Noob building his first speaker

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
re:"jaycar has a diy crossover building table" - NOOOOOO!!!! these tables are based on the nominal impedance of a driver, which is nothing like the ACTUAL impedance of the driver at the crossover frequency you're using. To find that, you need an impedance graph of the driver, (I'll see if I can dig mine out for that driver, may take a couple of days), then use that value in a crossover calculator such as this:
Crossover Design Calculators - that'll get you a much better crossover
To work out tweeter attenuation use this: L pad calculator - attenuation dB damping impedance decibel loudspeaker speaker voltage divider - sengpielaudio Sengpiel Berlin

haha ok thanks, I suspected it would be only very basic. Jaycar actually has the impedance graph for both the 8" and 5" (assuming i build a 3-way speaker). Assuming ive understood everybody correctly, Im just reading off here at the point of crossover (300-400Hz for woofer) and about 5kHz for the mid and use something like this to calculate. If this is the case, could someone explain bandpass gain and which type of crossover would be chosen?

I have been thinking about this on your behalf (when I should have been plastering and painting LOL) perhaps you could return the tweeter and use the money to purchase a halfway decent 3 inch full range. Then all you need is one crossover in the range 300-400 or so.


3 inch drivers have a reasonable top end and the woofer you have will fill in the bottom quite well.

http://www.fostexinternational.com/docs/speaker_components/pdf/fe87.pdf

eBay Australia: Buy new & used fashion, electronics & home d?r

There are many other good cheapish 3 and 4 inch fullrange drivers around
I see lots of cheap computer speakers at our local recycle centre going for $5 or $10- a pair and some of them have quite reasonable little drivers in them

That would definitely be something to consider, thanks

Building a pretty decent sounding speaker system isn't too hard, with a little trial and error experimenting. Building a really good one is much more complicated. Internal cabinet shape and damping, and external listening room acoustics have a huge influence on what you will hear in the end. After reading the advice others have given you, I noticed several cases where the advice is actually quite complicated (Get ARTA and use to measure impedance, for ex.).

For a newbie, I'd actually recommend a closed box woofer, perhaps with a passive radiator made from a regular woofer so you can short the voice coil out for better damping if that works better. I've had great luck with that. It pumps up the low end in the least complicated way. For midrange, I'm thrilled with the $12 Peerless 3 inch drivers I've been playing with lately (check Parts Express or Madisound websites here in the US). They're great from about 200HZ to 12kHZ+. I'd put in a 3/4 inch dome tweeter crossed at 7kHZ or higher for the top end.

It's true that nominal impedance of any driver and actual impedance at the frequency you want to cross it over at will be way different. That's one of the reasons to keep a passive crossover simple. If you can figure out how to measure the impedance at these frequencies, you can be more accurate. For a newbie who can't find a published impedance graph, I might suggest assuming the driver has about twice the nominal impedance at the frequency you want to cross it over at (based on my experience), build a crossover based on that, and then vary the values if necessary until it sounds the best to your ear. Many people out there will condemn me for saying that, but in the real world, that may get you closer to accurate than any other approach, if you're a newbie. And yea, going from an 8 inch to a 1 inch tweeter will have pretty substantial off axis problems. Difficult to listen to over time. Too much upper midrange. There is a lot to be said for having good tone controls in you preamp. Those who hate tone controls are fools, or have had a very bad version of them. Hope this helps.

Check out my audio hobby website if interested. It's pretty advanced, but perhaps relatively understandable to a newbie. My "Aurium Waveguide" speaker project has a lot of discussion that you could learn from. Bob's Website
Bob's Website

Thanks for understanding that yes, I am a complete noob haha! You explained pretty well how to go about designing the crossover so hopefully I understood it right when I posted the graphs for the drivers i'll probably use just above.

