Trillium build thread

[Paul Ebert]

This thread will be a rather sporadic build thread for my M8ta inspired design. I've spent several years now learning, which has been a lot of fun, but it's way past time to stop being an 'armchair' designer and actually get something built.

Common sense would indicate that I start out with something simple (perhaps building a tried and true MT or the like), but that's not how I work. Hey, what's the worst that could happen? 8)

So, this design will feature the following drivers:

• Seas 27TBCDGB-DXT tweeter

• Dayton RS52 midrange

• Dayton RS225-8 woofer (times two)

all put in a M8ta type enclosure, with (hopefully) a Duelund crossover. It will be ported.

So far, I've got the baffle pieces cut, the backs and the top and bottom pieces. I had the sides done, but I'm going to make it deeper, so I need to do those over. I've routed the driver holes in the front baffle pieces and installed the threaded inserts that will hold the drivers (see attached pictures).

Hopefully, as I progress, I can pass on things I learn that could be of help to others doing their first build. So, here's the first such tidbit: when I drove the first threaded insert, it caused the MDF around it to be pushed up - enough that the driver was no longer going to be flush. If you look closely at the attached pictures, you will see that I added a bit of a countersink using a larger drill bit. This solved this problem very nicely (though it was a good bit of additional work).

Anyway, so far so good. I did install one of the inserts poorly positioned by about 3/32 - enough that the bolt wasn't even close to going in. So, I removed that insert, drilled out the hole and used a plug cutter to make a plug. The plug went in nicely and it's drying now. I expect that this will fix this first mistake (nice to have that out of the way).

Well, I was hoping to attach some pictures, but they are way too big (I'm not good at this digital photography business). So, here's a link to my flickr account where you may find them:

https://www.flickr.com/photos/80104762@N00/sets/72157607167568936/

Please feel free to tell me what you think, ask questions, tell me I'm crazy, whatever.

[Paul Ebert]

Oh, I forgot to explain the name. To me, 'Trillium' sounds like some sort of unstable element or some other cool thing befitting such a large and ambitious three-way design, but actually, it's a small three petaled flower. No matter, I like the sound of it.

[Paul Ebert]

Oh, another thing. Actually, a gripe. I'm not keen on the position of the frame holes on the RS52. They are quite close to the edge of the cut-out. It was nerve wracking driving the inserts in and one did bulge out the side a bit. Hopefully, in the large scheme of this as a completed project, it will be a inconsequential nit.

[Paul Ebert]

What diameter port should I use?

I'm getting ready to order the foam for the walls, some acousta-stuff (as per John Krutke's recommendation for a tall cabinet) and the ports.

At maximum power for 2 RS-225s (160 watts), Unibox recommends a 4" diameter port. I'm tuning to 25 hz which comes out to 18" in length, or so.

I plan to use the 'Precision Port' from PE, so it will be flared on both ends. Will this work OK, or should I go down to a 3" or, perhaps, a pair of 2"?

Here are the Unibox graphs (by the way, how do you paste images in the text?):

[ThomasW]

(by the way, how do you paste images in the text?)

Open each attachment then copy and paste the URL into the body of your message.

[JonP]

On the port issue, generally the larger diameter the better, for less chuffing/noise due to the lower air velocity, and also less compression effects.

Sounds like a 4" Precision Port (max of 17"?) would be just about perfect...

There is a certain type of thin walled white sewer pipe, thinner than sched 40 by quite a bit, that fits fairly well with the PP flanges, if you need longer. Or, get in touch with the PP guys, I believe they might sell you another center tube w/o flanges, that you could splice with.

[Paul Ebert]

Thanks, Jon! I don't think the max of 17" will be a problem because the port feeds into the slot formed by the base, which, now that I think about it, will be quite long (10" or more?). I may find I don't need (much of) a port.

I have heard that it is possible to have too large of a port. I believe it has to do with the fact that the larger it is, the longer it is - resulting in pipe resonances. Is this correct?

Perhaps I should make the bottom removable so I can experiment. But, I think I'll start with the 4 inch.

[Sakura]

I bet, with a 3" port, listening to music, you won't hear chuffing.

[JonP]

Thanks, Jon! I don't think the max of 17" will be a problem because the port feeds into the slot formed by the base, which, now that I think about it, will be quite long (10" or more?). I may find I don't need (much of) a port.

I have heard that it is possible to have too large of a port. I believe it has to do with the fact that the larger it is, the longer it is - resulting in pipe resonances. Is this correct?

Perhaps I should make the bottom removable so I can experiment. But, I think I'll start with the 4 inch.

