diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Multi-Way (http://www.diyaudio.com/forums/multi-way/)
-   -   Two Boxsim projects, which one would you build? (http://www.diyaudio.com/forums/multi-way/224822-two-boxsim-projects-one-would-you-build.html)

Kaspari 2nd December 2012 01:29 PM

Two Boxsim projects, which one would you build?
 
5 Attachment(s)
Hello Guys.

I've been busy on designing a speaker in Boxsim.
The basic Idea is to make a speaker with the best possible value for money, hence the name Top Dollar. Right now the box and drivers are set. I choose two W170s in parallel as midwoofers and the ferro fluid cooled G25FFL as tweeter, both from Visaton. In general I like the sound of coated paper and soft dome drivers, if crossed properly their sounds match up very nicely IMHO. I Put them in a 30L box with a port tuning freq. of 32Hz. The total can be built for around 600,-, cost of the drivers alone is 280,-

Here is a pic of the design.
Attachment 315327

Now I end up with a dilemma. I made two possible crossovers.
This first one is a series design I used from a Troels Gravesen project and optimised myself with help from the XO-optimiser built in Boxsim. I mainly just removed a few networks I found obsolete and kept running the XO-optimiser until it looked decent. I love it's simplicity :)
Freq/Imp response:
Attachment 315328
Crossover:
Attachment 315329

The second option a calculated myself from scratch. It's an 2nd order parralel crossover with tweeter resonance compensation, sens. matching and a zobel on the woofers. The Q was calculated at 0,5 for a Linkwitz slope. I ran it trough the boxsim optimiser and I'm quite content with it. The impendance plot is not as flat as I hoped though, I'm starting to think Boxsim has little priority with keeping it flat. Also the green peak of the tweeter worries me a bit.
Freq/imp response:
Attachment 315330
Crossover:
Attachment 315331

Which one would you guys build?


Some sidenotes on Boxsim:
  • It tends to give quite big impendance swings in the freq spectrum when running the optimiser. In general the Freq response gets better but I worry the bigger impendance swings will mess up the speaker IRL. I strongly favour an as flat as possible imp plot. What do you guys think?
  • Looking in the Boxsim database website I see quite a lot of designs with big Impendance swings. Are they badly designed or is it just not that big of a deal?
  • Is there a place where I can find Freq, phase and imp response plots for drivers from different vendors to load in boxsim?

system7 2nd December 2012 02:58 PM

1 Attachment(s)
I did this one a while back when I was interested in a Morel MTM design. I found a little notch on the bass worked wonders for phase and reducing cone breakup from the woofers. Since it is a Qts=0.5 woofer with Vas of 38L, a big closed box made sense to me. Maybe 50L.

I don't really see what you are trying to do with the undamped tweeter notch. It's plain wrong for a Fs 1600Hz tweeter. :confused:

For all that, the tweeter does geta bit of a hammering in this design, but that's how it goes with loud MTM designs around 90dB SPL. In the end you build stuff because it appeals to you. I certainly wouldn't worry about some impedance peaks, transistor amps don't mind them. Just as long as it doesn't go too low.

Tricky blighters, serial crossovers. Don't think these drivers are suitable. Boxsim is limited in importing other drivers data, but it can be done.

a_tewinkel 2nd December 2012 03:55 PM

You need measurement gear if you want to design your own speaker. Boxsim is a nice tool but it is not accurate enough for a "final" design. I've seen many simulations with Visaton drivers which differ from measurements by > 1 dB, which is a lot if it covers an octave or more.

As for value for money, I'd choose a Seas tweeter. The woofer is ok but I think the price/performance ratio of the Peerless SLS line is better. You could also look on the web for low-budget designs. There's lots of those.

Dissi 2nd December 2012 04:37 PM

The simulation seems to assume a volume of 30 litres per driver. Therefore your speaker will need a total volume of 60 litres.

Kaspari 2nd December 2012 04:38 PM

Quote:

Originally Posted by Dissi (Post 3267266)
The simulation seems to assume a volume of 30 litres per driver. Therefore your speaker will need a total volume of 60 litres.

