Dipoles sorta like bob/arvo - long

[Davey]

C,

A preamp wouldn't really be necessary in this case. The receiver would do all the source selection, etc. You would attach the DCX to the tape-out of the receiver and then connect the six outputs back into the direct inputs on a receiver like the Marantz. The DCX would then receive full, line-level, signals that aren't volume-controlled and the receiver does everything else.
I know this seems really "non-hifi" to a lot of folks but it actually works pretty well.
The caveat that Dennis mentioned is valid. If any of the drivers is a very low ohm load (or need more power) then that would have to be evaluated and alternate (separate) power amplifier used.

That makes for one heck of a versatile setup with minimum $$$$ outlay. And it uses the DCX in a much more optimum condition than many users do.

Cheers,

Davey.

I think I missed it being a receiver. I would be using it with a proper pre-amp, you know? Otherwise, you would still put a simple stereo volume control somewhere in front. Seems like the power here is the low point in this system proposal?

Besides that, I'm surprised volume comes *after* the XO? I thought a pure power amp was the norm there.

C

[Dennis H]

With that many channels to work with, going active all around might make a fair amount of sense- it partly depends on how capable the DCX is - what kind of filter options are possible (I wonder if one can download a PDF manual - ought to try google tonight for that). If it can do tunable elliptical filters, cool. (I'll be surprised, though ) If it can't, I'd probably recommend a passive upper crossover (midwoofer to tweeter), and the rest active.

Jon, you can download the DCX PC software from the DCX page on the Behringer site. That will let you know what filters you have to work with. Then you could see if you could build a Cauer with the digital filters in LspCAD. I guessing you could do pretty well with a Butterworth 4th order somewhere close to the target Fc plus several notch filters (I think each one will only go to -15dB.) Load up your filters and let LspCAD optimize to LR8.

[Paul H]

Seems to me, in for a penny, in for a pound. Once you accept an active XO anywhere in the signal path, there's no real penalty (other than the extra amps) for going active all the way.

Dennis,

That comment is assuming dsp crossovers where an additional a/d and d/a conversion is performed?

If using analog active filters is the above statement accurate?

Paul

[sfdoddsy]

If you're talking receivers, I confess to being a fan of the Panasonic digital ones. It'd be nice of the DCX had a digital out so you could have an all digital chain, but maybe the replacement will.

Cheers

Steve

[mante]

Hi Steve,

I also have the dcx/Panasonic set up and like it for my two way speakers. The issue I see with the Panasonic is the ability to use only 5 channels freely. Though I believe if the xover used in the 6th channel is below 120hz, (is that right) then the Panasonic's six amps can be used.

Does a low ohm load become more of an "problem" when the volume is increased? I got to get a good beginners book.....

Craig

p.s. I found Hank's new speakers. They are hanging in our new $200 million Overture Center and are at least twenty feet tall. :yeah:

[Paul H]

Why are passive crossovers better?

I've been reading much material on the choice between active and passive crossovers, and finding mostly support for using active crossovers where possible.

The reasons given for using passives seem to be that its simpler (which may or may not be true but I think more people are used to designing these) and cheaper (because of less amp channels required).

My question is, assuming I'm willing and ready to spend the time and money to get enough channels of amplification to feed my 4-way fronts, and develop a custom active crossover, are there any benefits to going passive?

Thanks,

Paul

[JonMarsh]

Paul, there's a lot of different feelings about that, and for many folks the over riding consideration might be the flexiblity, ease of setup and modification of transfer function in the case of hte digital xovers, etc.

On the passive side supporters, there are those who feel that the available active crossovers don't provide the last word in transparency and neutrality (even the DEQX), and for one reason or another may prefer not to have multiple channels and asociated cables, amplifiers, etc.

One can make cogent arguements for the advantages and disadvantages for either type. Maybe it's as simple as what appeals to your sense of engineering esthetics?

