How do you baffle compensate a full range?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So, I've built me a quick little set of speakers from the famous Tymphany/Peerless TC9FD18 3 1/2" full range driver. They sound good, but a bit thin, so measured them and got a couple dB loss from (pulling from memory) 300Hz to 900Hz, which I presume to be baffle loss. My question is, without an eq before the speaker, how do you generally compensate a full range driver for baffle loss? Just pad ~1k and above by ~3dB to match?
 
Padding over / above a particular bandwidth requires filtering of some type, as opposed to simple broad-band attenuation, and is EQ - whichever side of the driving amplifiers on which it's performed - exactly how elaborate is a different story. Many computer based music players / sound card drivers have sufficient EQ function to achieve this, and can save a range of EQ profiles for different speaker type / music genres.

Of course if baffle is wide enough, or enclosure is close to back wall, the step loss frequency can fall as low as the driver can deliver, or at least be far less noticeable

Then of course the brute force method would be a bi-pole (front to back) configuration of drivers - no EQ required, but you'll still be constrained by the drivers' capabilities and cabinet baffle dimensions.
 
Last edited:
Padding over / above a particular bandwidth requires filtering of some type, as opposed to simple broad-band attenuation, and is EQ - whichever side of the driving amplifiers on which it's performed - exactly how elaborate is a different story. Many computer based music players / sound card drivers have sufficient EQ function to achieve this, and can save a range of EQ profiles for different speaker type / music genres.

Of course if baffle is wide enough, or enclosure is close to back wall, the step loss frequency can fall as low as the driver can deliver, or at least be far less noticeable

Then of course the brute force method would be a bi-pole (front to back) configuration of drivers - no EQ required, but you'll still be constrained by the drivers' capabilities and cabinet baffle dimensions.

You can do it passively using an inductor and resistor in parallel on the positive line to the speaker. The inductor value is a function of the baffle dimensions, the resistor value determine es the amount of baffle step attenuation. Lots of info on the web, do a search. So.e typical values could be 1.5 mH for the inductor and from 4 to 6 ohms for the resistor. Search for the French consortium they have free Excel based software to help calculate the values.

Pjn
 
Hi kcducttaper,

If measurement you have is good enough and can be saved as frd-file there is possibility tune precise component values over in free XSim, below example show how perfect flat frq/phase driver S1 will look when adding components R1 and L1, grey curves is components bypassed and blue curves is within those values shown. So if you doubleclick S1 driver and load your measurement in that dialog menu frq response will reflect your measurement and you can thereafter sweep component values up and down in value until frq response suits best your situation.

XSim is to find here: http://www.diyaudio.com/forums/software-tools/259865-xsim-free-crossover-designer.html
 

Attachments

  • 2.png
    2.png
    102.9 KB · Views: 797
It could be argued that signal level EQ could be as efficacious a treatment for BS loss as with passive components at line level - but maybe it's time to let that point go.

Oh, for sure and I'd totally go that route for a more permanent install. I'm just after a 'plug-n-play' speaker that has baffle loss compensated for regardless of where or what I hook it up to.
 
Formulas, schematic:
https://trueaudio.com/st_diff1.htm
Baffle Step Compensation
Loudspeaker Diffraction Loss and Baffle Step Compensation Circuits

0-6dB is the available range to correct baffle step diffraction, and the correct amount varies with the room size, the woofer size and your preferred tonal balance. 0dB is what you have now. 6dB is the theoretical max, which you would need to get flat FR in a true anechoic space. Room reflections support bass freqs so you need less than 6dB correction when playing in a room. Smaller rooms support bass into higher freq, so smaller room gets less baffle step correction. 3-4dB of baffle step correction is typical amount applied for 2way bookshelf speakers with similar size woofer as typical FR driver.

The frequency of the filter is determine by formula based on the width of your baffle. But different box shapes can effect the audible FR so some tweaking is usually needed to get it to feel just the right amount of bass.

Passive baffle step correction filter actually cuts the amplitude of the frequencies above the baffle step frequency. This reduces the speaker's sensitivity and the amplifier's headroom. So if you are using a low power tube amp, consider this in your decision. To apply 3dB of bafflestep correction will require double the output power of your amp to produce the same SPL at 1kHz as with no BSC filter. But it will sound fuller and more satisfying even at lower volumes when the bass is compensated. Amps with low feedback like low power tube amps frequently use with FR speakers will usually have rising harmonic distortion with power output. This could affect the tonal quality of your system after BSC is installed.

You'll want to minimize DCR (DC Resistance) of the inductor in your passive baffle step correction filter. Added resistance between the amplifier and the voice coil reduces the electrical damping of the driver causing lost detail and bass impact. You can minimize DCR either by using a magnetic core inductor (steel lam type,) which will add some low order harmonic distortion at higher frequencies due to the magnetic hysteresis of the core material, or you can use an air core inductor which doesn't add magnetic hysteresis but is a longer winding which increases the DCR. Air core coils are typically used for higher frequency drivers, but magnetic core coils are cheaper which can be significant at the inductance values needed for low frequency filters. Since your FR plays high frequency you should use air core inductor. Since it also plays low frequency, you should use thickest gage wire coil you can afford. 12ga is probably the biggest you'll find available from stock, but 14ga is probably big enough for your application. Smaller ga number = bigger wire diameter.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.