Hello everyone.
First I may be off topic, so don't hesitate if it's better that I open a new thread, but I figured there is a knowledgeable crowd here 🙂
I'm starting the design phase of my next main monitors (Flush mounted) for my mastering studio. Right now I have an old pair of PMCs (which I have stuffed the transmission line to render them aperiodic and have been properly re-crossed over with DSP). That pair features the famous ATC SM75-150S and I have high expectations 🙂
Anyway, I'm currently leaning to 4x Purifi PTT10 per speaker, mated with Bliesma M74B and T34B in a dual waveguide (like the KH420 or big genelec speakers).
I've setup fusion, gmsh & Akabak, and I'm going down the pithole of iterative simulations.
I've asked bliesma for the correct profiles and should get them soon to model properly. In the meantime, I started on the midrange driver (as the tweeter has already a few published designs), with a "seems-to-be-matching-visually" dome profile and simple conical waveguides + radius at the mouth (like the KH420).
From the few simulations I've done, it seems to me that :
Am I on the correct path ?
edit : BTW I did a simulation of the midrange in the KH420 waveguide with the 3D model than you can download on Neumann's website. Their model seems to have a bigger radius than the real one so I guess I should take that 3D model with a grain of salt anyway. I'll post these simulations when I'm back on my work computer.
First I may be off topic, so don't hesitate if it's better that I open a new thread, but I figured there is a knowledgeable crowd here 🙂
I'm starting the design phase of my next main monitors (Flush mounted) for my mastering studio. Right now I have an old pair of PMCs (which I have stuffed the transmission line to render them aperiodic and have been properly re-crossed over with DSP). That pair features the famous ATC SM75-150S and I have high expectations 🙂
Anyway, I'm currently leaning to 4x Purifi PTT10 per speaker, mated with Bliesma M74B and T34B in a dual waveguide (like the KH420 or big genelec speakers).
I've setup fusion, gmsh & Akabak, and I'm going down the pithole of iterative simulations.
I've asked bliesma for the correct profiles and should get them soon to model properly. In the meantime, I started on the midrange driver (as the tweeter has already a few published designs), with a "seems-to-be-matching-visually" dome profile and simple conical waveguides + radius at the mouth (like the KH420).
From the few simulations I've done, it seems to me that :
- mouth size of the waveguide kind of controls the lower frequency cutoff (where it starts to beam)
- depth of the waveguide alter the throat angle and so kind of controls the overall beam width
Am I on the correct path ?
edit : BTW I did a simulation of the midrange in the KH420 waveguide with the 3D model than you can download on Neumann's website. Their model seems to have a bigger radius than the real one so I guess I should take that 3D model with a grain of salt anyway. I'll post these simulations when I'm back on my work computer.
Last edited:
It's a mastering studio, non-environment room (heavy absorption, 1.1 meter deep backwall with waveguides and membranes), and the speakers will be sealed.
For reasons of size of the already existing flush-mounting cavity, I wanted to go to 2x15 sealed but it takes too much vertical space.
Vertically 2x12 fit, and horizontally, 2x 2x10 fit.
The 4 Purifis have slightly more SPL capabilities than 2x Volt 15" (for expl), so they win against 2x12 🙂
edit : a smal sketchup of the idea :
The top woofer will either play as the bottom ones (to better mitigate the floor bounce), or lowpassed lower to only help in the bottom octave if it causes too much vertical dispersion problems. But that will be later considered, when the waveguides will be characterized.
For reasons of size of the already existing flush-mounting cavity, I wanted to go to 2x15 sealed but it takes too much vertical space.
Vertically 2x12 fit, and horizontally, 2x 2x10 fit.
The 4 Purifis have slightly more SPL capabilities than 2x Volt 15" (for expl), so they win against 2x12 🙂
edit : a smal sketchup of the idea :
The top woofer will either play as the bottom ones (to better mitigate the floor bounce), or lowpassed lower to only help in the bottom octave if it causes too much vertical dispersion problems. But that will be later considered, when the waveguides will be characterized.
Last edited:
Hi everyone,
I’m looking for advice on the best way to mount a waveguide. It needs to be mounted flat and without screws in the housing. If you have any suggestions or references on how to achieve this, I’d really appreciate your input.
This is how it looks at the moment:
Kind regards,
Poiy
I’m looking for advice on the best way to mount a waveguide. It needs to be mounted flat and without screws in the housing. If you have any suggestions or references on how to achieve this, I’d really appreciate your input.
This is how it looks at the moment:
Kind regards,
Poiy
Glued? To be filled, primed and painted with the enclosure? If so be prepared to get some small cracks or other traces at the material joint.
Or should it be removable, screwed from behind the baffle? If so I would make the rebate inside.
Here is a very rough test example I did. The WG is partially CNC routed into the MDF baffle and the tweeter adapter piece is 3d printed. The joint is still very visible, because I did not finish the filling, priming, painting.
Eventually I will cover the joint with a glass fiber mesh or similar, to avoid cracks appearing.
Or I will CNC route the WG and just attach a very "thin" tweeter adapter piece.
