Microfarads- Nano Speaker using PS95 and W5-2053

I wanted to share a decent build I finished, what I'm calling the Microfarads!

4245a7b41d412a0a4c09d2a6a8bc88e9.jpg


Project Description:

This is the most challenging smallest speaker I've built to date. Due to the strict enclosure volume restrictions, this is also the most precisely built speaker I've made with each cut toleranced to less than 1/32".

Total internal volume (not air space) is 3 liters including the PS95 sub-enclosure, I was pretty frustrated about some of the nonsensical restrictions on this project initially but ultimately grew to appreciate Ben's sense of neurotic restricted creativity =)

The subwoofer is the Tang Band W5-2053 with 2 Peerless 5.25" per enclosure.

Crossover frequency is about 300hz

Design Goals:

1 - Build a full range speaker capable of solid bass with smooth response to 20kHz

2 - Total internal volume of 3 liters including mid enclosure

3 - Unique striking look

4 - Ensure every possible square inch is covered by a sound radiating diaphragm =)

Driver Selection:

I tried a few different drivers but always knew I'd end up with the W5-2053, this thing is a monster and the bass it puts out from only 3l is astounding. Was an interesting driver to work with as it's response falls off a literal cliff at around 1kHz, so I used that to my advantage in the XO.

I had a pair of PS95's I'd been wanting to try out and was shocked by how good they sounded, so after trying some different drivers I settled on those. They were also welcome with the limited space I had to work with.

Enclosure design:

This was purely determined by 2 factors...driver and PR mounting dimensions and the strict internal volume requirement. As you can see it's all a tight fit.

Precise calculations were made initially using some triangle area calculators to ensure internal volume was at or less than 3l.

Passive Radiators ended up with no weight which resulted in a 56hz tuning frequency, more boosted than I wanted but adding any weight maxed out the excursion that much sooner. I was able to add 15 grams to bring it down to 50 but the passives were straining much sooner, as is the W5 easily over powers 2 5" PR's, really 3 or 4 of these would better keep up with 1 W5.

4fc1564dda2859958b7ce1e753d3f9ee.jpg


Another passive radiator issue I dealt with early on was "enclosure walk", using dual opposed passive radiators solves most of that. Also Sorbothane feet in 30 durometer further eliminated any chance of movement.

Enclosure Construction:

Construction is solid mahogany, angled joints are mitered, all glue construction.

4c8f2e5d69d3f00ca8f8cdeed49950b1.jpg


I made this CAD (Cardboard Accurate Dimension) to help with the initial concept, I also used it to check test cuts to ensure things were aligning and helping with angle calculations:

81b8cdfd423aa56d63b26530fe479d60.jpg


Plywood test cut that was perfect, this locked in all my angles on the table saw:

1268acea905adbcac527e6180faf938d.jpg


Tape held it all together:

85f0b69016f6c1abaa57afe5fde51d44.jpg


In Mahogany:

8e7f8039e6dbb6e14c3819c7f4a226af.jpg


a5cdf2c6e89df81a47d6b5317e373c41.jpg


9da95a960463cec299a8aec365712423.jpg


I made the cut out of one board to match the grain:

2fa788ad1ab2ae44b614553509096577.jpg


Attaching front baffle:

b9cc02427e7f96da66522fe28a281757.jpg


My cut list:

97dac6c51eafaf6fe52018f47294e8e8.jpg


Test fit:

4cabc1cf6c36a0cd6b57acf895ac4a54.jpg


0deb65e447b8ac65222341300f7204a8.jpg


Clamping and gluing the baffle:

54fd1782c2cad5d4ecb88d60e847d736.jpg


Before I attached the sides I wanted to finalize the mid enclosure, since the PS95 needed no airspace to play to 300hz, I molded a tight fitting fiberglass enclosure.

