Philharmonic Audio BMR Speaker Discussion

[GelbeMusik]

... figure out a way to make it work, be it with a port, transmission line, or ...

Since I use digital room correction, I don't care about bass alignment too much. Reflex always wins. Only after I correct in-room response. Tuning is all the same every time: just about 35 to 45 Hertz for reasons You know all to well.

I was referring to "chuffing" of the port once the excursion with the Purify multiplies.

You were clearly promoting the Tymphany TC9FD18-08 as a step up from the BMR, and I was pointing out that you might wish to do some cross checking on sensitivity. Also, can you point me to documentation for your claim that the little Tymphany has "drastically less distortion"?

Nope. It is DIY. The TC was advertised by so many, You couldn't have missed it. Regarding sensitivity I could tell a lot. Only that most of the readers here wouldn't be able to follow it. Let's try: The SB17NAC35-4 (4Ohms) would go with the TC. With 90dB@2,83V on infinite baffle, corrected for baffle step only 84 would be left. The TC would have a) bandpass gain, b) gain from baffle directivity. And it is smooth as moonlight, so no delicate corrections, hence not much of reserve needed. The TC with bare 84,5dB fits easily. Even more so with the 8-Ohm variant of the SB. You know that.

You have many followers here. I don't want to learn more.

Regarding dispersion BMR wins, also on power handling/compression (32mm voice coil in BMR vs 20mm in TC9). Given that crossover frequency went up tp 900Hz, it may not be an issue, but pairing tc9 with RAAL 64-10 at 3.5kHz-4kHz might.

Nope. No considerably better disperion than a larger wideband TV speaker, the TC9. Available volume is limited by distortion / intermodulation, not by power. Regarding XO: with waveguide I dare to use even 25mm tweeters from 1kHz on. Only lately I explored "wide dispersion". 7" bass/mid and a naked 1" tweet on top--again from 1kHz on. No distortion, no burnt voice coil with elevated levels.

 

[Zvu]

When i say that BMR wins, it means that it has wider dispersion. And no surprise there since we are comparing 3" driver (tc9; Sd=36cm^2) and 2" driver (BMR; Sd=19cm^2).

I guess there's a language barrier because i don't quite understand what are you saying. I was talking about power compression and power handling. Not maximum volume at which the BMR and TC9 will sound clean. I would have to measure them both in same conditions to know that.

And about that 7"/1" two way, you should give some more data. What drivers at what voltage and with what signal? At this point it sounds quite unbelievable.

 

[GelbeMusik]

When i say that BMR wins, it means ...

But the TC9 remains my personal favorite. Even with dispersion.

I guess there's a language barrier because i don't quite understand what are you saying. I was talking about power compression and power handling. Not maximum volume at which ...

Exactly. Only that I explicated on that in saying, that cmprssn and pwrhndlng were limited by dstrsn,

And about that 7"/1" two way, you should give some more data. ... At this point it sounds quite unbelievable.

What would You do with the info? Adire Audio 6.8 closed with Peerless DX25TG09-04 XO'ed at 1kHz 12db/oct Butterworth (digital) arbitrarily augmented by two JBL 2206 subs. It worked for high fidelity-ish spl, say 84db (A) plus dynamics of unknown value, analogue. Example Meridith Monk, Dolmen Music, early, authorized recording (ECM, Germany). So? The tweet ran at about 0,1..0,2 Watts on average ...
as far as my estimative calculations hold.

 

[Dennis Murphy]

Since I use digital room correction, I don't care about bass alignment too much. Reflex always wins. Only after I correct in-room response. Tuning is all the same every time: just about 35 to 45 Hertz for reasons You know all to well.

I was referring to "chuffing" of the port once the excursion with the Purify multiplies.



Nope. It is DIY. The TC was advertised by so many, You couldn't have missed it. Regarding sensitivity I could tell a lot. Only that most of the readers here wouldn't be able to follow it. Let's try: The SB17NAC35-4 (4Ohms) would go with the TC. With 90dB@2,83V on infinite baffle, corrected for baffle step only 84 would be left. The TC would have a) bandpass gain, b) gain from baffle directivity. And it is smooth as moonlight, so no delicate corrections, hence not much of reserve needed. The TC with bare 84,5dB fits easily. Even more so with the 8-Ohm variant of the SB. You know that.

You have many followers here. I don't want to learn more.



Nope. No considerably better disperion than a larger wideband TV speaker, the TC9. Available volume is limited by distortion / intermodulation, not by power. Regarding XO: with waveguide I dare to use even 25mm tweeters from 1kHz on. Only lately I explored "wide dispersion". 7" bass/mid and a naked 1" tweet on top--again from 1kHz on. No distortion, no burnt voice coil with elevated levels.

