The dome midrange thread

[howardg]

I just got ahold of a pair of these DSM50FFLs in a trade deal

Do you have any way of knowing if these would be representative of a random new pair from an official distributor? You liked the older versions, right?

 

[profiguy]

@howardg Yes, I like the older version better. This newer redesign shows some skimping in important areas. The faceplate is flimsy and the back chamber is resonant at Fs, but that can be easily dealt with using some surface dampening. The FF is easy to deal with, as I'd remove it regardless of application. It can always be added in later. As far as the overall FR and performance above 1k, it's the same as the old one. It does sound very good overall. If I'd compare the DSM50FFL to the Dayton 2" metal mid dome, its quite a step up from that one. It doesnt have the typical Ti breakup most other domes have. Be aware that the replacement domes they sell don't fit the older version. The connector tabs are placed differently. I dont know the cutoff date for the changeover of diaphragm styles.

 

[Azrael]

The FF is easy to deal with, as I'd remove it regardless of application.

I'm also thinking about this for my pair, especially as the finished speaker can't handle brutal levels as it will only be equipped with an 8-inch woofer in BR. The cooling properties of FF may then not even be necessary.

In any case, I'm curious to see what will become of your pair. 🙂

As far as my pair is concerned, I'll see if I can at least test the match of my two specimens with impedance measurements in the next few days. Incidentally, I bought the two DSM50FFLs second-hand, but so far I've only measured RE to make sure they're not broken.

I find the shape of the faceplate comparatively difficult to rout in. I could probably manage it somehow, but I could also look around for two defective G50FFLs, as their round front panel supposedly fits the naked DSM50FFL. Or I could ask nicely at Visaton.... 🙂

 

[howardg]

I find the shape of the faceplate comparatively difficult to rout in. I could probably manage it somehow, but I could also look around for two defective G50FFLs, as their round front panel supposedly fits the naked DSM50FFL.

I do not assume that anybody in particular, (including you, Azreal😉) do not know this technique; but I put it here for those who have not tried it yet. This is a small shop technique for those of us without a CNC router:

A router flush trim bit that has its bearing above the cutting edges can be called a "pattern bit". If one makes a pattern of sufficient thickness, a driver rebate (or other cut-out) can be made by attaching the pattern on top of the work and running the top bearing along the pattern. These are also good for exactly repeating a cut-out multiple times.

Patterns of unusual shape can be tedious to make----much more skill and patience involved (and sometimes some plastic body filler to fix mistakes) than round patterns; but tracing the face plate and using appropriate wood bits for the radiused corners can get one most of the way there.

 

[Bryguy]

 

[mV8]

@IamJF We definitely disagree on the T34A/B HP. A.good number of speaker designers would side with you on this one, but I won't. In an elaborate way of putting things, here's why....

My big beef with lower tweeter HP points is I can hear the tweeter do things I don't like even though the HD measurements on paper. The T34B doesn't sound good to me under 3k... at least my own personal ones don't. Same is true with many other larger domes. Just because they're large domes doesn't make them capable of playing lower/louder without some form of audible issues, even if the HD data says to the contrary. There are many possible reasons for this, not just related to a large diaphragm.

A big issue with large domes is they tend to do non linear things which can't be easily observed with the common driver testing methods and devices. Laser interferrometry can detect this, but its not a very accessible test method to most designers. Simply put, I don't like a single suspension driver moving alot. Thats why I prefer the most emissive surface area possible with the least diaphragm movement. "Low enough" HD isn't "good enough" to me, which is still noticeable with specific musical content. This is unlike tone/pink/white noise bursts or static test signals which don't exactly/fully represent all of the possible negative performance traits and behavior from a driver. There are some things we can't detect with typical measurment gear and methodology.

Next issue for me is crossing a driver in a sensitive region, denying its full bandwidth potential with flatter phase trend, specifically having a single driver delivering the entire sensitive range uninterrupted. This is a huge benefit on a 3 way, likely the biggest benefit of all. Sticking a crossover point in there, especially with a steep filter that throws the phase around is exceptionally destructive to imaging, phantom center and soundstage. Percussive transients will suffer as well,.depending on how accurate they're preserved in the recording.

Directivity isn't everything in the real world to me and its not as critical as one would think - certainly not worth negating all the benefits I listed above. People make such a huge deal about this, it becomes an obsession. You can still have good (non-optimal) off axis performance using the tricks I suggest and reap the rewards of a more time domain accurate speaker without the need for DSP or hammer fisted, sharp filter cutoffs. Yes, you can have phase correct filters this way, but there's no free lunch ie. pre-ringing and artifacts related to processing errors.

The way a speaker performs playing back music is significantly different than reproducing test signals. Specifically transient IMD isn't well represented with just a few test frequencies. The full audio spectrum from music is a far more accurate indicator of IMD in every way. Yes, chances are, if there are problems in the usual test signal domain, there will be even more issues with complex music containing countless harmonic structures and their relationships.