(BTW, could I implement a cauer-elliptic in the available digitals? I haven't heard anything to indicate I can). There are other problems I can think of with the high slope digitals which are not obvious in casual examination, but we don't know enough about the way the human ear responds to speak with authority on what works or doesn't.

In an ideal world, try both. In a more realistic world, listen to one or more of each which others use, and if you like what you hear, go for it.

My preference is to encourage anyone to try out and develop what appeals to them in concept- it's probably the quality and care of the implementation which counts at least as much as the principle behind it.

~Jon

[RonS]

(BTW, could I implement a cauer-elliptic in the available digitals? I haven't heard anything to indicate I can).

This is something that I've been wondering as well, and also wondering if there was any point to it. Currently(I'm using a Behringer DCX2496) I'm using 8th order L-R between the XT25 and M8A. Would there be any point in going to a 4th order L-R and adding a notch to emulate the caur filter? If so, how deep would the notch have to be, and what is the Q? The DCX has maximum -15dB for the notch filter, but I think I could overlay more than one on top of each other. The question for me is, what would be the benefit of doing so, if any?

Cheers,
Ron

[JoshK]

I plan to build Arvo-like dipoles and I am soon to receive my DEQX. I am wondering if one can emulate C-E filters or whether some other steep slopes would work equally as well. Side note, is there a thread here where you note the advantages/disadvantages and corresponding properties of the C-E filter? Besides the steep slope in the xo range what advantages does it have over say an 8th LR?

I know you can do an 8th LR in the DEQX, and you can do many others as well. If you can superimpose one transform onto another you could easily replicate nearly any transform you want. I will ask the DEQX forum whether this can be done.

[cjd]

8th LR has been reported by many to sound pretty poor - and there's the group delay issues (or perhaps that's WHY they sound worse).

CE that emulates an 8th order has group delay much like a 4th LR.

I have a suspicion that 4th order + notch (or otherwise, a single point steep filter EQ combined with 4th order) might emulate a C-E filter, but the question is not really entirely whether you can get the graph to look the same, but whether the actual transfer function is the same. Though maybe that means the same thing as having the graph look the same. Beyond me at this point still.

Why not experiment?

C

[RonS]

8th LR has been reported by many to sound pretty poor - and there's the group delay issues (or perhaps that's WHY they sound worse).

CE that emulates an 8th order has group delay much like a 4th LR.

I have a suspicion that 4th order + notch (or otherwise, a single point steep filter EQ combined with 4th order) might emulate a C-E filter, but the question is not really entirely whether you can get the graph to look the same, but whether the actual transfer function is the same. Though maybe that means the same thing as having the graph look the same. Beyond me at this point still.

Why not experiment?

C

I'd like to experiment with it. That's why I asked how deep the notch has to be, and what is the ideal Q. I guess we'll have to wait until Jon weighs in on this.

Cheers,
Ron

[DIY_newbie]

Guys,

It seems to me like you might be looking at this the wrong way.. Jon sets up his C-E filter to mimic the slope of a LR-48 filter. If your doing a digital filter why not just set it up for LR-48 slopes? With most digital/active filters there is no phase difference or group delay to worry about. All the outputs of the active filter should be in phase and with 0 delay with respect to each other..

I'm using LR-48 db slopes on my mid/tweeter crossover and using delay correction to time align the driver's accoustic center and it sounds better than any passive crossover I've ever heard.. The only problem with using the DCX like I am is that it has a relativly loud noise floor.. The noise not audible at 3M on my current bass drivers (89db sensativity) however it is audible at 3M with my mid/tweeters (100db,98db sensativity)

YMMV,

--Chris

[RonS]

Chris, two things come to mind. First, you are correct that Jon is setting up his c-e filter to mimick an L-R48dB filter with fewer parts. However, I seem to recall that Jon said that the c-e filter would have less ripple than an 8th order L-R filter. If this is the case, then that would be a reason to use the c-e filter. Second thing is that the DCX filters mimick analog filters, meaning that they will have phase delay just as their analog equivalent filters would. Meaning that you would need to introduce a delay not only for the physical offset of the drivers acoustic centres, but also for the phase delay introduced by the filter itself. Or so my understanding of the topic suggests.