I doubt that many people are reading this exchange. You seem more intent upon showcasing your knowledge and equipment than providing constructive criticism of the BMR. The statements you're making are much more complex in real speaker life. For example, a speaker designer would almost never build in a full 6 dB of baffle step compensation. That would give you very heavy bass in a real room with normal midbass gain. I certainly don't use that much, and I use more than most commercial designs. Also, you usually can't rely on factory specs for woofer sensitivity. So much depends on the baffle they're using, at what frequency it applies to, how smooth the inherent response is, and their resistance to marketing pressures for inflated specs. And, depending on where you cross in a 3-way, some or even all of the baffle step compensation will have to be applied to the midrange, not the woofer. So starting out with a mid with 84.4 dB sensitivity (assuming the spec is reliable) will usually yield a low system sensitivity. But enough. I think I've been able to meet my design goals with the BMR, and life's too short to second or third guess every detail. There are only so many drivers I have time or the budget to work with.

 

[Zvu]

.........What would You do with the info? ..........

It worked for high fidelity-ish spl, say 84db....

Analyze it. When you said elevated levels i thought something more like 95dB/m.

...For example, a speaker designer would almost never build in a full 6 dB of baffle step compensation. That would give you very heavy bass in a real room with normal midbass gain. I certainly don't use that much, and I use more than most commercial designs....

How is something measuring like this not a full BSC design ?

 

[HooStat]

Just a naive question here, but would it be possible to use a more traditional tweeter in the BMR and still get similar results? Or is the ribbon + BMR an ideal combination? I am asking because the ribbons are expensive and it would be nice to play around with something less expensive.

 

[Zvu]

......would it be possible to use a more traditional tweeter in the BMR and still get similar results?

If wide dispersion over 10kHz (low directivity) is what you're after, than no.

But you can always play and no one says it will not sound good if done properly.

 

[Dennis Murphy]

Analyze it. When you said elevated levels i thought something more like 95dB/m.



How is something measuring like this not a full BSC design ?

That's an excellent question. BSC refers to how much you level out the 6 dB rise in a woofer's (or in come cases a mid's) anechoic response between around 400 Hz and 1 kHz. If you level out all of it, then the simulated anechoic response of the woofer will look flat up to it's low pass roll-off point, but the room system response will be elevated in the bass, partly from any bandpass gain from the mid and mostly from room reinforcement as you start to get into the bass region. The first plot plot you posted is my Praxis measurement of a BMR. Praxis starts to transition from an anechoic measurement to a room measurement at about 500 Hz, which accounts for the rise in output at that point. If I had used 6 dB of BSC and matched the mid and tweeter response to that level, the midbass response would look considerably more elevated on that graph. The second set of plots was made by James Larson of Audioholics outside with the speaker on a tall platform. It shows the bandpass gain from the midrange, but no room effects. I have no experience with this kind of testing and don't really know how to compare it one-on-one with the Praxis measurement. About the only thing I know for sure is that the indicated 2 dB rise at 9000 Hz doesn't show up in the NRC anechoic chamber measurements or in my Praxis plots, and I suspect that is a calibration issue with James' mic.

 

[Zvu]

..... The second set of plots was made by James Larson of Audioholics outside with the speaker on a tall platform. It shows the bandpass gain from the midrange, but no room effects. I have no experience with this kind of testing and don't really know how to compare it one-on-one with the Praxis measurement...

Second measurement is free field quasi anechoic and shows how the loudspeaker measures without any boundary reinforcement. In that measurement everything is flat down to 250Hz which looks like full baffle step to me since baffle step effect for that width of the front baffle will extend down to about 300Hz.

I'm not worried about 9kHz rise since you've explained earlier that it may well be measuring artefact.

 

[Zvu]

Same flatness is observable in your NRC measurements down to 200Hz, if we disregard measurement artefacts of their chamber happening bellow that.

 

[Dennis Murphy]

Same flatness is observable in your NRC measurements down to 200Hz, if we disregard measurement artefacts of their chamber happening bellow that.

View attachment 67274

Same flatness is observable in your NRC measurements down to 200Hz, if we disregard measurement artefacts of their chamber happening bellow that.

View attachment 67274

That's not really true. There is a definite rise in the response up to 1000 kHz. As for the Audioholics plot, I'm certainly pleased with the results, but I can't attest to the absolute validity of the measurements in the midbass. The BMR isn't a very good vehicle for a discussion of woofer BSC, since the Scan is crossed over to the midrange before the baffle step occurs (it's 6 dB down by the time it reaches the Xover point of 600 Hz). . This discussion would be much cleaner for a 2-way crossed around 2 kHz, where all of the BSC would be applied to the woofer and there would be little or no bandpass gain from the mid in the 450 Hz - 1 kHz region.