I guess I need to look at your design, not getting the bottom slot concept. If you don't have the port fairly unobstructed, there will be added effects.
If the port's coming out the bottom, if you have enough height from the ground you'll be OK... (or only tuned different a little bit)

You can get Unibox to show the port resonance effects... it's a checkbox on the right, up in the top design area. When you get a long pipe going, you get a lower freq effect. I think the effect is smaller with a larger pipe, not sure if those two are tightly tied. Mainly look at your resonance vs your vent velocity when going between the port sizes... you'll be trading off shorter lengths for smaller diameters, and higher resonances, against wider pipes with lower velocity, and longer lengths... and lower resonant points.

And yes.. I'd guess that a 3" port with 1 RS225 is probably going to be good.

[Paul Ebert]

The design for the base (with the slot) is the same as what's found at the lower left of the attached drawing (Tony Gee's Soup).

The port fires down into the slot which is formed by a U shaped base. The slot opens in the back of the base.

Last night I took a look at my dog-eared copy of Loudspeaker Design Cookbook and Vance recommends a 4" port for 8" and 10" drivers, calling it marginal for a 12". The effective diameter of my 2 RS225s comes out to 11.3", so even a 4' may not be enough.

I have to admit that I'm a bit leary of this aspect of the design. I would think it would be a problem if the speaker was on carpet (I currently have wood floors). I may change it so the port fires out the back, but this would probably require an elbow.

Jon, perhaps that thin walled pipe you mentioned has an elbow. I'll have to look for it.

[Paul Ebert]

Some progress ... finally!

I finally got some time over Thanksgiving and this last weekend to work on my cabinets. Both have all their pieces cut and one has been assembled, including the resonance reducing slats and the sound absorbing foam.

I decided to cut the facets using a handsaw. It's an elcheapo saw with horrendous ergonomics (I'd never use my antique Disstons or my Japanese saws on MDF like this!), but it is working adequately. I'm using pieces of MDF with miters on the edges as saw guides as you can see in the picture below.

One of the cuts was a bit off and when I tried to fix that I ended up rounding the edge a bit. This is barely discernible in the circle on the second picture (sorry about the lousy pictures). There are also a few saw gouges that are too deep to sand out. I guess I'll have to fix those mistakes with some bondo.

It won't be long before I can start taking measurements. I'm looking forward to that!

[Paul Ebert]

A question: should I measure with the port installed or does it not matter?

[Paul W]

One suggestion...check the actual port tuning (with any fill) to make sure it is where you want. A slot that shallow may make the port "look" longer than it is. Not necessarily bad, but different than intended.

[Paul Ebert]

True and a good suggestion. I'm thinking of making two bases to the enclosure - a ported one and a sealed one. The purpose would be to see which I like better. While I want something reasonably full-range, I'm not a bass freak (he sayes while listening to his ProAc Tablettes).

But, for the purpose of designing the crossover, will it matter if I take my raw, mounted in enclosure, driver measurements with the port or sealed?

[Paul W]

Crossing high enough for an RS52, ported or not shouldn't make any meaningful difference. On the other hand, since setup is more effort than the actual measurements, maybe measure one ported and one plugged. That would also give you food for thought about your later ported/sealed decision.

[Paul Ebert]

Thanks, Paul. I've made the bottom of the enclosure removable. I'm thinking of making two - ported and sealed. Or, I could just make a 'cover' for the port.

I've mounted the drivers in the one that's further along and moved it up to the living room to get some sense of how they are going to fit in. Personally, I think they dominate too much, but the wife likes them (a lot!).

You can see a photo here:

Trillium in living room

Time to let my back recover from hoisting this thing around and build my Arta measurement jig.

[Quwiksilver]

Paul,

Those speakers look absolutely dominant! I look forward to following this one.

Scott

[AJINFLA]

Looking good Paul :T

[Paul W]

You're well on your way...enjoy the ride!

[Paul Ebert]

Thanks, guys! I am having fun (well, except for my back and the MDF dust). I hope to get to taking measurements and the crossover design soon. That will be a completely new experience for me.

[CraigJ]

Paul,

Looking very very good, I think you're going to hit a home run! Can't wait to see the Trilliums with veneer, then they will really come to life. I'm saving to make a pair.

Oh, is that you and your wife in the Provo River?

Craig

[dlneubec]

Nice looking speaker, Paul. :T Keep up the good work.

[Paul Ebert]

Looking very very good, I think you're going to hit a home run! Can't wait to see the Trilliums with veneer, then they will really come to life. I'm saving to make a pair.

Thanks, Craig! I'm really looking forward to seeing them with veneer as well, but I anticipate it will be a while. I'd guess sometime around summer or fall. I don't get to work on them much, so my pace is very slow.

