My first project is a 4 way speaker...

[TEK]

I have used glue-gun and think that works great. Have to rub the components a bit to ensure that it get holds - but if you do that I think it works nicely.
From your requirments Epoxy based glue springs to my mind. You can get that with different dry times and it is quite sticky.

[cochinada]

But has anyone of you glued a coil with 1.5 Kg hanging upside down without any other attaching? This is the test that separates men from the boys ;lx

[BobEllis]

Not inverted but vertical using a huge smear of hot melt glue. We tend to underestimate the holding power of various glues. When I disassembled that crossover it took a chisel to remove the inductor from the board and a lot of wood still came with it. If you don't want to go the hot melt route, epoxy should be more than strong enough. Aircraft are glued together these days.

[cochinada]

Epoxi is that dual thing that we have to mixture to activate it, right? For sure my coil is not as heavy as an aircraft.

[BobEllis]

Yes, epoxy is a two part system. Mix thoroughly, at least a minute. It won't cost you much open time and thorough mixing is important to bond strength. There are straight resin types and thickened epoxies. A thickened epoxy will allow you to build a pad for your inductor more easily than straight resin. If all you can find is the straight resin, you can mix in talcum powder to thicken. The WEST System fillers include colloidal silica, microspheres and micro fibers. Unfortunately the smallest WEST System packages are far more than you will need, but it gives you some idea what you can do with epoxy.

[cochinada]

Thank you. Looks like a crash course in chemistry is also needed to make a pair of speakers... :scratchhead:

[BobEllis]

Don't overthink it.

Just get some epoxy putty and glue your inductors down.

[csmielke]

Guys,
Have any of you tried cable ties? I have used them successfully with silicone underneath the components. Requires drilling some additional holes in the board. The large capacitors and coils still have little wiggle room if needed.
Chris

[cochinada]

Guys,
Have any of you tried cable ties? I have used them successfully with silicone underneath the components. Requires drilling some additional holes in the board. The large capacitors and coils still have little wiggle room if needed.
Chris

Yes I have but if I can avoid drilling holes better yet.

[Steve Manning]

Very nice indeed! :T
I was double measuring again and I'm clear to use a 10mm board if I decide to. By the way, as my components are also going to be upside down I'm facing the same problem. I intend to attach the big capacitors and all the coils with plastic braces but for the others I will have to use glue. What glue did you use?

I used this stuff, http://eclecticproducts.com/products/amazing-goop.html. They have something for about any application, it seems to have held up well.

[TEK]

Yes I have but if I can avoid drilling holes better yet.

For your use I think cable strippers would work great - maybe with some hot glue under...

[cochinada]

The epoxi was too expensive so I bought this one instead:



I've now glued all the components and although the resistors and the capacitors wobble a little they say that the glue will reach maximum efficacy only after 24 hours so let's wait and see before turning it upside down. For now the one I was most afraid seems to be quite well fixed.


Images not available

[Matt M]

I've silently followed this thread so far.

I'm surprised: I would apply that much glue if it were some kind of final crossover.
Your simulations look great, so that is a good start. But usually it takes quite a few iterations. So, I am wondering how you are going to approach this.

-Matt

[cochinada]

Hi Matt! You are right. This is the final XO, unless of course it sounds horrible to my ears but I'm quite confident it would be OK.

Besides, I'm so tired of having my living room upside down that when this is finished I don't want to hear anything about speakers in the next few years (at least mine).

[Matt M]

I can understand the part about ending the construction site status. But the messy part is over anyway, isn't it?

Some advice, if I may? Make sure that you can easily access and remove the crossovers. Best: keep them in external boxes. If you want to improve the sound later on, it's a good move to start out making a two-way crossover just using the mids & tweeter section. That way you will get a feeling how the various choices (topology, crossover-frequency & BSC, on-axis response vs. off-axis, effect of better phase tracking) affect the voicing. It can be really fun, and you will learn a lot. Give it time, enjoy the process, don't force it.

[cochinada]

Thanks for the advise Matt. Unfortunately the messy part is far from being over and it's at its peak right now as I have the four boxes lying on the middle of my living room with boxes and tools all over the place with the subwoofers and other stuff. I can't barely sit on the couch :-y

[Steve Manning]

Your friend did a nice job, those look great.