I started by mounting the driver into a board just like I would in the final enclosure, then I built a makeshift mold using layers of alum foil and masking tape:

a73fff4bee3aa7a63c9390c061fe8975.jpg


Fiberglass matte:

7fba5e98bbd1588342d75a66b7787f70.jpg


Polyester resin:

b51008b2988a675e45e5d8a2b24be1cb.jpg


Wala, about 5 layers:

0eeb43e6854b70e245c663eb4be2e696.jpg


Removed from the mold:

6df5f152534df84893aed73ed88ed14d.jpg


Trimmed up and painted:

04f3ee81c202e35b57ca10a18158a260.jpg


8387332c78df5c4de633e75b73140370.jpg


Built into the mold was a 3/16" Sorbothane disc which creates a CLD treatment to the back of the enclosure, as the driver is screwed down it loads up on this viscoelastic dampening medium and greatly acoustically dampens the mid enclosure. I tested this and it works well:

db60538a1bc7e45f7e8bcd4a413482d2.jpg


33d5c3b95fb58f5d984608208331cdd4.jpg


c7f0a5c2cdf3ce081af99dacf290c83d.jpg


Mounted, it is attached with silicone adhesive and some screws for good measure:

697d23b887fa7d0f96c4c9dee318a1ba.jpg


4cee916f504a2618ae59f2b1253b199a.jpg


50b5039c9f6580fb7e92cbf6d2956c4a.jpg


Time for some sides:

076ed62e8b4ec41e7b7b27e025e8cbb1.jpg


All panels glued and flush, ready for some roundovers:

fb3d22b04dfebe273888b12a44610792.jpg


I didn't have enough room for a full 1" round over like I wanted to do, but fortunately the W5 wouldn't benefit from it, so I developed a way to do a tapered round over using a wedge for the router bearing to ride on:

ada9a83f721c2cc14cc55e30b1ef2bce.jpg


f5b076ec57774a1f2503bb6ffe29fd4c.jpg


3aaba84327fbbd1e72357d666ded2759.jpg


This created a precise contour that cleared all driver frames.

And all round overs and treatments done:

6ed4a70f54381641477767bafa96fd59.jpg


a271fbe1a7805f6a0aa99fbf06986ba3.jpg


And a coat of Danish Oil, BOOM! There's the grain!

2e343d87a9ef8b38dbf5379908ea73d5.jpg


25fb16c2765d91a0e2d05a1f34599015.jpg


And after soaking in:

5e82d51e2812b997b954339f6aa365f6.jpg


Crossover Design:

I tried many designs including various frequencies and electrical slopes before finalizing on what I have here. I ended up with 18db on the mid and 12 on the sub. Since I was crossing over at 300hz I used an Lpad to not only balance out the more efficient PS95 but also as a baffle step function.

The PS95 also displays break up bad above 15k but it starts around 9k, I implemented a 12db low pass sound 10kHZ which rolled off the top end nicely until it was voiced the way I wanted. Not sibilant but with air and space. Indicators are 18ga, all capacitors are electrolytic except for poly in the PS95 top end roll off and the 1uf listed in the schematic.

I tuned this until it had a flat response but like that the highs just fell flat, despite loving the FR plot it just didn't sound rift, so what sounded best is what's below (near field at about 12" centered on the mid.

51c7e01586a2e208006688a128e0793d.jpg


6ca7aee46c4d3a42185d2aeab1ad0ffe.jpg


ce57fb1ffb5fe517440f1b7045ab53a1.jpg


Installed, tight fit!

a19756f3eabdfee990ca2050935b0034.jpg


Conclusion:

a9a3af37b153d7699328bd4fdafcc99b.jpeg


facf36d638030fb23a97b0a34caa1fb5.jpeg


93265556d4f0ee37b63b44a564d42080.jpeg


43be77243e0c35b941a769886a99b2d8.jpg


cceee327debd1b88a1355c9cca20a21b.jpg


6fde07870c8fe1e96c174534bdbcc1fc.jpg


These were a blast and they've been my go to speaker for casual listening the last few weeks! Everyone that has heard them has fallen love with them, so I'll be very interested to hear how they hold up at Indiy. The bass is shockingly powerful for 3 1/2 inch diaphragm, they put out comparable bass to many 10 inch sub offers that I've heard and built in the past, they will easily trounce most 8" on the market.

