Balancing SPL of Woofer and Tweeter

[SpeakerEasy]

Hi, all!

I have two questions. The first one is theoretical because I am a newbie with electronics and want to make sure I am understanding that side.
1) Can I use a resistor in parallel with a tweeter to lower its SPL to be closer to that of a woofer?

The second question is because I am trying to build speakers the right way....
2) Even if I am understanding the electronics theory correctly, is using a resistor in this way acceptable/recommended?


I have attached a simple (possibly crude) diagram showing a first-order high-pass filter for an 8 Ohm woofer rated at 97 dB/W and then a first-order high-pass filter for a 4 Ohm tweeter rated at 104 dB/W. My thought was, if I have the tweeter in parallel with a 6 Ohm resistor, the power going to the tweeter will be lowered and the SPL of the woofer and tweeter will be close.

I'm in over my head, so please correct anything I have wrong - if it's too much, I'll hold off on the speakers and do more learning :D

Circuit resistance = 10.4 Ohms
We'll say the amplifier is delivering 20 W to the circuit.
The voltage for the circuit is 14.42 V

Voltage across SPKR1 = (8/10.4) * 14.42 = 11.09 Volts
Voltage across SPKR2 = (2.4/10.4) * 14.42 = 3.33 Volts

Power to SPKR1 = 11.09²/8 = 15.37 W
Power to SPKR2 = 3.33²/4 = 2.77 W

SPKR1 = 99.2 dB SPL at 10 ft
SPKR2 = 98.7 dB SPL at 10 ft


I plugged the relevant values into this website: https://myhometheater.homestead.com/splcalculator.html and it seems like the apparent SPL of the two drivers are very close.

I went through the same steps with the arbitrary value of 150 W to the circuit
Power to SPKR1 = 115W
Power to SPKR2 = 20.79W

SPKR1 = 107.9 dB SPL at 10 ft
SPKR2 = 107.5 dB SPL at 10 ft

Thank you in advance for your time and responses! :D

Jess

[Chris Roemer]

Howdy.

I suggest you get/read/learn "SpeakerBuilding 201" to get started. Your idea is correct; paralleling a resistor w/a tweeter WILL drop the tweeter output, BUT ...

this is MOST commonly done by either adding series resistance to a tweeter, or by using an "L-pad" on the tweeter ( which uses a series resistor (SR) and a parallel resistor (PR) ).

Your schemo won't work as drawn. Only a tweeter would have a high-pass (HP) filter, which CAN be just a simple cap (and some attenuation), but a 2nd order filter works much better.
A woofer has a low-pass (LP) filter, which blocks highs. A single series coil won't due much to roll most woofers off due to their rising impedance (above about 200-300 Hz). The coil will work better if you use a "Zobel" across the woofer (which flattens that impedance rise), but more people will just use a 2nd order filter on the woofer as well.

You've basically drawn two drivers in series w/some series capacitance. All the lows will be filtered out. Both drivers will get the same freq. range.

MOST people on here (at least the good ones) use one of the (free) crossover simulation softwares that you can find links to here on PETT.
They use individual driver .frd (freq. resp.) and .zma (impedance) files (many of which you can also get here at PE). It IS a fun (and addicting) hobby.
Whatever you do, don't buy anything yet (except maybe that book).

[billfitzmaurice]

A parallel resistor won't lower the tweeter level. Wired parallel the tweeter and resistor will both receive the same voltage as the tweeter alone, so the tweeter output is unchanged. The impedance load on the amp is lowered, which means current delivery goes up, and that can cause problems. It also changes the knee frequency of the high pass, so you have to recalculate that. A series resistor causes a voltage drop in the tweeter, and therefore lowers tweeter output, but this also changes the knee frequency of the high pass, so again you have to recalculate the filter. The correct method is to use an LPad, which maintains the same impedance load as the tweeter alone. BTW, first order filters are seldom used. They don't provide adequate protection for tweeters, and allow far too much pass band overlap between drivers.

[fpitas]

Buy the book Chris suggested, and get a crossover SIM program. Then, perhaps look at some crossovers from the threads here, and get an idea what yours might end up like.

