Feedback on drivers for 3-way bookshelf

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am considering a 3-way bookshelf (maybe better called a mini-tower as it will be about a 24" high front ported design) to be used as mains in a 5.2 system (for both HT and m-ch music).

The speakers will be crossed to sub(s) at 80Hz, so I'm thinking an f3 of about 60Hz should be OK.

Current driver thoughts are:

- woofers, a pair of Seas L16 (crossed to mid at 400 - 600Hz)

- midrange, Audax HM100Z0 (crossed to tweeter at 2.5 - 4 kHz)

- tweeter, Hyquphon OW4

These will be placed within 6" of the back wall, so this will be considered in BSC calculations.

Any thoughts, concerns, reactions from those with experience to this concept (driver selection, crossover ranges, etc)?

Thanks (a humble rookie)

PS. What impact would replacing the Seas L16 with Audax HM170C0 have? Higher sensitivity and lower power handling - anything else?
 
Last edited:
The mid will have to be padded down heavily. I take caution and parallel resistors to make sure they wont see to much power when padding the mid of a 3 way.

The CTC distance of your mid/tweet will limit how high you can cross. A neo tweeter tends to make more sense with a small mid because of this.
 
The mid will have to be padded down heavily. I take caution and parallel resistors to make sure they wont see to much power when padding the mid of a 3 way.

Would most of the padding be due to lower sensitivity of woofers (but doesn't having two of them help here a lot), or the tweeter?

The CTC distance of your mid/tweet will limit how high you can cross. A neo tweeter tends to make more sense with a small mid because of this.

If the flangeless OW tweeter is used then the CTC would be under 4". Also, isn't CTC more of an issue in horizontal designs?

PS. Thanks for educating a newbie.
 
Would most of the padding be due to lower sensitivity of woofers (but doesn't having two of them help here a lot), or the tweeter?



If the flangeless OW tweeter is used then the CTC would be under 4". Also, isn't CTC more of an issue in horizontal designs?

PS. Thanks for educating a newbie.


I missed where you said a pair of woofers, still even with minimal BSC you're looking at around 88db from the Seas (and a low impedance). The mid is in the low 90's, plus there's band pass gain for the mid in a 3 way. A pair of those Audax woofers would get you better match.

Drivers should be within a minimum of 1 wavelength at the crossover point.

Sorry didnt know there's a flange less OW4, all I've seen is this.

An externally hosted image should be here but it was not working when we last tested it.


How are you designing the crossover?
 
How are you designing the crossover?

I actually don't have DIY skills, but do enjoy doing the research to come up with the high level design for the speakers.

Ultimately this will end up being a commission to Salk Sound (he builds gorgeous speakers and is open to custom projects), so the crossover would get designed by Dennis Murphy (he designs great crossovers).
 
I missed where you said a pair of woofers, still even with minimal BSC you're looking at around 88db from the Seas (and a low impedance). The mid is in the low 90's, plus there's band pass gain for the mid in a 3 way. A pair of those Audax woofers would get you better match.

Could you clarify my minimal understanding of band pass gain?

I'm starting to think the Audax woofers may be a better match.

Drivers should be within a minimum of 1 wavelength at the crossover point.

So if my math is right, 3kHz max distance is 4.5" and 3.4kHz max is 4.0"

Sorry didnt know there's a flange less OW4

It's an option with all OWs.

Is the efficiency of the tweeter apt to be an issue here?
 
Could you clarify my minimal understanding of band pass gain?

I'm starting to think the Audax woofers may be a better match.



So if my math is right, 3kHz max distance is 4.5" and 3.4kHz max is 4.0"



It's an option with all OWs.

Is the efficiency of the tweeter apt to be an issue here?


Bandpass gain is basically where the summing of the overlapping outputs of the woofer/mid/tweeter are greater then the output of the mid alone.

Your CTC numbers are good.

The tweeter sensitivity could limit you, depending on the amount of BSC with those Audax woofers. I would consider the SB29RDNC personally.
 
diyAudio Moderator
Joined 2008
Paid Member
...the in-room response will have reinforcement at lower frequencies so the amount of compensation will be different than if you were designing for flat frequency response in open space.
The wall will only be exposed to sound that the baffle has already sent in that direction so the baffle is already out of the picture, at least until some low frequency where the gap disappears. The spacing will cause a series of related peaks and dips as they come back.

The wall should therefore give a boost at some point which might, for example, happen just below the baffle frequency giving the illusion of it being extended. Above and below this there may be a cancellation.

If I had a point it would be that to some degree the wall is a separate issue, and you might at some later point want to make some adjustment with it.
 
Bandpass gain is basically where the summing of the overlapping outputs of the woofer/mid/tweeter are greater then the output of the mid alone.

OK - understood (just didn't know the phrase).

Your CTC numbers are good.