Bottom line is, the SOLE purpose of a 3 way speaker IMO is so the critical midrange from circa 500 hz - 4k can come from ONE single driver. If you split that range, you may as well build a 2 way. Even if you don't have perfect off axis performance with a tighter sweet spot, you can still treat the room properly so it doesn't disrupt soundstage, imaging and phantom center, even with a relatively (on paper) less than ideal power response.

As I said before, unless I'm building a flat sounding monitor or PA speaker (which I'm not), I'm always going to put my priorities first before perfect off axis coverage in multiple, acoustically compromised listening environments. This compromise would involve hacking up other areas of performance which are IMO far more important to enjoyable, non fatiguing, yet convincingly accurate sounding music reproduction for the majority of scenarios and listeners tastes. Yes, I'm mainly approaching this from a psychoacoustic perspective. And I'm not going to quote GL metrics, etc... sorry. Its a bit more complicated that that.

Alot of this obsessive off axis perfection trend is spilling over into PA design. I believe its a big part of why most live concerts sound like **** these days despite all of the so called "advances" in sound reinforcement we've supposedly made. These advances are shifts in priorities that have pushed the most important aspects of SQ aside when it comes to the execution of most modern PAs. Its supposed to be important to provide even, intelligible coverage across the venue, but its come at a huge cost and consequence to the integrity of the actual SQ. Its far from actually delivering transient correct, truly low distortion sound that doesn't make peoples ears bleed. Then there's the LF... thats a whole different mess thats just pathetic. I'm a bassist, and when I go to a concert, I want to be able to discern the individual notes, not have my bowels oblitered with a muddled LF cacophony....

Anyways, there's way more to accurate sound reproduction than mainly even coverage, and certainly not at the cost of increased harshness, fatigue and odd types of distortion which don't exist in even the most primitive 2 way speakers. Combing, smearing and transient distortion from highly diffracting WGs is apparently more acceptable...

I design and build speakers for the purpose of enjoying most types of music, delivering the most convincing accuracy possible while having the lowest possible amount of relative distortion, fatigue and stress to hearing. That doesn't mean putting a smily face EQ on everything and calling it good. Its much more complex than that. There is no cookie cutter formula for this. Specific parameters are more important than others to make this work successfully and convincingly. This is why I look at designing speakers from a wide perspective, not just from gospel coming from engineers wanting to hear themselves talk about how they figured out some special formula that explains it all black and white. Sorry, that simply doesn't exist. Sure, I have my own prejudices and methods of subjective approach, but I try to focus on the most important stuff first. So, is off axis response really that holy grail of speaker design?

Sorry, I had to take the ball and run with this one. I just don't agree and thats why I'm trying to present other options to people who want accurate sound that doesn't offend in other ways. Speakers are like chairs for people... there's one for every @$$.

The audio world is a world of the greatest contradictions. I have indeed heard a lot of different top hiend systems, with very different solutions. I've also been building speaker systems myself for years, based on speakers like and top speakers like Scan Speak Ellipticor or Bliesma beryllium drivers, but by far, the absolute best speakers I've listened to were built on Accuton diamond domes, and the crossover filters used were ultra-steep. Below is a test of the exact same pair of Gauder speakers that I had the great pleasure of listening to.

https://soundrebels.com/gauder-akustik-berlina-rc11-black-edition-english-ver/

 

[Bmsluite]

I find the shape of the faceplate comparatively difficult to rout in. I could probably manage it somehow, but I could also look around for two defective G50FFLs, as their round front panel supposedly fits the naked DSM50FFL. Or I could ask nicely at Visaton.... 🙂

My solution for this has been 3d printing a template. I then stick it to the wood and then just sink in my router bit by bit. Perfect size/shape every time

 

[howardg]

My solution for this has been 3d printing a template. I then stick it to the wood and then just sink in my router bit by bit. Perfect size/shape every time

If designing and printing is easier for the crafts-person, than is creating the template by more manual methods, I can see how it would be preferred. For me either type of template seems easier to use/create than the one in the video, but each person approaches a task in the way that appeals to them. The method in the video did seem to get good results.

 

[Bmsluite]

If designing and printing is easier for the crafts-person, than is creating the template by more manual methods, I can see how it would be preferred. For me either type of template seems easier to use/create than the one in the video, but each person approaches a task in the way that appeals to them. The method in the video did seem to get good results.

I do CAD daily and 3d print daily for my job. I could see how learning these skills would be more difficult to learn, however, it takes me under 60 seconds to model up a template, slice it, and send it to the 3d printer. Then I just go do something else for 20 minutes, come back, and I have a perfectly accurate template. Once you have the skills, its impossible to beat when comparing work put in to effect.