"Bottleships", In progress micro 3way build.

Wolf

I owe Nick for the light bulb on the name. He said they 'resemble putting a ship in a bottle'. Mix with Battleship, and voila!

So, if you thought the Purveyor were small, these are shoehorned into HALF that volume. You likely now think I've sincerely lost my mind. These are the smallest 3ways I've both seen and heard (of) in DIY circles. The Purveyor were 5.25"/2.5"/1.0" arrangement, and the new 'ships are 4"/2"/0.75". There is no way a 5.25" woofer would fit and perform well in a total of 3.5 ltrs, nor will the drivers fit on the baffle without using a 4" or smaller woofer. If I want the bass extension, a 3" with good bass performance would be too low in sensitivity for my requirements, so a 4" it has to be. Next problem is that the number of 4" with great bass performance is even fewer than that of the 5.25" pool of options. I settled on the Peerless SDS 830855, $19 at PE, and a supposed 86dB sensitivity. The obvious next choice is that of a 5.25" PR to align the stars for the best maiden voyage into bass dredging. (Looking around, I feel that the Dayton TCP115-4 could be a viable woofer substitution with some xover rework, as long as it fits on the baffle.)

So, putting this info above into practical use, the mid I decided on is the Audible Physics 220CP1 8ohm from Newark/Farnell for $22. It's a carbon paper cone with an aluminum dusctcap, cast frame, 86dB sensitivity, and a large Neo motor. Being that Chuck and others have really liked the AP stuff, I didn't even second guess my choice at the cheap price per. I fiddled around with 4-5 different chamber arrangements empirically modeling and physically testing them until I came upon another of my 'ice-cream sandwich' type alignments, or Multichambered Aperiodic sealed. Being a 2", and being primarily used from 700Hz+ as an estimate, the amount of volume required is small. However, the Fs magnitude and frequency is fairly high and could stand to be minimized a bit to aid in xover simplicity and damping of the rolloff. Using a 2" PVC coupler and endcap, a 1.5"-2" adapter, 2 equal pieces of Bonded Logic or Ultra Touch, a ring of wool batting, and a bit of E6000 adhesive; sufficed for a 0.13 ltr chamber.

Further on the boxes... Being of minimal midrange volume, this lent the woofer volume to be ~3.2 ltrs. Using the DSA135 PR without added mass models to Fb/3/10 of 60/68/48Hz, or a very gradual rolloff. Trying to conserve sensitivity, using a steel-laminate coil creates a bump in response around the Fs of the woofer of a dB or two. Then using a 730uF Passive Assist capacitor in combination with the highish Fs and the higher Q of a vented box, while gaining another order on the rolloff, actually also extends the lowend a smidge here. (I have a HUGE box of 330uF polarized caps, and I have been making large NPE caps with them for awhile now.)

I am using the Dayton ND20FA-6 because I have them, and it fits atop the mid on the baffle. They are tiny, but look big on these small speakers. The main positioning of the drivers on the baffle actually involves the midrange chamber as the most influential information. Acoustic offsets are at 2" from midrange to tweeter, and 4" from midrange to woofer. Unfortunately, this also means that the truncated Peerless MUST be mounted vertically to get everything to fit.

The xover as modeled and refined represents 17 individual components. I can hear the gasping and see the whites of your eyes now.... In this portion I am using the smaller and cheaper parts I've acquired over the years, like 22AWG coils for the tweeter and midrange shunts, small mylar caps, and as few resistors as possible. "How you gonna fit all that $#!+ in that tiny box, dude!?" was the usual reply in conversation with the closest of you.
Well- it ended up requiring a 3D xover board assembly. I've thought of doing this before, and have done daughter boards before a few times but never to this nth degree. It almost operates as a brace that press-fits vertically inside. The simulated response shown is using the spec FR for the woofer, as the response of the driver as measured was only good to about 600Hz. The model with the capacitor in Xsim is also shown.

More later as things are solidified;
Wolf

Wolf

Now, a lot of you are likely thinking, "Why!? would you have a 3way with a 4" woofer?" The initial goal was to use the AP mid or any 2" mid effectively. Going further, the tweeter is good to 4kHz before HD rises. The AP has HD above 1% below the 750-800Hz region. The 830855 has dominant 3rd order HD above about 1.2kHz, even though it does not rise above 1% in amplitude. According to these frames of reference, it makes sense why the HD would be lower or better for these particular drivers utilized together.

ND20FA6:

220CP1:

830855:

More later, including construction images...

Steve_Lee

Wolf has gone to the woodshed and carved a battleship in a bottle for us.
Love it!.

Wolf

Wolf

Wolf

Wolf

Wolf

Wolf

The mid takes #4 flush-head screws as the mounting holes are beveled. I bought some torx head "for plastic" screws that worked really nicely. The woofer/PR have #8 torx sheet metal screws. The tweeter is press fit and sealed with E6000 on the back. Woofer inside edge is beveled.
This was used a lot to speed up the progress:

The largest is 4.125" for tweeters' rebates. It's faster to drill the circumference, and then router out the depth of the rebate. I also have a rabbet bit set of bearings for the adjustment of the initial hole.

