Completed project: "Fenrir"

So I'm building the Fenrir, in an MLTL enclosure designed for the woofer by Paul K. The baffle dimensions are the same, superimposed onto a floor stander. I'm now at the point of verifying everything is wired up right before gluing the crossovers into the box, and have run into a little glitch in my measurements.

Here's what I've got:

An externally hosted image should be here but it was not working when we last tested it.


DATS impedance sweep - tuning frequency was meant to be 32hz so this looks pretty good I think. General impedance/phase shape looks close to Wolf's.

Now my FR measurements. Conditions:

40" distance, mic on tweeter axis
Each speaker in exact same location, mic didn't move
2.5ms gating (using step response to find first reflection)

An externally hosted image should be here but it was not working when we last tested it.


Both have a narrow dip @ the XO frequency. I tried reversing phase on the tweeter, and the result was:

An externally hosted image should be here but it was not working when we last tested it.


I'm wondering if any other builders did their own measurements and saw something similar? Or if anyone can see something obvious I may have done wrong in the XO or otherwise.
 
Same artefact is observable on the measurements in first post of this thread - just more obvious because you are using gated measurements and different grid resolution.

Try to measure it 30 degrees off axis. It could be diffraction related issue.
 
Same artefact is observable on the measurements in first post of this thread - just more obvious because you are using gated measurements and different grid resolution.

Try to measure it 30 degrees off axis. It could be diffraction related issue.

HobbyHifi baffle H40xW20cm Wavecor WA06...
 

Attachments

  • Wavecor 30.jpg
    Wavecor 30.jpg
    712.4 KB · Views: 182
Thank you gentlemen. I see the dip there in the hobbyhifi measurements too.. My baffle is 10" wide x 16.75" tall. The overall tower is 41" tall, same width as baffle. 3/8" roundover on sides of baffle, sharp edges on top/bottom of baffle.

Here's the same on-axis plot from before, with a 30deg off axis measurement overlaid:

An externally hosted image should be here but it was not working when we last tested it.
 
I should be clear here that I have complete faith in Wolf's design, and that the speakers (even in incomplete form) sound great as-is.

As ever though I don't trust my ears, and I haven't assembled an XO in a decade, so thought I would verify I had everything correct before final assembly.

Is it safe to say that what I'm seeing is not an error in my XO assembly, but rather a diffraction artifact that can't be avoided given relatively sharp baffle edges I used?
 
I should be clear here that I have complete faith in Wolf's design, and that the speakers (even in incomplete form) sound great as-is.

As ever though I don't trust my ears, and I haven't assembled an XO in a decade, so thought I would verify I had everything correct before final assembly.

Is it safe to say that what I'm seeing is not an error in my XO assembly, but rather a diffraction artifact that can't be avoided given relatively sharp baffle edges I used?

Yes, it's definitelly a diffraction on axis. It would need much more than 19mm roundover or 45 degree chamfer to mitigate it. I wouldn't hack the cabinet just yet. I'd see what can be done with some felt. There is really interesting read at David Ralph's speaker pages about diffraction control with felt pads.

David Ralph's Speaker Pages - Felt Effects on Baffle Diffraction

It is a bit ugly but can be incorporated in speaker grills.
 
Last edited:
I have a half-inch roundover on my baffles, followed by a step of about another half inch total width, so it might just be compensated more in my slightly larger baffle dims. I have roundovers on all edges of the baffle, not just the sides. That also will have an effect.

Also note the xover you built has not been shown, so we cannot verify the connections presented. I agree that your impedance looks very similar, outside of your TL tuning, and that likely means it's pretty close.

Later,
Wolf
 
Yes, it's definitelly a diffraction on axis. It would need much more than 19mm roundover or 45 degree chamfer to mitigate it. I wouldn't hack the cabinet just yet. I'd see what can be done with some felt. There is really interesting read at David Ralph's speaker pages about diffraction control with felt pads.

David Ralph's Speaker Pages - Felt Effects on Baffle Diffraction

It is a bit ugly but can be incorporated in speaker grills.

