Hi all new here,
Recently heard a real Synergy and was very impressed with the overall response, clarity, and imaging. So I have been reading up on them through Tom's white papers, patents, this forum, etc. However, more than likely I am completely off on a few things
Here is what I think I understand
- General premise is to create a speaker that has an even response, is time aligned, and has good directivity all from one box to create essentially a point source?
- Efficiency is increased given it is a horn design and the fact that every time we double up the matching drivers we gain about +6dB. This is due to how well they couple when done right?
- In order to get the desired coupling, the matching drivers need to be physically spaced within 1/4 wavelength of their upper cutoff frequency? So a 3kHz upper limit would mean the drivers need to be within 1.115" of each other?
- This 1/4 wavelength rule also applies to the distance from the driver to the throat?
- Generally speaking, smaller the better for most drivers to accommodate the 1/4 wavelength rule?
- Because the mid ranges and woofers do not enter at the apex, some of the sound will fire back towards the throat before it goes out? This phase/timing difference is part of the reason we need to follow 1/4 wavelength?
Here are the things I definitely do not know
- Where do you measure the distance for 1/4 wavelength? Center of driver? Center of Port? Some other location?
- How are the port sizes calculated?
- Do ports need to be in the corner?
- Could you use a Planar tweeter with Synergy? Would it need a rear chamber or cup?
- In the patent Tom talks about the crossover frequencies and the physical distance between them and trying to find the "correct" cutoff. Is this to due with time delays? Why would I want my cutoff here?
Sorry for the wall of text. Just very curious!
Recently heard a real Synergy and was very impressed with the overall response, clarity, and imaging. So I have been reading up on them through Tom's white papers, patents, this forum, etc. However, more than likely I am completely off on a few things
Here is what I think I understand
- General premise is to create a speaker that has an even response, is time aligned, and has good directivity all from one box to create essentially a point source?
- Efficiency is increased given it is a horn design and the fact that every time we double up the matching drivers we gain about +6dB. This is due to how well they couple when done right?
- In order to get the desired coupling, the matching drivers need to be physically spaced within 1/4 wavelength of their upper cutoff frequency? So a 3kHz upper limit would mean the drivers need to be within 1.115" of each other?
- This 1/4 wavelength rule also applies to the distance from the driver to the throat?
- Generally speaking, smaller the better for most drivers to accommodate the 1/4 wavelength rule?
- Because the mid ranges and woofers do not enter at the apex, some of the sound will fire back towards the throat before it goes out? This phase/timing difference is part of the reason we need to follow 1/4 wavelength?
Here are the things I definitely do not know
- Where do you measure the distance for 1/4 wavelength? Center of driver? Center of Port? Some other location?
- How are the port sizes calculated?
- Do ports need to be in the corner?
- Could you use a Planar tweeter with Synergy? Would it need a rear chamber or cup?
- In the patent Tom talks about the crossover frequencies and the physical distance between them and trying to find the "correct" cutoff. Is this to due with time delays? Why would I want my cutoff here?
Sorry for the wall of text. Just very curious!
Yes, this is a common experience to hearing well implemented multiple entry horns (MEHs)...
Yes.
Horn loading is the major effect...the acoustic transformer effect is the dominant factor, while array gain is secondary--similar to boundary gain close to a room boundary.
If you were trying to cross to a midrange driver at such a high frequency as that...yes. But I don't recommend crossing at that high frequency in a MEH, however. It makes a lot more sense to use a coaxial driver (like a Faital Pro 6HX150) or dual diaphragm compression driver (like a BMS 4592ND). The reason for using a midrange driver in commercial MEHs is power output, not acoustic fidelity for home hi-fi application.
More reasonable crossing frequencies on a multiple entry horn are in the range of 1000-1500 Hz, IMO, and lower frequency crossovers are definitely an advantage--like around 500 Hz or even lower. There are a couple of reasons for this.