 

[Zvu]

That's not really true. There is a definite rise in the response up to 1000 kH.

If there is a rise on NRC measurement i can't see it. Maybe that's a product of their grid resolution but i really can't.

.... the Scan is crossed over to the midrange before the baffle step occurs (it's 6 dB down by the time it reaches the Xover point of 600 Hz). This discussion would be much cleaner for a 2-way crossed around 2 kHz, where all of the BSC would be applied to the woofer and there would be little or no bandpass gain from the mid in the 450 Hz - 1 kHz region.

Here's baffle step effect on my measurement of raw 6.5" woofer from Kef R300 done in factory cabinet with baffle dimensions HxW - 38.5x21 centimeters. Dimensions of BMR are HxW - 50x20,3 centimeters so the width is almost the same hence very similar baffle step effect. The response because of baffle step is about -3dB at 450Hz and not going to -6dB until 230Hz. Gate is a bit over 10ms so we get good resolution to about 150Hz.


https://i.postimg.cc/gcyPzPfG/image.png

In my book, full baffle step loudspeaker is one that measures flat from 200Hz-1000Hz anechoic regardless of crossover points and mutual influence of drivers. That is precisely what i see on BMR Philharmonitor measurements so that's why i call it full baffle step design. Your terminology may differ, of course.

 

[GelbeMusik]

That's not really true. There is a definite rise in the response up to 1000 kHz. As for the Audioholics plot, I'm certainly pleased with the results, but ...

All the back and forth is spawned by a simple question: would a midrange driver of some 84dB output for 2.83V input match a regular 6..7" bass driver in a regular enclosure?

Answer: may be, or not. It comes back again to a matter of taste. It cannot be disputed. Thank You!

Background is the decision to use the eponymous BMR driver despite its higher distortion and hence a quite high xo.

This again was spawned from the question, if the BMR driver has better dispersion in comparison to a regular wideband driver.

Every design can be discussed to death. Because there are so many compromises, and, admittedly, taste comes into play.


My final point:

The measurements, that were quoted show, in the details, kind of arbitrary results. Missing calibration, reference axis etc.. How could one "optimize" crossovers with changes of about +/-1dB? Wasn't it just an adoption of the speaker to a specific measurement environment?

(It is a rhetorical question, answer not expected.)

I strongly recommend again as several pages before, to buy, steal or borrow a measurement microphone, and adjust the in-room response using an EQ to ones liking. Taking the measurements as a guideline only. Taking the Olive tilt as a guideline only. In case the overall design of the speakers is o/k-ish, that will surely work. Critical parameters for the design: dispersion and available sound power output.

Dispersion: most seem to like wider dispersion. But still "wide" can be implemented wide to very high frequencies, or narrowing increasingly--I prefer a bit more narrow dispersion from the lower mids on, but that being maintained constant up high

Power: for most of the audience 84dB is quite loud. It is about four times the level of a personal conversation. But additional 20dB dynamics could be achieved with very rare, exceptional recordings. Clearly neither with conventional pop nor rock music. A well implemented 8" bass / mid driver should do it for most people. To go smaller in size is starting to cut corners. A sub wouldn't help too much, since the critical clarity of the lower mids cannot be supported by that.

 

[bogart]

@Dennis Murphy, thanks again for engaging in such depth on these questions. As the OP, it has been super interesting to learn more about this design; I have been slow to reply as there's enough technical content folks have brought up that has sent me to read other resources that I have been slow to comprehend everything.

At the moment, I am completely lost in the discussion of baffle step correction - it is a phrase that I have encountered several times before, but haven't been able to wrap my mind around.

It's interesting, another thread I've watched with some interest is the discussion of a ASR community collaboration on a DIY speaker - it seems they're looking to solve similar questions as your design. I am enjoying peeking over the shoulders of more qualified folks as they engage the in the development process.

 

[R Swerdlow]

At the moment, I am completely lost in the discussion of baffle step correction - it is a phrase that I have encountered several times before, but haven't been able to wrap my mind around.

