Desktop OB computer speakers

[merajsalek]

Ok, I've been playing around with this idea for a while now. Have made a couple of posts on it too. The idea have evolved much since conception, albeit in my mind only :-)

At present I'm looking at Hawthorne Audio Solo/Duet style desktop monitors coupled with Rod Elliot's Hi-Fi PC Speakers style (sub)woofer. I couldn't find a suitable co-axial, so I'm going down the conventional TM path..

Objectives:
>Lowest possible distortion
>As low-cost as possible (Jaycar Drivers)

Drivers:
>Jaycar CT2005 Soft Dome Tweeter
>Jaycar CW2190 4" or CW2192 5" midwoofer

XO:
>4th order active LR
>Filters as required

Baffle:
>Shallow U Frame

Positioning:
The monitors would sit on my desk (120cm x 75cm) about 60cm away from my ears. The back would be quite close to the wall, about 15cm.



Now for the burning questions:

*Which midwoofer should I choose? The 4" or the 6"?
*Is adding a rear-firing, phase-inverted tweeter a good idea?
*given the intended positioning of the speakers, would adding sound absorbing material at the back of the mid negate the benefits of OB?



These would be my first speakers developed and built from scratch. I've been closely following the OB exploits of the forum members, especially Gainphile and Moondog, and have realized how demanding a task it can be. My limited knowledge and skills means I'll be asking a lot of stupid questions - my apologies and thanks in advance.

Cheers

(HAPPY AUSTRALIA DAY)

[merajsalek]

The baffle I have in mind:

[valleyman]

If I were building desktop speakers to go 15cm from a wall I would go sealed rather than OB... they really need more room than that IMO.
If you are set on OB, I would try angling them inwards and stuffing some damping material between the drivers and those brace columns.

I would probably go 6" rather than 4". I think the 4" in such small baffles might even struggle to make it low enough to cross sensibly to a sub

[ttan98]

Desktop OB without a proper woofer9seal or ported) you don't get any bass.

[Dr.EM]

I would have thought since it is very nearfield you will still get bass response, but not much if you're away from the desk.

[Moondog55]

I've no experience with those particular drivers, but up close bass should be OK, box is too much for the application, simpler would be better.
To be honest with you the statement >Lowest possible distortion
>As low-cost as possible (Jaycar Drivers) is a tautology, the 8 inch bass driver is OK but i have not yet heard any good reports on the 5inch or smaller.
You can make up your own co-axial "A La SEAS "

Whereabouts are you?? I may be able to help with a small dedicated OB full range, an old Panasonic 6inch with an alnico slug magnet, usable from 250 thru to 15000, see if you can get any of the old 8inch papercone woofer, it can handle 5 or 6 watts OB needs a low XO point because of the HF noise it makes tho

[merajsalek]

Thanks for the replies, guys!!

Thanks for the suggestions, Valleyman. I'll keep them in mind when I build the speakers.

Ttam98, I'll be adding a sealed (sub)woofer crossed over around 300Hz with active xo and Linkwitz transform.

DrEm, do you suppose I'd have bass problems even with a sub? if so, how can I combat the lack?

Ted, you are spot on with the 'tautology' bit :-) but I'm out of options mate. A student budget restricts a lot of my aspirations. I was avoiding the 8" as that'd roughly mean a 16" baffle and that's way too big for my desktop. The WAF or rather the GAF is a concern here.

What aspect of the baffle seemed 'too much'? I was taking Andy G's comment on Gainphile's Low Distortion OB into consideration: http://www.diyaudio.com/forums/showt...89#post1705589 Did I do something wrong here?

You can make up your own co-axial "A La SEAS "

Sounds interesting. Could you elaborate on that?

I'm from Brisbane, and the Jaycar here seem to have more dodgy chinese stuff (The amazing dancing coke bottle or something along the line) than serious stuff.

The Panasonic alnicos sound like winners, do you have any pics?

Cheers

[Moondog55]

More that the two posts at the back are simply not needed for structural reasons.

I don't think that for near feild listening that baffle width has much meaning when you are making a U-frame, perhaps Rudolph could comment on this??

I've made a comment in another thread about making co-axial speakers.
It was concerning an old SEAS 12inch Co-ax I bought as an experiment.

Make up a frame from some aluminium flat, that runs from one side of the driver to the other, mark the centre of the bar. find a small tweeter that fits in the bar and just screw the bar across the woofer, personally I thought 12 inch was too big; might work with an 8inch mid-range KEF of course use a pole mounted Neo tweeter

[gainphile]

Happy Australia Day ! ... we went to cape paterson yesterday, beautiful beaches and great weather.

