Usher 8945A +9950-20tweeter soft dome

[gabdx]

Hi, this is a project for anyone to copy for home use.

First, there is one design flaw, I use 2 8945 per enclosure , one only as passive radiator, this could be substituted with a decent passive radiator for a nice cost decrease.

Box is 9wx14dx23tall, at edges (roughly) Material is 3/4 high quality spruce plywood with a oak finish.

For sound treatment: front baffle has a sound absorbing outside vinyl like sheet.

Bottom is the passive radiator, middle is the tweeter, top is the active woofer.

Inside is 1/2 bag of poly fiber , + vinyl sound dampening on all surfaces.

Tweeter is inserted flush, woofers are inserted only 2-3 milimeter into front baffle and have Inside 'wave guide' routed to help back air flow.

Box is constructed like a classic BBC box with multiple screws and sealing strips.

XO : bold, parts which sould be exact.

Woofer: 3rd order:
2.4mh (0.15R) bypassed with 10R
13uf to ground
0.39mh

Tweeter 3rd order:
13uf
0.3mh to ground,
7uf,
+ ( Z series: 7.5R+1.7mh+13uf, to ground)
and last, db Divider: 17R to ground, 3R in series.

Personal tweaking and tolerance to suit your room:

woofer: main inductor 2.4mh +/- 0.1mh , last inductor 0.3mh+/- 0.1mh (I use 0.39mh for my room.)

tweeter: 13uf +/- 0.5uf.
7uf +/- 0.5uf
the 13uf Z filter cap can be +/- 0.5 uf depending on your liking

The L pad is supposed to be 14R +/- 1R, however I like 17R, experiment.

 

[TMM]

Most expensive 7" passive radiator ever haha. The size enclosure seems adequate to have both drivers active, ported low. You'd pick up some sensitivity and lower the bass distortion to extraordinary levels.

 

[gabdx]

Hi TMM,

to me this was sounding more relaxed.

This is the last and final installation, first I went like you proposed with both drivers.
Could never make it work.

Then I went one driver with only a bass filter.

I did 2 drivers with a port. 1 driver with a port.

I did then 1 driver with sealed, and now with one passive radiator.

 

[gabdx]

in MTM, the stiffness is too much, sound constricted.

even with one driver with a bigger inductor. it is a small improvement.

In ported there is a boom bass which is very uneven.

With only one driver the bass is shy. with the passive one this improves. the cross over has plenty of baffle compensation..

the critical design point, if the builder has the equipment is to cross the woofer and tweeter at approximatively 2250 +/- 25 hz, done by ear, this is best. this implies tweaking the values for which I gave a range of tolerance.

 

[gabdx]

when you consider the effort to build this speaker and the kit which is around 900$ with everything. this is cheap for the sound it gives.

I don't mind to waste a driver of 100$ , I connect the coil to the ground wire, like a feedback for EMF. I think it might help the sound to have a voice coil..

I would venture to say this sound as good as 3500$ commercial speakers. and the best is that it sounds good unlike so many internet kits. It's got a good bass with détails, a good midrange, good détails, good high pitch with plenty of space and Imaging, more over it has this shouting voices controled in the 500 - 1000 hz , very hifi in terms of frequency response, no major dips or hips. Timing is also very nice. It can play loud.

It took me 6 years to develop the XO, first I thought I would never make this sound right, I keep trying with periods of rest until it sounded acceptable. That was my only goal, making an acceptable speaker.

 

[qguy2000]

Hi

Where did you get the Usher 8945a ? Parts express no longer carries it

Thanks

 

[ernperkins]

Unfortunately, Usher has left the DIY market. All drivers will be NLA after existing inventory is exhausted.

 

[gabdx]

I got it 7 years ago, this is how long it took me to design my XO !!!!

They are an incredible bargain in audio and perfect for amps such as tube 50 - 100 watts like Audio-Research

They still sell the tweeter and the Kevlar version, I am sure you can still get used parts on ebay or somewhere else.

 

[gabdx]

I did a little search of 2 min on internet, you can buy the drivers there:

Opentip.com: Usher 8945A 7" Carbon Fiber/Paper Woofer

Usher 8945P + HDS Peerless / ScanSpeak MTM Vifa Speakers Mid Century Style HiEnd | eBay

P version, better suited for midrange, 3 ways

 

[gabdx]

News of the project, System7 commented on the impedance. And I thought I could redesign the XO with help of computer simulation for better phase and better fidelity.

However, the new XO has a 3 db dip at 190 hz which isn't very good and a 3 db bump at around 2800hz caused by the driver resonance...

This needs to be fixed sooner or later

tweeter 3 order :
5 uf , 0,4 mh, 10 uf
L-pad 4R + 20R //

woofer input // filter : 2.4mh +5R + 30uf(this should be lowerer I think to boost 190 hz dip)
series filter to gnd : 0.3mh 5R 50uf

3 order = 1.8mh 9 uf 0.2 mh

 

[gabdx]

New XO:

Tweeter , 10uf, 5mH, 10uf (3rd order) L pad , serie 7.5R, Ground : 18R

Woofer: 4mH (0.3R) --- ( filter, 9.1R+68uf to GND in serie) ---- 8.2 uf to GND

Tweeter and woofer polarity reversed.

I tried many types of filters etc, notch, pass bands, Z, and this is better with the brute force 4mH inductor, it does the work.

 

[gabdx]

smooth sounding accurate phase and no low R xo with less parts possible:

 

[wolf_teeth]

I would not place a resistor to ground on a woofer that is that low of value unless it is in series with the cap that is also there. The resistor will shunt a lot of current and likely get warm.

Later,
Wolf

 

[gabdx]

Hi wolf, I looked into it, the system is around 12 ohm impedance, and it wont get warmer than the one for the tweeter, it is essential to have exactly 20 OHM for that resistor, a normal 10W is more than enough.

Ill play loud music and touch if it gets warm, I have a 0.5Watt for testing purpose, it should get warm if there is any power dissipated there...

 

[gabdx]

I just tested with sin waves at maximum cone excursion and the 1 W resistor barely gets warm with 2.5 V across it P=2.5*2.5/20 = 0.3 Watts

with music playing normal level I cannot feel any heat from the 0.5 W or 1 W ...

I think the most important is to have zero induction on that resistor and it is fine, if it is inductive maybe a 10 W is best.

 

[gabdx]

Without the resistor all the energy goes into the woofer at 2khz creating a massive peak.

With the resistor to proper value it brings down that peak. It is the most important part of the design.

 

[gabdx]

temporary setup, to be replaced soon, however it sounds very good... my dogs name is Mika, unfortunately at 14 + and two traumatic dog attacks, she is almost deaf, still looking good!