Phew, my math degree wasn't useless after all.

The tweeter sensitivity could limit you, depending on the amount of BSC with those Audax woofers. I would consider the SB29RDNC personally.

So, in doing a 3-way, is there an ideal sensitivity relationship between the 3 raw drivers in order to get a flat FR from the final system? (i.e. if the mid is 94db, then an ideal tweeter would be xx and woofer yy).
 
The wall will only be exposed to sound that the baffle has already sent in that direction so the baffle is already out of the picture, at least until some low frequency where the gap disappears. The spacing will cause a series of related peaks and dips as they come back.

The wall should therefore give a boost at some point which might, for example, happen just below the baffle frequency giving the illusion of it being extended. Above and below this there may be a cancellation.

If I had a point it would be that to some degree the wall is a separate issue, and you might at some later point want to make some adjustment with it.

Thanks.
 
OK - understood (just didn't know the phrase).



Phew, my math degree wasn't useless after all.



So, in doing a 3-way, is there an ideal sensitivity relationship between the 3 raw drivers in order to get a flat FR from the final system? (i.e. if the mid is 94db, then an ideal tweeter would be xx and woofer yy).

The woofers are what set the system level, figure out that number will be and you can make sure the mid and tweet are at least that level. With full BSC and two parallel woofers it would be the same as a single woofer with no BSC. 1/2 BSC would be +3 db to the single woofer sensitivity number.

Nothing wrong with padding down the mid/tweet, you just cant pad the woofers.
 
The woofers are what set the system level, figure out that number will be and you can make sure the mid and tweet are at least that level. With full BSC and two parallel woofers it would be the same as a single woofer with no BSC. 1/2 BSC would be +3 db to the single woofer sensitivity number.

Nothing wrong with padding down the mid/tweet, you just cant pad the woofers.

Apologies, but I'm still missing it here :eek:

BSC is to deal with the fact that high frequencies will be reinforced by reflection off of the baffle (question: this will happen for all frequencies whose wave length is less than 1/2 baffle width, is this right).

So these higher frequencies need to be padded so that with reflection they will match lower frequencies output (right).

My comment earlier about the wall is that, wouldn't frequencies that wrap back and hit the wall be reflected (just out of phase by the distance of baffle to wall); so the only frequencies that have no reinforcement are the ones whose wavelengths hit the side of the speaker (ie, greater than 1/2 the baffle width, but less than 1/2 baffles width plus baffle-to-wall distance).

How out-to-lunch am I?
 
In terms of your woofer choice, I think you're going to have to choose between transient response and sensitivity unless someone who has heard both of them can chime in with a favorite.

Personally, I would have a preference for the better transient response of a sealed woofer to match with that Audax mid. Sealed and using Zaph's measured data ( Zaph|Audio), the Seas L16RN-SL will hit an f3 of 66Hz in 8L with Qtc of .711 and heavy fill. Sealed, the Audax HM170CO only hits 93Hz in 7.6L, Qtc of .707 and minimal fill. To go lower the Audax must be ported, f3 of 63Hz in 10L tuned to 60Hz but then the transient response isn't as good. The Seas also has very low 3rd and 5th harmonic distortion between 80 and 700Hz, (see below) altho if the Audax data is to be believed, the HM170 also shows outstanding cone control via it's CSD.

The big difference is that with the Seas' 84dB sensitivity, 2 of them in parallel gets you to 90dB and then minus say 4dB for BSC, leaves you with 86dB sensitivity. That means you have to pad down the mid (93dB) by about 7dB. That may or may not affect the overall sound depending on who you are talking to, but it'll certainly mean that your amp will have to do more work than with the Audax woofers who'll end up at about 92dB sensitivity after the same calculations. I'm assuming btw that your amp will be happy with the 4 ohm load of 2 woofers in parallel. If impedance is a factor, it looks like the Audax is the one with the slightly higher impedance minimums.

Notice however that your tweeter choice won't work with the Audax woofers since it's only 90dB in sensitivity. (You don't really want to be padding down your woofers) No great loss since I think you could do better for less money too. I would possibly think of a planar (B&G Neo3) or ribbon (maybe a Fountek) to match with your mid or look at Zaph|Audio again for more choices. Directivity starts to change at 2000Hz with that mid so I'd be looking for a tweeter that will be happy crossed at 3000Hz. But you'll need to decide on the woofers first.

Since you're looking at a bookshelf speaker I would think that box size might be a consideration but the enclosure sizes of the Seas and Audax are pretty close so that might not be a factor. Sealed in 8L with an f3 in the 60's is already pretty darn good. I guess if anything, you can push the Seas into a smaller sealed box than you can the Audax into a smaller vented box if you really had to.

Oh and in terms of max SPL and xmax, both are again about similar as long as they are both crossed to a sub around 80Hz.