It should be removable. Thank you for the suggestion. I’ll need to examine some references more closely, such as this one.
I got this from the thread, and it looks great.
EDIT: Wasn’t that someone from this thread? Unfortunately, I can’t find it anymore. I’m curious to know if these are two different materials or just different colors.
Last edited:
A few questions and suggestions:
*) How far do you sit, how loud will you need it? Only mastering - cause that's less loud as recording can be?
*) Is your current tweeter of the PMCs loud enough?
- Think about using a T25B instead of T34B. It has a higher membrane resonance frequency and is one of the lowest THD tweeters available. And for me the best sounding 1" tweeter I know. Also used in an non environment room with Hypex amps - resolution and natural reproduction is better than everything else I heared.
*) Do you NEED waveguides?
-With a NER you should have very little reflections from walls and ceiling. Without waveguides tweeter and midrange get way closer together and get a "coax like" behavior. I did some listening trials for a MTM vs MT setup and prefered only one midrange cause of the better point source.
I even developed a frontplate to get them closer together and have less influence of the midrange membrane to the tweeter ... what all other designers "forget" about ;-)
The benefit - more EVEN radiation! It's very wide over the whole range. Vertical as horizontal. The wavegudies behave pretty different over frequency and hor/vert.
*) Your wall opening is fixed?
-The 2 10" nest to each other will very likely do some horizontal beaming already at 450Hz. The waveguide doesn't do a lot at this frequency, so it's possible you get some unsteadyness here? I would opt for a 3.5 way - but that's probably not that easy to measure in a room, even in a NER.
But 4 of the Purify 10" ... that's nice 😎
*) What leads to the next question - did you think about a single bass array?
- My main speakers are lay out so the low end drivers are close to the position where they need to be for a bass array - and this works perfectly to chancell the first horizontal and vertical modes of the room! For a top grade listening system it really is a great benefit, I would never go back or design a room without it. When you share a scetch of the front wall/room we can have a look how close you are already with your position.
I'm just redesigning my main speakers from T25B/MD60N/SS 20m to Bliesma 3" midrange and 2x 12" SB34NRXL75 drivers and doing a new frontplate for the midrange (the new M74T or the A version which I used before a few times).
We can probably team up if you are interested? I have experience with the drivers and what they can do and room integration but didn't get my feet wet with gmsh & Akabak so far.
For the Bliesma M74B and according to Hificompass, it seems hazardous to remove the protection grill as it is embeded in beetween the 2 layers of front plate and there is a risk to off-center the voice coil. Maybe you should ask Mr Malikov from Bliesma about that
Also what is the projected cut-off frequency between the M74B and T34B ? it will define the size of the mid/high wave guides and therefore the ctc distance of the M74B/T34B.
You can order them without the front plate from Bliesma, then you are free with your design.
It was printed on an SLA printer and then painted black. Here is a link to the pdf of the assembly process
https://drive.google.com/file/d/1hTBLKORgGTOltmUg4CpEhTOgF5y6VoL5/view?usp=drive_link
So few quick answers in order to avoid hijacking the thread, later in the design phase I'll do a dedicated thread.
- Somewhere between 2 and 4k. I'm not totally set right now. In a studio as deeply treated, you kind of don't care of the ceiling bounce, and the floor bounce is mainly destructive at it's fundamental frequency (around 200 Hz with my listening distance), above in the range it starts to get blocked by the desk.
So I'm not dead set on everything, that's why I want to experiment / modelize dual waveguides, to check what different compromises they bring vs a direct radiator. The tweeter / midrange baffle can be removed on my PMCs, so I can even replace it with a waveguided one to test with my actual drivers to hear if I like it in contex or not.
Anyway, without the information everywhere on diyaudio, I would not have started that endeavour, so every good results I get, even if not for my project, I inted to share them to the community.
So now, back to my question which concerns that thread a bit more 🙂
From the few simulations I've done, it seems to me that :
- Actual listening position is 2.7 to 3 meters from the speakers
- Tweeter is loud enough for mastering, but I'm also using the room for producing electronic music, teaching sound design / production, mixing, and I have kind of a big bass preset in my actual crossover to simulate club feel. In the near future, I'll also rent the room to freelance engineers / producers. So I though 34 instead of 25 to align a bit better the max SPL on all the spectrum and sleep on my two ears knowing eardrums will die before the speakers.
- The "console" is based around a motorized sit / stand desk so I also have linear & bass boost presets aligned at a standing position in my crossover (Linea Research ASC48F, which has LIR filters (basically equivalent to LR4 linear phase), lot of EQ, linear phase high shelves and also custom FIR slots). I'm only using minimum phase EQing of the drivers then the LIR filters. Stand position adjust only the delays between drivers. The crossover feeds 6x monoblocs based on NC500.
- Waveguides. Errr. I don't need them life and death, but anything that goes into the absorber is still wasted power. If I could get something constant directivity whit flat response from 500/600 to 20k which is flat within -30° to +30° (at 35 degrees you are already standing against the wall and it's not a usefull position) and nothing farer than that, it would be perfect (yes I know I'm dreaming). It would also mean most of the power would go directly into the backwall absorber which is more beefy than the sidewalls (1m10 vs 0.73m).