Ron

[JonMarsh]

Guys,

It seems to me like you might be looking at this the wrong way.. Jon sets up his C-E filter to mimic the slope of a LR-48 filter. If your doing a digital filter why not just set it up for LR-48 slopes? With most digital/active filters there is no phase difference or group delay to worry about. All the outputs of the active filter should be in phase and with 0 delay with respect to each other..

I'm using LR-48 db slopes on my mid/tweeter crossover and using delay correction to time align the driver's accoustic center and it sounds better than any passive crossover I've ever heard.. The only problem with using the DCX like I am is that it has a relativly loud noise floor.. The noise not audible at 3M on my current bass drivers (89db sensativity) however it is audible at 3M with my mid/tweeters (100db,98db sensativity)

YMMV,

--Chris

Rons' on the right track there. The advantage to the CE implmentation of what is effectively an L-R 8 (48 dB/octave) is that the group delay and transient response is very similar to the LR4, not the LR8, with different phase characteristic, and IMO this sounds better... This is why I'm not keen on implementing a conventional digital filter approach, unless I can program it completely; i.e., emulate the cauer-elliptic which I use passively.

Makes me kind of passive-agressive, doesn't it? :B


~Jon

[Dennis H]

Yep this phase stuff is confusing. As Ron and Jon said, most digital boxes, such as the Behringer, use minimum phase IIR filters that behave just like analog filters so most likely emulating a Cauer would sound better than an LR8. A few expensive ones like TacT and DEQX use linear-phase FIR filters, with no extra phase shift or group delay due to the rolloff, so with them you'd probably just go straight for LR8 because the problems the Cauer "fixes" don't exist.

[DIY_newbie]

Yep this phase stuff is confusing. As Ron and Jon said, most digital boxes, such as the Behringer, use minimum phase IIR filters that behave just like analog filters so most likely emulating a Cauer would sound better than an LR8. A few expensive ones like TacT and DEQX use linear-phase FIR filters, with no extra phase shift or group delay due to the rolloff, so with them you'd probably just go straight for LR8 because the problems the Cauer "fixes" don't exist.

My understanding was that the DCX outputs where all in phase.. I've posted a question via online-support on the behringer website.. I'll let you guys know what I find out..

IMHO, the results I got where much better with LR48 than LR24. I'll try to get a friend to help me with a double-blind listening test while I wait for a response from Behringer...

--Chris

[john k...]

(JPK) If I may, I think it would be beneficial to those who are interested in CE type filters if you obtained a copy of Thiele's paper, Loudspeaker Crossovers with Notched Responses. Thiele treats the subject in depth giving active and passive topologies and design tables for 6th order active and 4th order passive filters. The crossovers sum flat. Odd order filters reduce to Butterworth filters when the notch frequencies are moved far away form the crossover point. Ever order filters reduce the LR type.

I've attached a figure of the passive 4th order type.


While not the "best" digital crossover in the world I believe that both Sound Easy and LspCAD digital filters are capable of emulating these.

[RonS]

Rons' on the right track there. The advantage to the CE implmentation of what is effectively an L-R 8 (48 dB/octave) is that the group delay and transient response is very similar to the LR4, not the LR8, with different phase characteristic, and IMO this sounds better... This is why I'm not keen on implementing a conventional digital filter approach, unless I can program it completely; i.e., emulate the cauer-elliptic which I use passively.

Makes me kind of passive-agressive, doesn't it? :B


~Jon

Jon, you can certainly program a 4th order L-R filter and a notch filter (as I said, up to -15dB any Q you want) to emulate the transfer function of the C-E filter. My question is, how deep must the notch be and what Q? Or is this dependant on the actual drivers rolloff, whatever is necessary to "force" the driver into an 8th order rolloff using a 4th order L-R and notch?