That's saying the same thing as above for crossing to a midrange driver from a high frequency compression driver.
I don't know how you equate this--other than expediency in implementing the design. I personally avoid using a midrange driver and instead would use a full-range compression driver, then cross over to a pair of woofers. It's simpler, more effective acoustically, and probably cheaper overall (considering the crossover problem and implementation costs).
"By George, I think he's got it..."
The acoustic center of the apex-mounted compression driver and off-axis acoustic drivers (including woofers). Generally, this is fairly close behind the plane of the port penetration on the front surface of the horn. What goes on behind that point--stays behind that point and the horn. Rejection of higher-order acoustic dynamics behind the acoustic center is surprisingly complete. That's one of the strengths of the whole design approach. This of course doesn't include non-linear pistonic motion of the driver's diaphragm, which does come into play on the front side of the horn.
I use a rule of thumb: a compression ratio of no more than 10:1. YMMV. There are implications of efficiency loss and viscous effects of air through finite-length ports if taken too far. Read your Beranek and Olson on this subject of horn efficiency for a "first-order understanding" of why I mention this.
If you worry about things like consistent off-axis coverage vs. frequency, the answer is yes...in my experience.
Why? Compression drivers were invented to handle the more robust acoustic loads seen by horn-loading.
Generally speaking, planar drivers aren't up to the task: search on this topic--you'll see it discussed elsewhere. There are other issues related to using planar drivers on horns. They're not up to the task, in my experience. Another point of the discussion is "why?"...what are you trying to gain?
I think that Mr Danley didn't actually put any details of his MEH crossover design in his second MEH patent (US8284976) or his first (US6411718...now expired).
In any case, crossover filters need to be designed to handle the physical acoustics, including the delays so that there isn't a time delay issue at the crossover interference band. IIR filters introduce phase delays on the lower frequency drivers that is a function of their "order": minus 90 degrees of phase for every order past first order. These delays must be accounted for in the crossover filter design and the relative distance of the MEH ports (axially) with respect to each other.
Chris
One thing that I didn't mention: the notch effect of the off-axis ports begins at the edge of the port closest to the throat of the horn and extends downward in frequency until the far edge of off-axis port is reached. For elongated off-axis ports, this means that you place the slot starting at the 1/4 wavelength highest frequency that you intend to use as the low pass point (acoustically) and design your electrical crossover based on the point, i.e., no higher in frequency.
You can also move that crossover frequency lower to have the higher frequency driver overlapping the 1/4 wavelength port point, but no higher. This is calculated as the 1/4 wavelength along the central axis of the horn (i.e., NOT the slant range along the wall from the throat to the off-axis port).
I've found that this works quite accurately: I created a 1/4 wavelength port corresponding to 475 Hz to the leading edge of the off-axis port, and that's exactly what I got acoustically.
You can also move that crossover frequency lower to have the higher frequency driver overlapping the 1/4 wavelength port point, but no higher. This is calculated as the 1/4 wavelength along the central axis of the horn (i.e., NOT the slant range along the wall from the throat to the off-axis port).
I've found that this works quite accurately: I created a 1/4 wavelength port corresponding to 475 Hz to the leading edge of the off-axis port, and that's exactly what I got acoustically.
The sound fire backwards is insignificant. This is because the horn will not support these frequencies. If you were to measure with a microphone inside the horn, the only frequencies that you will measure will be those that the horn will support at that point in the horn. Even if you have a compression driver that will go to 300hz, you will not be able to measure 300hz until you get to the point in the horn that will couple that frequency.
I disagree, my opinion is that the 4592 sounds better at 400hz than a cone.
This is because the 4592 is lighter, faster, and has less front to back than a cone, so a more refined sound. If you like *warm* or *phat* I suppose a cone is for you. I like refined, quick responding, and dynamic.
