I think the crossover is spot on! Good job.
As for the raised response in the upper treble, just aim the speaker forward rather than at the listener. That should smooth out the treble.
I think the crossover is spot on! Good job.
As for the raised response in the upper treble, just aim the speaker forward rather than at the listener. That should smooth out the treble.
For over 15 years I've been working for Dell, so anything I say about computers, mobile phones, printers, networking gear, or anything Dell sells must be taken with a grain of salt.
I had shelved this for later work so I could get back to my "real" work.. but it was always there a little bit, when I had a quick thought.
I thought the Baffle Edge Diffraction was more of a tweeter issue rather than a woofer issue. When I originally figured out the driver layout, I don't think I even included the woofers in my calculations (note: I used Baffle Diffraction Simulator spreadsheet from FRD).
I've played around with my edge diffraction spreadsheet -- it doesn't seem to do much in the 100-1000hz range. I'm thinking I'll re-measure the speaker using the ground plane technique and see if that alleviates the MMmm issue.
I also measured my driveway -- it looks like it should be wide enough to get accurate readings down to 125hz (it's about 18ft wide). That should work for most of my MMmm range. With that in mind, I'll forego making the trek to my friends house with his bigger driveway.
Finally, putting the parts list together. For the project, part of the plan was to keep costs low. With that in mind, and also with the thought that I'm buying everything from PartsExpress, I thought I'd go with their regular Dayton Metalized Polyprolene caps for everything but the cap listed as being 6.61 on the pic below. For that one, I'd go with the better quality Solen. Inductors would be the Jantzen 18awg versions save the one listed as .463 below, for which I'd use a 15awg instead. For the crossover prototyping, everything was 18awg. I'm assuming the jump up to 15awg won't have an adverse effect. The DCR does change from .32 to .20. Finally, I'd stick with the Dayton non-inductive resistors.
As is, the crossover part cost is ~$54/speaker. (For those keeping score, the drivers are ~$150/speaker). It seemed to me to be a reasonable way to go.
As for getting the parts I wanted, I lucked out on the resistors -- Dayton makes exactly the values I wanted. The inductors have some that were really close -- enough that you could get away without unwinding to get to the "ideal" value but still be within 2%. The caps, however, weren't so fortunate. Of the 4 caps needed, only one could be completed with a single cap if I stick to the Dayons and the 1 Solen. Doubling up on the Daytons is still cheaper than going with a single cap with a closer value from one of the other brands PE sells.
Anyway, wanted to get a sanity check before I have my friend buy the parts.
It all makes sense to me. Sanity check is passed.
For over 15 years I've been working for Dell, so anything I say about computers, mobile phones, printers, networking gear, or anything Dell sells must be taken with a grain of salt.
So, finally, the confluence of events occurred in which I was able to finally do the ground plane measurement. (it's been raining and we've had company)
Here is the setup:
And here are the results
I still get the MM response in the 4th-6th octaves, but it doesn't look quite as bad as before.. even with the one big spike/dip between 100-200hz. On the other hand, I'll admit I was hoping for something better. That being said, I still don't completely trust my sound card. I'm also assuming that I ignore the response after~200hz.
I does show the extension I expected though. I know some of that has to be because the speaker is on the ground, but it's not completely in 2pi space.. not 4pi either.. Anyway, it more looks like what I'm hearing (as poor as my uneducated ears can tell).
And I've played around with my baffle edge diffraction spreadsheet -- I'm not seeing much change in the modelled frequency response in that octave regardless of speaker placement.
Thoughts? Comments?
First, lay the speaker on its side for this measurement. You want both of the woofers as close to the ground as possible to reduce any potential reflection issues.
Chech the input levels on your sound card. If you are clipping the input, even a little bit, the results could get weird. Other issues can arrise of the levels are way too low.
On the enclosure, is there dampening behind the woofers, or anywhere inside the box? If there is no absorption going on, you will have standing wave and transient dynamics issues which could cause harmonic based issues in the amplitude response.
For over 15 years I've been working for Dell, so anything I say about computers, mobile phones, printers, networking gear, or anything Dell sells must be taken with a grain of salt.
You know, I thought about that and was going to do it -- not that I had the reasoning, just a "why not" kind of thing. Did I do it? Of course not.Originally Posted by IrritateGuy
As far as I know, I'm not clipping. When I do that loop back test, I make sure the clipping light doesn't come on. Other than that, the sweep isn't getting super loud, and the preamp is set to -11. I'm assuming I'm good here.
Ok, here may be an issue. I had the acoustastuf in there when I brought it home. However, when I had to diagnose the wiring problem, I removed all that stuffing from the lower half of the speaker. I didn't put it back. All of my tests have been with a box that has only the upper half with stuffing. Alright then.. new tests to come later.
As usual, thank you for all your help.
The loopback test levels have nothing to do with the Mic levels. They are independent of each other as the mic will have a different output level than the output of the sound card.
I recommend adjusting the levels so the signal light comes on, but not so the clip light comes on.
For over 15 years I've been working for Dell, so anything I say about computers, mobile phones, printers, networking gear, or anything Dell sells must be taken with a grain of salt.
Ahh.. misunderstood but see it now. I'll check.
Thanks again.
Ran the test again, this time with the speaker on its side.
The top result I realized that the tweeter wasn't attached. It turns out one of the alligator clips disconnected.
Reconnected it and ran three tests. Those are the next three. I then added the acoustastuff and ran the bottom three responses.
Here are the results.
Looks like I may have the polarity backwards again on the tweeter. I'll have to revisit that later. I also assume the (maybe) low end roll off with the latter result is a result of the box looking bigger due to the stuffing.
Unfortunately, it doesn't look like the stuffing or the different testing smoothed out the MMmm issue. I guess it is what it is.
Posting Permissions
Reply With Quote