[a4eaudio]

I have the Speaker Building 201 and I like it, but unless you are going to start this as a hobby you can find all of the information you need on the internet, it is just not all in one nice convenient place. There is a "Sticky" thread on the main page of this forum, "The Speaker Building Bible." that is an okay place to start but it is old and a lot of the links are dead and there is new and better information available. For crossover information a few places to start are:

https://www.audioholics.com/diy-audi...speaker-design

https://www.diyaudioandvideo.com/Tut...uildCrossover/

https://www.diyaudio.com/forums/mult...asurement.html

[Chris Roemer]

CORROBORATE, please, anyone?

My "instinct" tells me that paralleling a tweeter (in the OP's schematic - which is NOT how XOs work, no one's questioning that) will drop it's output.
Bill says "no" (and as someone here just recently said "Bill knows a heck of a lot more about speakers than I do").

I AM aware that voltage divides in series, and that the tweeter and bypass resistor will see the same voltage, but current divides in parallel, doesn't it?
I'm asking someone here to run the same sim that I did to verify my results, only because the sim software I use is NOT one of the "main stream" packages.

If I put 2 AURA NS3-193-8 drivers in a series string (for now, leave the caps from the OP's schemo out, or set them to 100000uF), the pair has the same output as a single driver, but ea. one drops -6dB down, and (of course) the string's load rises from 8n nom. to 16n(ohms).
If I add an 8n bypass resistor across ONE NS3, the total SPL (for both drivers) stays (pretty much) the same, BUT...
Instead of both drivers being -6dB down (from their combined output), the "woofer" is down only -3.5dB (approx.) whereas the "tweeter" (the one bypassed) is about -9.5dB down, making the "tweeter" about 6dB "quieter" than the "woofer".

The reason why I think this is so is that the bypassed NS3 is effectively only 4n "nominal" (when bypassed), whereas the initial NS3 (the "woofer") remains near 8n nom. So now the voltage IS dividing (in series) w/the "woofer" dropping about 2/3rds the voltage, while the "tweeter"(& resistor) drop the other 1/3rd. Any takers?

(btw - the 2 NS3s end up w/pretty different xfr fns: the tweeter is actually rolled OFF on the top end (prob. by the "woofers" v.c. Le) to the tune of -16dB down @ 10khz, and it's output @ Fs (near 80Hz) is about -20dB. The "woofer" OTOH, is only down <-1dB @ Fs. The 2 drivers xfer fns are (perfectly?) reciprocal to the point where their combined output STILL (perfectly?) matches that if the bypass resistor is removed. Very interesting.)

If you want/need my F/Z files, just ask.

[billfitzmaurice]

current divides in parallel, doesn't it?
.

It does, and that's why the current draw on the amp will go up. The output of the tweeter doesn't because excursion is relative to voltage swing.

[PWR RYD]

It does, and that's why the current draw on the amp will go up. The output of the tweeter doesn't because excursion is relative to voltage swing.

That's not covered in Speaker Building 201? Maybe in 101?

[SpeakerEasy]

Thank you, everyone!

I do plan on taking this very seriously and have been trying to learn, but mostly using a random combination of whatever I'm finding online (and I did read through parts of the Speaker Building Bible here) and I definitely have gaps in my learning, even at the 101 level.

I'll get Speaker Building 201 and look through the rest of the links shared here.

I see now what I did wrong with my schemo as far as the XO layout goes. One related question about the woofer's filter.... I have been practicing some theoretical speaker designs in WinISD and that shows Cone Excursion for even woofers being too great with low frequencies, so do woofers need a pass-band filter to protect them (with a high-pass for cone excursion) *and* a low-pass (for the crossover with the tweeter)?

Jess

[billfitzmaurice]

That's not covered in Speaker Building 201?

I never read it. It's covered in Elements of Acoustical Engineering by Harry Olson (1940), and undoubtedly in previous works as well, but that's the oldest I have in my library.

I have been practicing some theoretical speaker designs in WinISD and that shows Cone Excursion for even woofers being too great with low frequencies, so do woofers need a pass-band filter to protect them

That's an option, but it's more practical not to drive it with enough voltage to exceed xmax. The primary use of the xmax graph is to determine how loud you can safely go, and if it's not loud enough for the intended application you either have to go with a different driver or you need to use more than one.

[wogg]

CORROBORATE, please, anyone?