Wolf

The finish is 2-3 coats of Watco Danish Oil, then Minwax Golden Oak stain. Being the grain is a bit open, I did something different this time. I used Permatex black silicone to cover the surface, and then slowly wiped off the excess. After that dried, I steel wooled the surface to remove the crumbs and excessive bits. Then a coat of polyurethane finished the front and back off. The chatoyance is pretty nice, and I really like the edge look.

tajanes

so what did you do for the binding posts? can you show some detail- thx

squamishdroc

All hail the zip tie! Cool project - love how you made your skeleton for the crossover - good idea.

Wolf

Andy, those are rear mounted binding posts that did not have the front nut to clamp down. I grabbed a bag a couple years ago at a DIY event. Basically, I drill a 19/64" hole, recess the entrance from the rear so it's closer to the front, and lightly bevel the front entrance. Then I seal it with E6000 to be air tight and keep it strong. I even made/altered a custom socket to drive them from an 11/32" Craftsman deep well socket as I have a LOT of them. I apply solder to the tips to pre-tin them, reheat and insert the wire.
If you mean the 3 terminals, I plan on measuring the difference nearfield with and without the 730uF in play. The Blue terminal is direct to the lowpass coil and bypasses this cap whereas the red terminal is input to the 730uF capacitor bundle. I can just move the banana plug and use it the other way.

tajanes

Thx. I’m looking at doing something similar with my current build- not much room and I like simple (sans the third rail).

6thplanet

Huh, yeah, I guess "10 lbs of shit in a 5 lb bag" would have been to long of a name.
I like it🤘🏼 Nice build.

Tom_S

Nice! That's some crafty crossover layout there.

Wolf

Well, they are running! I'll break them in for a couple days and see what I think. Right now they seem balanced, and rather clean.
More later...

tktran

Sounds insane for a 3 way in 3L, but if it's not a crazy idea, it's probably not worth doing.

I have my own 2L 2-way, except it uses that notorious Purifi 4" midwoofer that blows the budget sideways, but I'm not sure if it would be able to blow up your battleship. Well done!

How thick if your timber? Maybe time for 6mm MDF or 3D printed...

Wolf

This is formica on 3/4" particle board, so it's ~7/8". Good, solid, dense, and well finished material.

Wolf

My Synchaeta are a 4" 2way in 2 ltrs, so I've done that as well.

Wolf

Mid and tweeter sound perfect together, but maybe a decibel hot over the woofer level. That said, these are actually really impressive WITHOUT the big cap. With was okay, but has some insertion loss for sure.
I could take them to -10 on the pre to where they started to complain a bit without the cap engaged. There doesn't seem to be much difference in spectral balance with the cap though. They really can kick up a storm in the bass too. I have a roll of wool batting wedged between the woofer and PR. I have yet to measure tuning, but there is no added mass on the PR. I really like these little mids.

tajanes

and the woofer / PR may 'appreciate' the extra internal space...(sans the 750+uF cap)

Wolf

I think depending on application, it might be required. It's not huge, maybe the size of 4x 10W resistors.

Wolf

Tuning verified:

It was supposed to be 60Hz, but came out 50Hz. Not really sure how this happened. I only have the dinky supplied washer and bolt that came with the PRs attached to the mounting hole. Maybe the Mms on these units was higher than spec, or the volume somehow is larger than modeled- and I know if anything it's less volume. So- a better prediction of the response (pre BSC) is like this:

Yields an F3/10 of 80/42Hz, and that is not really all that bad. Bass sounds great for a 4" design.

Wolf

Here is the current set of system FR and HD of the right channel:

I have since modeled that an additional 20uF on the midrange highpass cap will yield about +2.5dB to fill in at 700Hz area. As if it wasn't hard to fit in there what is already in there. I did measure off axes, and the dip remained if not worse, so it is definitely there. If it sounds worse, I'll adjust as needed. Vocals don't really have any warmth, and sound kind of thin or sterile at the moment, so this is kind of expected.

It's nice that there is reasonable output at least to 60Hz, even though the gating skews that a bit.

Wolf

I was able to parallel another 15uF to the inherent HP cap, and then went on to remeasure:

I was lucky to find I had one more pair of those Eton caps, as I don't prefer to mix caps if at all possible.

Wolf

I tried to isolate the project further from reflections this time too, as they are so broad response in terms of dispersion. The 15uF cap addition yields roughly xover points of 900 Hz on the lower and 4000 Hz (as before). It also turns out that the dip at 700 Hz was not just caused by inherent response, but also some by room interference. Measuring the system higher in my room moved the dip lower to about 500Hz, as well as cleaning up the dip in the woofer rolloff. This shows closer to what is actual response of the woofer, and looks a lot more like that of the more recent box simulation. HD is still not bad and below the -45dB point for the most part.

Time for some more listening, but I really like what I'm seeing above! In theory I should be able to leave them alone as-is.

tajanes

also nice to see the issue around 160 +/- was room / floor related

Wolf

I knew that was for certain, I guess I did not say so earlier.

rstillin

Would there be a major disadvantage of leaving the tweeter out of this and using the AP as the tweet? Seems like you would lose a decent amount off axis above 10k, unless you stay within +/- 30 deg., but the CP220 looks like it has lower distortion from 8k up vs the ND20. Parts count would drop as a plus. Just curious if you've tried such a format before (small FR as the tweet)?

Wolf

I have, but the off axis suffering would be the disadvantage. I'm not usually a WAW or FAST kind of designer, nor do I favor them. Hybrid Vigor is one I did though. There also might be a loss of bandpass gain if not using the tweeter.