actually be passing through the notch filters. With the switch "ON" the signal will be bypassing them, and the upper mids will be louder.
Here is the list of parts; you can substitute equivalents from other sources:
75 uF polypropylene capacitor (Dayton 5%)
2 x 40 uF polypropylene capacitors (Dayton 5%)
5.1 uF polypropylene capacitor (Dayton 5%)
.20 mH inductor, 20 gauge or 18 gauge aircore (Jantzen)
.05 mH inductor, 20 gauge aircore (Jantzen)
3 x 20 ohm, 20 watt sandcast resistors
8 ohm, 20 watt sandcast resistor
SPST toggle switch, heavy duty
toggle switch boot
dish for toggle switch
gasket for dish
And, here are the pages where you'll find them:
[Invalid or Expired Link Removed]
Invalid Link Removed (scroll down)
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
[Invalid or Expired Link Removed]
The hole in the dish is smaller than the 1/2" diameter needed to mount the toggle switch, so you'll have to drill it out. Mount the dish from the inside of the cabinet; if you mount it from the outside, the toggle switch will stick out too far and it can be broken of easily. So, you want to place the gasket on what would have been the upper surface of the "lip" around the dish, so that it's between the dish and the MDF.
The parts cost for the whole shebang, assuming that .20 mH inductor is 20 gauge, is about $51.56 if my arithmetic is correct. Ouch!
Well, we can omit stuff. We can omit the switch, and that brings us down to $44.09.
That may still be on the high side. Okay, the lower peak is going to be the more audible one because it's over twice as wide. Narrow peaks are generally not as audible as gentler, wider ones. So if we just do the 1.5 kHz filter (and no switch), that would be about $19.93.
Incidentally, if I showed these filters following convention, each would look like one capacitor, one inductor, and one resistor, for reasons described earlier. In other words, this circuit would be said to consist of seven elements (including the switch), even though physically there are eleven elements. The crossover for my 1x12+horn cab consists of twenty-three elements. My point is not that more = better, but that in some cases the crossover is eating up a bigger share of the budget and design effort than you'd expect.
That's a lot of work you did on designing the filter, Duke. Thank you! I'm sure some folks will experiment with the filter out of pure curiousity and maybe b/c they have the parts laying around. The cost of the filter knocks it out of consideration for me, and that's OK. I'll be interested to read the inevitable comparos b/t the "with/without filter" sounds. I'm sure this little cab will sound great stock. My parts are on the way and I'll make a "build post" here in this thread when the time comes. Thanks again.
Mike
Mike
o0o, that 1.5Khz filter would be nice and do-able. i might include that in my budget if i do indeed decide to get this off the ground.
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Let me explain briefly how a parallel notch filter works:
The signal from the amplifier has three possible paths it can take: Through the capacitor, through the inductor, or through the resistor. So, let's take a detour and see what each of these components will do:
A capacitor blocks low frequencies. Now it's not a "brick wall"; it reduces low frequencies by 6 dB per octave below a "corner frequency" that is a function of the capacitor value and the impedance of the speaker.
An inductor blocks high frequencies, and is likewise not a brick wall.
A resistor attenuates all frequencies (not necessarily equally, but we won't get into that here).
At low frequencies the capacitor won't let anything through; the resistor is putting up some resistance, and the inductor is happy to let everything through unimpeded. The signal naturally takes the path of least resistance, which in this case is the inductor, and the signal passes through the filter at full strength (through the inductor).
At high frequencies it is the capacitor that is happy to let everything through unimpeded, so once again the signal passes through at full strength, this time through the capacitor.
But we've carefully chosen our capacitor and inductor values so that, at the desired frequency, they both strongly resist the passage of the signal. So in this frequency region the path of least resistance is through the resistor. And the resistor is going to attenuate that signal somewhat, and that attenuation is what causes a dip or "notch" in the response. My job here has been to shape and align the notches so that they have the most net beneficial interaction with the peaks in the speaker's frequency response curve.
Now one other thing we need to look at is the power handling of our components. For example, why am I only using a single 20-watt resistor in that upper notch filter? The reason 20 watts is sufficient is twofold: First (and foremost), the signal will be passing through that resistor only over a tiny portion of the spectrum. Second, up at 2.5 kHz, there isn't very much energy from a bass guitar anyway, compared with much lower down the spectrum.
Notice that I've built a bit more power handling into the 6.7 ohm resistor (formed by three paralleled 20-ohm resistors) in the lower, wider notch filter. Because this notch is wider we have more bandwidth passing through the resistor, and a bass will have more harmonic energy down there anyway.
The signal from the amplifier has three possible paths it can take: Through the capacitor, through the inductor, or through the resistor. So, let's take a detour and see what each of these components will do:
A capacitor blocks low frequencies. Now it's not a "brick wall"; it reduces low frequencies by 6 dB per octave below a "corner frequency" that is a function of the capacitor value and the impedance of the speaker.