This is the first time I've used a "full range" driver and I'm very pleased with it, working with the top and was challenging but once I got it rolled off and minimized the break up it really smoothed out and is quite surprising how much crisp clear treble this driver can produce. Imaging is excellent, they sound great near field as well as 6 to 8 feet away, they have no trouble filling a room with sound and I have even fooled a couple people into thinking they were hearing the Rally Sports when in fact it was the microfarads.

Most notable, ironically, is their bass performance. In fact they have too much bass, due to the limited excursion capabilities of the passive radiators as well as the very small air volume, tuning frequency is high and the bass response quite boosted below 120Hz. Though they rolloff below 45 Hz, that's enough to encompass the majority of all content in just about all my reference music and other music that I enjoy listening to regularly.

This has been an eye-opening venture into the micro speaker category, and I'm looking forward to doing some more small builds in the future, and another version of these with slightly more airspace with a more capable passive radiator.

Thanks please let me know your thoughts and if you have any questions!

Javad
 
Last edited:
The wood is absolutely lovely looking with so much effort invested in building an unusual shape along with taking photos of the process. This is just the kind of speakers you usually start to appreciate once you are passed showing off with big boys systems. Congratulations!
 
I've tried to simulate your speakers and it appears now that the issue at about 870 Hz isn't due air resonances, rather odd filter value of full range HP. This is your filter impedance. Can you measure to confirm that? I'd suggest a little modification to improve that, once we're sure it's what I'm suspecting.
 

Attachments

  • Microfarads original XO filter.jpg
    Microfarads original XO filter.jpg
    156.1 KB · Views: 93
I've tried to simulate your speakers and it appears now that the issue at about 870 Hz isn't due air resonances, rather odd filter value of full range HP. This is your filter impedance. Can you measure to confirm that? I'd suggest a little modification to improve that, once we're sure it's what I'm suspecting.

I spent over 30 hours tuning the XO on these, this is some anomoly with the PS95 interacting with the raised frame of the W5 woofer as if I removed the W5 from the baffle the response would smoothen out here. Ultimately it's quite a small bump, EQing it out digitally made little difference that I could hear.

Javad
 
...this is some anomoly with the PS95 interacting with the raised frame of the W5 woofer as if I removed the W5 from the baffle the response would smoothen out here.
Javad

The difference in impedance plots of your and my minor filter modification is evident in peaking of FR. Further FR plots show summing in and out of phase.
 

Attachments

  • Javad's filter in phase and out of phase.jpg
    Javad's filter in phase and out of phase.jpg
    137.1 KB · Views: 127
  • Lojzek's filter in phase and out of phase.jpg
    Lojzek's filter in phase and out of phase.jpg
    137.1 KB · Views: 103
  • HP IMP-Javad Grey-Lojzek Blue.jpg
    HP IMP-Javad Grey-Lojzek Blue.jpg
    142.4 KB · Views: 102
Absolutely beautiful and impressive both in design and implementation! I wanted so much to be able to attend InDIYana so that I could hear the contest entries in person.
While the "playoff" nature of the speaker contest had its benefits I can't help but wonder if some of the top designs (perhaps including yours) got knocked out of the competition early because of unlucky match ups. Think , "rock breaks scissors, scissors cut paper etc." Anyway, congrats on a gorgeous project.

Best,
Jay
 
Absolutely beautiful and impressive both in design and implementation! I wanted so much to be able to attend InDIYana so that I could hear the contest entries in person.[quote/]

Glad you like them, wish you could have made it out!



While the "playoff" nature of the speaker contest had its benefits I can't help but wonder if some of the top designs (perhaps including yours) got knocked out of the competition early because of unlucky match ups. Think , "rock breaks scissors, scissors cut paper etc." Anyway, congrats on a gorgeous project.



Best,

Jay


It was a fair matchup, actually seemed like a great way to do it.

Thanks!
Javad