I don't follow all of the synergy threads, so this is the first I have heard of the 10/1 ratio. Good rule, if sound. I'm assuming this is cone area vs. port area. I remember Tom chatting about port configuration some time ago, and opining about oval vs round. Has this conversation evolved? Or do most people use round? (It seems most of the pictures I have seen are round.)
No need to overthink this one, just copy what works. Here are the locations of the woofer and midrange taps in the SH-50:
Midrange taps are 3/4" in diam, 3.5" from throat
Woofer taps are 2.5" in diam, 10.5" from throat
Woofer ports are 2.5" in diam, 14.5" from throat
If you'd like to buy the speakers that I measured, they are for sale:
Pair of (2) Danley SH50 Speakers | eBay
(not mine; I rented these.)
Midrange taps are 3/4" in diam, 3.5" from throat
Woofer taps are 2.5" in diam, 10.5" from throat
Woofer ports are 2.5" in diam, 14.5" from throat
If you'd like to buy the speakers that I measured, they are for sale:
Pair of (2) Danley SH50 Speakers | eBay
(not mine; I rented these.)
Did you analyze the passive crossover network in the SH-50?
Since most of the threads that I read on DIY MEHs use active crossovers, this isn't a big deal. But it might bring a little more illumination to this subject area which was not actually described in detail in either patent.
The late great Jean LeCleach solved that one for us:
http://www.melaudia.net/zdoc/jml_crossovers_etf04.pdf
If you want to make a Synergy Horn, the solution is in that pdf.
That's a big part of the 'secret sauce.' I see people obsessing over the small stuff, but a big part of getting it to work is that crossover. IMHO, the crossover is the biggest difference between the Synergy Horn and Unity Horn.
So you're implying that you know the passive network design...but aren't going to describe it for the SH-50...other than alluding to phase/group delay flattening using the common devices at our disposal (inverting phase of LF driver, perhaps adding HF channel delay and also use of all-pass filters).
That's okay, I don't believe that it is really that important what the specific implementation actually is, but rather the net effect that is being cultivated or maximized by the network.
I think that the audibility differences of phase distortion are much better described by David Griesinger than in the presentation by Le Cléac'h (YMMV). There are a couple of really important concepts related to preserving phase in listening spaces, starting at about slide 3, then slide 17, and especially at slide 21 of the linked presentation. The other approaches used by Blauert & Laws (et al.) I find don't describe the "why" in this subject area like Griesinger has described. It's the "why" which is most important. The "how it's implemented" can vary enormously I've found and still achieve the same results--in this subject area.
Flattening the group delay curve is a big deal as I found using DSP-based crossovers having easily selectable crossover filter type/orders and delay settings on the K-402-MEH for immediate comparisons. If you've got access to a DSP-based loudspeaker processor, I would recommend other approaches in addition to the ones found in Le Cléac'h's presentation.
It is possible to hear those difference in not only clarity but the pleasantness of the resulting timbre shift, which can be described more in terms of a shift in the loudspeaker's character and neutrality of its sound.
It's not really all that subtle to my ears once you get control of the phase growth of the loudspeaker. Getting an MEH dialed in and compared to say, a pair of Jubilees on each side of the loudspeaker in terms of timbre and clarity, is something that really separates the wheat from the chaff.
But I find in most things related to hi-fi loudspeakers, driving electronics, and listening rooms, there seems to be a large fraction of the listening public that is really quite insensitive to these effects (as I've witnessed in listening group sessions). Being a musician by training, I know firsthand that the human hearing system is actually trained. In music, it's called "ear training" for detection of musical intervals and perhaps absolute pitch, and I've found that other aspects of hearing are also trained, as the Harman/JBL training of "certified listeners" has demonstrated. Some/many people cannot detect these effects, I've found. Given training however I think that the story would be much different.
Chris
That's okay, I don't believe that it is really that important what the specific implementation actually is, but rather the net effect that is being cultivated or maximized by the network.