Would I be right to use the calculator from above and enter something like 9ohms for the woofer with 350Hz crossover, 15ohms for the mid with 5000Hz crossover and just double the tweeter like you said (16ohms)? Have I understood this right? Im I even on the right track? haha

thanks
 
diyAudio Member
Joined 2007
Seeing as how Pete gave me the woofers and I need to pay it forward, would you like a pair of reasonable 3 inch full range drivers?
I have several sets here and I cannot use them all.
Shielded LG and quite reasonable
You would need to put what is called a Zobel on the woofer ( this is just a small capacitor and a resistor in parallel ) and a second order XO on the drivers around 300Hz. Then take back the tweeters and use the money to buy the XO components from Speakerbug

You would use what is termed a Linkwitz-Riley second order crossover and honestly I think you could treat this as a "Near enough-good enough" design treated as a learning process
 
Last edited:
diyAudio Member
Joined 2007
Not selling these, yours if you want them.
it's called pay it forward for a reason, if you keep going with speaker building and develop some skills; you help others.
They cost me nothing as people keep giving me stuff and I have far too many projects half finished at the moment.
These will need a very small inner box of about 1/2 litre MAX, even half that will work and I do not have the parameters for these but Fs is somewhere around 120Hz so second order at 300 is as low as you can go with the XO
 
So I had a go at doing a 3-way crossover with the advice of using all frequencies at read from impedance graphs for the drivers I chose (jaycar for now as a project for a bit of practice) at the crossover frequencies I chose. These are some numbers

Jaycar 25mm tweeter (no impedance graph, assumed twice nominal impedance as suggested by bob)
*high pass*
Impedance: 16 ohms
crossover: 5000Hz

Jaycar 5" woofer (mid) (bandpass)
*Low pass*
Impedance: 16 ohms
crossover: 5000Hz

*high pass*
Impedance: 7.5 ohms
crossover: 400Hz

Jaycar 8" woofer
*low pass*
Impedance: 9.5 ohms
crossover: 400Hz

***all impedance values used in online calculators where read from available impedance graphs supplied by jaycar, read at the crossover frequencies chosen***

I've run all the calculations under these assumptions and have a circuit diagram with all capacitors and inductors for a 2nd order crossover.

Is this the right approach so far? Assuming it is I also did calculations for l-pads for the tweeter and woofer (bring everything to 88dB) and and notch filter for the tweeter for 1300Hz (free air resonance from the jaycar tweeter specs)

Could somebody tell me if i've calculated this correctly and properly understood and interpreted the process to building a 3-way crossover? Or have I just made a bunch of mistakes?
 
Hi,

Just to throw in my 2p, I think it'd be more educational to pull apart and experiment with an existing design before trying to do a 3-way from scratch (they're difficult, at best).

For that, I'd recommend getting a pair of Mission 760i speakers. You can pick them up for £20-30 over here. If they're available to you, I'd say get those and work on turning them 3-way with the 8" woofer, and building a better cabinet.

The crossover in there is fairly simple, so it should be easy enough to understand what its doing, and play around with it if you so wish.

The important bit here is that you effective set off with a known-good, and can experiment from there. Throwing a semi-random set of drivers and crossover points together will give results that, while operational, will be far from optimal.

Chris
 
diyAudio Member
Joined 2007
As a general rule speakers sound better when the midrange crosses as low as possible and has the widest spread and where the drivers cross over before the wavelength of the sound being reproduced is smaller than the diameter of the driver

The tweeter may not need a resonant peak filter if using a second order XO at 5k but 5k is fairly high for a 5inch midrange.
As well as the cross-over points you need to consider the electrical order of the cross-over and those jaycar drivers will probably need second order crossovers as they do have some HF stuff that needs to be excluded especially the woofer.
All in all I do still think a woofer plus full range is a viable beginners project.
If you want to make a 3-Way then just do it, the experience will be valuable
 
So I had a go at doing a 3-way crossover with the advice of using all frequencies at read from impedance graphs for the drivers I chose (jaycar for now as a project for a bit of practice) at the crossover frequencies I chose. These are some numbers

Jaycar 25mm tweeter (no impedance graph, assumed twice nominal impedance as suggested by bob)
*high pass*
Impedance: 16 ohms
crossover: 5000Hz