[TEK]

Those look very nice!
Tell me, is that 440 uF? Could you share the crossover schematics?

[kimmosto]

Could you share the crossover schematics?

At least my original X/O design is public. There are few minor differences to final. For example big cap for woofers was not connected and only two actual resistors needed (3.9 and 1.2 ohm 5W, in red).

Image not available

[cochinada]

Yes, this is the final xover except for some tweaking in a few components.

[TEK]

That's a couple of really nice speakers, congrats! arty:
Hope you are pleased with the result after what must have been a exhausting, but hopefully also fun and rewarding, journey :blink:

[cochinada]

Thanks a lot! It was really exhausting in all aspects and although I still need to tune the subwoofers and play with the DSP of my Crown amplifiers as at the moment I didn't even connect my PC to it I can take a well deserved rest. Finally I see some order on my living room at last 8)

[JonMarsh]

Very nice looking crossover board builds, and beautiful cabinets in your room! :T

[cochinada]

Thank you.

To be honest they look much nicer in pictures than on live due to many building imprecisions even with CNC. As some of you might remember I had the bad luck of the CNC being out of tune. This alone would explain many of the problems I had to cope with, namely the ellipse shaped cutouts for the drivers. The other problems like some visible big gaps I simply cannot justify as easily.

Anyway the most important is that they're finished at last and now I have to focus on the subwoofer tuning. Any suggestions how to do it?

[---k---]

Trailblazing is never easy or perfect. Glad to see them done. Many people have undertaken similar projects only to have abandoned them half way.

[Steve Manning]

Thank you.

To be honest they look much nicer in pictures than on live due to many building imprecisions even with CNC. As some of you might remember I had the bad luck of the CNC being out of tune. This alone would explain many of the problems I had to cope with, namely the ellipse shaped cutouts for the drivers. The other problems like some visible big gaps I simply cannot justify as easily.

Anyway the most important is that they're finished at last and now I have to focus on the subwoofer tuning. Any suggestions how to do it?

Very nice looking set of speakers ...... if they sound half as good as they look, they should be killer. :T I can relate with the cnc issues, one would think you ought to get perfect cuts from something controlled by a computer ..... not so sometimes. But you kept at it and they look great.

[5th element]

By subwoofer tuning are you meaning port tuning? I don't recall seeing any ports on this.

[cochinada]

By subwoofer tuning are you meaning port tuning? I don't recall seeing any ports on this.

No, no. There are no ports. I'm just figuring out how to measure the subwoofer in room in order to optimise the DSP. At this moment I don't even have a LP filter.

What I'm thinking is to place the mic at let's say 2m from each speaker, or even better at the listen sweat spot and do a gated impulse measure, first without the subwoofers turned on and then with only the subwoofers turned on. I know there are going to be lot of reflections so the curve will not be 100% true but this should give me a good starting point anyway. What do you think? Is there any other way?

[5th element]

Bass doesn't really work like that, due to the massive wavelengths, which is a good thing as it makes some things a lot easier.

First of all, if you want to apply corrective EQ, such as a Linkwitz Transform, or a low pass filter, then all you need are near field measurements. Get that mic 1cm from a woofer cone and start making some measurements. Due to the huge wavelengths involved there are literally no effects caused by the cabinet here that you need to worry about.

Considering applying some EQ to help integrate the subs with you room, this is best done at the listening position. Here you want to make ungated measurements. This seems contrary at first but all a gate does is remove the rooms influence from what you are measuring. Typically you apply a gate to remove reflections but due to the large bass wavelengths nothing is really reflected down there, you can't even fully form most bass waveforms within the typical living room because it's too small. If you want to measure the rooms modal region you do it without a gate. This usually looks terrible because we see the crazy high frequencies marred by reflections but this is because, most of the time, we don't ignore the higher frequencies.

If you do an ungated frequency sweep and then display only 20-100Hz things stop looking so bad now because you're not looking at reflections any more just the rooms modal region.

Each loudspeaker will have a different response. You will need to measure each one individually and apply different EQ to the left and right channels, at least preferably.

How are you powering the subs?