DIY_newbie, the DCX output is in phase as much as it doesn't rotate phase 180 degrees on it's output unless you tell the unit to do so. However, as it emulates it's analog equivalent, there is inherent phase shift. Whatever phase shift would be present in a passive filter will be there in it's digital equivalent. As mentioned, some very expensive units don't do this (on the basis of using linear-phase FIR filters).

JohnK, I'll leave it to others to delve into Theile's paper, I suspect the math will be over my head. Perhaps you'd be willing to give us a quick summary of the paper?

Cheers,
Ron

[Paul H]

Paul, there's a lot of different feelings about that, and for many folks the over riding consideration might be the flexiblity, ease of setup and modification of transfer function in the case of hte digital xovers, etc.

On the passive side supporters, there are those who feel that the available active crossovers don't provide the last word in transparency and neutrality (even the DEQX) ....

~Jon

Jon, your response has created some lively interest and debate on digital crossovers. However, my 'custom crossover' that I was referring to creating would be an analog one.

Paul

[RonS]

Wow, when did AES preprints jump to $20 for non-members?! Last time I ordered it was $5, and half the time you actually got a hard copy of the article. It's a real shame that the AES has increased their prices so dramatically. They were a real good source of information, but at $20 a crack, I don't think I'll be ordering many papers. Too bad, at $5 a pop, it kind of made it easy to take a risk on a particular topic.

Oh well, maybe the reference library has the articles. Not that I have the time to go down there and look for them.

Ron

[Dennis H]

John, Thiele's implementation is slightly different than Jon's although they probably get the same thing done. The highpass is the same. For the lowpass, Thiele uses an extra L is series with C2. Jon uses an extra C in series with L5. Jon's travelling on business so he probably won't be able to comment for a while.

DIY_newbie, be careful about semantics when you read the response from tech support. Some people will say that the woofer and tweeter in an analog LR4 or LR8 crossover are "in phase." They are sort of right except the phase has rotated 360 degrees (LR4) or 720 degrees (LR8.) Kinda like if you spin around once or twice you're still facing the way you started but your motion isn't the same as the other guy who never moved.

[john k...]

Jon uses an extra C in series with L5.

(JPK) In series with L5 or in parallel? Thiele gives a total of 7 different topologies for the 4th order crossover, 3 different LP configurations and 4 different HP. One of the LP configuration places a cap parallel to L5 instead of in the shunt element. Similarly, for the HP the cap can be removed from the shunt element and an inductor placed in parallel with C13 in my figure.

To answer Ron's questions, Thiele basically develops the transfer functions for notched responses that sum flat from 3rd to 8th order. He introduces a parameter, K, which is equal to Fnhp/Fx for the HP and Fx/FnLP for the LP section where fx is the crossover frequency and Fn is the notch frequency in the hp or lp response. Thus he can shift the notch frequency at will. As k gets small (notches far from Fx) the even order responses reduce the LR filters and the odd numbered to Butterworth. With his formulation the filters are symmetric and have the same GD for the HP and LP sections. Like what Jon described, the GD for the 4th order notched response is very similar to the LR4. As I indicated in my initial post, Thiele provides design tables for the 4th order passive crossovers at 4 different values of k (0, 0.4472, 0.5, 0.5774) (k^2 = 0, 1/5, 1/4, 1/3). He also provided design tables for active implementation of a 6th order crossover at 3 different k values. In that case two stages of ordinary 2nd order Sallen & Key filters are cascaded with a 3rd 2nd order Sallen and Key type filter with a bridged T network to provide the notch.

Is the paper worth $20? I guess it depends on your interest. I got it when it first came out in 2000 when it was still $5. I don't think it would be found in a library since is a preprint, at least my copy is. Perhaps it found its way into the journal.

[Dennis H]

Oops, you are right, John. The cap is parallel to the coil. Here's the schematic and electrical transfer function for the Modula project. Jon should speak to this himself but I believe the design goals include LR8 initial acoustical rolloff and stopband ripple at least 40dB down because these filters have equal voltage ripple in the passband and stopband. Did Thiele address the ripple issue?

[john k...]

