Bento Box Build Thread

[JasonP]

Ok finally, with much delay, they are up and at'um for further testing! Still unfinished, but I'm putting that off a bit until they are fully designed and I'm busy doing something else so I don't miss my stereo



So I can work on the XO, its outboard and mounted with hot glue on a piece of ply. This is ugly, but functional for now. If I get more serious about doing this stuff, I'll look into making a xo board from prototyping like I've seen in threads here



I reduced the XO by several components and went with the other PDF curves for the tweeter (not the one on PEs site, since it reads like its preliminary data, but the one of Typhany and Madisound's, vastly different FR curves). Here is the one I'm voicing now, drawn up on TinyCAD (thanks for the link Greywarden!) I'm calling this version UNAGI, keeping with the Bento Box theme



I didn't have the right 8uf for the high-pass second order, so I'm using a 6.8uf and a .47uf combined but that is still only 7.27uf and its audible as a sort of darkness to some voices. I'm going to attribute that darkness to that mistake for the moment anyway (heh). Here is what PCD was showing me with this XO, just above axis (5 degrees):



Here is the same with power response, driver output, and phase:



Based on limited listening impressions, I've start work on prototype XO 3, codename: EBI (mmm, shrimps!). I'll make another short reply post with some listening impressions in a moment.

[JasonP]

Ok, so now the important stuff! Listening!

I was really worried when I first hooked them up. Song one has a strong bright female vocal and man they sounded like garbage. They sounded tinny and harsh, way worse than the Klipsch Reference bookshelves they replaced. However, not to be dissuaded I turned it up and listened on. It took about a hour of high-volume material for them to settle in, I assume that is actual break-in? I don't have too much experience with it to have a frame of reference.

After the first hour I found myself just enjoying the music. The system has great detail and the treble is so balanced it doesn't sparkle or sheen, its just there. People have talked about the SG shielded version of this tweeter, and man I can see why. I'm not sure what to say beyond the fact that I'm running it with a 2nd order electrical XO at just under 2k and its resonance is at 1130 or so and it doesn't seem to mind at all. I'm impressed.

At the moment they sound a bit dark, I'm certain there is something going on there with my room, because its a small 11' x 13' (8 foot ceilings) theater room with a serious amount of reflection. I was shocked at first, given how bright they would appear to sound from the PCD charts. It'll be interesting to measure these once I get the gear this month. Meanwhile I'll tweak and listen and reap the benefits of this low-cost speaker

[JasonP]

Finished the boxes, and while it may not quite be up to the snuff of the other master woodworkers of the site, I'm quite pleased considering the limited cost and effort it took to achieve. Pics!









Now to just finalize the XO design. I've already gone from a 2nd order electrical on the tweeter to a 3rd in EBI and increased the BSC inductor a bit on the woofer, plus totally redesigned the tweeter shaping/attenuation circuit. Its almost there, still sounding a bit bright at the moment, but I'll solve it!

Oh, and part of the lag on the XO design is a glaring mistake on my part, I forgot to extract the phase on the impedance of the Vifa Tweeter before going into PCD <.< It made a couple DB reduction in the crossover range, enough to make a lot of female vocals sound hazy. Whoops!

[dthomas]

Utter Junk... where is the bare MDF and gouges from your belt sander at. This is supposed to be DIY if we wanted speakers that looked all pretty and sissy like that we could just go to the store and buy some shiny pretty speakers.

Seriously looking good man.... I always like the way a nice wood finish contrasts against a black cabinet.

[davepellegrene]

I think they look really nice. Remember all those master woodworkers started somewhere.

Dave

[dcgold2000]

Utter Junk... where is the bare MDF and gouges from your belt sander at. This is supposed to be DIY if we wanted speakers that looked all pretty and sissy like that we could just go to the store and buy some shiny pretty speakers.

+1

[JasonP]

Utter Junk... where is the bare MDF and gouges from your belt sander at. This is supposed to be DIY if we wanted speakers that looked all pretty and sissy like that we could just go to the store and buy some shiny pretty speakers.

Seriously looking good man.... I always like the way a nice wood finish contrasts against a black cabinet.

lol at the first part, thanks for the second These are no small part inspired by Brewski's Levitations (thread here: http://techtalk.parts-express.com/sh...d.php?t=224505) that look nothing short of fantastic, also the designs share sister tweeters - if my guess if correct based on their PDFs and naming, BC25TG15-04 and BC25SG15-04.

Now if I can just get this crossover nailed... they still sound a bit bright. Its amazing how different XOs sound even when they have similar responses, I'm always blown away by that.

