Curious what experienced eyes make of this graph

[Roy_D]

I will share the specific model later - curious what experienced eyes make of this before, like that directivity crunch up at around 12k?
Its tweeter is set in a waveguide. The measurements appear to have been made in a livingroom/listening room situation.
Bass extension is noted as F3 at 51Hz. I would place these as I would my kef coaxials I reckon - dead on facing the other wall considering that brightness way at the top end
(but my hearing rolls off at 17kHz or so so anyways at regualr listening levels of 55-60db or so anyways lol)

 

[a4eaudio]

(Not sure I count as "experienced" eyes, but since no one else has replied...)

Since the measurements get pretty wavy below 700Hz and only show down to 200Hz we can presume farfield measurement with no nearfield to merge. Assuming the woofer was modeled well and the box designed accurately, things should at least be okay on the low end, even though we don't know for sure. (And I saw an example recently where things went really bad on the low end even though it should have been easy to build something that worked.)

On axis looks pretty good. At least nothing that looks "bad".

The line charts and polar map show the same thing - some bunching and excess energy between 3kHz and 5kHz. This is pretty typical of two-ways with LR4 crossovers, so doesn't necessarily indicate they would sound bad, other than they would sound bad in the same way that a lot of speakers like this sound similar. I wouldn't worry about the bunching at 12kHz - there is very little music energy up there and it is quite narrow (probably not even 1/3 of an octave).

The objectivists can draw a lot of arrows at the Directivity Indices and label them "directivity errors" but they are the same errors typical in two-ways like this. I would pay more attention to the In-room Response and Power Response - they are not very smooth/linear and the bump at 3-5kHz reflects the same excess energy as the Line charts and Polar Maps.

Woofer appears to be SBxxPFC and the passive radiator is SBA "racetrack" PR. Not sure about the tweeter, although it looks like maybe a tweeter put in the Visaton WG148-R. Having mentioned a few times that things look "similar" to what we see in typical 2-way bookshelf speakers, this is a little surprising. The main point of the waveguide would be to AVOID those typical directivity errors.

Why do I say that? Here is a speaker using the SB16PFCR25-8 and a Peerless tweeter in the Visaton WG. Compare the Line Chart and Polar map below to your images. (In fact, looking at this image and thinking about the waveguide, I'm not sure how you could get that bunching at 12kHz in a waveguide like that.)

 

[Roy_D]

Thank you! Some handholds for me to learn more and yet more questions that bubble up.
Yes the tweeters Visaton advises both appear to have domes that stand quite proud even with their faceplates - the Seas one used in this design is not that. Makes me wonder if perhaps removing the seas' faceplate and mounting directly to the wg might have been the better way. In this caise mounting was done using thin doublesided tape on top of the faceplate to hold in position then screws but perhaps also that additional half a mm extra gap is enough to cause some issues. On the other hand doing away with this seas' face-plate opens up a can of worms with regards to mounting and maybe turning that ring around the tweeter dome and the harder edge around that into diffraction ally.

Building Stand Mount Speakers with WAVEGUIDE & PASSIVE RADIATOR |

soundblab.net

Right on the money on the wave guide! and woofer and radiator.

 

[kemmler3D]

@a4eaudio has a good response here, my drive-by comment is that the match between WG and tweeter seems to be just okay, my guess is it wasn't designed specifically for it. Could use a phase plug / coherer to handle the top octave, but we know those aren't exactly easy to design right. I think your idea of toeing out a bit is probably good, maybe a little toe in but not much.

Overall doesn't look too bad but some of the issues that do exist might be tricky to EQ.

 

[Roy_D]

It has been 5 years with my current coaxials and looking for something to best them for a few months now. I happened across some terrific ready made designs such as those Korean made speakers of similar design as this one above here but on a different level. But the fun in discovering what makes it all work and getting sparked into perhaps taking on a build project I did not expect..

 

[kemmler3D]

Thank you! Some handholds for me to learn more and yet more questions that bubble up.
Yes the tweeters Visaton advises both appear to have domes that stand quite proud even with their faceplates - the Seas one used in this design is not that. Makes me wonder if perhaps removing the seas'faceplate and mounting directly to the wg might have been the better way. In this caise muonting was done using thin doublesided tape on top of the faceplate to hold in position then screws but perhaps also that additional half a mm extra gap is enough to cause some issues. On the other hand doing away with this seas' face-plate opens op a can of worms with regards to mounting and maybe turning that ring around the tweeter dome and the harder edge around that into diffraction ally.

Building Stand Mount Speakers with WAVEGUIDE & PASSIVE RADIATOR |

soundblab.net

Right on the money on the wave guide! and woofer and radiator.

What I've seen on DIY Audio threads with waveguides is that people do typically remove the faceplates. I think the extra gap does tend to create problems at high frequencies.

 

[Roy_D]

What I've seen on DIY Audio threads with waveguides is that people do typically remove the faceplates. I think the extra gap does tend to create problems at high frequencies.

Well then perhaps one mystery solved there after all.=] Now for that 3-5k hump. And the jump in brightness up top but that too may be from the wg mounting method a ways.
I will go have a dive into those threads with waveguides for a bit!

 

[kemmler3D]

the jump in brightness up top but that too may be from the wg mounting method a ways.

The directivity increases but on-axis also goes up, to me this says "excessive beaming" and you might want to tone the on-axis down with the crossover. But if you are planning on taking off the faceplate, I'd leave that alone until you can re-measure.

The 5khz hump could be diffraction, maybe? But it looks a little wide for that. Not sure.

And the little bump around 500-700hz seems to line up with a wiggle in the impedance, which says "resonance" to me, so maybe the cabinet is not stuffed / braced enough?

