I did some testing over a few weeks, fixed L-Pads made from 2 watt resistors, and at levels so loud I can't hear anyone talk, or myself, they did not even get warm on the midrange and tweeter. Just find an L-Pad calculator online, and find the values for example, -2, -3, -4 db. They are about $.50 each, so $1 for each db step.
Please be sure the 2 watt resistors do not get hot at your listening levels, please. I would not want to hear of anyone's speakers catching fire. Then, it is fixed and will not crackle or able to be messed with.
Please be sure the 2 watt resistors do not get hot at your listening levels, please. I would not want to hear of anyone's speakers catching fire. Then, it is fixed and will not crackle or able to be messed with.
1 Watt resistors can be used with some efficient tweeters/super tweeters.. but unless sure, I'd agree with Steve.
There can be a negligible improvement in using some types of resistor and many good ones come in about a 1W package, however the improvement isn't worth compromising the required values. Ie: I may have only 3r3 of a good resistor and try to use that value when possible. It's normally better to use components of different/lesser quality to make another value when needed.
Sometimes L-pads aren't as critical to the absolute values used. The impedance can end up a little higher or lower than 8 ohms most of the time, as long as it has been flattened a little and you have the right L and C. I'll sometimes tweak the two resistors to suit some value of inductor I have to use.
There can be a negligible improvement in using some types of resistor and many good ones come in about a 1W package, however the improvement isn't worth compromising the required values. Ie: I may have only 3r3 of a good resistor and try to use that value when possible. It's normally better to use components of different/lesser quality to make another value when needed.
Sometimes L-pads aren't as critical to the absolute values used. The impedance can end up a little higher or lower than 8 ohms most of the time, as long as it has been flattened a little and you have the right L and C. I'll sometimes tweak the two resistors to suit some value of inductor I have to use.
If we use an assumed 8 ohms and we seek -2dB, -3dB, & -4dB, here are the values I get -
-2dB -
Rs = 1.6454 ohms
Rp = 30.8969 ohms
-3dB -
Rs = 2.3364
Rp = 19.3822
-4dB -
Rs = 2.9523
Rp = 13.6777
Off the top of my head, I would use these values assuming they are available -
-2dB -
Rs = 2x 3.3 ohm in parallel for 1.65 ohms total
Rp = 1x 10 ohms + 1x 20 ohm in series for 30 ohms
-3dB -
Rs = 2x 5 ohm in parallel for 2.5 ohms
Rp = 1x 20 ohm
-4dB -
Rs = 1x 3 ohm
Rp = 1x 8.2 ohm + 1x 5 ohm in series, or 1x 10 ohms + 1x 4 ohm in series for 13.2 ohms to 14 ohms total
I think this might be what AllenB is getting at, you use a combination of standard values in series parallel combinations to come reasonably close to the calculated value.
Prices for 10 watt 5% are about $0.65 each. Cheap enough.
Actually there are probably resistor values that would work better but Parts Express doesn't list them in order of size, so exact values are hard to track down.
I think, matching left speaker/right speaker attenuation is more important than precise attenuation. If the attenuation is -1.9dB or -2.1dB, is less important than both side being the same.
Just a thought.
Steve/bluewizard
-2dB -
Rs = 1.6454 ohms
Rp = 30.8969 ohms
-3dB -
Rs = 2.3364
Rp = 19.3822
-4dB -
Rs = 2.9523
Rp = 13.6777
Off the top of my head, I would use these values assuming they are available -
-2dB -
Rs = 2x 3.3 ohm in parallel for 1.65 ohms total
Rp = 1x 10 ohms + 1x 20 ohm in series for 30 ohms
-3dB -
Rs = 2x 5 ohm in parallel for 2.5 ohms
Rp = 1x 20 ohm
-4dB -
Rs = 1x 3 ohm
Rp = 1x 8.2 ohm + 1x 5 ohm in series, or 1x 10 ohms + 1x 4 ohm in series for 13.2 ohms to 14 ohms total
I think this might be what AllenB is getting at, you use a combination of standard values in series parallel combinations to come reasonably close to the calculated value.
Prices for 10 watt 5% are about $0.65 each. Cheap enough.
Actually there are probably resistor values that would work better but Parts Express doesn't list them in order of size, so exact values are hard to track down.
I think, matching left speaker/right speaker attenuation is more important than precise attenuation. If the attenuation is -1.9dB or -2.1dB, is less important than both side being the same.
Just a thought.
Steve/bluewizard
Last edited:
I tend to find that component tolerances of 10% can be ok while first prototyping a crossover, enough of what's important is still coming through, such as whether speaker issues are being rolled off or reduced in level (eg: making up values from odd resistors). Being precise with values won't fix the acoustic problems, but it's more important later.
