S.E.X.y speakers

Doc B. · 105511

0 Members and 2 Guests are viewing this topic.

Offline Doc B.

  • Administrator
  • Hero Member
  • *****
    • Posts: 9643
    • Bottlehead
on: October 02, 2009, 08:07:36 AM
NOTE: This is a very old thread and the links are no longer good.

(http://www.bottlehead.com/loosep/S.E.X.speaker_filtered.jpg)

Here's the formula for a splendidly simple DIY monitor/woofer speaker system designed to work with the S.E.X. amp. Total cost is less than $500! We don't sell any of these parts, but we have listed sources for each component. This should make a killer dorm room or office speaker setup, and it should work great in that first apartment, too. The small pre-finished cabinets and pretty Fostex drivers have a very high WAF, too.

 
   
(http://www.bottlehead.com/loosep/S.E.Xspeaker2-filtered.jpg)

Monitors

The monitors operate from about 125Hz up. They are composed of Fostex FE166E full range drivers in pre-finished Dayton .25 cubic foot sealed cabinets from Parts Express. The only scut work is that you have to cut holes in the front baffles for the Fostex drivers. Here's the basic parts list -

Drivers - Fostex FE 166E - from Madisound - one pair needed
http://www.madisound.com/cgi-bin/index.cgi?cart_id=7741375.12505&pid=324

Cabinets - Dayton .25 cu.ft. pre-built cabinets - from Parts Express - one pair needed
http://www.partsexpress.com/pe/showdetl.cfm?&DID=7&Partnumber=302-702

Cabinet Damping - Acousta-stuff polyfill - from Parts Express - 1lb. will be more than enough
http://www.partsexpress.com/pe/showdetl.cfm?&DID=7&Partnumber=260-317

Binding posts - Dayton knock-in binding post - from Parts Express - two pairs needed
http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=091-1245

Assembly is pretty self-evident. You will need to cut holes in the front baffles to mount the drivers, using the template provided with the drivers. We used an adjustable fly cutter that we got from Home Depot on a 1-1/2 HP floor standing drill press, and just centered the holes on the baffles. If you don't have that kind of equipment, maybe you can cut a round enough hole with a saber saw??? We'll leave any alternative hole making approaches up to you!

Stuff each cabinet with about 1/8 lb. of the Acousta-stuff. Install the binding posts and wire the speaker terminals to them - red to + and black to -, with your favorite flavor of wire. You may want to solder the connections, or one could install crimp-on quick connectors. Use long enough wires to make it easy to connect the wires with the drivers sitting outside the box. Mount the drivers and their gaskets, and the monitors are done!

Woofers

The woofers we originally spec'd are a pair (yes, you must use two because of the high 125Hz crossover frequency) of Parts Express subwoofers. However you may use whatever generic active subwoofers you prefer. Just be sure what you get can be set to a 125Hz or so crossover frequency. We call them woofers rather than subwoofers in this case precisely because they are running up so high. The Dayton subs mentioned here were chosen on the basis of cost, and because you will be ordering a most of the rest of the project from PE (man, they should give us a commission!). They are down firing subs. In theory a front firing sub will integrate better - and typically cost more.

Woofers - Dayton 10" 125W active subwoofer - Parts Express - you will need TWO.
http://www.partsexpress.com/pe/showdetl.cfm?DID=7&PartNumber=300-633

Hooking it up

Run the signal from your CD player/sound card/Ipod/whatever into your S.E.X. amp with interconnects. Connect speaker cables from the speaker output binding posts on the S.E.X. amp to the left and right monitors. Then run more speaker cables from the binding posts on the monitors to the input binding posts on the corresponding subwoofer, with the woofer connections reversed (red post on monitor to black post on woofer, black post on monitor to red post on woofer). Connecting the system this way will avoid running the signal through a 125Hz filter that's built into the monitor output of the subwoofer. That filter would interfere with the designed roll off of the monitors and mess up their bass response. 

Keep the woofers close to the monitors. They are running up to 125Hz and at that range you can definitely localize where the sound is coming from. Aligning the monitors so they are directly over the woofers will probably give the best integration of midrange-to-bass. Try to set the monitor height so that the center of the Fostex driver is close to level with your ear when seated. The Fostex drivers are very directional, and you may find that the the highs will sound best with the speakers aimed directly at you. They will also squawk until they break in! Be patient, they do tone down with use. Start with the woofer crossover frequency set to about 125Hz, the output turned all the way down, and the phase set to 0 degrees. While playing music bring the woofer gain up slowly until it sounds balanced with the midrange emanating from the monitor. There may be a slight "hole" in the lower midrange.  Fiddling with the phase switch (or knob) may help. Be careful to set the phase the same way on both woofers, or your bass may disappear! A great way to get the system to sound more balanced is to incorporate a baffle step correction filter (BSC), which compensates for the lower midrange/upper bass rolloff which occurs due to the narrow baffle width of the monitor cabinet.