The plan would work ok. It will be more complex than simple sealed/ported box but you get no box coloration in return. 15cm from the wall is really pushing it too, I would not use the u-frame as it will create cavity resonance and takes more space.


Good to know you're going active xo. More easier to design.

- Go with 5". Baffle width should not be more than 2x cone diameter. Generally the narrower, the better (but a tradeoff with freq. extension).

- Since you are using active, bass is easy. Do a shelving low-pass from about 500Hz down to 200Hz. The low end doesn't matter as we will cut it with LR filter.

- The midrange will neen dipole peak notch filter. Usually in the order of 600-700Hz, -4 to -6 db down, and medium Q (0.9 - 1.2). This will need measurements or perhaps a software parametric eq. to test. Without this notch it will sound shouty.

- xo between midrange panel and bass, just use 12db/octave at 300Hz. Simpler and better group delay. I think I can hear ringing with 24db although SL says it's not audible. You can experiment later to cut them even lower, 120Hz if possible so localization is avoided.

- xo between mid and tweeter should be 24db. This is because you want to cut the tweeter as low as possible, so that the midrange would not go into peak-and-dip territory of OB. 1.5kHz is good if the tweeters can handle it.


btw. this is quite a complexity as you need 3 amps, measurement etc. Other option is just use a full-range and accept the lack of HF. My work in progress is 3.5 Full-range with 10" jaycar woofer. Not too big and don't need subwoofer (I've tried from 5" to 15" on my desk!, u-frame, h-frame etc.) but still need 2 amps. The crossover is very simple.



The tangband is very hard to find now down under, but HiVi B3N/B3S is cheap and available at darcheraudio ($19). They measure and sound great as per Zaph's investigation (http://zaphaudio.com/audio-speaker18.html). I think B3N and 10" Jaycar would make a great combination.



There was an old thread here which may not be relevant anymore as things have been learnt since.
Bookshelf dipole speakers - is it possible ?

[merajsalek]

To Moondog:

I'd added those posts at the back because I'm thinking of making the front baffle out of polystyrene - a Decibel Dungeon idea - http://myweb.tiscali.co.uk/nuukspot/...speakers2.html They'd also come in handy for mounting a rear-firing, phase-inverted tweeter and terminals/binding posts. So the top, bottom and possibly the posts at the back would be wood.

I'll keep the DIY coaxial in mind. I don't have access to metal working tools at present, so I'll have to stick to a TM arrangement.

To Gainphile:

Happy Australia Day!! WOOT WOOT!! :-) Cape Paterson, eh? Good place to work up that tan ;-)

Cavity resonance issue with the U frame noted, will ditch the idea then.

- Go with 5". Baffle width should not be more than 2x cone diameter. Generally the narrower, the better (but a tradeoff with freq. extension).
- Since you are using active, bass is easy. Do a shelving low-pass from about 500Hz down to 200Hz. The low end doesn't matter as we will cut it with LR filter.
- The midrange will neen dipole peak notch filter. Usually in the order of 600-700Hz, -4 to -6 db down, and medium Q (0.9 - 1.2). This will need measurements or perhaps a software parametric eq. to test. Without this notch it will sound shouty.
- xo between midrange panel and bass, just use 12db/octave at 300Hz. Simpler and better group delay. I think I can hear ringing with 24db although SL says it's not audible. You can experiment later to cut them even lower, 120Hz if possible so localization is avoided.
- xo between mid and tweeter should be 24db. This is because you want to cut the tweeter as low as possible, so that the midrange would not go into peak-and-dip territory of OB. 1.5kHz is good if the tweeters can handle it.

Noted!

I'm looking fwd to the learning experience that this project would afford me. I plan to use ARTA for the measurements and design filters based on the measurements. This would be my first speaker done this way, so complexity I'm willing to tackle. BTW, do you still have any more of those Panasonic WM-16A's? I'd posted a comment on them in your BLOG, under Cheap and Accurate Speaker Measurement...do let me know.

As a 10" baffle is the largest I'm prepared to put on my desktop, paring Hi-Vi B3S's with 10" jaycar is not an option. The Hi-Vi's are a good choice for drivers tho. I'm drifting off on a different tangent now...

How about paring a B1S or a B2S with the 5" jaycar? Is a rare-firing, phase-inverted driver required? Or should I mount 2 B2S's back to back in FR mode and have a forward firing augmenter tweeter?

What say ye?