(JPK) Yes Thiele addresses the peaks. The k values he specified, 1/3, 1/4, 1/5 are for stop band peaks after the notch of -30, -35 and -40 dB in the 4th order alignment. Those seem to be the peak levels he targets throughout the paper, regardless of order. The phase after the notch is 180 degrees compared to the complimentary filter so, for example, if the peak above the notch in the LP is at -40dB, the HP section will have a peak of 1.01 or 0.086dB at the LP notch frequency. I also note that Jon's HP section LC shunt in the second leg where as Thiele's has the LC shunt in the first leg. Some of Thiele's other topologies are interesting since they involve coupled pairs of inductors.

FYI, the peaks are at -40dB for the plots I posted.

[Digisan]

John and others, Thiele's preprint is available online at aes.org .

[Paul H]

Progress

I've built and stained two cabinets for the dipoles, like the one shown here. Measurements in the actual cabinets are redone, and design layout for the 4-way active (analog) crossover is nearing completion.






I'm still very much wrestling with logistics and practical details of building a large crossover system - I'm trying to keep it relatively simple without much in the way of notch filters, etc, etc at least until I get it running - but it's still large and appears complex - time will tell.

I was thinking of using perforated boards like these to at least initially piece these circuits together on:

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=320-470&scqty=10

Will these work - anyone have suggestions for something better?


Paul

[Saurav]

You're using opamps, right? You can get boards that are better suited for opamps. Specifically, you want some rows of holes connected, so you can use the board to actually make connections. With this one, each solder point is isolated, so you'll essentially have to do a point-to-point job. It can be argued that PTP is better for sonics, but for quick prototyping, it'll be a pain.

For example:



Each vertical column of holes is connected on the solder side. With this one, you can place the IC pins on 2 sides of the gap in the middle (it's sized to fit the pin spacing on a regular 8-pin IC). That gets you 4 extra holes per IC pin to connect stuff to, while of course keeping the pins isolated from each other. That board looks like it's from Radio Shack - it's pretty cheap, the traces lift if you overheat them even the slightest bit. I'd recommend getting a better quality one, it'll save you time and frustration in the long run.

For even faster prototyping, you could get one of these (we called them breadboards in India, not sure what they're called here):



No soldering, just insert the pins into the holes. Works the same way, horizontal rows are connected internally. The two columns on each end are also connected (within each column, not with each other), which makes them convenient for V+/V-/ground busses running down the length of the board. Even worse sound quality, but best for when you're making lots of changes.

[Dennis H]

SL's circuit boards aren't all that expensive and could easily be combined to do everything you need.

http://www.linkwitzlab.com/pcb.htm

[JonMarsh]

Do you have PCB layout software your self?

I use Advanced Circuits in Colorado for my work and personal stuff. They have a special "division" offer, called Bare Bones Proto PCBs, which is great for prototyping, due to low cost. Of course, there are some rules/limitations:

Only 2 layer FR-4, 0.062"

All holes plated

No mask, no legend

6 mil line/gap minimum (this used to be state of the art, but now it's "relaxed" spec)

no internal routing or scoring- cut to rectangle shape only

Gerber 274X or 274D

Excellon drill files only

Max size 12" X 16" or 192 sq. inc.; min size 1.25" X 1.25".

Min hole size 0.015".


Bare Bones PCB at Advanced Circuits

[Paul H]

Do you have PCB layout software your self?

No, I'm afraid not.


This will likely end up being wired point-point, after using a large breadboard to check, test and evaluate the crossover.

I've looked at many online offerings for PCB's and kits, but everything I look at is too small, missing something or will cost 3x as much as a total diy. I figure if I have to seriously modify a kit I might as well build from scratch.

To say this is a learning experience for me would be a drastic understatement - I'm in way over my head 8O That being said, thanks to SL and a bunch of others on the net I've gathered enough knowledge to put together a design in LspCAD which I'm optimistic about. Now I just need to get the logistics of physically building the crossover sorted out.