[cochinada]

I'm using a Crown XTi 2002 which has DSP built in.

So, if I understand correctly we can say this is a two step approach; first measure the subwoofer near field in order to apply a LP and/or LT and then at the sweet spot without any gating for room EQ purposes, each subwoofer at a time.

One question: how can I best define the xover point for the LP? Shouldn't I go back to the simulation program with this new curve for the subwoofer?

Thanks!

[cochinada]

Well, I was busy this weekend tuning the active subwoofers. I don't find the Crown very user friendly or intuitive and some settings are very confusing. For instance, with 'Polarity' is not clearly understood when it is inverted or not. The difference in the GUI with this and other options set is barely noticeable.

I placed the mic at the sweet spot, disconnected the right channel from my sound card and run a sweep with the Crown turned off and another with all hands on deck, so to speak. The words of 5th element came to mind "This usually looks terrible because we see the crazy high frequencies marred by reflections but this is because, most of the time, we don't ignore the higher frequencies." but I still could see the resemblance with the simulated curve in an 'average' way, that is, if one ignored some peaks and dips.

Then I defined a low shelve filter and since the Crown doesn't allow a frequency below 20Hz I had to stick with this and just entered a gain of +8dB. I confirmed also that the measurements are not consistent below a certain frequency (circa 50Hz) so I didn't pay much attention to the readings below 30-40Hz. Sometimes I got readings that increased below 20Hz and sometimes they were decreasing. Go figure...

Anyway, this LS sort of levelled the curve in a broader spectrum so I focused on the xover point. It was immediately clear that I needed to invert the subwoofers or otherwise I would get a huge dip in the crossover point. After some trial and error I settled with 140Hz which may come as a surprise but not as much when you look at the final simulation curves I got, with the 400uF capacitor in series. Then I applied some EQ with only 3 bands in the region between 40 to 60ish if I remember correctly to get rid of some peaks. Here I may have gone too far trying to level also one dip or two and perhaps I should review this as soon as I find the will again.

Finally, I followed the same procedure for the right speaker which is the one closer to the window and needing more EQ. This was a difficult case. I noticed for instance that I have a dip at 100Hz that absorbs all the energy I throw on it. I didn't see this on the left speaker so I assume it has to do with the positioning and perhaps the close vicinity of the window and the respective reflections. I gave up trying to compensate this dip and remembered that one should better stay away of the temptation to apply EQ to dips (the problematic ones at least) and instead focus on the peaks.

In my haste to set this properly I didn't save the impulse responses but if I repeat the measurements I will do it then and publish the curves.

After finishing setting up the main speakers I finally run the ARC with all the artillery turned on. To my dismay, after the measuring was done I noticed that one of my cylinder shaped subwoofers was not producing any sound and it took me about one hour to discover the culprit: one cable of the MiniDSP had one wire loose. So I had to repeat the process again but this time I got another problem... After the 5 runs, as Anthem requires a minimum of 5 positions, I got the message that there were more than 10dB difference in the subwoofer readings, being the reading on the left side closer to the center of the room the lowest one and the one closer to the right corner of the room the highest. I kind of cheated and repositioned the mic at two different but close positions until ARC was happy. Theoretically I should repeat the whole process again but it was already too late and I was tired. Besides, nothing would guarantee me that I would not have problems again placing the mic at those positions without any mark.

The 'good' thing is that this showed me the kind of acoustic problems I have in the room which I intend to minimise as much as possible in due course. For now my next goal is to rearrange the furniture but first I need to get rid of my big CD cabinet.

[JonMarsh]

To my dismay, after the measuring was done I noticed that one of my cylinder shaped subwoofers was not producing any sound and it took me about one hour to discover the culprit: one cable of the MiniDSP had one wire loose. So I had to repeat the process again but this time I got another problem...

Not to be a kill joy, but this is so typical of the kinds of problems I’ve had in the past with active sysetms, and friends have- cable and signal path integrity at the line level is an issue more often than you’d expect!

At least you’re making some solid process on the setup. Congratulations.

[cochinada]

Not to be a kill joy, but this is so typical of the kinds of problems I’ve had in the past with active sysetms, and friends have- cable and signal path integrity at the line level is an issue more often than you’d expect!

