Speaker Frequency Response

[Frosty_47]

Komrades,

I can't seem to find any companies that will provide frequency response graph (bode plot) of their speakers. I want to start designing an active cross-over but cannot proceed due to lack of technical data on woofers, mid range, and tweeters.

If you know of any manufactureres that provide this information, please share.

Thanks

 

[dknguyen]

Wasn't it 40Hz for a subwoofer? I forget why, but it didn't have anything to do with the speakers thesmelves as much as it had to do with the human ear.

 

[Frosty_47]

Wasn't it 40Hz for a subwoofer? I forget why, but it didn't have anything to do with the speakers thesmelves as much as it had to do with the human ear.

No, Frequency response varies a lot depending on woofer. Yes hearing is a great factor however, nothing is more clearer than a bode plot of a particular speaker because a cross over system can be designed to compensate the non-linear response of a particular speaker. A speaker with a flat response will not have unpleasant (to the ear) peaks or dips at certain frequencies. And if it does, it can be compensated for with a properly designed active cross-over.

 

[dknguyen]

You can use an equalizer to fix that kind of thing. I read on on an audiophile forum it had something to do with the ear being able to tell which speaker the sound is coming from and below 40Hz it can't.

 

[Hero999]

Lots of speaker datasheets give frequency response graphs, even if they are fuzzy.

**broken link removed**
**broken link removed**
**broken link removed**
**broken link removed**

The frequency response can depend a lot on the enclosure.

You can easily create your own bode plot using a signal generator, an oscilloscope, an electret mic and preamplifier.

Connect the signal generator to the speaker, set to a sinewave of a known frequency and peak output voltage, check with a 'scope to make sure. Connect the mic to the preamplifier and connect to 'scope, record the amplitude of the waveform. Repeat the procedure for frequencies between 40Hz and 15kHz using a logarithmic frequency scale, testing the input voltage with the 'scope each time and adjust the signal generator if it changes.

Convert all the recorded voltages to dB using the sensitivity on the mic datasheet. Electret mics tend to have a flat frequency response curve, look at the datasheet and take adjust your bode plot to account for any peaks an troughs.

 

[Frosty_47]

Lots of speaker datasheets give frequency response graphs, even if they are fuzzy.

**broken link removed**
**broken link removed**
**broken link removed**
**broken link removed**

The frequency response can depend a lot on the enclosure.

You can easily create your own bode plot using a signal generator, an oscilloscope, an electret mic and preamplifier.

Connect the signal generator to the speaker, set to a sinewave of a known frequency and peak output voltage, check with a 'scope to make sure. Connect the mic to the preamplifier and connect to 'scope, record the amplitude of the waveform. Repeat the procedure for frequencies between 40Hz and 15kHz using a logarithmic frequency scale, testing the input voltage with the 'scope each time and adjust the signal generator if it changes.

Convert all the recorded voltages to dB using the sensitivity on the mic data sheet. Electret mics tend to have a flat frequency response curve, look at the data sheet and take adjust your bode plot to account for any peaks an troughs.

Thanks for advice. I will hook up speakers to function gen and a fixed gain amplifier. I will see if I can find a software that will plot the frequency vs. amplitude on the Tektronix's oscilloscopes we have in school.

Howsoever, I fear that I will get my ass kicked in the lab during testing due to unpleasant sounds.

 

[Speakerguy]

I was a loudspeaker design hobbyist for a while. You need good measurement equipment and you need the actual loudspeaker units in the intended cabinet. You can achieve reasonably accurate quasi-anechoic results 200Hz-20KHz, and can assume things are linear below that (just use T/S modeling). Below 200Hz the listening room and its effects will dominate the observed frequency response.

Are you trying to do a full-range speaker? Two way, three way, two way intended for use with sub? What are your drivers? I've got about a dozen designs under my belt and have several friends who own speakers of my own design. Good links:

Zaph|Audio
Madisound Speaker Components | Assisting speaker builders for more than 25 years.
Parts-Express.com - Speakers, Speaker Building, Home Audio and Video, Pro Audio, Electronic Parts & Accessories PARTS EXPRESS, Speakers, Speaker Parts, Guitar speakers, Bass speakers, Woofers, Drivers, speaker upgrades and replacement speakers. Emine
Linkwitz Lab - Loudspeaker Design

Personally I would design an active system based on one of the passive designs on Zaph's site for a first attempt, maybe his SR71 project. The drivers and cabinets can be purchased as a kit at Madisound.com. Measure the complex impedance of the drivers in the actual cabinet, simulate how the passive components affect the signal at the driver terminals when terminated by that complex impedance, and then go about making an active filter that has that transfer function.

Easier said than done. I have a few thousand in speaker measurement and design gear, and that's separate from my electronics gear.

 

[Speakerguy]

Also, since you are in Canada, Solen is a good supplier up there. Whereabouts up there are you? I can maybe refer you to someone on one of my forums if someone is local to you.

 

[Bob Scott]

Komrades,

If you know of any manufactureres that provide this information, please share.

Dayton?

You can get Thiele/Small parameters from the catalog at www.qcomponents.ca

As mentioned, www.partsexpress.com has on-line data sheets with curves for some concientious manufacturers like Dayton, plus the Thiele/Small parameters.

However, if you want to be accurate with your crossovers, you have to mount the woofer in the intended enclosure and then take impedance measurements. The enclosure and resonance has a very large effect on impedance.