Jaycar 5" woofer (mid) (bandpass)
*Low pass*
Impedance: 16 ohms
crossover: 5000Hz

*high pass*
Impedance: 7.5 ohms
crossover: 400Hz

Jaycar 8" woofer
*low pass*
Impedance: 9.5 ohms
crossover: 400Hz

***all impedance values used in online calculators where read from available impedance graphs supplied by jaycar, read at the crossover frequencies chosen***

I've run all the calculations under these assumptions and have a circuit diagram with all capacitors and inductors for a 2nd order crossover.

Is this the right approach so far? Assuming it is I also did calculations for l-pads for the tweeter and woofer (bring everything to 88dB) and and notch filter for the tweeter for 1300Hz (free air resonance from the jaycar tweeter specs)

Could somebody tell me if i've calculated this correctly and properly understood and interpreted the process to building a 3-way crossover? Or have I just made a bunch of mistakes?
I give the following advice having not seen any specs on the drivers involved. All I know is their approximate sizes and nominal imedance ratings.

I like that the tweeter will be highpassed at 5kHZ. It's a rare 1 inch tweeter that sounds good much below that (with a 1st order xover), regardless of what a graph might seem to imply. Forget any notch filter on the tweeter. I wouldn't go 2nd order on the xover. It's hard enough to get very accurate with a 1st order in the real world. I'd only put an L-pad on the tweeter. I'd low pass the five inch at 1 or 2kHZ initially, so the harsh, hard to listen to, uppermid frequencies will be reduced (attenuated) a bit in that range. Also because a 5 inch gets more directional than you want above about 1-2kHZ. Using a 1st order crossover will keep this uppermid valley from getting too deep, if you get the phase differential right (try both ways, reversing the wires at the tweeter, and pick the one that sounds louder in the upper mid freqs). I'd take the 5 inch down to 200 - 300HZ before hi-passing it, again with 1st order. I'd let the woofer have more SPL (which it probably will inherently) with it's low pass at around 100-300HZ, so it can give you a warmer sound overall. I'd align all the drivers vertically, especially the 5 inch and the tweeter. When you put an 8 ohm L-pad into the mix, (not a bad idea), I would calculate for 12 ohms, rather than 16 ohms, with a nominal 8 ohm tweeter. The L-pad will average things out a bit.

The formula for a 1st order filter for the cap is:
impedance as seen by the cap(Xc) = 1/2 pie f C(in farads), or transposed;
C (in farads) = 1 / (6.28 x freq x Xc)

The formula for the coil in a 1st order xover is:
L (in henries) = XL / 6.28 x freq
Where XL is the impedance as seen by the inductor.
 
As a general rule speakers sound better when the midrange crosses as low as possible and has the widest spread and where the drivers cross over before the wavelength of the sound being reproduced is smaller than the diameter of the driver

The tweeter may not need a resonant peak filter if using a second order XO at 5k but 5k is fairly high for a 5inch midrange.
As well as the cross-over points you need to consider the electrical order of the cross-over and those jaycar drivers will probably need second order crossovers as they do have some HF stuff that needs to be excluded especially the woofer.
All in all I do still think a woofer plus full range is a viable beginners project.
If you want to make a 3-Way then just do it, the experience will be valuable

Yeah, im still considering returning the tweeter and getting a full range like you suggested earlier...

I give the following advice having not seen any specs on the drivers involved. All I know is their approximate sizes and nominal imedance ratings.

