Help me make my first project.

[milsman2]

So my other thread has reached a point that I kind of hoped it would when I began. It seems like now is as appropriate time as any to give some context and constraints to this project given that this has turned into a full on build thread. So let me give some background. I am currently a Mechanical Engineer at Duke University and this summer I had a really fantastic job where I worked with a rather large amount of very skilled electrical engineers. The experience really lit the fire beneath my electrical side of engineering and this sort of transformed to me finding a real passion in DIY audio.

Budget:
Total at the moment is probably about $500 but I'm willing to spend a relatively sizeable amount per month to continue building the project (in the $100 range)

Experience:
I'm going to be a senior this coming year so I have taken plenty of classes that are principle intensive but this year will be a lot of design. I have also taken a few electrical classes as no motor runs without power and power is relatively useless unless applied to something else. I am knowledgeable about most of the basic principles involved in both the electrical part of speaker design (e.g. filters, op amps, power, grounding,) as well as mechanical (e.g. doubling the resonant frequency, what a cut off means for the speaker, woofer vs. tweeter, etc.).

Resources:
As I said earlier, I will be at university so I will have many of the machine shops as well as professors and engineering friends who will want to help out possibly and pitch in for the greater good. I personally do not have any software currently that is all too helpful in terms of speaker or electrical design but I can have access to basic stuff (i.e. AutoCAD, SolidWorks, MAPLE, MATLAB). If a piece of software costs money and diyAudio seems to think it is truly worth it, I will most likely buy it for the sake of saving myself from future headaches. I have also ordered Loudspeaker Design 201 and Loudspeaker Recipes Vol. 1. Laying around are a couple of Forest Mims books, which is where I got started this summer with breadboarding and the sort. I have so far bought myself a pretty decent adjustable Weller soldering station as well as a few miscallaneous supplies like wire strippers, nippers, the basics. I can solder and do a few tooling tasks without an issue so at this point my biggest challenge is the design and ordering of all the parts rather than the actual construction.

I know this is sort of an uphill battle and I AM WILLING TO USE KITS FOR THE ACTUAL DRIVER AND BOX, but this is only because I have had my fill for mechanical design for now and my ambition has turned mostly to doing my own electrical work and design. I am aware of all the things that can be bought assembled or pre designed but this has become a labor of love and I feel like the learning experience is the biggest thing I will get out of this as clearly for $500 I can buy myself a pretty decent set of speakers but hopefully I will learn much and gain the ability to build and design some mean speakers as well as end up with my own set built by me. I really don't want to order a PartsExpress kit, assemble it passive crossover and all included, plug it in and call it a day.

Requirements:
1.) 2 way bookshelf speaker.
2.) Drivers to be actively crossed over.
3.) All drivers powered by kits from ChipAmp.com as the site is extremely well reviewed, provides a pretty good balance of DIY and things already thought out and tested.

I do not have any test equipment but again for recommendations or actual work for me ( please?) I turn to diyAudio.

At this point all my time has been spent getting myself together in terms of what I want and gethering together the most basic of supplies to kick off such a project and I feel like I have a pretty solid foundation albeit a pretty small one to do what I want. Where do I go from here? Thanks for all the input. I suppose the first step is picking the speaker kit or drivers. Help?

I am currently thinking about following this plan but with a lower powered tweeter and like I said doing my own active crossover as well as zobel network and baffle step correction as needed.

Parts Express DIY Project

[DrDyna]

If you're planning on going active, using something like a MiniDSP will allow you to BSC and EQ functions actively and might be easier than a good passive crossover, as well as give you another electronic gidget to play with and tune.

As for a speaker kit, if you're going active, you could realistically select whichever drivers / kit you liked. I'm going to go out on a limb and say a HUGE percentage of the "ease of a kit" is the passive crossover, which you'll overcome easily if you're going to go active.

I'd say, select a kit you like based on performance / price / looks, grab a MiniDSP kit, which you select will depend on how many drivers are in your design, replicate the crossover point with the active crossover in the miniDSP and once it's playing, tinker with it until it sounds pleasing.

Paul Carmody has lots of good 2 way designs that come in well under $500 that you could snag and actively cross over.

[PeteMcK]

have a wander round this site:
Zaph|Audio

[milsman2]

When I said active I actually meant like implement my own filters via op amp configurations, but I appreciate the tip anyway.

I am currently looking at the classix by Paul Carmondy as a kit that I could actively crossover, but my only issue is finding the correct crossover point and slope. Assuming I could figure that out, would this be a good build to bi amp with op amp filters and power with LM 3886 gainclones?