Apologies for the delay in getting photos of a line-level BSC installation on the page. Meanwhile here's a circuit:

 

IN ------------------- 0.0012uF -------------- OUT
      |                      |
      --- 10k ohm -----
                      |
                      |
                    10k ohm
                      |
                    0.033uF
                      |
GND ------------------------------------------- GND

The 0.0012uF is optional and provides the treble boost if your room and taste calls for it. You can go up to twice that value if more treble is desired.

This circuit can be placed at the input of the power amp (e.g. SEX amp).

 

 
« Last Edit: July 22, 2016, 10:07:37 AM by Doc B. »

Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.


Offline Thoburn

  • Full Member
  • ***
    • Posts: 188
Reply #1 on: October 24, 2009, 01:30:02 PM
Hi Bob,

Is this the ciruit that Paul Joppa is refering to in his reply to my Lowther post over in the Paramount forum?

Thanks,
Thoburn (Toby)

Dynavector DV-20X2L > VPI Scout II > Musical Surroundings NovaPhonomena
Mac Mini > USB DACiTx
Stereomour > Lowther Medallion DX4 and Rythmic Subs
Monster Power HTS3600


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #2 on: October 24, 2009, 02:26:06 PM
Yes it is, but the formatting did not come across. Here's another try at the circuit:


Paul Joppa


Offline Thoburn

  • Full Member
  • ***
    • Posts: 188
Reply #3 on: October 25, 2009, 07:33:35 AM
Thank you Paul for the clarification. The values are very different than the circuit used inside my DX4 cabinet. And there is no inductor in your circuit. Is there an equation that gives starting values for the components based upon the driver and cabinet being used? The Fostex 166 in that skinny cabinet and subs to carry the bass is signifcantly different from my single Lowther in a 12 inch wide MLTL.

Toby

Dynavector DV-20X2L > VPI Scout II > Musical Surroundings NovaPhonomena
Mac Mini > USB DACiTx
Stereomour > Lowther Medallion DX4 and Rythmic Subs
Monster Power HTS3600


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #4 on: October 25, 2009, 11:15:29 AM
The circuits are quite different indeed! What's in the cabinet is at speaker level, meaning a load impedance of 8 ohms (or 4 or 15 or whatever the speaker is). At line level, between preamp and power amp, most devices want to see at least 10000 ohms load impedance. You can get away with RC circuits (no inductors) here because there is no need to maintain a low resistance in the bass for good speaker damping.

The circuit as drawn is optimized for an 8.5-inch baffle width. A wider baffle will want a larger value for the 0.033uF capacitor, in proportion - so for example, use 0.047uF for a 12-inch baffle. Of course this is a theoretical value assuming perfect speakers and an anechoic chamber for a listening room, so feel free to experiment with different values! The value does not need to be precise because the step response function is very gradual; it will be difficult to hear changes of less than 20% in value.

The baffle step function occurs at several hundred Hz, well above any subwoofer, so the presence or absence of a sub will make little difference in the optimum BSC component values.

Paul Joppa


Offline rchristie

  • Newbie
  • *
    • Posts: 5
Reply #5 on: November 08, 2009, 06:16:46 AM
Doc,
 
Congratulations on the new site, it looks great.  I thought I'd try out the new forum to ask a question that has been bugging me for a while about the "hooking it up" instructions on the SEXy speakers:

Quote
Hooking it up

Run the signal from your CD player/sound card/Ipod/whatever into your S.E.X. amp with interconnects.  Connect speaker cables from the speaker output binding posts on the S.E.X. amp to the left and right monitors.Then run more speaker cables from the binding posts on the monitors to the input binding posts on the corresponding subwoofer, with the woofer connections reversed (red post on monitor to black post on woofer, black post on monitor to red post on woofer). Connecting the system this way will avoid running the signal through a 125Hz filter that's built into the monitor output of the subwoofer. That filter would interfere with the designed roll off of the monitors and mess up their bass response.

What is the rationale for reversing the connections from the monitor to the woofer?  Is it that the reversal of the woofer connection results in the bypassing of the built-in filter?  Or are there two different things going on here, one of which is the bypassing of the internal filter by hooking up the monitor and woofer in parallel, and something else to do with reversing the polarity of the signal? 

At any rate, thanks again for the great music, and great fun, with the kits!

RChristie

Rich Christie

Rich Christie


Online Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19702
Reply #6 on: November 08, 2009, 08:01:33 AM
You could run the speaker cables out of the SEX amp into the subwoofer, then out of the subwoofer to the monitors.  This hookup style will impose a highpass filter on the monitors.  Running the signal to the monitors first will let them play the entire spectrum. 

Hooking the speaker leads up out of phase puts the woofer and monitor drivers out of phase with each other, allowing for better blending between the drivers. 

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #7 on: November 08, 2009, 09:29:39 AM
Just to augment CB's response, the effective acoustic crossover is approximately a second order Linkwitz-Riley, which calls for the LF and HF sections to be out of phase. If your sub amp has a phase switch, you can use that instead of reversing the leads. If the phase is wrong, there will be a big suck-out at the crossover frequency.