I also decided I'm going to use black screws for the drivers and just leave the six on the front baffle in brass, its just too much brass otherwise, total overkill.

[JasonP]

Ok, this looks like the final XO (I think). When the parts come tomorrow I'll build and listen for a week or so and see. Man, I'm going to have a stack of XO parts for my next project The EBI XO is quite reduced, down to 5 components on the woofer and 4 on the tweeter for 9 total.

Here is the summed response, 10 degrees off-axis towards the woofer, with the predicted phase null overlayed:


The big picture of what is going on with the XO, driver phase, power response, filter graphs, and overall impedance.

Yes, it does go down to 2.9 ohms from 200 to 300 hz, these aren't cheap receiver friendly and that was never a goal The shelf on the tweeter's filter is due to the 1100 resonance, which is down enough (-27db) I can't hear it so I didn't LCR it.

The low pass part in detail:

The slope fits the target LR4 at 2200 nearly perfect, and the breakup modes are down down down by more than 30db. I don't think I can hear them, though I'll admit my lack of experience with metal drivers makes me wonder what I should be listening for...

The high pass part in detail:

I left the bump in at 16khz, because frankly just playing with the EQ on my Sound Blaster I could barely tell anything adjusting, so it seems a waste on me to add more components to solve something I may or may not notice.

TinyCAD circuit as it is proposed for the final version:


So I'm crossing my fingers and plan on long nights of critical listening to great music. Oh no, so much work to be done now

[JasonP]

Oh, and that whole darkness on the female vocals deal, turned out not to be an XO issue at all. When I was listening to them before they were stained and painting, I didn't glue and seal and front board on... which left a gap all around the plate of 1mm or more. This made the box a funky vented enclosure, which caused a boost in the 200ish range, enough to make vocals often seem dark or hazy. Crazy. I hadn't noticed till I was standing next to them and felt the air pushing out from the gap <.<

Well, not that the fronts are finished and glued, that effect has gone away. Its interesting how that changed their total tonal character.

Edit Whoops! that .15mh in the woofer XO is .20 I'll update the diagram.

[JasonP]

So let me put another vote in for adding this snippet to the guides for using response modeller and pcd: Always extended the tweeter's response downward at 12db per octave before extracting minimum phase (As noted in another thread). This made me flip the design tweeter up, and build a new XO at 1800 Hz with the driver's in phase on axis, standard LR4 acoustical phase. Man, talk about a change in sound. They are getting there now

[JasonP]

Ok, these babies are complete. Here is the final XO circuit, with more better modeled diffraction effects on the tweeter ( ). They sound good, a little forward at times, like say when I turn up Dark Side of the Moon. For most recordings though, they sound simply awesome. They don't handle volume well, but I'm not sure if that is the speakers themselves or because of: my small live room or my weak t-amp. When turned up loud, they tend to sound overcast and I notice voices seem to be missing something in the upper range. Since they sound good at moderate to low levels, I'm assuming its not the XO. Here are some details.

Here is the final PCD response, 10 degrees towards the inside, with phase, drive outputs, and power response:


Here is the response on-axis and 10 degrees towards the inside (meaning the tweeter placed on the outside of the listening setup), plus impedance, phase, and filter functions:


You can see the baffle effect at 2-3 khz, which I'm not happy about and will solve in the second version of this design. I'm sure its audible and why some female voices sound overcast - not terrible but just a hint of something going on there on careful critical listening.

This is the circuit making it all happen (the coils 18AWG Jantzen in the woofer path, and 20AWG Jantzen shunting the tweeter, DCR is listed):


Given how good they sound, I will indeed (as mentioned earlier in this thread) give this pair to my Dad and end up building a second for myself. I want to do a ported one, using these cheap CNC knock-down cabinets by parts-express: http://www.parts-express.com/pe/show...umber=300-7048 So far unibox has them playing nicely down to 50 hz at 40W with just 5mm excursion (+1mm from their spec'd 4).

I'll post a cost breakdown next in a couple minutes.

[JasonP]

Just for a test I put the entire ported design using the knock-down cabinet (with everything like terminal plates and some sonic barrier/foam to line the inside walls) into a cart and it tops out at 251.07, not bad at all IMHO. So that'll be for my next build - Bento II - the ported reckoning

The drivers: 103.48
The XO:

- 57.56 using non-polarized shunt caps in the woofer circuit without bypass.
- 90.80 using all dayton metallized poly. caps.

Seems reasonable enough.