Here are several links to articles I've found helpful at understanding baffle step and it's correction (BSC). The first two are found here: http://www.salksound.com/blog.php

This is the simplest treatment to the subject, written by Jim Salk:
http://www.salksound.com/blogtopic.php?id=2

Next is in more detail, written by the late Jeff Bagby:
http://www.salksound.com/blogtopic.php?id=7

I think the single best explanation of baffle step and BSC can be found here:
http://www.quarter-wave.com/General/General_Articles.html

It's written by Martin J. King:
http://www.quarter-wave.com/General/BSC_Sizing.pdf

And even though it isn't about BSC, here is article written by Dennis Murphy that outlines his general ideas of how he goes about designing crossovers. I think you may like reading it:
http://www.salksound.com/blogtopic.php?id=5

I hope that isn't too large a reading assignment .

 

[bogart]

Here are several links to articles I've found helpful at understating baffle step and it's correction (BSC). The first two are found here:
http://www.salksound.com/blog.php

I literally finished reading that blog post right before seeing your message. Most excellent coincidence that I started in a great spot! Overall, the Salk blog seems to be infrequently updated but with very excellent content. Thanks for the reading list suggestions.

This is definitely spot on. It's also indirectly useful, but I watched a video last night from @Joseph Crowe that was really interesting, but is more to do with baffle geometry: Circular vs. Round baffle

 

[R Swerdlow]

I literally finished reading that blog post right before seeing your message. Most excellent coincidence that I started in a great spot! Overall, the Salk blog seems to be infrequently updated but with very excellent content. Thanks for the reading list suggestions.

This is definitely spot on. It's also indirectly useful, but I watched a video last night from @Joseph Crowe that was really interesting, but is more to do with baffle geometry: Circular vs. Round baffle

Talk about serendipity .

Be sure to read the last two links I mentioned.

 

[Dennis Murphy]

Here are several links to articles I've found helpful at understating baffle step and it's correction (BSC). The first two are found here:
http://www.salksound.com/blog.php

This is the simplest treatment to the subject, written by Jim Salk:
http://www.salksound.com/blogtopic.php?id=2

Next is in more detail, written by the late Jeff Bagby:
http://www.salksound.com/blogtopic.php?id=7

I think the single best explanation of baffle step and BSC is found here, written by Martin J. King
http://www.quarter-wave.com/General/General_Articles.html

http://www.quarter-wave.com/General/BSC_Sizing.pdf

And even though it isn't about BSC, here is article written by Dennis Murphy that outlines his general ideas of how he goes about designing crossovers. I think you may like reading it:
http://www.salksound.com/blogtopic.php?id=5

I hope that isn't too large a reading asignment .

Thanks Swerd. That's very helpful. The baffle step and its correction is much easier to understand in the context of 2-way speakers where the woofer is fully responsible for the rise in output between around 400 Hz to 1 kHz. Once you get into 3 ways, things get more complicated depending on the crossover point and which driver will need to be targeted most for correction. In general, many commercial speakers don't apply enough correction because that requires larger-value components, or more of them, and it reduces the sensitivity measured at 1 kHz. On the other hand, as I stated earlier, a full 6 dB of compensation is is generally too much given the effect of room reinforcement in the midbass region. I try to base my BSC on how I think the particular speaker is behaving in a typical room.

 

[GelbeMusik]

@Dennis Murphy, thanks again for engaging in such depth on these questions. As the OP, it has been super interesting to learn more about this design; ...

So, why is it, that You want to know more. Aren't You satisfied, so that You want to roll Your own? Is it curiosity? In any case, stop reading the internet. You need a sound foundation in physics. That's a first. College physics, not more, but that should be present every time.

 

[bogart]

So, why is it, that You want to know more. Aren't You satisfied, so that You want to roll Your own? Is it curiosity? In any case, stop reading the internet. You need a sound foundation in physics. That's a first. College physics, not more, but that should be present every time.

Bit of a strange question and directive, but I'll assume it was asked in good faith.

I should say that I don't think that "In any case, stop reading the internet and go learn physics first" is particularly helpful or polite.

I am not a speaker designer, but have enjoyed hearing the variety of designs that people thought were good enough to bring to market. I am a trained engineer, but not an electrical or acoustical engineer. While I do not wish to design my own speakers, I enjoy learning what engineering problem solving took place - what were design goals, what choices were made to achieve them, and so forth.

To make an analogy: I won't design a mechanical watch, but I enjoy knowing more about the mechanics in a watch that I like and what challenges were solved to make it work. Ultimately, if it keeps good time and I like its appearance, that's all that I *need* to know. But, I enjoy learning more about the technical challenges solved and clever ideas that went into it for the sake of appreciating that specific design more.

In short, I wish to learn more because I enjoy this pursuit of faithful audio reproduction and the competing schools of thought on how to achieve it. I am grateful for the chance to discuss it directly with someone experienced like Mr. Murphy, and to observe discussions of people who have more depth in the subject as a way to understand areas for me to invest in more (applied technical) research on my own.