My "instinct" tells me that paralleling a tweeter (in the OP's schematic - which is NOT how XOs work, no one's questioning that) will drop it's output.
Bill says "no" (and as someone here just recently said "Bill knows a heck of a lot more about speakers than I do").

I AM aware that voltage divides in series, and that the tweeter and bypass resistor will see the same voltage, but current divides in parallel, doesn't it?
I'm asking someone here to run the same sim that I did to verify my results, only because the sim software I use is NOT one of the "main stream" packages.

If I put 2 AURA NS3-193-8 drivers in a series string (for now, leave the caps from the OP's schemo out, or set them to 100000uF), the pair has the same output as a single driver, but ea. one drops -6dB down, and (of course) the string's load rises from 8n nom. to 16n(ohms).
If I add an 8n bypass resistor across ONE NS3, the total SPL (for both drivers) stays (pretty much) the same, BUT...
Instead of both drivers being -6dB down (from their combined output), the "woofer" is down only -3.5dB (approx.) whereas the "tweeter" (the one bypassed) is about -9.5dB down, making the "tweeter" about 6dB "quieter" than the "woofer".

The reason why I think this is so is that the bypassed NS3 is effectively only 4n "nominal" (when bypassed), whereas the initial NS3 (the "woofer") remains near 8n nom. So now the voltage IS dividing (in series) w/the "woofer" dropping about 2/3rds the voltage, while the "tweeter"(& resistor) drop the other 1/3rd. Any takers?

(btw - the 2 NS3s end up w/pretty different xfr fns: the tweeter is actually rolled OFF on the top end (prob. by the "woofers" v.c. Le) to the tune of -16dB down @ 10khz, and it's output @ Fs (near 80Hz) is about -20dB. The "woofer" OTOH, is only down <-1dB @ Fs. The 2 drivers xfer fns are (perfectly?) reciprocal to the point where their combined output STILL (perfectly?) matches that if the bypass resistor is removed. Very interesting.)

If you want/need my F/Z files, just ask.

A parallel resistor across any driver will only reduce output if there is also some resistance in series. If I'm reading your scenario correctly, the other driver is effectively the series element in the makeshift L pad, so the parallel resistor does reduce the power to that driver.

If there are no series elements, the parallel resistor simply heats up and the driver sees the exact same voltage as before.

[AEIOU]

CORROBORATE, please, anyone?

My "instinct" tells me that paralleling a tweeter (in the OP's schematic - which is NOT how XOs work, no one's questioning that) will drop it's output.
Bill says "no" (and as someone here just recently said "Bill knows a heck of a lot more about speakers than I do").

I AM aware that voltage divides in series, and that the tweeter and bypass resistor will see the same voltage, but current divides in parallel, doesn't it?
I'm asking someone here to run the same sim that I did to verify my results, only because the sim software I use is NOT one of the "main stream" packages..

Your instinct has merit. Theoretically a parallel resistor isn't supposed to, but the reality can be quite different depending on the specific amplifier. Yeah, not all amps have sufficient headroom to drive a lower impedance load that would be presented by a tweeter paralleled with a resistor, so there can be an apparent reduction in tweeter output simply because the amplifier cannot keep up. Limitation of the amplifier, especially with inexpensive entry level Receivers. Of course it would be easy to test, demonstrate, by simply switching in a parallel resistor on a test bench.

[scholl]

One of my favorite sites http://www.mh-audio.nl/Calculators/att.html

Of course PE has lots of resistors to dial in the right balance.

The real question is what balance do you like. Some like hot tweeters others like them a little lower than the midbass.YMMV

Beware the variable L-Pad some can degrade sound. you can use one to find the balance you like than measure the values and replace the variable with fixed resistors.

[rpb]

MOST people on here (at least the good ones) use one of the (free) crossover simulation softwares that you can find links to here on PETT..

The really good ones don't need the software, but they do enjoy playing with it.

This thread is hysterical! Don't stop.

[AEIOU]

One of my favorite sites http://www.mh-audio.nl/Calculators/att.html

Of course PE has lots of resistors to dial in the right balance.

The real question is what balance do you like. Some like hot tweeters others like them a little lower than the midbass.YMMV

Beware the variable L-Pad some can degrade sound. you can use one to find the balance you like than measure the values and replace the variable with fixed resistors.

Yes, stay away from variable L-pads, instead used fixed resistors in an L-pad configuration.