An inductor blocks high frequencies, and is likewise not a brick wall.
A resistor attenuates all frequencies (not necessarily equally, but we won't get into that here).
At low frequencies the capacitor won't let anything through; the resistor is putting up some resistance, and the inductor is happy to let everything through unimpeded. The signal naturally takes the path of least resistance, which in this case is the inductor, and the signal passes through the filter at full strength (through the inductor).
At high frequencies it is the capacitor that is happy to let everything through unimpeded, so once again the signal passes through at full strength, this time through the capacitor.
But we've carefully chosen our capacitor and inductor values so that, at the desired frequency, they both strongly resist the passage of the signal. So in this frequency region the path of least resistance is through the resistor. And the resistor is going to attenuate that signal somewhat, and that attenuation is what causes a dip or "notch" in the response. My job here has been to shape and align the notches so that they have the most net beneficial interaction with the peaks in the speaker's frequency response curve.
Now one other thing we need to look at is the power handling of our components. For example, why am I only using a single 20-watt resistor in that upper notch filter? The reason 20 watts is sufficient is twofold: First (and foremost), the signal will be passing through that resistor only over a tiny portion of the spectrum. Second, up at 2.5 kHz, there isn't very much energy from a bass guitar anyway, compared with much lower down the spectrum.
Notice that I've built a bit more power handling into the 6.7 ohm resistor (formed by three paralleled 20-ohm resistors) in the lower, wider notch filter. Because this notch is wider we have more bandwidth passing through the resistor, and a bass will have more harmonic energy down there anyway.
Thanks for the detailed explanation on how all the components on a crossover work! I always wondered what job each component had and how they interacted. I learned a lot by reading that. 
man, this stuff is a true science. i'll just stick to Computer Forensics (atleast there's no math) and let you experts do what you do best...
i'll be over here, coloring
really though, that filter sounds like it could be a neat tool to include in this box if guys have some resistors and whatnot hanging around. a true art, for sure, designing that type of thing...
i'll be over here, coloring
really though, that filter sounds like it could be a neat tool to include in this box if guys have some resistors and whatnot hanging around. a true art, for sure, designing that type of thing...
Could I just build the response shaping filter using small cheaper surface mount parts and put it in the effects loop of my amp? I'm guessing the impedance difference would mess with the parts values, though. And might have to be active to not kill the signal level?
Just as an FYI, Google's picasaweb.com is pretty good and has 1 gig for free. That's what I'm using for all my TB hosting.
Would you be interested in a preamp side filter with the roughly the same response? Might be a cheaper option, I can throw something together if it would be of use.
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
I just figured I'd shoot the idea out to someone who knows rather than just blabbing and not knowing if what I'm saying is right. "I need a 305.00 hz crossover and a tweeter with a range of 1-4500000000 hz, right?"
But seriously, I thought it might be worth a shot. If you know of any driver that'd be worth using (you seemed to hint at a few?) feel free to send me in that direction. I'm not afraid to attenuate it a tad if needed either. Subtle highs are still okay by me. I'd rather make two small cabs that I love than two that I like.
Arrrrgh, I really didn't want to get into tweeters. But you make a very valid point; I'd rather you have two small cabs that you love as well.
Here's the Parts Express dome tweeter page, which has a pretty wide variety:
Invalid Link Removed
Looks to me like a 2.5" outer flange diameter is about the biggest we can get away with. Aside from efficiency and power handling and bandwidth considerations, we'd like for the tweeter to be physically protected against accidental pokes.
I see a good possibilty, but it's not perfect:
[Invalid or Expired Link Removed]
Apparently decent power handling, decent efficiency, decent bandwidth... but inadequate protection from finger pokes. Okay the shallow horn profile will help somewhat in that respect. Unfortunately Parts Expess doesn't sell any protective grilles small enough.
Designing a crossover that results in a smooth transition with attention to the off-axis response would take more time than I want to put in, sorry. I have customers who are probably firing off angry e-mails even now, Duke get back to work on the cab I'm paying you for! So, here's a quick-n-dirty set of suggestions for this tweet:
Coming from a wire connected to the amp side of the woofer's notch filter "suite", you will have these four components in series: A .10 mH 18 gauge inductor; a 3.3 uF capacitor; a 4-ohm 20-watt resistor, and another 4-ohm 20-watt resistor. Thence onward to the + terminal of the tweeter. The - wire coming from the tweeter can connect to the - wire coming from the woofer, or connect back at the input terminal.
Normally we couldn't get away with something this simple, but in this case I think we can. The ferrofluid in the tweeter's magnetic gap reduces the magnitude of the impedance peak at resonance to the point where I think a simple filter, consisting of only a capacitor, will work... provided the crossover frequency is fairly high. In this case, I think it's high enough.