 

[Ktacos]

The 5khz hump could be diffraction, maybe? But it looks a little wide for that. Not sure.

looks basically exactly how the dispersion did on a wg300 + DX25 combo I had looked until I added some round overs. Waveguides still need some sort of edge diffraction treatment IME.

That looks like the WG148 to me, honestly not a fan of them for the same reason I don't like the WG300, there is a pretty tall wall surrounding the dome which tends to generate some complex reflections. I believe you can see these reflections as response issues at ~7khz and ~12khz in the data. I think waveguide tech is kind of past these sort of waveguides where they aren't designed around the driver. Stuff like waveguides from somasonus tend to not have these problems and I would pursue those OP if you're after a waveguided tweeter speaker.

 

[a4eaudio]

...Makes me wonder if perhaps removing the seas' faceplate and mounting directly to the wg might have been the better way...

Yes, i think that is probably a good idea.
Secondly, he puts some gasket tape between the faceplate and the waveguide which is going to create even more distance. I would run the gasket tape around the outside of the tweeter body (not the surround) and the waveguide just to make sure there are no air leaks.
These are VERY small distances and he argues in the video that they probably don't matter...but they could be the cause of the anomaly at 12kHz. Also, to be fair, it may not be audible.

...The 5khz hump could be diffraction, maybe? But it looks a little wide for that. Not sure....

I hate to "guess" at what someone else is doing, but...
I see "xsim" in the screenshot of the impedance but the directivity measures are clearly being created in VituixCAD. I wouldn't be surprised if the crossover was designed based on the on-axis response in xsim and then measurements taken after the fact. On axis everything looks fine, and I actually would expect the speaker to sound fine (but maybe not as good as it could) - we are just nitpicking over the details. But if he would have focused a little more on the Predicted In-Room Response and Power Response (assuming he didn't, which could be completely wrong) he could have sacrificed a little on-axis linearity for better off-axis behavior.

 

[ThatSoundsGood]

In addition to what everyone else has said here, there's a bump in the impedance plot around 450Hz the could indicate some sort of issue; maybe the box isn't sealed correctly or something else.

 

[Roy_D]

In addition to what everyone else has said here, there's a bump in the impedance plot around 450Hz the could indicate some sort of issue; maybe the box isn't sealed correctly or something else.

I don't see any kind of sealant material used in mounting the drivers etc used in the build video. Also no inside chamfering of the driver mounting holes and not much on the bracing. From what I understand polywool is not all that effective in combating standing wave resonances - not in the amounts used here. Oldskool rockwool or fiberglass is. And the addition of bitumen pads as Troels Gravesen does, or self adhesive solid butyl rubber deadening has been proven to be very effective without eating too much interior space. As much as possible of bad behaviour should be eliminated at the source though. Let's see if I can find raw data on these woofers

 

[Roy_D]

That looks like the WG148 to me, honestly not a fan of them for the same reason I don't like the WG300, there is a pretty tall wall surrounding the dome which tends to generate some complex reflections. I believe you can see these reflections as response issues at ~7khz and ~12khz in the data. I think waveguide tech is kind of past these sort of waveguides where they aren't designed around the driver. Stuff like waveguides from somasonus tend to not have these problems and I would pursue those OP if you're after a waveguided tweeter speaker.

oh thank you for the link to somasonus !
The SB Acoustics SB26STAC tweeter looks like a very good value performer. Add the somasonus 6" circular wg with ps.


Good grief what I am setting myself up for

 

[Roy_D]

ZinAmp | Crossovers

interesting

 

[Roy_D]

Is there a webpage or book that sort of guides one along the tiers of importance in how to go about assembling a speaker?
I think I may purchase the Soundblab plans as a base to work from but to work through the hierarchy of where to start and finish from bafflestep to diffraction, the damping and tuning of the passive radiator, how to take the raw driver data and re-calculate the raw driver data into a crossover.
I was looking at that better sb acoustics tweeter, the somasonus 6"wg with ps, up one tier of woofer to match its performance and control.
I have time but not unlimited funds ^^ Chronically ill stay at home dad on disability with some skills building, soldering etc..

 

[Ktacos]

You might wanna just start with a kit and learn the construction process that way. Something like a c-note would offer great performance for very minimal cost and teach some basic assembly skills and some finishing skills depending on what you go with. Other than that, you mostly just look at as many builds with the construction process documented and try to learn from them. Troels Graveson has some really good pages with tips on how to build.

The hard part is really analyzing the speaker for filter design, that can take a long time to develop a good consistent process.

 

[Roy_D]

You might wanna just start with a kit and learn the construction process that way. Something like a c-note would offer great performance for very minimal cost and teach some basic assembly skills and some finishing skills depending on what you go with. Other than that, you mostly just look at as many builds with the construction process documented and try to learn from them. Troels Graveson has some really good pages with tips on how to build.

The hard part is really analyzing the speaker for filter design, that can take a long time to develop a good consistent process.

Wow those C Notes by the looks of it perform amazing for the price!

I am spoilt by having enjoyed some great speakers over time that all cost me less than 3 of these c note kits ; B&W Matrix 3, Kef Reference Three ~Two, and currently now 6 yo tiny in comparison Kef Q100's and a little Audio Pro sub with their push pull system.
Coaxials are fantastic. The F6B are so tempting but a difficult proposition currently and so my thoughts wandered to self built. Construction would be enjoyable not a challenging. But the formulaic design works I have no experience with yes.. Perhaps the Q1 Meta deal I passed on recently and the mods to that are more attainable in the shorter term and do as you say meanwhile absorb the other aspects over sometime before taking on more