I think that the audibility differences of phase distortion are much better described by David Griesinger than in the presentation by Le Cléac'h (YMMV). There are a couple of really important concepts related to preserving phase in listening spaces, starting at about slide 3, then slide 17, and especially at slide 21 of the linked presentation. The other approaches used by Blauert & Laws (et al.) I find don't describe the "why" in this subject area like Griesinger has described. It's the "why" which is most important. The "how it's implemented" can vary enormously I've found and still achieve the same results--in this subject area.
Flattening the group delay curve is a big deal as I found using DSP-based crossovers having easily selectable crossover filter type/orders and delay settings on the K-402-MEH for immediate comparisons. If you've got access to a DSP-based loudspeaker processor, I would recommend other approaches in addition to the ones found in Le Cléac'h's presentation.
It is possible to hear those difference in not only clarity but the pleasantness of the resulting timbre shift, which can be described more in terms of a shift in the loudspeaker's character and neutrality of its sound.
It's not really all that subtle to my ears once you get control of the phase growth of the loudspeaker. Getting an MEH dialed in and compared to say, a pair of Jubilees on each side of the loudspeaker in terms of timbre and clarity, is something that really separates the wheat from the chaff.
But I find in most things related to hi-fi loudspeakers, driving electronics, and listening rooms, there seems to be a large fraction of the listening public that is really quite insensitive to these effects (as I've witnessed in listening group sessions). Being a musician by training, I know firsthand that the human hearing system is actually trained. In music, it's called "ear training" for detection of musical intervals and perhaps absolute pitch, and I've found that other aspects of hearing are also trained, as the Harman/JBL training of "certified listeners" has demonstrated. Some/many people cannot detect these effects, I've found. Given training however I think that the story would be much different.
Chris
Thanks for posting that. While we're here. I was really impressed by that little LeCleach/Synergy horn you posted on the Tube. Even on my puny little laptop speaks it was clear that it had a richness I wouldn't have anticipated by looking at its size. Did you take it any farther?
Great question!
Theoretically, one of the advantages of a LeCleach horn versus an oblate spheroidal horn is improved impulse response. You can see this in hornresp; model the two horns and you'll see that the impulse of the LeCleach is way better.
But in the real world, I found that the JBL PT waveguides had very good impulse response.
I wish I had the measurements handy, but IIRC the PT waveguide had superior impulse compared to an OS waveguide, and not quite as good impulse response as a LeCleach.
Basically the PT waveguides are a nice compromise between the two designs. These days nearly all of my waveguides are clones of the PT waveguides or OS waveguides. They just work.
I wasn't trying to be coy, just suggesting that when I read the LeCleach paper it was one of those "aha" moments.
Sixteen years ago I started studying the Unity horns when Tom Danley was on the basslist. I made my first one five years later. I didn't begin to comprehend the crossover requirements until about five years ago.
I'm not the sharpest tool in the shed, so many will figure it out in less than ten years, but it took me a while.
Note that you CAN make a phase coherent Synergy horn using a conventional topology. When the drivers are so close together, it's quite easy to get the phase right.
The nice thing about the LeCleach crossover is that it's a recipe, just follow his instructions and you will get the results you are looking for. There is virtually no trial and error.
BTW, I don't think there are any all-pass components in the Synergy Horn crossover.
@ Patrick Bateman
Thanx for the LeCleach PDF link Interesting reading etc, in spite of spelling mistakes & a few funny phases that don't translate well But he is French
Re the pic in Post #16
Is that your build, Or ? The coils are Very close together, so @ mid to high SPL there is a chance they could interact with each other. Even though some are positioned differently !
Still, good info etc as usual
Thanx for the LeCleach PDF link
Interesting reading etc, in spite of spelling mistakes & a few funny phases that don't translate well But he is French Re the pic in Post #16
Is that your build, Or ? The coils are Very close together, so @ mid to high SPL there is a chance they could interact with each other. Even though some are positioned differently !
Still, good info etc as usual
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.