At least you’re making some solid process on the setup. Congratulations.

Thanks. Yes, true. The amount of variables is so high that the probability of something to go wrong is very high as well. :-y

[5th element]

My apologies for not getting back to this sooner I've been busy with sick family and then getting the same damned thing myself a week later or so.

Essentially though, yes. You use near field measurements for applying the LT and the low pass filter. Due to the wavelengths involved diffraction essentially goes out of the window with standard cabinet dimensions. You will still have to balance the levels of the subs with the other drivers and the proximity to boundaries will have an impact on the overall levels required, but beyond that it should be fine. Phase-wise, as you've experienced yourself, having the phase right (or wrong) is usually pretty obvious, once you've got the filters dialled in correctly.

When I mentioned that an ungated measurement would look like a mess if the high frequencies were left in.

For example below is the ungated measurement of one of my loudspeakers in my room in the listening position. It looks a mess with the usual peaks and dips all over the place.



If however we take a different approach to this and cut out all the frequencies above 200Hz, things don't start to look so bad any more.



Yes there are a few peaks and dips from 40-200Hz, but this is actually what we want to see. This is showing you the effects of room modes on the loudspeakers frequency response. For example you've got peaks at 40, 60 and 80Hz, but you've got also a large suckout/loss of energy around 100Hz. In this case you can easily EQ back some of the peaks, but there's no way you can try and fill in that suckout with EQ and even if you did you'd no doubt create a huge peak in the frequency response elsewhere in the room. This is where the multiple bass source thing comes in and you hope that some of your other bass sources will fill in the the dip at 100Hz.

[cochinada]

Hi, I hope you are all right now. This is that time of the Year at least on the Northern Hemisphere, that one should take care of himself for not catching anything.

Very interesting to see your curves. Mine resemble yours even with about the same variance in amplitude and I also have a dip at around 100Hz on the right channel. I was not able to correct it with EQ as I told.

I will definitely measure the responses again since I'm going to rearrange the furniture tomorrow and expect some changes in the acoustics as well. Maybe not dramatic but some changes still. This time I will take a more conservative/realistic approach and will not try to correct some (if not all) of the dips but just focus on the peaks below 100Hz.

I will post the results here with and without the subwoofers connected and with and without EQ as well.

[sdl2112]

Those look magnificent :T. You should be very proud.

[Steve Manning]

I agree with Scott ..... a very nice job on the speakers and room setup.

[cochinada]

Thank you guys.
I'm very curious to see what differences I'll have after the hole in the wall is covered for good. It should be done next week.

[5th element]

Those certainly are room dominating and I bet they sound as impressive as they look!

[cochinada]

Yes, everyone that comes to my house cannot refrain from making a remark.

I'm waiting for the construction work to end and then perhaps for some acoustic treatment before making the final conclusions but all I can say at this moment is that I'm very pleased with the way they sound and even 'disappear' despite their size.

Does anyone know if and where I can get some grills just for the Transducer lab tweeter (N26MGR-G)? This is so fragile that I'm a little worried to leave it unprotected. I'm looking for something that just sticks to it without any drilling or so.

[BobEllis]

Last time I looked, Meniscus had grills for the TLs. I wonder if my fat fingers can control them well enough to put the grill in place safely.

[cochinada]

Last time I looked, Meniscus had grills for the TLs. I wonder if my fat fingers can control them well enough to put the grill in place safely.

Found it!

http://meniscusaudio.com/tranducer-n26grille-p-1528.html

Thanks a lot! I think I'll order a pair of these. It will need a steady hand to place it in position I reckon as the magnetic force is considerable...

"...Extreme care must be taken when adhering to the tweeter as the magnetic force can pull the grille from your fingers and damage the dome! " :-y

BTW, is this one unit or a pair?

[BobEllis]

I assume that's each, but ask them.

[cochinada]

Good Googaliemoogalies! Those are rather impressive.

Never heard this expression before! What does it mean exactly?

[Srixon]

Never heard this expression before! What does it mean exactly?

Not sure, but it fits in this thread. Download "Nanook Rubs It" from Frank Zappa's 1974 album, Apostrophe, and play it through those beasts and you'll hear it. Great looking speakers.