I like that the tweeter will be highpassed at 5kHZ. It's a rare 1 inch tweeter that sounds good much below that (with a 1st order xover), regardless of what a graph might seem to imply. Forget any notch filter on the tweeter. I wouldn't go 2nd order on the xover. It's hard enough to get very accurate with a 1st order in the real world. I'd only put an L-pad on the tweeter. I'd low pass the five inch at 1 or 2kHZ initially, so the harsh, hard to listen to, uppermid frequencies will be reduced (attenuated) a bit in that range. Also because a 5 inch gets more directional than you want above about 1-2kHZ. Using a 1st order crossover will keep this uppermid valley from getting too deep, if you get the phase differential right (try both ways, reversing the wires at the tweeter, and pick the one that sounds louder in the upper mid freqs). I'd take the 5 inch down to 200 - 300HZ before hi-passing it, again with 1st order. I'd let the woofer have more SPL (which it probably will inherently) with it's low pass at around 100-300HZ, so it can give you a warmer sound overall. I'd align all the drivers vertically, especially the 5 inch and the tweeter. When you put an 8 ohm L-pad into the mix, (not a bad idea), I would calculate for 12 ohms, rather than 16 ohms, with a nominal 8 ohm tweeter. The L-pad will average things out a bit.

The formula for a 1st order filter for the cap is:
impedance as seen by the cap(Xc) = 1/2 pie f C(in farads), or transposed;
C (in farads) = 1 / (6.28 x freq x Xc)

The formula for the coil in a 1st order xover is:
L (in henries) = XL / 6.28 x freq
Where XL is the impedance as seen by the inductor.

thanks! so youd suggest have a first order crossover for the tweeter/mid (while dropping the crossover point of the mid to 2kHz) with an l-pad on the tweeter and drop the mid/woofer crossover to 200-300Hz (maybe keep this a second order?)

The new crossover would be the tweeter at 5000Hz (first order)
mid to cross at 2000Hz (first order) and 300Hz (second order?)
and woofer at 300Hz (second order)

*using your formulas*
So the capacitor for the tweeter approx= 1.99uF
and the inductor approx= 0.9mH
Where the l=pad (using 8ohms) = 1.65 and 30.9 ohm resistors (for 2dB)


my numbers for the high pass on the mid and low on the woofer would be the same. Looking better?


Thanks for your critique!
 
Last edited:
Looking better?

The crossover calculation method proposed by Bob is very basic and does not consider actual driver frequency response and baffle step. A simulation based on manufacturer data predicts the following result:

SimulationDijaam.jpg

CrossoverDijaam.jpg

The simulation is not meant to be accurate at all, as driver data are insufficient and questionable. But it gives an impression, what the outcome may be, and it shows the very limited effect of first order filters regarding attenuation of unwanted resonances. Judge yourself if you really want to build that speaker.
 
Yeah, im still considering returning the tweeter and getting a full range like you suggested earlier...



thanks! so youd suggest have a first order crossover for the tweeter/mid (while dropping the crossover point of the mid to 2kHz) with an l-pad on the tweeter and drop the mid/woofer crossover to 200-300Hz (maybe keep this a second order?)

The new crossover would be the tweeter at 5000Hz (first order)
mid to cross at 2000Hz (first order) and 300Hz (second order?)
and woofer at 300Hz (second order)

*using your formulas*
So the capacitor for the tweeter approx= 1.99uF
and the inductor approx= 0.9mH
Where the l=pad (using 8ohms) = 1.65 and 30.9 ohm resistors (for 2dB)


my numbers for the high pass on the mid and low on the woofer would be the same. Looking better?


Thanks for your critique!
My formulas are for 1st order only. Although a 2nd order between the woofer and mid is a great idea, I don't recommend it for a newbie since it's hard to get right without the ability to measure impedance of drivers at the frequencies in question. For the L-Pad, I was recommending a variable pot version of an L-Pad, not just two fixed resistors. That way you could adjust the tweeter level without throwing off the crossover function dramatically, as you would with a simple variable resistor.

Again, I haven't seen the specs on these drivers, and apparently they aren't very complete, and are therefore questionable all around IMO. When you consider the amount of time it takes to build speakers, it just doesn't make sense to use questionable drivers. Giving a newbie poor quality drivers for his first project is like giving a 15 yr old a really bad first guitar; it will likely teach him to never want to play guitar ever again.