Thanks Zvu, I've seen that page before, impressive results! No grilles for these speakers, and I don't think I could bear the sight of the felt strips on the baffles. Perhaps a shroud or some sort of star pattern could work, but even then I suspect that at the thickness required it would be an eyesore.
 
Last edited:
40" distance, mic on tweeter axis

I suppose that is one 1m? In that case, dip is gone at listening distance and room filled up the valley...

Yes that's right, just a bit more than 1m. Listenint to the speakers in stereo at normal listening distance, they sound wonderful, I don't notice any issue. That being said, I wasn't critically trying to zone in on the dip, looking for instruments in that FR range etc.
 
I have a half-inch roundover on my baffles, followed by a step of about another half inch total width, so it might just be compensated more in my slightly larger baffle dims. I have roundovers on all edges of the baffle, not just the sides. That also will have an effect.

Also note the xover you built has not been shown, so we cannot verify the connections presented. I agree that your impedance looks very similar, outside of your TL tuning, and that likely means it's pretty close.

Later,
Wolf

Thanks for chiming in Wolf! That's quite possible... and I do have a step at the bottom of the baffle, sharp edge to the rest of the (floorstanding) cabinet, so another difference there.

I have pics of the XOs but am a little embarrassed to post them publicly due to sloppy workmanship ;) I'll PM them to you though if you wouldn't mind taking a quick look.
 
Morbo
I would remeasure and move the microphone up and down until I would get maximum null, and then try to correct the freq dip in the crossover: correct the crossover frequency until satisfied. Maybe there's some variations in production samples, but from the measurements it would be a really big difference. I bet the roundover on the baffle edges contribute a lot the flatness, at least when I measured my own spkrs.
Can you post individual Freq resp for tweeter and woofer on top the summed freq response. This way we can compare to Wolf's original design.
 
Last edited:
Morbo
I would remeasure and move the microphone up and down until I would get maximum null, and then try to correct the freq dip in the crossover: correct the crossover frequency until satisfied. Maybe there's some variations in production samples, but from the measurements it would be a really big difference. I bet the roundover on the baffle edges contribute a lot the flatness, at least when I measured my own spkrs.
Can you post individual Freq resp for tweeter and woofer on top the summed freq response. This way we can compare to Wolf's original design.

Thanks for the suggestions pleidtka. I agree it could be production variances but I would be surprised, I'd expect wavecor and rival to have pretty tight QC, and the responses across both speakers are fairly consistent so it would have to be batch-batch consistency not driver-driver consistency that is the issue.

I'm not clear on your suggestion to move the mic up/down; surely if I want tweak the crossover, I should be measuring on the intended design axis? Using the reverse null makes sense for some designs, but my understanding is that some experienced designers don't shoot for a deep null @ XO frequency, so without knowing if this was Wolf's target, I'm not sure it's the appropriate thing to look for.

At this point, Wolf has been good enough to look at my XOs and verify that the wiring looks correct. My measurement conditions are really not great right now, so I'm not terribly confident in my data at this point, at least not enough to get good raw measurements of the drivers, model the crossover, and start making changes.

Over the next couple of days, I have a good deal of free time on my hands. The speakers really do sound excellent as-is (whether this speaks to my tin-ears, I'll let you decide), so what I'm inclined to do right now is complete the speakers while I have the time to do so; this project has been in progress for months and I'd like to get some closure on it. I will however make the bases/crossovers accessible rather than gluing them in as I had planned. Around the end of November I should have access to a larger space with fewer reflections to measure in, so I may revisit then.
 
Morbo
I would remeasure and move the microphone up and down until I would get maximum null, and then try to correct the freq dip in the crossover: correct the crossover frequency until satisfied. Maybe there's some variations in production samples, but from the measurements it would be a really big difference. I bet the roundover on the baffle edges contribute a lot the flatness, at least when I measured my own spkrs.
Can you post individual Freq resp for tweeter and woofer on top the summed freq response. This way we can compare to Wolf's original design.

If it is diffraction and you fill in the dip on axis, you will have a peak off axis potentially making it much worse.