Last edited:
I've run a comparison where: the upper grey trace is a plain 6dB slope, the red trace is an 8 ohm parallel and 4 ohm series L-pad with a 10uF capacitor using the shown impedance.
The green trace is from reducing the parallel resistor to 2.5 ohms. The series resistor needed to be reduced to 1.5 ohms and the capacitor went to 25uF.
The green trace is from reducing the parallel resistor to 2.5 ohms. The series resistor needed to be reduced to 1.5 ohms and the capacitor went to 25uF.
Attachments
Just to summarise my understanding.
A series resistor changes the impedance seen by the amp. This effects the frequency response of the tweeter.
An Lpad flattens the impedance. This does not effect the frequency response.
A follow up question if I may. Theoretically, if there was a choice of 2 identical tweeters, with differing sensitivities, and one needed an Lpad and the other needed no attenuation, would it be advantageous to use one instead of the other?
A series resistor changes the impedance seen by the amp. This effects the frequency response of the tweeter.
An Lpad flattens the impedance. This does not effect the frequency response.
A follow up question if I may. Theoretically, if there was a choice of 2 identical tweeters, with differing sensitivities, and one needed an Lpad and the other needed no attenuation, would it be advantageous to use one instead of the other?
"...if there was a choice of 2 identical tweeters, with differing sensitivities, and one needed an Lpad and the other needed no attenuation, would it be advantageous to use one instead of the other? ..."
If it were only as simply as that. Treble can be piercing. That is, it can sound louder even at the same measured level.
So, as an example, if you have a woofer with a Sensitivity of 89dB and a Tweeter with a Sensitivity of 89dB, that is no guarantee that the tweeter will not still need a bit of attenuation. Though, it will be very little.
To the extent that it is possible the two part L-Pad is impedance stable which is an advantage to the Tweeter crossover, as you design it either for the actual tweeter impedance at the crossover point, or you design it to the generic rated tweeter impedance.
However, if you use a Series only resistor, then you have to design to that resistor plus the impedance of the tweeter at the crossover. And, neither the attenuation or the impedance is as constant as with an L-Pad.
That said, there are many speaker crossover design that use a Series Resistor. But that is typically done by people who know what they are doing and who are specifically trying to achieve the effect that a series resistor will give them.
But, in my opinion, for some one designing a basic crossover, and need control fixed impedance attenuation, the simplest solution is a common L-Pad. Resistors are cheap, buy a bunch of them to cover a range of attenuation, ultimately choose the one that works best for you.
In my opinion, there is a time and place for a Series Resistor, and that time and place are in the hands of an experience crossover designer.
But then that ...is... just my opinion.
Steve/bluewizard
If it were only as simply as that. Treble can be piercing. That is, it can sound louder even at the same measured level.
So, as an example, if you have a woofer with a Sensitivity of 89dB and a Tweeter with a Sensitivity of 89dB, that is no guarantee that the tweeter will not still need a bit of attenuation. Though, it will be very little.
To the extent that it is possible the two part L-Pad is impedance stable which is an advantage to the Tweeter crossover, as you design it either for the actual tweeter impedance at the crossover point, or you design it to the generic rated tweeter impedance.
However, if you use a Series only resistor, then you have to design to that resistor plus the impedance of the tweeter at the crossover. And, neither the attenuation or the impedance is as constant as with an L-Pad.
That said, there are many speaker crossover design that use a Series Resistor. But that is typically done by people who know what they are doing and who are specifically trying to achieve the effect that a series resistor will give them.
But, in my opinion, for some one designing a basic crossover, and need control fixed impedance attenuation, the simplest solution is a common L-Pad. Resistors are cheap, buy a bunch of them to cover a range of attenuation, ultimately choose the one that works best for you.
In my opinion, there is a time and place for a Series Resistor, and that time and place are in the hands of an experience crossover designer.
But then that ...is... just my opinion.
Steve/bluewizard
You guys have it wrong, it's the speakers dc resistance that is used as the 'impedance' in the formula and the L-pad is calculated so that the speakers 'impedance' remains constant. 5w resistors are sufficient here as the tweeter gets less than 15w in a 3-way system from a 100w amp.
For two sources of good info:
http://www.bcae1.com/lpad.htm
http://sound.westhost.com/lr-passive.htm#s6.0
For two sources of good info:
http://www.bcae1.com/lpad.htm
http://sound.westhost.com/lr-passive.htm#s6.0
Last edited:
Here's a power distribution graph. That is an instantaneous or continuous power graph. Actual dissipation will be as low as 10% of the graph since music is pulsed power. A steady sine wave at any of the frequencies for example would match the graph.
Why Do Tweeters Blow When Amplifiers Distort?
