First speaker build 2 way micro bookshelves - need crossover help

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm working on some 2 way powered speakers; making a bluetooth music player basically. I'm using a 5/8" tweeter and 3.5" full range woofer from parts express. The speaker boxes are micro sized, .04cu ft, so I'm using a first order 6db slope, since I don't think I have room to fit a second order of higher in the boxes. My tweeter sensitivity is around 89db, with my woofer around 83db. I have the exact specs of the speakers if they are needed. The woofer is a 4 ohm woofer, and the tweeter is a 6ohm. The amp requires 4ohms+.

I have a 50ohm resistor wired directly to the positive terminal on the tweeter. This is for the tweeter level, 30 ohms was still way too loud. On the other side of the tweeter resistor, wired directly to the resistor, aka: the amp side of this connection, I have the crossover, a 0.33uF capacitor. Is this correct? When figuring the crossover frequencies, I entered in that the tweeter has a total resistance of 56 ohms. Or should the capacitor and resistor be flipped around? Are they crossing where I calculated it to cross?

I have the same wiring setup with my woofer also, along with a zobel network on it. I'm just wondering, if when figuring crossover frequencies, when it asks for speaker impedance, is that ACTUAL speaker impedance that it wants? Or is it asking for total impedance between the crossover and the speaker, including the speaker (or any resistors added to adjust the level)?

Thanks a bunch guys, I'm super new to this and have googled a ton, but still can't find an answer to this specific question.


edit: the woofer is a 3.5" Dayton Audio ND91-4, the tweeter is a 5/8" Dayton Audio ND16FA-6. They are only going to be receiving 8 watts. One more question, regarding 1st order crossover calculators. Every 1st order crossover calculator I see, you enter the crossover frequency, and it spits out a cap and resistor value for each driver. But shouldn't a 1st order have the woofer and tweeter crossed at 2 different frequencies pretty far apart (like 4500hz for the woofer and 10k for the tweeter) so that when they merge with such a low slope, there isn't any peaks? Those are the frequencies I have mine set at so far, and so far I'm impressed, or can I do better?
 
Last edited:
frugal-phile™
Joined 2001
Paid Member
I have a 50ohm resistor wired directly to the positive terminal on the tweeter. This is for the tweeter level, 30 ohms was still way too loud. On the other side of the tweeter resistor, wired directly to the resistor, aka: the amp side of this connection, I have the crossover, a 0.33uF capacitor. Is this correct?

This suggests tome that the tweeter could be disconnected with little or NO loss of top end.

Do you have a choke on the FR?

dave
 
This suggests tome that the tweeter could be disconnected with little or NO loss of top end.

Do you have a choke on the FR?

dave

That's true, the full range sounds great on it's own. The reason I added the tweeter is because cymbals needed that extra sparkle. I can hear a pretty decent difference covering the tweeter with my hand, the music loses its clarity basically. From the calculators I used online, it said I need a 30 ohm resistor on the the tweeter just to get the spl's equal between the two drivers. But a combination of it being too loud and my first crossover point being two low, I ended up at 50 ohms.
 
This suggests tome that the tweeter could be disconnected with little or NO loss of top end.

Do you have a choke on the FR?

dave

Sorry, got off topic lol. The full range driver has a .5 ohm resistor and a .15mH coil to give it a 6db low pass at ~4500. So total resistance on the woofer is 4.66 ohms. (.16 for the inductor, 4 for the driver, and .5 ohm resistor). The tweeters total is 6 ohms plus 50 ohm resistor, so 56 ohms. Together, this should present a total load of around 4.16 ohms to the amp.
 
Can you go to mh-audio.nl? this was a site recommended by Madisound. It's an education just to struggle with it! I looked up a sub bessel 4000Hz 2nd order. and got 4.5 u .270mh for a tweeter, (4.5ohms) I think second order here is mandantory. Possibly .35 mh for the 4 ohm woofer. You MUST do an Lpad calculation for the TW. I used a rough 7 parallel 1 series resistor combo, (getting 4.5 ohm adjusted impedance) Play with it! Don't skip the math! Don't fudge the woofer impedance. It's value rises with frequency, anyway.
 
It's usual to wire the tweeter components in the order: resistor, cap, tweeter. That way, any tweaks to the resistor won't affect the crossover frequency i.e. you don't need to include the series resistor when calculating the cap value
re:'when it asks for speaker impedance, is that ACTUAL speaker impedance that it wants?" - yes, read the impedance value at the crossover frequency from an impedance plot of the driver
 
It's usual to wire the tweeter components in the order: resistor, cap, tweeter. That way, any tweaks to the resistor won't affect the crossover frequency i.e. you don't need to include the series resistor when calculating the cap value
re:'when it asks for speaker impedance, is that ACTUAL speaker impedance that it wants?" - yes, read the impedance value at the crossover frequency from an impedance plot of the driver

Does this mean that all of my crossover values are incorrect then? Or since I figured for the resistor values being included in the crossover math, they are correct but need to be recalculated if I change the resistor for the drivers level?
 
