Anyone build DIY "mains" not from a kit? (1 Viewer)

[Seth_L]

Maybe I've just missed it, but has anyone built a pair/set of DIY speakers from scratch? What I mean not using plans on the web, or buying a kit from someone else. Something more along the lines of picking your own tweeter, crossover, mid, and woofer (or maybe a mid/woofer) and seeing what happens.
I've built a few DIY subs and plan on building a monster sub around a Tumult, but I'm interested in potentially building something DIY around a Vifa XT tweeter ( http://www.partsexpress.com/pdf/264-555.pdf ). Is it best to stay within a the same manufacturers line? (like all Vifa components) or can I go hog wild mixing and matching anything I want? Or is this entirely too complicated/risky to try building a set of DIY speakers from scratch?
Seth

 

[Patrick Sun]

I did. The gory details are in the DIY project link below.

 

[Seth_L]

I gather from the description you have that without LEAP and the ability to make detailed response measurements there's no point in even trying to design my own speakers.

Seth

 

[Patrick Sun]

Until you have an infinite amount of patience and access to lots of different values for inductors/capacitors/resistors, or go the active crossover route, yeah, it's a tough job to get it right.

But if you choose to make the initial investment in a measurement system (doesn't have to cost an arm and a leg) and the rather inexpensive LSPCad (or just play around with the free Speaker Workshop software, which is what I did with my current center channel), you open yourself up to a larger world of drivers and speaker possibilities.

 

[Mark_E_Smith]

I gather from the description you have that without LEAP and the ability to make detailed response measurements there's no point in even trying to design my own speakers.

I disagree. It makes life easier, and you might HIGHLY consider some crossover modeling software. If you match spls and dont get into a 3 way design you can have good results. I built my own center channel, then I bought a clio system and found out I came really close, so close I may not change anything. Pretty good for first time! One big tip is to use the more expensive drivers 1. that you dont have to do alot of special filters and 2. the frequency responses overlap. Look at the response graphs not the range. Often the driver tends to break up at the ends of the response range. Example a mid-woofer that has a response from 42Hz to 5kHz you would cross at 2kHz because at 5kHz it has a nasty rise that would require a notch or impedance equalization circuit. Also look for tweeters that have a flat and smooth response out to 20kHz. These tend to be the ferrofluid enhanced ones. Get some books on speaker building and read them first. David Weems' and Vance Dickason's are both great, with Weems' being more "hands on" and basic, Dickason's more theoretical.

 

[VinhT]

I built DIY satellite speakers and a PR sub of my own design from scratch.
The sub was relatively easy to design, build, and test.
The satellites on the other hand (5 identical two-ways) were much more difficult. I researched everything thoroughly, but the lack of sophisticated testing equipment was a great hinderance. I used the Dayton 5-1/4" woofer and a cheap $15 dome tweeter from Parts Express. My first crossover was designed based on their given specifications. Well, a peak in the roll-off of the woofer combined with the sibilance of the tweeter resulted in an uncomfortable amount of sibilance. So I went back and obtained rudimentary response measurements of the drivers using Speaker Workshop and the Radio Shack SPL meter(yes, I know this is not a very accurate method even with correction, but it's better than nothing). I then made new crossovers and tested the speakers. Are they completely flat? Of course not. You have to be pretty elite to build flat speakers from scratch. Me, I was just a college student eager to design and build more speakers after my success with my sub. But in the end, the sibilance is gone and I am very satisfied with my little DIY speakers and won't be replacing them for a long time (or until I purchase a real mic and discover their true response ). It was definitely a great learning experience as well.

 

[Seth_L]

From what I can tell LSPcad doesn't have anything beyond Unibox 4.00 for box design. When I try modeling some of the various mid/bass drivers out there they seem to push past their cone travel limits at 100Hz in reasonable configurations. Is Unibox not valid for modelling small cones, or is there something else at play, or should I not be worried about cone travel?

Seth

 

[Bryan_G]

I built a set of mains using the vifa xt tweeter. I did however fallow a plan on the internet, but I ended up changing most of it around. I am now getting ready to start designing a center channel from scratch to match. My mains are too big to work as a cc so I have to change something . If you will have to order your all your crossover parts, it will be extreemly hard to find the right values the first time around. Unless you have a lot of extra money or a nice store close by I might try to look for a plan and maybe modify it a little to fit your exact needs. If you are going to spend $500 + on a pair of speakers then you want to know that they are going to be the best they can be.

 

[Mark_E_Smith]

Cone travel is determined by amp power input, lower the power and the cone travel will reduce. Most of the programs use the max power of the driver as the power input. Unibox is a good program to start with for box design. I use it and Bass box/X-over Pro.

 

[Seth_L]

About Unibox...

For example if I take 2 of the Vifa XT18WH09-08 7" WOOD PULP WOOFER in series and put them in a 2cu-ft (56.6L) enclosure with a Fb of 45.69Hz (No leaks, "walls covered" fill) the simulation shows that the woofer is done at 90Hz when it runs out of travel with 100W. I've tried other mid/woofers and they all seem to have similar problems.

I've seen towers with 2 lesser 7" drivers that quote number down into the high 30s for frequency response. How is my design any different?

Am I just missing something?

