speaker frequency responce
We often look at driver frequency responces. Some of them may be very flat but what amplitude do the manufactorers measure them at? Is the frequency responce of a speaker generally the same at all amplitudes? I never thought about this before but maybe someone can answer this question. thanks
sardonx
We often look at driver frequency responces. Some of them may be very flat but what amplitude do the manufactorers measure them at? Is the frequency responce of a speaker generally the same at all amplitudes? I never thought about this before but maybe someone can answer this question. thanks
sardonx
The frequency response charts you see generally are measured at one meter with 2.8 volts input.
For 8 ohm speakers, 2.8 volts corresponds to one watt.
However, for 4 ohm speakers, 2.8 volts equals 2 watts. This is 3 dB higher than 1 watt.
Some 4 ohm speaker manufacturers do give the response at 1 watt, (2 volts into 4 ohms). However, many simply run 2.8 volts through their 4 ohm speakers and never mention that for a 4 ohm speaker, 2.8 volts is 2 watts, not one.
A frequency response chart will generally tell you what the midrange sensitivity is. Often, it will be about 88 dB at 2.8 volts.
For 8 ohm speakers, 2.8 volts corresponds to one watt.
However, for 4 ohm speakers, 2.8 volts equals 2 watts. This is 3 dB higher than 1 watt.
Some 4 ohm speaker manufacturers do give the response at 1 watt, (2 volts into 4 ohms). However, many simply run 2.8 volts through their 4 ohm speakers and never mention that for a 4 ohm speaker, 2.8 volts is 2 watts, not one.
A frequency response chart will generally tell you what the midrange sensitivity is. Often, it will be about 88 dB at 2.8 volts.
Re: speaker frequency responce
Well, that is the general idea.
Mostly, yes. However, a speaker driven at 100 or 200 watts will likely produce certain distortions that will show up as frequency changes. Hopefully, they won't affect the response too much.
In vented enclosures, the behavior of the port can change as air comes rushing thorugh it. The higher the amplitude, the more interference in the port, the more noise it produces, and eventually the port inhibits the flow of air through it, changing the frequency response.
This effect can be minimized by making your port nice and large, and adding flares to it.
I would say that the ideal is that frequency response remains the same no matter the amplitude, but under the strain of high power and high output, the response can change.
sardonx said:
Well, that is the general idea.
Mostly, yes. However, a speaker driven at 100 or 200 watts will likely produce certain distortions that will show up as frequency changes. Hopefully, they won't affect the response too much.
In vented enclosures, the behavior of the port can change as air comes rushing thorugh it. The higher the amplitude, the more interference in the port, the more noise it produces, and eventually the port inhibits the flow of air through it, changing the frequency response.
This effect can be minimized by making your port nice and large, and adding flares to it.
I would say that the ideal is that frequency response remains the same no matter the amplitude, but under the strain of high power and high output, the response can change.
Accurate?
Do you ever listen to the weather report on the television or radio? There is a saying that the reports are provided for entertainment purposes only.
I fear that same saying should be applied to too many frequency response graphs. I have some familiarity with testing. I know how you can change test system boundaries or limits (often miscalled parameters) to change the look of the frequency response. Even diy folks do this. Check the Parts Express diy web pages. Many of the response graphs will tell you that they are "smoothed." Al least most of these diy folks will tell us that they are smoothing their responses.
Still, the alteration of the response leaves little validity except entertainment purposes.
There is also the problem that some manufacturers do not know how to use their own test equipment. A number of years back myself and one other person were sent to a manufactuer to do a magazine article (this was in the early years of computerized acoustic measurement systems). While there I sat down with their test equipment and produced much different results than they did. Their engineer was away at the time (problems with vapor deposition and he was at the vacuum chamber) and he would have been the person doing all the in house testing. I did not get a chance to talk with the person responsible for testing during the visit. Several days later I get a phone call from the engineer asking how I produced the results. It became clear that he wasn't all that confident about how to use his own test equipment.
I guess the lesson here is to examine frequency response graphs, but not take them too seriously.
Just as an aside, I do impulse testing. My average power is low, but my peak power is high. My typical test session will be run with about 75 watts of peak power. Sometimes I go even higher. When I do low frequency woofer testing, my impulse duration is fairly long (this is important to mantain frequency resolution). I can easily see the woofer diaphragm and surround deflect during testing. I can also test to Xmax without thermal concerns. Just one of the many reasons I choose impulse testing and deal with the low rejection of ambient noise.
Great question, just no great answers,
Mark
Do you ever listen to the weather report on the television or radio? There is a saying that the reports are provided for entertainment purposes only.
I fear that same saying should be applied to too many frequency response graphs. I have some familiarity with testing. I know how you can change test system boundaries or limits (often miscalled parameters) to change the look of the frequency response. Even diy folks do this. Check the Parts Express diy web pages. Many of the response graphs will tell you that they are "smoothed." Al least most of these diy folks will tell us that they are smoothing their responses.
Still, the alteration of the response leaves little validity except entertainment purposes.
There is also the problem that some manufacturers do not know how to use their own test equipment. A number of years back myself and one other person were sent to a manufactuer to do a magazine article (this was in the early years of computerized acoustic measurement systems). While there I sat down with their test equipment and produced much different results than they did. Their engineer was away at the time (problems with vapor deposition and he was at the vacuum chamber) and he would have been the person doing all the in house testing. I did not get a chance to talk with the person responsible for testing during the visit. Several days later I get a phone call from the engineer asking how I produced the results. It became clear that he wasn't all that confident about how to use his own test equipment.
I guess the lesson here is to examine frequency response graphs, but not take them too seriously.
Just as an aside, I do impulse testing. My average power is low, but my peak power is high. My typical test session will be run with about 75 watts of peak power. Sometimes I go even higher. When I do low frequency woofer testing, my impulse duration is fairly long (this is important to mantain frequency resolution). I can easily see the woofer diaphragm and surround deflect during testing. I can also test to Xmax without thermal concerns. Just one of the many reasons I choose impulse testing and deal with the low rejection of ambient noise.
Great question, just no great answers,
Mark
The simple answer is that the frequency response will be identical
at all power levels the device/ loudspeaker is linear, ie. does not
distort.
So testing below the nominal 2.83V rms should be fairly pointless.
At this level and above distortion graphs are much more helpful
at pointing to trouble / potential trouble than looking for changes
in the frequency response.
/sreten.
at all power levels the device/ loudspeaker is linear, ie. does not
distort.
So testing below the nominal 2.83V rms should be fairly pointless.
At this level and above distortion graphs are much more helpful
at pointing to trouble / potential trouble than looking for changes
in the frequency response.
/sreten.i thought that's about the answer i'd get. i figured that responce should stay the same unless the driver is being over driven or in some cases just playing very loud. Another reason why linearity is important! I suppose that's why drivers with SD motor assemblies and careful attention to linearity are the more expensive ones. Among other reasons too...
sardonx said:
SD, (Symmetrical Drive), motor assemblies used to be just on the most expensive speakers, but now affordable model lines are beginning to use them. The Peerless CSC, CSX and HDS are inexpensive lines that use them. So are the Blaupunkt polymer car subwoofers. These are all speakers that retail for well under $100.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.