These are the specs that I used for all the calculators I've used, and so far, I'm pretty happy with how it's going. This will be my 4th revision to the crossover, if I do need to revise it. I'm basically looking for some more knowledge in the topic, and for someone to double check my work.
When you say you want to show your work or "for someone to double check" it, it's good to post the crossover layout for ease of understanding.
You can do some testing, with more accurate simulations if you use drivers FRD & ZMA files. This is a good way to be closer to your objectives, after your "4th revision to the crossover". :)
 
Assuming a crossover frequency of 5 kHz and a woofer impedance of 16 ohm, then the coil would be around 0.5 mH IMO. However, theoretical crossover calculations neglecting actual driver frequency response and baffle step are of limited benefit.

I've been playing around with my simulation program and agree with colwaiting, who said the tweeter needs a second order high-pass filter. Attached a basic crossover proposal consisting of only 3 components. It's a series crossover providing a second order tweeter filter. The little trick here is, the impedance peak of the tweeter filter acts as notch for the woofer as well.
 

Attachments

  • SimulationUsername8611.jpg
    SimulationUsername8611.jpg
    113.5 KB · Views: 321
  • CrossoverUsername8611.jpg
    CrossoverUsername8611.jpg
    34.1 KB · Views: 323
  • DriverVoltagesUsername8611.jpg
    DriverVoltagesUsername8611.jpg
    96.3 KB · Views: 314
An externally hosted image should be here but it was not working when we last tested it.



Sorry if it's a bit messy, but I think it shows what I'm doing, obviously this is only showing one channel. Anyways, I figured that with my values, I should have a 6db low pass on the woofer at around 5500hz, and a high pass 6db filter on the tweeter at around 9000hz. Is that right? Are they actually being crossed where I think they are?

Modeling software would be cool, but I've tried it and am immediately lost. This is just a hobby so I don't have the time to devote to learning the software, but I do know that people can get a dang good sounding speaker even without modeling software.

Sorry if that drawing is a bit elementary, I made one of these for a my brother in law and he requested at schematic also.
 
Last edited:
Assuming a crossover frequency of 5 kHz and a woofer impedance of 16 ohm, then the coil would be around 0.5 mH IMO. However, theoretical crossover calculations neglecting actual driver frequency response and baffle step are of limited benefit.

I've been playing around with my simulation program and agree with colwaiting, who said the tweeter needs a second order high-pass filter. Attached a basic crossover proposal consisting of only 3 components. It's a series crossover providing a second order tweeter filter. The little trick here is, the impedance peak of the tweeter filter acts as notch for the woofer as well.

I'm confused, where are you getting 16 ohms for the woofer? Is that the impedance of the woofer at the crossover frequency of 5kHz? Also, the resistor in parallel with the woofer, what does that do? It looks like the woofer is played straight through in your schematic, with no filters, or am I reading that wrong?

Edit: Do you recommend crossing the tweeter with a second order for sound quality, or just to protect the tweeter? I'm curious as I figured it should be safe with a first order at such a low wattage.

Edit2: If anyone is wanting to kill some time, it would be really cool if they could use some modeling software to input what I have on my schematic, and then with what they think is the best crossover system, to see how far off my guesswork really is.
 
Last edited:
Sorry for the confusion, I did overlook you mentioned a zobel in the OP. 16 ohm would be the impedance without the zobel.

Attached a simulation of your current setup. I would say the woofer filter basically works as intended, but the tweeter response is more or less flat from 3 - 10 kHz and rises above. A second order filter is required to get some level at 10 kHz and an attenuation at 3 kHz.
 

Attachments

  • SimulationUsername8611Original.jpg
    SimulationUsername8611Original.jpg
    113.4 KB · Views: 301
  • CrossoverUsername8611Original.jpg
    CrossoverUsername8611Original.jpg
    45.4 KB · Views: 140
  • DriverVoltagesUsername8611Original.jpg
    DriverVoltagesUsername8611Original.jpg
    96.4 KB · Views: 111
Sorry for the confusion, I did overlook you mentioned a zobel in the OP. 16 ohm would be the impedance without the zobel.

Attached a simulation of your current setup. I would say the woofer filter basically works as intended, but the tweeter response is more or less flat from 3 - 10 kHz and rises above. A second order filter is required to get some level at 10 kHz and an attenuation at 3 kHz.

Thank you, thats really cool to see my guesswork actually plotted out. I can actually hear the 10kHz peak, the highs sound almost too high, if that makes sense. How would I wire in a second order filter? Can I leave the woofer circuit untouched, and just use a calculator like at the12volt.com to find the values for the second order high pass filter? Or does it need to be a specific type? It also looks like the tweeter level is too low. If I were to redo my crossovers completely what would you use?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.