More XO tweaking fun with the SunOnes.

[Patrick Sun]

An update on the 2.5-way speaker project called the SunOnes:
6/15/01 - One last try with the current topology (2nd order on the tweeter, 4th order on upper midwoofer, 2nd order on lower midwoofer):

Graph 19

While it's got less glare from the midwoofer's output in the 3000-5000Hz range, it doesn't sound right. I'm taking some time off this XO and going to work on my other speaker project, the SunTwos (a 2-way speaker).

7/9/01 - 7/10/01 A change of scenery:

After spending some time tweaking the SunTwos, I decided that I needed to revisit th 4th order filters for both tweeter and upper midbass for the SunOnes. I initially got the impedance for the upper midwoofer wrong (thinking that with 2 midwoofers, I would have to half the impedance from their parallel connection, but as it turns out, this was not the case, once I was getting into the crossover region of 2KHz, the lower midwoofer was not making much effect on the overall woofer impedance, so I had to adjust my capacitor and inductor values accordingly.

Graph 20

Obviously the filter values for the upper midwoofer are wrong with the response that I measured (and yes, the speaker did not sound good at all with all the midbass sucked out). My inductors were too small (by almost 50%), and my capacitors were too large in value (by over 200%).

Graph 21

So I make some educated guesses at what I think the revised values for the C's and L's, and it raised the output of the upper midwoofer. But I didn't think a 4th order should have such a shallow slope roll-off, and I didn't quite like how the tweeter filter was kicking in around 2500Hz. I could still hear the glare, and it didn't sound so good either. I knew I was close. I made some more educated guesses (played around with modelling 4th order filters at certain frequency points to get a feel of how the C's and L's values moved as I went up or down in frequency).

Graph 22

After making some more changes to the C's and L's, I hit the jackpot. I do what I thought I needed to do to move the tweeter break frequency towards 2000Hz, and the same with the upper midwoofer. The slopes for both the tweeter and upper midwoofer filters are now behaving as I wanted them to perform. They each have nice steep slopes, and the glare is finally gone from the speakers. I listened to some music, and was greeted with a nice wide soundstage, and was certainly able to pick out the location of the instruments mixed into the tracks. The 500Hz-2KHz region is a little flat and I think I will try one more tweak tonight (increasing the inductor pair on the filter for the upper midwoofer to gain just a little more output). But it's amazing how much the inductors can affect the audio quality of the upper midwoofer.

Hopefully the end is near with the major tweaking on the SunOnes.




PatCave; HT Pix;Gear;DIY Projects;DVDs; LDs

[Patrick Sun]

7/11/01:

Okay, just to play around some more, I decide to see what happens to the midwoofer response if I increase the values of the inductors about 15%.

Graph 23: Well it isn't pretty. As you can tell, the larger inductors totally drag down the response before the tweeter kicks in.

I quickly restore the upper midwoofer's XO to its previous condition. Then I see if I can move the tweeter slope closer towards 2KHz, so I increase each capacitor by about 0.47uF.

Graph 24: This indeed produced the desired results. It's getting there, but I know there's something amiss. Actually it's the bass in the midrange, it's just not there. I sleep on this for a couple of days.

7/13/01:

I made 2 decisions while I slept on my current situation: I upped the tweeter padding 1dB more to tame a bit of peakiness, and I also decide to throw in the upper midwoofer filter from the SunTwo into the SunOne. The exact same filter (requires me to reduce the inductor values by 20%). I do this to try to regain some of the midrange that's being dragged down by the current inductor values. Well, this works, but I discovered something that's probably known to most speaker builders: if the output near the crossover point is lacking, reverse the polarity of the tweeter connection and measure again. I did this as a lark and was pleasantly surprised at what I heard and measured:

Graph 25: Not only do I get back my midrange in the 1KHz-2KHz range, I get nice output around the 2KHz crossover point. If you compare this graph with the previous one, you'll see what I mean. Also, the bass response is now operating really well, I'm getting some good bass from the 2 midwoofers now. Also, the tweeter response is more controlled than before, but I may go back to old padding level if I want to play with the "tweeter imaging height" for these speakers.

BTW, the other line is the response when the tweeter is wired in-phase with the midwoofers, as you can see it's got that dip that tells me the phase tracking is decent (probably not exceptional, but I'm not complaining).

I really hope this is the end of the road (being optimistic here!) for this project. We'll see.




PatCave; HT Pix;Gear;DIY Projects;DVDs; LDs