They share the same 30L housing :)

Kaspari 2nd December 2012 04:48 PM

Quote:

Originally Posted by a_tewinkel (Post 3267197)
You need measurement gear if you want to design your own speaker.

I'm planning to do so indeed, do you know a good place near Enschede for some cheap but good starting equipment?

Quote:

Originally Posted by system7 (Post 3267116)
I did this one a while back when I was interested in a Morel MTM design. I found a little notch on the bass worked wonders for phase and reducing cone breakup from the woofers. Since it is a Qts=0.5 woofer with Vas of 38L, a big closed box made sense to me. Maybe 50L.

I don't really see what you are trying to do with the undamped tweeter notch. It's plain wrong for a Fs 1600Hz tweeter. :confused:

For all that, the tweeter does geta bit of a hammering in this design, but that's how it goes with loud MTM designs around 90dB SPL. In the end you build stuff because it appeals to you. I certainly wouldn't worry about some impedance peaks, transistor amps don't mind them. Just as long as it doesn't go too low.

Tricky blighters, serial crossovers. Don't think these drivers are suitable. Boxsim is limited in importing other drivers data, but it can be done.

Thanks for the tips!
I'll try your crossover, I'm curious :)
-I worry about impendance peaks because when you cross at that point the rising impendance will shift the crossoverpoint. Or am I mistaken?
-What makes a driver suitable for series networks in comparison with parallel? I always thought the drivers didn't matter.
-I tried dampening the tweeter resonance peak but for some reason it didn't work. I'll redo the math soon.

system7 2nd December 2012 05:40 PM

W 170 S - 8 Ohm

30-40L reflex or closed box looks OK. :)

That circuit I showed you is not a million miles from the SEAS Odin 3 which uses metal woofers, but the principle of notching the woofer still applies.

Serial filters are just picky about drivers. Polycones work best. Crossover point is not affected by impedance, though a valve amplifier might behave differently. I would do things differently if designing for a SET valve amplifier with a highish output impedance. I'm not sure you should even try to notch the tweeter Fs resonance. Not necessary really.

Kaspari 2nd December 2012 05:53 PM

2 Attachment(s)
Quote:

Originally Posted by system7 (Post 3267350)
W 170 S - 8 Ohm
That circuit I showed you is not a million miles from the SEAS Odin 3 which uses metal woofers, but the principle of notching the woofer still applies.

That circuit works quite well indeed! :)
I only added a little resistor to up the minimum impendance to 4Ohm and damped the tweeter a tiny bit. I think it's the best so far, I hope if you don't mind me using it?

Freq response:
Attachment 315391
Crossover:
Attachment 315392

Worthy a build and a measure?

a_tewinkel 2nd December 2012 06:09 PM

Quote:

Originally Posted by Kaspari (Post 3267269)
They share the same 30L housing :)

This means that each driver should have 15 litres each in Boxsim. For a W170S it would be better to use one and put it in 30 litres BR.

As for measurement gear, I've got a Dayton EMM-6 microphone which I've calibrated (the original calibration was a mess). You'll also need a decent sound card (preferably with ASIO support) and mic amp with phantom power. If you want to measure impedance you could build a simple circuit with a few resistors yourself.

system7 2nd December 2012 07:01 PM

I've gotta agree that going below 40L here might get a bit problematic. You can reduce the bass bafflestep coil to compensate, but that would have knock-on effects elsewhere.

Adding resistance (that 0.33R) around the bass coil has effects too. It increases the overall Qe of the driver and makes the bass woolly. Since you are around 4 ohms already, you really don't want more than 0.4 ohms on the bass coil to avoid making the box even bigger.

Explained here by the brilliant Morgan Jones: http://www.diyaudio.com/forums/diyau...udspeaker.html

But a single bass would be rather dull, wouldn't it? Not D'Appolito at all. :D

Mate, I have no patent on these circuits. Feel free to use them.


All times are GMT. The time now is 11:49 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2