Paul Joppa


Offline rchristie

  • Newbie
  • *
    • Posts: 5
Reply #8 on: November 08, 2009, 11:10:58 AM
Thanks to both of you for the replies.  Paul, should that phase relationship with the subwoofer also hold for the Climax  (using the passive crossovers only, without the line-level tweak)?

Rich Christie


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #9 on: November 08, 2009, 12:43:39 PM
nt

Paul Joppa


Offline Thoburn

  • Full Member
  • ***
    • Posts: 188
Reply #10 on: November 11, 2009, 05:54:32 AM
The circuits are quite different indeed! What's in the cabinet is at speaker level, meaning a load impedance of 8 ohms (or 4 or 15 or whatever the speaker is). At line level, between preamp and power amp, most devices want to see at least 10000 ohms load impedance. You can get away with RC circuits (no inductors) here because there is no need to maintain a low resistance in the bass for good speaker damping.

The circuit as drawn is optimized for an 8.5-inch baffle width. A wider baffle will want a larger value for the 0.033uF capacitor, in proportion - so for example, use 0.047uF for a 12-inch baffle. Of course this is a theoretical value assuming perfect speakers and an anechoic chamber for a listening room, so feel free to experiment with different values! The value does not need to be precise because the step response function is very gradual; it will be difficult to hear changes of less than 20% in value.

The baffle step function occurs at several hundred Hz, well above any subwoofer, so the presence or absence of a sub will make little difference in the optimum BSC component values.


Hi Paul,

What would changing the resistor values do?
Thanks!

Dynavector DV-20X2L > VPI Scout II > Musical Surroundings NovaPhonomena
Mac Mini > USB DACiTx
Stereomour > Lowther Medallion DX4 and Rythmic Subs
Monster Power HTS3600


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #11 on: November 11, 2009, 09:22:11 AM
Changing resistor values would change the frequencies just as the capacitor values do - the frequency of any response corner is inversely proportional to the RC product. But changing the resistor values will also change the impedance of the circuit, smaller values placing a more severe load on whatever is driving the circuit, and larger values requiring a greater minimum impedance of the device that follows the circuit.

Right now, I would recommend that the source driving the BSC be capable of driving 10K ohms and have a source impedance no greater than 1K; and the load (amplifier input impedance) be no less than 100K ohms. These are not absolute requirements, just general recommendations.

Paul Joppa


Offline nullspace

  • Jr. Member
  • **
    • Posts: 12
Reply #12 on: February 05, 2010, 04:42:55 AM
Hi Paul,

With the S.E.X.y speakers, what impact does the 3-4ohm output impedence of the SET amp plus the impedence peak of the FE166s have on the input signal seen by the subwoofer amps? I modeled the FE166 in a .281cu ft. and it looks like the impedence peak is at ~120hz. So, wouldn't there be a bump up in voltage-out from the SET centered on ~120Hz, and consequently also in the input signal to the sub amp? And, if that's the case, does that have a small impact or a large impact?

Thanks for the help. I've been wondering about this for a while, and I haven't been able to wrap my head around it.

Regards,
John

John Bauman


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5823
Reply #13 on: February 05, 2010, 05:34:15 AM
Yes, you are correct. It will create a bump of about 3dB, right at the crossover point. The subs I am familiar with have a Linkwitz-Riley second order function which is 6dB down at the crossover point, so you are left with a net 3dB down, i.e. a Butterworth filter function.

The electrical Q of the FE-166 in the box is also greater with an SET driving it - it comes out to approximately a Linkwitz-Riley with high damping, and Butterworth with an SET. But that does not account for the large amount of stuffing, which will decrease the mechanical Q and the net Q. I haven't measured this myself so I can't quantify it, but the effect will probably be modest, around 0.1 change in the total Q.

In an anechoic chamber, this will give a 3dB bump at 125Hz if the sub driver is very close to the FE166 (a tenth of a wavelength is about one foot). For large separations, the average power into the room is flat for a Butterworth. But room effects are much larger than this, so I have not worried about it excessively. For perspective, note that open-baffle designs, which are quite popular these days, have a 3dB bump at their bass cutoff, usually around 200Hz.

At the time this was developed, I was expecting we would have a crossover product soon. One application for that would be to add a second order Butterworth function to both the sub and the fullrange, to produce a net fourth order Linkwitz-Riley crossover. That task is proving more complex than we anticipated, and I have not yet had a chance to audition the result with a S.E.X.y system so I can't even say I prefer it - though in theory it would be a big help.

Paul Joppa


Offline nullspace

  • Jr. Member
  • **
    • Posts: 12
Reply #14 on: February 05, 2010, 06:33:19 AM
Thanks very much for the thorough answer, Paul; I really appreciate it.

The more I look at other people's work, the more I realize how well thought-out and sneaky-smart a lot of it is.

Regards,
John

John Bauman