Anyone know approximately how large a breadboard I should buy to set up a 4-way crossover of medium complexity?

Paul

[JonMarsh]

I've used protoboards like he describes above, where you plug in the comonents and plug in wires point to point. AP made some good ones, I've still got one or two around.

BTW, the cabinets/panels are beautiful- you've done a very nice job, Paul! :T

Next question for you is, how much of a hurry are you in? If you have an "op amp" circuit in LSPCAD, can you post it or send it to me by email? Building it on a PCB is a LOT easier and more reliable than point to point. I do a lot of PCB work from time to time, maybe I can squeeze this in while on planes, trains, or other "spare" time. Let me take a look at your schematic just so I have an idea what it would involve. There are demo or student versions of some PCB software available, but OTOH, it probably doesn't make a lot of sense to lern a package like that just for one project. I use Protel DXP2004, or as it's now called, Altium Desinger. Not super high end, but not cheap- ~$8,000. You don't want to know what the "high end" stuff costs....

There's even a place that give you software to design PCB's which they fab.

http://www.pcb123.com/pcb123benefits.php?WT.mc_id=psgoo00020

I've never tried this, but just browsed through their basic tutorial, and it looks surprisingly decent. Now, I doubt it has some of the features I'm used to, like complex polygon pours, programmable design rules, signal integrity analysis, a SPICE3 simulator, etc, but it it should be pretty easy to learn. There's online tutorials about some of the major editing operations, showing how the software works in screen playback. THere's a very good third party PDF tutorial on their site also, which explains a lot of the background information about designing and fabricating PCB's. Well worth reading, even if you try to build something by hand (you want to plan it out- a PCB editor is a good way to do that).

Take a look at this PDF.

http://www.pcb123.com/tutorials/PDF%20Documents/Web_Schematic_Layout%20Tutorial.pdf

See what you think.

Express PCB has something similar, but the PCB editor doesn't look nearly as powerful.


~Jon

[Dennis H]

BTW, the cabinets/panels are beautiful- you've done a very nice job, Dennis!

Credit where credit is due, those are Paul's. And yeah they look awesome!

[JonMarsh]

Boy, I can't keep anything straight lately! Gotta get my head from out of work a little bit!

But, Paul DID do an awesome job!

[Paul H]

Thanks Jon and Dennis,

Jon, ya got mail

I'm not in a hurry at all - anxious to build something and make some progress but not in a panic to finish.

I'm interested in a PCB eventually but I think I may need to build a crossover and experiment first, before committing to a PCB.

Thanks,

Paul

[Saurav]

Anyone know approximately how large a breadboard I should buy to set up a 4-way crossover of medium complexity?

Figure out how many opamps you think you'll use. It's 4 rows per opamp, and you'll want to leave 8-10 rows between opamps for all the caps and resistors, otherwise it could get a little crowded. I built a straight (i.e. no EQ or filters) stereo 3-way LR-4 on a 9"x6" perf board that had 3 sets of IC rows. It's a little busy on the board, but not unmanagable. Plan for more space if you're using larger/fancier caps, I just used plain old poly caps.

[AJINFLA]

Sweet

The dark color wouldn't fit my decor, size looks rather imposing :E . That's one heck of a nice looking speaker Paul :T! Measurements? I'm very curious as to what the off-axis plots look like with that driver arrangement/wide baffle.

Cheers,

AJ

[JoshK]

One word for you.... BADDASS! Love the dark wood! Even my wife thought they look'd nice.

[Paul H]

Thanks for the help and kind words.

These are fairly large (64" high overall x 28" wide baffle) but are intended for a 17' x 24' room in the basement which is pretty well dedicated to HT/audio.

... Measurements? I'm very curious as to what the off-axis plots look like with that driver arrangement/wide baffle.

Cheers,

AJ

The baffle size and layout was done with a disregard for esthetics and common (but possibly flawed) thinking that baffles must be narrow for visual appeal and/or stereo imaging.

