Hi All,
I am building a speaker which uses some LM3886TM chip amps. Its going well so far and have just ordered some HF drivers which are a little loud compared with the my LF drivers. I am hoping a limit their volume with some passive components but am unsure of the details of how this is done and not sure what this is referred to as?
Would someone be able to suggest some available methods and what they are called to allow me to research this further?
Many thanks
Mark
I am building a speaker which uses some LM3886TM chip amps. Its going well so far and have just ordered some HF drivers which are a little loud compared with the my LF drivers. I am hoping a limit their volume with some passive components but am unsure of the details of how this is done and not sure what this is referred to as?
Would someone be able to suggest some available methods and what they are called to allow me to research this further?
Many thanks
Mark
Gentle compression of average volume can be done with low voltage incandescent lamps in series with the driver. In some PA speakers these are also used as fuses. As the filament heats up the resistance increases.
A T pad could also be made for fixed attenuation and you can build several for different attenuations and use a switch or patch bay to determine the dB attenuation you like. It gives an impedance match looking both ways. Use non-inductive resistors.
This file discusses symmetrical attenuation networks.
https://www.americanradiohistory.co...Section-9-NAB-Engineering-Fifth-Edition-8.pdf
Page 9-9 may show you what you want.
A T pad could also be made for fixed attenuation and you can build several for different attenuations and use a switch or patch bay to determine the dB attenuation you like. It gives an impedance match looking both ways. Use non-inductive resistors.
This file discusses symmetrical attenuation networks.
https://www.americanradiohistory.co...Section-9-NAB-Engineering-Fifth-Edition-8.pdf
Page 9-9 may show you what you want.
Hi Opcom,
Thanks for your response. The T pad seems like a good option, from what I had read thus far it seems that this is something which would be best if applied to the input signal side rather than on the speakers themselves? (speakers being rated at 60watts RMS, 4ohms)
I assume i would need some quite serious resistors to handle the current if they were on the speakers themselves?
Many thanks
Mark
Thanks for your response. The T pad seems like a good option, from what I had read thus far it seems that this is something which would be best if applied to the input signal side rather than on the speakers themselves? (speakers being rated at 60watts RMS, 4ohms)
I assume i would need some quite serious resistors to handle the current if they were on the speakers themselves?
Many thanks
Mark
To limit the maximum volume in your system without distortion, you would reduce the gain so the loudest source would never be too loud. This could be done with a series resistor at each input, selected by trial. You can buy a series resistor attenuator to plug into an RCA socket, or make your own. A potentiometer at each input instead could also be used to make the initial setup simpler. You want to do this level attenuation at line level, not at speaker level.
Last edited:
I did wonder if a resistor in series would do the trick, I do have lots of different value resistors so this would be convenient.
Does either the T pad or series options have any particular benefits over the other?
Does either the T pad or series options have any particular benefits over the other?
You may want to head over to the Speakers (Multi-Way) forum, as what you have here is basically a crossover design question. Mismatched driver sensitivity is the norm rather than the exception, and the crossover is an integral part of loudspeaker performance.
There are a few possible incentives for using a mix of active and passive crossover:
* You might have a hard time keeping amplifier hiss in check. An LM3886 at 26 dB isn't terribly noisy (around 50 µV) and as such often swamped by preceding stages, but it does pose a limit to the effectiveness of preceding L-pad attenuation, with potentially audible hiss at driver sensitivities around 100 dB / 2.83 V / m and up. Believe it or not, audible hiss is a very common complaint in (inexpensive) commercial active monitor speakers, and the actual power amp seems to be the culprit in some of them.
* HF driver distortion levels may benefit from current driving. Then the higher the source impedance, the better. (You don't have to use a series resistor for this. You could also configure the amplifier as a current source.) Mind you, this also means that you have to deal with the entire fun factor of the driver's impedance response, up to the point of having to suck out its base resonance with a series LC in parallel to knock down source impedance in the offending frequency range. Of course all of this has to be taken into account in the speaker design process.
Tweeter series resistors tend to be rated on the order of 10 W, though of course the details will depend upon driver power rating, efficiency, crossover frequency, crossover order etc.
Audio is not controlled impedance but rather a voltage transfer system (short into open), and as such you generally won't find T-pads but rather L-pads there (one series R + one parallel R). Over the plain series resistor, L-pads offer substantially lower source impedance and a reduced influence of load impedance on actual attenuation, at the cost of higher power dissipation. These characteristics make L-pads a no-brainer when you need attenuation in front of the power amp, where power levels are low either way and overly high source impedance would negatively impact noise and distortion. In fact, your average volume pot is set up as an L-pad - an L-pad is nothing but a tapped voltage divider. Its ratio of input and output impedance becomes more beneficial as attenuation increases - usually a good compromise between previous stage loading and output impedance can be reached by about 10-12 dB.
I would suggest looking at levels throughout the entire system here, as it would appear that not much consideration has been given to this aspect so far. Crossover levels at full power should be chosen wisely - run it too hot and you risk excessive distortion, run it too low and you risk excessive noise, especially in the HF section. I would shoot for somewhere around 1 Vrms (+/-, depending on circuit specifics), padding down afterwards as needed.
There are a few possible incentives for using a mix of active and passive crossover:
* You might have a hard time keeping amplifier hiss in check. An LM3886 at 26 dB isn't terribly noisy (around 50 µV) and as such often swamped by preceding stages, but it does pose a limit to the effectiveness of preceding L-pad attenuation, with potentially audible hiss at driver sensitivities around 100 dB / 2.83 V / m and up. Believe it or not, audible hiss is a very common complaint in (inexpensive) commercial active monitor speakers, and the actual power amp seems to be the culprit in some of them.