[Loren42]

Originally Posted by milsman2

When I said active I actually meant like implement my own filters via op amp configurations, but I appreciate the tip anyway.

I am currently looking at the classix by Paul Carmondy as a kit that I could actively crossover, but my only issue is finding the correct crossover point and slope. Assuming I could figure that out, would this be a good build to bi amp with op amp filters and power with LM 3886 gainclones?

You are much better off going with something that is easily reconfigurable like DSP. There is so much tweaking that you will need to do to dial in the system that it will get downright painful using discrete components (either before the amps or after the amps.

[DrDyna]

Paul has 2 different crossover points for the classix that he speaks about, I think the second one came about largely because of the type of music that the speaker was being designed around.

So, you're planning on designing your own active circuitry? Very nice That's a big plus for the MiniDSP stuff, is that it's adjustable, however designing your own circuit may result in a higher quality end result if you decide which version of the Classix XO you're going to try and emulate. The classix "relaxed" version is probably the one I'd go for, as I listen to a wide variety of stuff, from classical, new jazz and Flecktones, right up to heavy metal like Buckethead.

The "relaxed" version seems to be crossed over around 3.3 or 3.4khz, I'm not sure what slope, but from the looks of the crossover I'm going to guess 24db electrical.

[DrDyna]

Originally Posted by Loren42

You are much better off going with something that is easily reconfigurable like DSP. There is so much tweaking that you will need to do to dial in the system that it will get downright painful using discrete components (either before the amps or after the amps.

Agree 100%.

[milsman2]

Now after looking at what a DSP does, this makes a lot more sense. Like I said, I really want to dive into the actual electronics of this, but relative to soldering and resoldering and going through that whole headache, I'm probably much better off tweaking a tool that is already made for my purpose than going through the headache of basically what would amount to making my own DSP. Do they do zobel networks and baffle step corrections as well? Also, can they take analog input e.g. 3.5mm cable to prepared lines or terminal? I am thinking the Classix with the "relaxed" version would be perfect for a college apartment smoky room setting.

[Loren42]

Originally Posted by milsman2

Now after looking at what a DSP does, this makes a lot more sense. Like I said, I really want to dive into the actual electronics of this, but relative to soldering and resoldering and going through that whole headache, I'm probably much better off tweaking a tool that is already made for my purpose than going through the headache of basically what would amount to making my own DSP. Do they do zobel networks and baffle step corrections as well? Also, can they take analog input e.g. 3.5mm cable to prepared lines or terminal? I am thinking the Classix with the "relaxed" version would be perfect for a college apartment smoky room setting.

Zobel is something that DSP does not do, but you should not need it. The primary reason for the Zobel is to keep the impedance flat so the passive crossover network works in a predictable way.

Since you have isolated that filter before the power amplifier, the impedance at the driver is not important.

Baffle step correction is easily performed using DSP.

[Arty]

I'm not that sure a first time project needs to be that advanced, but i might just be wrong.
Actualy if i was you i might just first try to build some low power amplifier modules.
That alone is a nice task.
MiniDSP indeed has it good use, but it might not be needed. Sure, its nice to play with it, but there are software that can turn a soundcard in your computer to act as a minidsp. And when you are happy with the result you can make a standalone device, so no need for the pc anymore.

Actualy i would highly recommend first building 4 amplifiers, a proper psu, and finding drivers to make an acceptible 2 way bookshelf speaker. Sounds easy, but this alone might cost you a half year or so.

On the question of what to avoid..
High FS tweeters are one. (prefer near or less than 1 Khz)
larger than 6" midbass (prefer something 4-5.5" or so, less problems when crossing them)
metallic cone midbass (has nasty breakup modes, prefer paper, or prehaps PP)

as for amplifier..
well i would avoid high priced high fidelity ultra super-duper mega cool stuff.
I would aim for something like tda1557Q or similar. The simpler the better, most probably you will not need more than 15 watt at most. I would say even 10 watt will be more than enough.
Later you can change to better gear if you like, but if you fry -or worse, can't even make a complicated kit work at all- you will be dissapointed.

Minidsp is handy, but.. there are software to act like it. I would consider to first just use that, and see if you need all of its functions at all. It can happen that eqvivalent circuitry will not consist more than a few passive components added to the power amps, so it woud be a waste of money. (sure, can be handy later on, but who knows the future for sure..)

I would avoid hunting for quality bass from bookshelf speakers.
My design goal would be decent crossing, level matching, and more or less smooth range from 100hz and upwards. Beuty of bookshelfs is that you can later on extend it downawards with a properly built and hidden subwoofer.

I think its better to set goals that are easy, and then set up harder goals then to set up a goal that might be just too hard for a first project.