Why Do Tweeters Blow When Amplifiers Distort?
Not quite. I'll summarise the four changes of interest, assuming A is a series resistor, B is a plain L-pad, and C is an L-pad that uses smaller resistance values than the calculators would give but is designed to give the same overall level of attenuation as A and B.
1. Changing the overall tweeter level: A, B and C will do this.
2. The frequency response will alter slightly due to the resistor interaction with the driver impedance: A, B and C will do this but A is the worst and C is the least. If someone is designing through measurements, this change may not be a problem otherwise it is best to leave the response closer to flat so there is less guessing.
3. The (overall) impedance seen by the amp changes: A and C will do this. A will give the highest impedance and C the lowest. Again this may not necessarily cause an unwanted problem.
4. The impedance will flatten so the crossover capacitor causes less unexpected change in the frequency response: A, B and C will do this. A is worst and C is least. Again this may not matter if designing through measurements but it helps if not.
Last edited:
I really do think an Lpad will dampen mechanical oscillations some, depends on the parallel resistor value being low. It's like an electric brake, take a motor, switch a small value resistor like a 10 ohm. Spin it with the resistor on and off. It can even get stiff to simply turn the motor with the resistor across it's terminals. If you are lucky at pushing a woofer in/out, try it with the terminals shorted and open, see if you feel any difference. It shouldn't be affected by applied power, but will stiffen it to moving by natural spring or inertia.
That would show up on things like kick drums and synthesizer music, slap bass, things with sharp fall time waveforms. An added bonus I have never heard any talk of.
The audition is the real answer, if it sounds ok and will not cause any harm, who cares if it doesn't follow the rules ?
That would show up on things like kick drums and synthesizer music, slap bass, things with sharp fall time waveforms. An added bonus I have never heard any talk of.
The audition is the real answer, if it sounds ok and will not cause any harm, who cares if it doesn't follow the rules ?
Last edited:
This is true but it's a bit of a red herring. A driver connected directly to an amp is also shorted as a voltage source is a virtual short circuit. A parallel resistor in this case would have no discernable effect. Change this to an L-pad and the series resistor will also be in parallel with the driver.
However it will also be changing the applied voltage vs frequency, and hence the Q factor of the rolloff.
However it will also be changing the applied voltage vs frequency, and hence the Q factor of the rolloff.
Thanks for all this. And from the 2nd link in reply 31 from balerit (http://westhost etc) there is a very uncompromising answer to my question.
Quote: "Attenuation networks are a necessary evil, it is immeasurably better not to use them at all". He goes on to say that using drivers with sensitivities that allow this is the ideal.
Quote: "Attenuation networks are a necessary evil, it is immeasurably better not to use them at all". He goes on to say that using drivers with sensitivities that allow this is the ideal.
I don't want to get into one of these but I see it as Zsource = Zload/DF. DF = Zload/Zsource, looks like a tail chase. DF is the Damping Factor. Wouldn't the actual output Z be dynamic and a on a single pulse, Zo goes very low then to some non-perfect residual conductance, mainly as output stage biasing error in DC amps, and the transformer DCR if transformer coupled, capicitor coupling complicates this.
Audio amps aren't my specialty but the same as a servo drive and sometimes used as one. In that case an electric vehicle would not coast at zero throttle. But they do and many electric vehicles use a switched resistor across the motor terminals as a brake assist.
At least we learn some stuff, and I'm learning that tri-amping might be an answer to all the minute details.
The parallel resistor can add some effect to damping. And unless I'm only wanting to see it, my midrange driver sure seemed to sound better with an Lpad.
Audio amps aren't my specialty but the same as a servo drive and sometimes used as one. In that case an electric vehicle would not coast at zero throttle. But they do and many electric vehicles use a switched resistor across the motor terminals as a brake assist.
At least we learn some stuff, and I'm learning that tri-amping might be an answer to all the minute details.
The parallel resistor can add some effect to damping. And unless I'm only wanting to see it, my midrange driver sure seemed to sound better with an Lpad.
Last edited:
Wavewhipper, if the voltage applied to a speaker is varied due to the voltage divider effect between the resistor and the driver impedance then it could be seen as a separate issue, independent of the effect you mention. Yet in combination it still modifies the Q of the rolloff.
I can't find the post that midrange refers to but I suspect the necessary evil bit could be seen overstated. It is more important to get the levels right and a good speaker can be made with attenuation included. I'd question becoming obsessed with minimalist crossovers due to concerns about how everything has a downside, (been there
).
I can't find the post that midrange refers to but I suspect the necessary evil bit could be seen overstated. It is more important to get the levels right and a good speaker can be made with attenuation included. I'd question becoming obsessed with minimalist crossovers due to concerns about how everything has a downside, (been there
).- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.