Also, I hate to dissuade you, but these are my first design and build. I hope they turn out well, but the Mt8a's are a much safer bet. I do plan to take them to events so they can be peer evaluated (oh, my aching back 8O ). We'll see how well I'm able to design.

Oh, is that you and your wife in the Provo River?
Craig

Yes! We went to Sundance for our 25th. It was fantastic. It was her idea to do the fly fishing (honest!). Neither of us had done so before.

[Paul Ebert]

Nice looking speaker, Paul. :T Keep up the good work.

Thanks, Dan, especially coming from you - one of my design and esthetics hero's!

To be honest, if I were starting today, I'd probably build a variation of your Soundrounds (different mid and tweeter and the mid enclosures vertical pvc tubes). That will be my next speaker.

I tried to sell my wife on doing the Soundrounds instead, even putting a grill over the upper portion and she wouldn't go for it. She just really likes the Avalon style cabinet, despite the size. I'm not about to push the subject. 8)

[Paul Ebert]

I thought I'd share a few design considerations, since these are getting some attention.

The driver choice was made on the basis of low distortion with cost being a significant secondary factor. I also required a tweeter with a protective grid. I really liked the idea of a dome midrange (for some reason) and the excellent off-axis response of the DXT is a nice bonus (hopefully, they will blend well).

I plan to use a Duelund crossover - hoping to get good integration of the drivers. Recently, there's been discussion comparing the RS52 to the MDM55 with the RS52 being criticized fairly strongly. I'm hoping that the fact that the mid is a 'filler' driver when used with a Dueland crossover (and padded down quite a bit) will help eleviate the issues with the RS52.

After much going back and forth, I ended up going with the 2 RS225s in parallel (as opposed to only 1). The reason was that the increase in sensitivity gives me more freedom with respect to baffle step compensation (or, that's the thought anyway...). Also, I went with the RS225s over the RS270 because I did not want the extra width. Finally, I suspect that the woofer will be the limiting factor with respect to distortion (given the cost constraints), so doubling them up would help with that.

[JonW]

Hi Paul,

This project looks great! :T

I also recently took on my first speaker design. And I also went for the faceted look. Go all the way. :B The facets really do have some significant design benefits, it seems.

Yes, watch the back. The cabinets do get quite heavy in order to have them thick enough to be able to cut the facets. Including crossovers, drivers, and everything else, my speakers came out to 87 pounds each. As my back ages, I think that my future speaker projects will get to be smaller and lighter.

You’ll have fun. Good luck with it. :T

-Jon

[dlneubec]

I plan to use a Dueland crossover - hoping to get good integration of the drivers. Recently, there's been discussion comparing the RS52 to the MDM55 with the RS52 being criticized fairly strongly. I'm hoping that the fact that the mid is a 'filler' driver when used with a Dueland crossover (and padded down quite a bit) will help eleviate the issues with the RS52.

FYI, Paul Kittinger's Budget class entry this year at Iowa, the cenTiLana, uses the RS52 with a Duelund crossover. Pete Schumaker did the crossover design. It was very well received. He also had it at the DaytonDIY this year.

[augerpro]

Looking good Paul. One thing to consider using two RS225's will be the ctc distance from the lower woofer to the mid versus crossover frequency, assuming the layout will be TMWW. The dome mid by nature will have to be crossed relatively high and the Duelund filter is nominally 2nd order for some time. I've thought about how to deal with this if I ever wanted two woofers in my Delphis, going WTMW is one choice, or more likely I would try a 3.5 way.

[Paul Ebert]

Yes, I may have a problem with CTC between the RS52 and the lower RS225. If so, I'll probably do a 3.5. I'm leaving that open until I do the crossover experimentation. I wonder what impact the .5 would have on the Duelund approach. I guess I'll see.

I sure wish I could have made Iowa or Dayton, but with two kids in college soon, it just couldn't happen.

[Paul Ebert]

Testing woes

Well, it appears that I'm stymied in my starting to do testing. :x

My (elcheapo) Behringer soundcard seems to have a dead right input. Very frustrating as I was almost able to get my first impedance measurement of a RS180 I have laying around. It basically worked, but the graph was very noisy (I guess the input was going bad at that point). But now it seems completely dead.

At least I got to play around with ARTA a bit. It seems like it will meet my needs very well. I'll have to use it in demo mode for a while, but I can live with that. In the mean time, the growing strength of the dollar helps bring its price down. :T

[ThomasW]

Pictures of the build

http://www.flickr.com/photos/80104762@N00/sets/72157607167568936/

[Paul Ebert]

Thanks for posting the picture, Thomas. I tried the instructions you gave earlier for doing so, but couldn't get them to work. I'll try again.

[Paul Ebert]

OK, let's see if I can get this to work...