My theories (right or wrong) were that the baffle needs to be wide enough for the 2-15" drivers, and large enough for decent bass response from the 2-10" without excessive filtering/electronic manipulation. Once that size was established I used a couple of baffle simulation programs (the Edge and one other that I forget right now) to establish the driver locations with the best (smoothest) frequency response.

The plot of the tweeter only measured in the finished baffle at various angles is shown below with 1/6th octave smoothing.






Paul

[Paul H]

Help with grounding

I finally got some time to start putting my analog active crossover together.

Using a breadboard, I put together a high and low pass for the woofers, fired it up, and hmmmmmmmmm - humming and more humming. I can play music through it, and it seems to be cutting off at the right frquencies, but I cannot get rid of the hum. It's very noticeable - 40-50 dB maybe, but doesn't get louder with increased preamp volume.

After playing with the circuit a hundred times to check connections, etc, etc it certainly appears to be a grounding problem - but I'm stuck there.

I've grounded the ac power cord ground, power supply ground & case, and signal line grounds together.

Either I've grounded the circuit improperly, or possibly I can't get good enough connections on the breadboard.

Suggestions to attempt to fix the hum are appreciated. Suggestions for good books/links regarding details of constructing active crossovers would also be helpful.

Thanks,

Paul

[cjd]

First thing to do is tryin separating the signal ground from power ground/earth.

You can also try lifting the earth at the electric socket. If ANY other component in your system is earthed, and that is connected to the signal ground as is usually the case, you'll have yourself a nifty ground loop.

When I run my amp off my computer for measurements, I have to lift the earth on my amp for this reason (since the computer must have it!)

I am also thrilled to say my wife liked the look of those speakers, and made a comment to that effect reading over my shoulder! Which means, I may be closer to my dipoles than I had previously considered.

C

[Dennis H]

I am also thrilled to say my wife liked the look of those speakers, and made a comment to that effect reading over my shoulder! Which means, I may be closer to my dipoles than I had previously considered.

Hehe, I'll bet you a beer she doesn't realize how huuuuuuuuge those things are.

[cjd]

Hehe, I'll bet you a beer she doesn't realize how huuuuuuuuge those things are.

Ugh. beer. :P Probably true, but there ARE visual cues in that photo as to scale. And my plans aren't *quite* that big. I don't think. :P

C

[Paul H]

I rewired the grounding to more of a parallel feed back to the single power ground - hum dropped noticeably, enough that I'll carry on with rest of breadboarding and finish the prototype. I should be able to 'dump' the rest of the hum when I permanently wire.

... there ARE visual cues in that photo as to scale. And my plans aren't *quite* that big. I don't think. :P
C

They're 2'4" wide x 5'4" tall. Dipoles with large bass extension before major filtering is applied = large cabinets. Your plans might end up that big - trick is to make them a really nice colour

My previous project was 180 lbs each mains and the current ones seem headed for about the same. I have promised myself that my next speaker project will be bookshelfs or small surrounds.

Paul

[Paul H]

Making progress .. I've got a working crossover for one channel - so I can play music from one speaker. The crossover is, as you can see, slapped together in a quickly made wooden crate with a plexiglass cover over the power supply (120 volt power feeds - safety first) and alligator clip feeds - a thing of beauty it is not.

Anyway, it works, and splits the frequencies 4 ways to pretty well where they were modelled by LSPCad :T . I've got much more testing and tweaking to do, (killing resonance frequencies in the mid, tweaking tweeter crossover frequency, extending low end bass are first on the list) but it's fantastic to finally have a working prototype - it's been 9-10 months since I started thinking about these.

Initial impressions on the sound - superb, clean and powerful. Dipoles are occasionally accused of not having enough 'oomph' or 'slam' - I think the 15's and 10's to cover the lower end have taken care of that. A well-recorded guitar riff doesn't sound like a well-recorded guitar riff - it sounds like there's a stratocaster and guitar amp actually in the room.
arty: I think I'm going to like these ...

[DearS]

so the current drivers are ss9700, w12,2-rs180, and 2-10"

atleast thats what it looks like.