Seth

 

[Mark_E_Smith]

Yes, thats over 110 dB spl at 90 Hz pretty tough for any 7" driver esp with only 5 mm of Xmax. But are you running a subwoofer? will you be expecting the mains to go below 100 Hz?

I've seen towers with 2 lesser 7" drivers that quote number down into the high 30s for frequency response. How is my design any different?

Speaker manufacturers quote ranges at 1 watt or 2.83 v NOT 100 watts

 

[Dan Wesnor]

First, designing a speaker even with software and measuring equipment is not apple pie. Sometimes moving a crossover point by 300Hz can change a speaker from boring to excellent. It generally takes me a month from start to finish with an XO before I'm happy with it. The main problem is nailing the baffle step compensation. It's impossible to directly measure without a chamber, and mess it up and the speaker sounds either too thin or too heavy.

Second, the Vifa XT is a tricky little bastard and not for beginners. It must be crossed at 3kHz or higher or it will be eaten up with distortion. There aren't a lot of woofers that will do 3kHz without distortion problems of their own.

Third, there is absolutely no reason to get 110dB out of a speaker at any frequency (HT subs excluded). Even if you do, VC heating and thermal drift in XO components will cause distortion. Excursion limited output of 100dB/1m @ 80Hz is good goal for a 6.5" driver.

If you really want to do this, I suggest you start with some easy drivers which are smooth through their pass-band with a controlled roll-off. The Vifa P17 woofer is a good choice, as are the Morel tweeters.

 

[Chris Tsutsui]

I think that if you have the time and money, then go scratch.

I am on a budget and don't have a lot of time so I went kit. I also don't have "the tools" or programs that would help with a speaker design.

When you buy a kit, it's just about following directions while a design takes some thinking and research.

 

[Mark_E_Smith]

So SethL maybe we should ask you a question? Why do you want to design your own. Is it for the thrill of accomplishment, pride of your OWN creation, a learning process? Damn the torpedoes full speed ahead!! I guess no body built DIY speakers before computers and testing equipment. :frowning: I had and am having allot of fun building my own designs. Yes I could have built a kit but what would of I learned (I already know how to cut wood or is mdf sawdust anyway) and would they really be mine? I know allot of you have experience and dont want to see people make mistakes. So lets help when we can. Dan is probably right about the drivers. Buy the books first, read them and then plan it out. Sometimes perfection is a goal not to be accomplished without the experience. :wink:

 

[Patrick Sun]

Actually, I've read about one DIY speaker design that crosses the Vifa XT25 around 1.2KHz (but the XO is ungodly - an acoustic 8th order LR slope roll-off). ThomasW/JonMarsh would be the guys to talk to about that design.

 

[Seth_L]

Exactly how accurate of frequency response measurements will I need to properly design the crossover? Will measurements taken from a Behringer Ultra-Curve® Pro DSP8024 ( http://www.behringer.com/02_products...P8024&lang=eng ) with it's matching measurement mic ECM8000 ( http://www.behringer.com/02_products...M8000&lang=eng ) be sufficiently accurate, or will this require something that's even more accurate?
That setup can be connected to a PC and the EQ curves can be edited and viewed on the PC. The Mic is really flat (see the PDF if you're curious) and the EQ can measure/set .5dB steps. Will this be good enough or do I need to buy something more accurate? (I already have both the mic and the EQ).
Also, not that this really means a whole lot, but I am an Electrical Engineer and I even took a class in analog filters in college (2 years ago). Maybe I can put the text book to some use with this project.
Seth

 

[Patrick Sun]

If you're an EE, most of this stuff should not be a big deal when talking about target slope roll-offs, target crossover points, etc. Just be prepared to read and grok the concepts and the pitfalls when designing a speaker (figuring out which drivers would work in a certain sized box, etc).

The mic is a good choice, and if the PEQ will allow you to power up the mic, and pass the mic data to something like JustMLS that comes free with the $129 version of LSPCad, you should be in business.

 

[Seth_L]

If you're an EE, most of this stuff should not be a big deal when talking about target slope roll-offs, target crossover points, etc. Just be prepared to read and grok the concepts and the pitfalls when designing a speaker (figuring out which drivers would work in a certain sized box, etc).

Personally I like 10th order crossovers with 60dB per octave slopes and a almost flat phase response. I'm a little curious why no one ever attempts DIY 3 ways? What is the explaination behind the avoidance of a 3 way system? Or are they just extremely hard to design/build?

Seth

 

[Dan Wesnor]

Patrick,

Someone over at Madisound posted frequency vs. distortion for the XT25, and it went through the roof below 3kHz or so. It's a pretty fresh post, so you should be able to find it on the main page.

Seth,

Because it's difficult to get accurate measurements below 300Hz, and because 3-ways often cross that low or lower, it's difficult to make a good 3-way. I will begin working on one soon, however. I have done one before, and it's not very difficult. In fact, it's easier than working with poorly matched drivers in a 2-way.

I think the reason most avoid 3-ways is not because of difficulty, but just because they're not really popular at the moment, and not really necessary given that nearly everyone owns a subwoofer.

 

[Seth_L]

Another question:

Why do more people not build sealed designs? It seems that the superior impulse response would be desireable. If most people already have subs, what's the point in going ported if you don't really need the extra low end extension?

Seth