:T

[---k---]

Wow. Those look great.

[Paul Ebert]

Thanks, Ryan! Having finished one of the enclosures (the other one is ready to glue up), I'm now on to the next challenge - measurement. I seem to be having some soundcard issues with ARTA which I hope will be resolved by the M-Audio Transit I just ordered. So, next big step - FR measurements.

[Paul Ebert]

So, I've made progress with the measurements (using ARTA). I'm now able to make what appear to be very reasonable impedance measurements. I've got everything set up to make frequency measurements and here is what I've got so far:



This is a measurement of one of the RS225-8s from about 4 inches or so from the baffle and on axis with the driver. It seems roughly OK, though the break-up seems too small compared to the PE graph and it's got that 60 hz bump. The 60 hz bump may well be noise from the room (for example, the transformer in my power amp hums). Any thoughts on this response?

Also, the pink noise is quite loud. I'm glad I tried it first on the woofer and I'm concerned about how the RS52 and DXT are going to handle it. Should it be so loud?

[Dennis H]

You can make the noise however loud you want it.

[Jed]

Should it be so loud?

No, adjust your mixer settings higher and lower the amp output. Otherwise you'll damage the tweeter. There is a meter that monitors the input level and output levels. They should be the same in the yellow range.

[Dennis H]

In the manual, he advises maxing all the mixer settings. If your amp doesn't have a volume control, you can set it in the generator window. I assume you're using pink periodic noise to generate an impulse response.

Another useful setting for tweeters is the pink cutoff frequency. He recommends setting it to the Fs of the driver. It doesn't affect measurement accuracy below that freq but it just means there is less LF content in the signal that might harm the driver. The software automatically compensates.

[Paul Ebert]

Thanks for the suggestions. I'll give setting the output in the generator window a try. I am using a power amp so there's no volume control on it. Well, actually, now that I think about it it has a rotary switch whereby the output can be attenuated. I'll give that a try.

I assume that I want to measure the drivers at the same settings so the responses are at the same relative level.

[Dennis H]

Same settings are good but not really necessary if you're doing a 2-channel measurement. It compares the SPL to the output voltage of the amp. If you have gone through the setup and calibration routine, and have a calibrated mic, it can plot everything at dB SPL / 2.83V. If nothing else, you'll probably want to change the pink cutoff for woofers and tweeters.

[Paul Ebert]

I do have a calibrated mic and I've loaded the calibration file. I have not, however, calibrated the rest of the system. I guess I should do so.

I'm finding myself doubting the approach I thought I needed to follow to design the crossover. Here's what I thought I needed to do:

1. Install the drivers in the enclosure so the measurements reflect the impact of the enclosure.
2. Position the mic a meter away from the speaker at the tweeter level.
3. Take measurements of each driver (or the two paralleled RS225s, in this case) at the same level and mic position.
4. Save the measurements in frd format, including phase data.
5. Load these data, along with the respective impedance data into PCD (it, and ARTA are all I have at the moment).
6. Design the crossover.
7. Build the crossover.
8. Iterate as needed.
9. Do final voicing.

Does this process make sense? Am I missing something? Is there a better way to do it (with these tools)?

[Jed]

I do have a calibrated mic and I've loaded the calibration file. I have not, however, calibrated the rest of the system. I guess I should do so.

I'm finding myself doubting the approach I thought I needed to follow to design the crossover. Here's what I thought I needed to do:

1. Install the drivers in the enclosure so the measurements reflect the impact of the enclosure.
2. Position the mic a meter away from the speaker at the tweeter level.
3. Take measurements of each driver (or the two paralleled RS225s, in this case) at the same level and mic position.
4. Save the measurements in frd format, including phase data.
5. Load these data, along with the respective impedance data into PCD (it, and ARTA are all I have at the moment).
6. Design the crossover.
7. Build the crossover.
8. Iterate as needed.
9. Do final voicing.

Does this process make sense? Am I missing something? Is there a better way to do it (with these tools)?

I usually splice nearfield data if measuring in a room. You're going to pick up a lot of reflections in the woofer response otherwise. This requires adjusting the baffle step and matching phase responses.

LSPCad manual shows how to do that... although you might be ok with the ARTA pink noise and not have to splice data like MLS.

[Paul Ebert]

Here's a measurement of the RS52 taken about an inch away and on axis:



The response looks OK to me, but what's up with the phase?

[cjd]

This is why when I measure a speaker, I set it up, set the mic up, and move NOTHING till I've measured everything.

Phase can look like crap but the key here is it's ALL EXACTLY RELATIVE so it lines up right, thus leading to the filters all being good.

There are all sorts of other options for extracting minimum phase and such to work with.