Peerless and Dayton drivers are cheap and very good. I can recommend specific drivers if you'd like. If you can't do a project well... sometimes you're better off not to bother at all. Any advice we try to give you may not actually make it better in the end. Dealing with Baffle Step is a bit more advanced and not something a newbie is likely to get right. A Bass control on the preamp can largely make up for that.
 
Slightly off subject, the next level: Bi-Amping, and Why?

If you were to put an active crossover using opamps ahead of power amps for the crossover at 300HZ, you would have very good accuracy, very sharp cutoff rates (4th order), completely non-interactive level adjustments, and once you've got a chassis and power supply, active woofer EQ can be added with not much more trouble that will make a closed box (only) woofer be acoustically flat to practically any low end frequency you might want (20HZ, 30HZ etc.). Getting the large amount of cone movement required for good bass away from the midrange frequencies is one of the best improvements you can make in a speaker system. It causes audible doppler effect distortion (frequency modulation of the higher frequency energies). If heavy bass causes an amp to clip (overdrive), and the same amp is handling the whole frequency response, then every thing is horribly distorted.

My tri-amp'd system measures acoustically flat from 20HZ to 20kHZ +/- 2-3 dB at the couch (listening location), and it sounds really dam good, especially when kettle drums or bass drums are in the program source.

These days many video/movie soundtracks have lots of special effect energy below 40HZ, and it's quite a thrill if your system can handle it well. Circuits for active crossovers and woofer EQ can be grabbed off the web and scaled pretty easily, verified with a free version of the SPICE program (why re-invent the wheel - the actual math is very complex). If you decide down the road to take it up a notch, this may be what you'll be looking at. It seems like too big a project until you've seen what a time consuming pain in the rump it can be to get a passive crossover right.
 
Last edited:
Bob that's a pretty big ask for a first project.
Mind you I agree that Bi-Amping is the best way to go for a cost effective solution below 120Hz or thereabouts
That's why I started with "slightly off subject". I'm not recommending it at this point. Just wanted to give you a quick nutshell view of the what and why, perhaps for down the road. A simple passive speaker system is often a "gateway drug"... Pretty soon you're doing the hard stuff... :(
 
My formulas are for 1st order only. Although a 2nd order between the woofer and mid is a great idea, I don't recommend it for a newbie since it's hard to get right without the ability to measure impedance of drivers at the frequencies in question. For the L-Pad, I was recommending a variable pot version of an L-Pad, not just two fixed resistors. That way you could adjust the tweeter level without throwing off the crossover function dramatically, as you would with a simple variable resistor.

Again, I haven't seen the specs on these drivers, and apparently they aren't very complete, and are therefore questionable all around IMO. When you consider the amount of time it takes to build speakers, it just doesn't make sense to use questionable drivers. Giving a newbie poor quality drivers for his first project is like giving a 15 yr old a really bad first guitar; it will likely teach him to never want to play guitar ever again.

Peerless and Dayton drivers are cheap and very good. I can recommend specific drivers if you'd like. If you can't do a project well... sometimes you're better off not to bother at all. Any advice we try to give you may not actually make it better in the end. Dealing with Baffle Step is a bit more advanced and not something a newbie is likely to get right. A Bass control on the preamp can largely make up for that.

Fair enough. Although to be honest I think id learn more building something to begin with and then be able to come back here and ask how I can improve certain parts after listening as opposed to just getting it right the first time or copying another design and never have to worry about it.. although the learning process might be different in speaker design, I dont know. it might just sound horrible and only be slightly better after fixing it ha!


Bob that's a pretty big ask for a first project.
Mind you I agree that Bi-Amping is the best way to go for a cost effective solution below 120Hz or thereabouts

Yeah im having enough trouble as it is haha..

Also, my amp has a 5-band eq on it so couldnt i just attempt a build as accurate as I can (using software with the specs im given) and just eq it to my room or come back here for suggestions to fix it? I think everybody is also overestimating how 'good' the finished product needs to be lol, Im building for a bit of fun mostly..

thanks guys
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.