* HF driver distortion levels may benefit from current driving. Then the higher the source impedance, the better. (You don't have to use a series resistor for this. You could also configure the amplifier as a current source.) Mind you, this also means that you have to deal with the entire fun factor of the driver's impedance response, up to the point of having to suck out its base resonance with a series LC in parallel to knock down source impedance in the offending frequency range. Of course all of this has to be taken into account in the speaker design process.
Tweeter series resistors tend to be rated on the order of 10 W, though of course the details will depend upon driver power rating, efficiency, crossover frequency, crossover order etc.
Audio is not controlled impedance but rather a voltage transfer system (short into open), and as such you generally won't find T-pads but rather L-pads there (one series R + one parallel R). Over the plain series resistor, L-pads offer substantially lower source impedance and a reduced influence of load impedance on actual attenuation, at the cost of higher power dissipation. These characteristics make L-pads a no-brainer when you need attenuation in front of the power amp, where power levels are low either way and overly high source impedance would negatively impact noise and distortion. In fact, your average volume pot is set up as an L-pad - an L-pad is nothing but a tapped voltage divider. Its ratio of input and output impedance becomes more beneficial as attenuation increases - usually a good compromise between previous stage loading and output impedance can be reached by about 10-12 dB.
I would suggest looking at levels throughout the entire system here, as it would appear that not much consideration has been given to this aspect so far. Crossover levels at full power should be chosen wisely - run it too hot and you risk excessive distortion, run it too low and you risk excessive noise, especially in the HF section. I would shoot for somewhere around 1 Vrms (+/-, depending on circuit specifics), padding down afterwards as needed.
Not in this case, just use a series quality resistor of the value that you prefer at the amplifier input.
If the series resistor you prefer is large, perhaps over 30k, use a smaller resistor (5k-10k) to ground
across the input, to allow a smaller series resistor. Bear in mind an input attenuator does not actually
limit the volume, it only reduces the gain.
Last edited:
Hi All,
Thanks for your responses, certainly quite a few terms to get my head around.
It sounds like an L-Pad is a decent place to start, I have a reasonable grasp of voltage dividers so should be able to manage this.
I'm currently using a pre-amp based on a single TI-NE5532 and plan on wiring this into 4 LM3886's, two per channel (one for HF one for LF) as I'm doubtful a single LM3886 will handle the load of both the HF and LF drivers i'm using.
Once i get my HF drivers in the post i'll have a play with an L-Pad on the HF side and hopefully I can adjust them to a suitable level.
If i install the L-pad between the output of the TI-NE5532 and input of the HF LM3886 are 1/4 watts resistors likely to be enough? I'm really unsure how I would calculate this the output current of a TI-NE5532 is 38mA (according the the datasheet).
watt = amp(0.0038) × volt(?)
http://www.ti.com/product/NE5532#
Not sure which value for voltage to use from the datasheet.
Thanks Again
Mark
Thanks for your responses, certainly quite a few terms to get my head around.
It sounds like an L-Pad is a decent place to start, I have a reasonable grasp of voltage dividers so should be able to manage this.
I'm currently using a pre-amp based on a single TI-NE5532 and plan on wiring this into 4 LM3886's, two per channel (one for HF one for LF) as I'm doubtful a single LM3886 will handle the load of both the HF and LF drivers i'm using.
Once i get my HF drivers in the post i'll have a play with an L-Pad on the HF side and hopefully I can adjust them to a suitable level.
If i install the L-pad between the output of the TI-NE5532 and input of the HF LM3886 are 1/4 watts resistors likely to be enough? I'm really unsure how I would calculate this the output current of a TI-NE5532 is 38mA (according the the datasheet).
watt = amp(0.0038) × volt(?)
http://www.ti.com/product/NE5532#
Not sure which value for voltage to use from the datasheet.
Thanks Again
Mark
Quick comment... apologies if I've missed it skimming the thread.
you do realise that you mustn't apply the full audio bandwidth to your HF drivers (assuming they are not Piezo types) as that will destroy them in seconds. If your driving them from their own LM3886 then you must have some kind of high pass filter in the chain such as a normal crossover type arrangement or perhaps electronic filtering of the input signal to the LM3886.
Just don't apply the full bandwidth to your tweeters.
Carry on 🙂
you do realise that you mustn't apply the full audio bandwidth to your HF drivers (assuming they are not Piezo types) as that will destroy them in seconds. If your driving them from their own LM3886 then you must have some kind of high pass filter in the chain such as a normal crossover type arrangement or perhaps electronic filtering of the input signal to the LM3886.
Just don't apply the full bandwidth to your tweeters.
Carry on 🙂
Haha, no I didn't realise that tbh i just assumed that I could limit the gain on the LF channels to address the sensitivity of the HF being greater. But thanks for pointing that out, good to know. I can add that to the list of things I've learnt from this. 🙂
I'm not sure what is meant by this. The pre-amp has a volume control but will be driving both the HF and LF, as I understand it the HF will be too loud in relation to the LF so i'm trying to address this. Am I missing something else?
OK, I see. Did not realize that was the issue. In that case you could use what is known as an L-pad to pad down the loudest driver. These are sort of low value potmeters, and are available in several forms.
See L pad - Wikipedia
Jan
- Status
- Not open for further replies.