For future atmos build: Large bookshelf
This is a design I have been working on for the past few months. I am currently without any decent speakers and plan to use two of these. Eventually, get a sub, and after my degree (which I haven't yet started it), I plan to save up and build a bunch more of these for an eventual 11 channel setup with 2-4 big sealed subs.
This speaker box isn't your average shape. I chose a sealed box as I wanted better transients and it would cross over much neater with a sub when I get one.
I will post screenshots of the SketchUp design later. I have already tested the internal volume will give a QTC of 0.707 once stuffed.
But currently, I have the crossover design ready to roll out to the public:
Here you go guys, it is a bit of a crazy crossover. The tweeter has a steeper roll of than the woofer, I was able to do this because the woofer doesn't break up until above 6khz but I wanted a low crossover point due to a weird resonance peak in the woofer's driver.
I started in passive crossover designer 2.0 and then created my design as a circuit in xsim. In xsim, I inputted the values but also the inductors' resistances. Weirdly this flattened the response further. YAY!
I also finetuned the design using different parts as ones you can buy are very different from your initial preference. All in all, I ended up with a response I much prefer.
The dip around 6khz is actually my preference, otherwise I can't turn it up as loud as I'd like because my ears are oversensitive to those frequencies.
Also the BSC isn't very intesnse at all because the speakers are going in the corners of the room. 🙂
The woofer: ND-140-4, Tweeter: ND-25FA
Both 4 ohms but in this parallel circuit I still get a 4 ohm impedence. My amp is rated for down to 3ohm.
Please could you help me with any mistakes I might have made. This is my first build.
It looks pretty nice. Just a quick question. What are you using the 2.2ohm resistor for? That is not a today placement. So I'm just curious what your accomplishing with it.
The tweeter has a much higher sensitivity than the woofer hence the resistor to flatten out the response. I currently read a couple of articles on the series of using tweeter waveguides to integrate the tweeter's off-axis response better into the woofer's. According to these articles, a response that follows the same pattern whilst going off-axis means the room effects do not colour the sound as much. So I am now trying to figure out if using the nd25FW instead of the nd25FA is a better option. The sensitivity is even higher meaning a bigger resistor. So far with my crossover designs, this seems to interfere with the tweeter's overall response (as it would) so I am unsure yet. Also, I don't like how the frequency response on the FW starts to drop before 20khz, I don't know if this will make a difference or not. Am I overthinking that?
Here we go. I did some tests before my half-space calculations to see what the crossovers might look like.
The first image is the new crossover design with the waveguide tweeter. The insane resistor value is for me to try and tame that sensitivity.
The second and third versions of that crossover design are me inputting parallel resistors as I want to create versions where the overall resistant value stays the same but the resistors aren't overwhelmed. For example, each resistor is rated at 10 watts, the tweeter at 25, because of them being in parallel it is like giant resistors rated at either 20 or 30 watts.
The fourth version is an impedance fix for the woofer as I realised I am not able to get that value inductor from the place im getting crossover components from.
Hopefully, I won't need those designs though because they will be much more expensive and look at that BBC dip I had to introduce in order to actually get the crossover to work. (not saying a BBC dip is a bad thing just would rather change that thing electronically instead of it being forced there).
Anyway, the last and final image is a reminder of how good the original looks. I mean look at that phase crossover, mmm. The only problem might be with it is the uneven off-axis response. I'll post again with half-space calculations to see if the extra money and complications are worth the waveguided tweeter edition.
Could I have some help with the half space response I'm really confused by WinPCD. I did the baffle step response and enclosure simulations with VirtuixCAD, that was fine. I figured out by testing different input wattages for the enclosure calculations (whilst the step response graph was overlayed and... well see for yourself: (black is the total to 2Pi, and brown is the total to 4Pi. Whatever that means)
I'm hoping the final responses to be similar to the black line
WOW, this is a happy accident.
In the enclosure simulations section for the woofer, I accidentally applied the crossover overlay to the tweeter instead of the woofer. The main crossover graph then changed the tweeter's response dramatically. This tells me that the baffle step and enclosure design has applied itself to the woofer's side of this graph. This means I can show you the 1/2 way done half-space response!!!!
Here you go!
Looking pretty good on the woofer. One thing I may suggest is look at working on the 500-1khz region. With that upward slope, especially at 1khz, that might not be the best for you. I'd probably first try to work with the Zobel you have or even turn it into a notch if necessary. It may not be an issue, but in my experience anything rising in the 1-2.5khz range can be quite audible and if you have sensitive ears become fatiguing. Of course that is all a designers choice.
After a bit of half-space testing, I think the waveguided tweeter will do more harm than good. I have decided to go with the "without waveguide" 25FA option as the waveguided option just messes everything up. It doesn't even help with the crossover, like many people said it should!
I'll post my final findings and final comparisons when I'm finished. I can't wait to start going to my local maker space to start building the first stereo pair. Unfortunately close is 1hr away!
Edit: without increasing the cost of crossover components, I am unable to reduce the bump between 500-1000hz without introducing a bigger peak in the 1khz-2.5khz region. I think I might just stick with this for now. Who knows?
Try something like this:
It should take care of that area while flattening out the response. Basically, we turned the zobel into and RLC notch filter. It maintains your impedance while also flattening out that area. The only added cost is a.1 Inductor for about $2.60 I am just using the Dayton files, if you have something else, let me know. Here is my projected response with the rework:
If you want to take it a step further you can add a cap to the inductor to get rid of that hump in the 4.5Khz range. If you leave it there, you will most likely have reconstructive waves which will affect your response. In this case, you will need to decrease the cap value a little to compensate for that.
That is so helpful, thanks Toid.
I'm just about to see how it affects my polar responses and I was just wondering, should I calculate Dd from WinISD or guess from the diagram. Because WinISD says the midwoofer has an Dd of 105mm which when looking at Dayton's diagram looks quite small. For the tweeter it says it is 30.9mm which seems about right. For now, I'll use these values, I just wanted your opinion! 🙂