The .10 mH inductor helps protects against clipping. The 8 ohms of series resistance pads the tweeter back, and results in less net variation in our impedance curve which makes our single-capacitor crossover even more likely to work well. The capacitor is chosen to give us a gentle rolloff starting at about 4.5 kHz (take another look at the curve of the S2010 and you'll see why I chose 4.5 kHz).
We're looking at padding back the tweeter quite a bit here, but remember that tweeter is going to have a fairly wide pattern relative to the woofer up in the crossover region, so to get a smooth net transition it's okay if the tweeter is down a bit on-axis because it will be up (relative to the woofer) off-axis.
Also the relatively gradual rolloff of the tweeter means that it's still making an significant contribution well south of the crossover point, effectively widening the radiation pattern of the cab a bit at the upper end of the woofer's passband.
No guarantees on any of this of course. And I have to leave it up to you to come up with a protective grill of some kind, if you think that's necessary.
Best of luck to you if you decide to go ahead with the tweeter mod.
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Could I just build the response shaping filter using small cheaper surface mount parts and put it in the effects loop of my amp? I'm guessing the impedance difference would mess with the parts values, though. And might have to be active to not kill the signal level?
Nope, sorry. Line-level filtering is out of my league, but the component values quickly become impractical into the sort of impedances involved (the inductor values anyway).
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Thanks for the detailed explanation on how all the components on a crossover work! I always wondered what job each component had and how they interacted. I learned a lot by reading that.
Thanks, Mike! I posted that part because I think it's quite interesting, and not too far "out there".
But as you can see, there are some soul-searching cost/benefit decisions that have to be made. Is increasing the build cost by nearly 50%, and maybe quadrupling the time investment, worth it? I dunno... especially when the most useful feature of the response-shaping circuitry might end up being the bypass switch!
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Thanks for the heads-up about picasaweb site! I'll check it out.
As for a preamp side filter, well that's up to others besides me as to whether they'd want it. But if it's a cheaper option, might make a lot of sense. If you can "throw together" something like that, despite your innocuous moniker you are obviously highly skilled in the dark arts!
Duke, thanks for putting this up.
Question: would taking only the 2.5k notch stack work for the EMDL22512 ?
Question: would taking only the 2.5k notch stack work for the EMDL22512 ?
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
I don't know what the EMDL22515 is. Can you provide a link?
These types of filters tend to be tailored to a specific speaker. I come up with the component values by trial and error. That being said, a notch filter stands a pretty good chance of being somewhat beneficial on another speaker with a similar impedance curve and a similar frequency response anomaly.
- Nov 24, 2008
- 4,343
- 7,163
- Disclosures
- Owner & designer, AudioKinesis; Auth. mfg, Big E (Home Audio only)
Ah, thanks. That's probably my favorite five-pound woofer; one of these days I might use it in a two-way commercial system.
In a tweeterless application, the 2.5 kHz notch filter would nip the top off that Mt. Everest peak. But I don't think it would be enough to really smoothe it out.
I picked up that little fake-barefaced 1x12 mini cab from RPSands a few months ago. loaded with the s2010 its a great practice/coffeehouse style cab. 2 of them would be fine for small gigs if volumes are sane. They weigh approx 3 pounds each, or so it feels.
that s2010 is a great "little" driver imo. You can get them for less than $74, for sure.
Duke, nice of you to link this DIY build.... Have you ever tried designing a cab from the ground up? ;-)
** i didnt read the whoel thread... looks like I was a little late to the party. ***
that s2010 is a great "little" driver imo. You can get them for less than $74, for sure.
Duke, nice of you to link this DIY build.... Have you ever tried designing a cab from the ground up? ;-)
** i didnt read the whoel thread... looks like I was a little late to the party. ***
Duke, the dual notch filter makes perfect sense now that you'd laid out the schematic. Thanks! 
UncleFluffy, pre-compensating wouldn't be a bad idea, you'd get rid of the inductor coils so there's an immediate bonus (no hum collection by the coils). Could it be done with a single or dual op-amp?
UncleFluffy, pre-compensating wouldn't be a bad idea, you'd get rid of the inductor coils so there's an immediate bonus (no hum collection by the coils). Could it be done with a single or dual op-amp?
aaiiee - the sound of a sherpa being hurled into the gorge to appease the gods ...
In english we translate that as 'jeeze, it's not even 8:00 and already my bubble is burst'
I'd be a charter member of that group. See when you show up to do the bluegrass / new grass thing, those guys hate to be reminded that it's an electric whatzat that you play ... so the more your rig disappears the better. So far the little Mojo Sonic is small enough - my 12.6 would blow their little mando-minds ... still a 10.3 fEarful could be killer for the wine bar, lounge, coffee shop, 'grass jam guys ...
ah well, I'll just start saving my sheckles for 2012 ...