Hope that helps.
 

Attachments

  • Seas-L16RN-SL-HD.gif
    Seas-L16RN-SL-HD.gif
    20.1 KB · Views: 185
diyAudio Moderator
Joined 2008
Paid Member
Rick, this is about right. The transition happens over a range of around a couple of octaves. The higher frequencies won't wrap. The lowest cancellation often happens below all this due to the greater distance to the wall and back and the fact that it only needs half a wavelength.

Some higher frequency sound will diffract around the cabinet anyway, especially without rounding. and this will likely be more detrimental than the subsequent reflection.
 
In terms of your woofer choice, I think you're going to have to choose between transient response and sensitivity unless someone who has heard both of them can chime in with a favorite.

......

Hope that helps.

Lots to think about in there, thanks. It was actually the L16RNX i was thinking about (not sure if that makes a big difference).

In terms of sound quality from a layman's terms, what would the better transient response translate to, and is sealed that much better?
 
Rick, this is about right. The transition happens over a range of around a couple of octaves. The higher frequencies won't wrap. The lowest cancellation often happens below all this due to the greater distance to the wall and back and the fact that it only needs half a wavelength.

So does that mean that if designing for rear wall reinforcement that BSC is actually not a 4db or so sensitivity ramp with lower frequencies higher, but a plateau, with only the ones that wrap to the side needing to be higher?

If that's correct then are the edges of the plateau marked by full or half-wavelengths and full or half baffle widths? (i.e. if the baffle width is 8", 3.35kHz wavelength is about 4", is that where BSC starts or is it closer to 1.7kHz where a 1/2 wavelength is about 4")

And yes, I understand lots of ,easuring will be necessary to fine tune - but I am trying to understand the science.

Some higher frequency sound will diffract around the cabinet anyway, especially without rounding. and this will likely be more detrimental than the subsequent reflection.

Fortunately, Jim always builds rounded front edges (no matter how much my wife would prefer the clean lines of hard edges).
 
It was actually the L16RNX i was thinking about (not sure if that makes a big difference).

L16RNX is designed more for a ported enclosure. L16RN-SL is designed more for sealed or ported. Personally, I like to kill the backwave as much as possible so it doesn't return back out thru the cone so I prefer to heavily stuff the cabinet. Density of stuffing will affect the f3 tho. Here are some numbers, but you should download a box sim program if you want to learn more (UniBox - Unified Box Model for Loudspeaker Design if have Excel, WinISD if you don't - LinearTeam ).

RNX sealed, light stuffing:
Qtc=.707
f3=78Hz
Vol=5.4L

RNX sealed, heavy stuffing:
Qtc=.707
f3=88Hz
Vol=3.3L

RN-SL sealed, heavy stuffing:
Qtc=.707
f3=67Hz
Vol=8.1L

RNX ported:
f3=41Hz
Vol=11L

RN-SL ported:
f3=28Hz
Vol=28L

In terms of sound quality from a layman's terms, what would the better transient response translate to, and is sealed that much better?

In layman's terms, first understand that a transient is the abrupt starting and stopping of a sound and it happens all over the place in music and movies etc. So a better transient response means cleaner, tighter and more detailed. It means less background noise. More controlled. It means the cone does a better job of starting and stopping when it's supposed to instead of continuing to move when the signal stops and of course, motion=sound.

Would everyone agree that it's noticeably better? Probably not, but it's my preference which is all I'm relating here.

Let me put it another way. The reason to go ported is for the added low frequency it provides. When you are using a sub, that can become a little redundant. Besides, the sealed box requires less volume and many people find the sealed LF roll off of 12dB/octave to blend better with the sub than the 24dB/octave roll off a ported enclosure, altho this is going to vary with the actual f3 and the xo point to the sub.

Here's the idea in graphic terms. We can roughly equate the driver's step response with it's transient response. First up is the Audax HM100 which you are trying to match the woofer to. Then follows the Seas L16RN-SL sealed with Qtc=.707 and then Qtc=1.25. Last is the RN-SL ported in 28L.

BTW, how many subs do you have?
 

Attachments

  • CB Step Response Audax HM100Z0 4 Qtc.894.gif
    CB Step Response Audax HM100Z0 4 Qtc.894.gif
    21.1 KB · Views: 159
  • CB Step Response Seas L16RN-SL (Z) Qtc .707.gif
    CB Step Response Seas L16RN-SL (Z) Qtc .707.gif
    21.7 KB · Views: 164
  • CB Step Response Seas L16RN-SL (Z) Qtc .12.gif
    CB Step Response Seas L16RN-SL (Z) Qtc .12.gif
    21.8 KB · Views: 154
  • VB Step Response Seas L16RN-SL (Z).gif
    VB Step Response Seas L16RN-SL (Z).gif
    23.3 KB · Views: 131
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.