- Waveguides could help narrow the vertical dispersion to avoid diffraction from the console. It is slotted to be acoustically transparent to the maximum I could, but it's never perfect, so that would help.
- Waveguides could help increase the gain of the midrange in the 400Hz + range, allowing to lower the distortion in the sub 1k range to the level it is above 1k. And the purifi woofer are excellent there as well, so I can try a lot of thing, and with linear phase, also running them in parallel for a part of the low mid range.
- My opening in the wall is fixed, and it's reinforced concrete, and I'm very happy with the italian plaster I did on my front wall, I don't want to touch that. I could make the waveguide as wide as 2x10 anyway, so I could match on that axis.
- I intend to run the Purifi sealed, LT transformed to pretty low, with the objective of getting 0 group delay @ 30 Hz. BTW, my room gain is already +9 dB @ 20 Hz vs the close miked response of a woofer in the wall. I also have an excursion limiter in my crossover that slides a highpass in case of too much excursion, meaning that I don't need a preset for extended lows and one for higher SPLs. These woofer have pretty good IMD distortion as well, which is very important when puting lots of powers on very low frequencies.
- I don't need a single bass array in that room (T60 decay is balanced at 150 ms from 20k to 30 Hz), and the room response is the same for stereo infrabass of mono one. (useful when mastering experimental ambiant for expl 🙂)
- Somewhere between 2 and 4k. I'm not totally set right now. In a studio as deeply treated, you kind of don't care of the ceiling bounce, and the floor bounce is mainly destructive at it's fundamental frequency (around 200 Hz with my listening distance), above in the range it starts to get blocked by the desk.
So I'm not dead set on everything, that's why I want to experiment / modelize dual waveguides, to check what different compromises they bring vs a direct radiator. The tweeter / midrange baffle can be removed on my PMCs, so I can even replace it with a waveguided one to test with my actual drivers to hear if I like it in contex or not.
Anyway, without the information everywhere on diyaudio, I would not have started that endeavour, so every good results I get, even if not for my project, I inted to share them to the community.
So now, back to my question which concerns that thread a bit more 🙂
From the few simulations I've done, it seems to me that :
- mouth size of the waveguide kind of controls the lower frequency cutoff (where it starts to beam)
- depth of the waveguide alter the throat angle and so kind of controls the overall beam width
This is very exciting, @Storm Mastering. I have acquired T34Bs wondered about a dual waveguide design like the KH420.
I look forward to watching your developments!
Bill
I look forward to watching your developments!
Bill
The size of the mouth sets the lowest frequency where the waveguide can control directivity. I would describe beaming for waveguides more as where the dimensions of the driver are the determining factor in overall directivity.
The other two factors are depth and nominal wall angle. You can only manipulate two of the three as the other one always has to follow.
In a 2 Pi environment it takes quite a large device to control directivity at 400Hz. As far as gain and supporting the low end that is best done by reducing the solid radiation angle i.e. narrowing the directivity.
There is not much wiggle room in profiles for larger domes if you do not want the CTC distances to become quite large. Making the mouth elliptical creates a steeper vertical wall angle which quickly limits the crossover point unless you can accommodate step crossover filters.
You can see this in a simulation I made for a 200mm wide waveguide for an M74. At 400Hz it is doing precisely nothing 🙂
To move the directivity features down by an an octave the size has to double.
https://www.diyaudio.com/community/...es-3-dome-midrange-lineup.380200/post-7178237
Yes.
Don't forget that the WG is way smaller vertically and will not control lower frequencies in that direction at all. So you get a mixed over all room frequency function. Also the 2 waveguides influence each other and I imagine it's not easy to get a linearly dampened table reflection.
But it's a totally cool project and worth to investigate! 🤓
*) 3m distance and renting the room ... big tweeter is for sure the save bet. I never compared the 2 1:1, I would probably start with that, build a test baffle and do a real comparison. With the complete cost of your project - I would put in both tweeters and switch between them in DSP setups. And techs will go crazy when they see 2 tweeters 😎
*) ASC48 is one of the best controllers for that use. Would really like to measure one of these units. Combined with NC500 you are done for a while.
*) Waveguides and Horns don't have sharp angles of projecting sound. You can get a side reflection myabe 4-5dB down from 6-800Hz upwards with some wiggles and a vertical reflection more uneven, starting at probably 1kHz and also 3-5dB down. You will for sure lower THD at the crossover frequencies for midrange and tweeter. MAybe not needed for the T34 but could be nice for the M74.
*) Getting rid of that reflection is tricky, esp when you change the height of the desk. I angled the "mirror point" of the desk and that helped but the floor reflection is the last one in my system.
*) Sealed LF drivers with enough power are the best thing to have in a room. You can literally choose your frequency response in the room. An array will keep the SPL level over distance better and always has benefits when moving through the room (horizontal/vertical with a sit/stand desk) but it's not always possible to implement. Do you put in a lowcut in your Club setting to simulate a "normal" woofer with it's longer decays?
*) Do you have a measurement of your speakers at different distances in your room? I have 20cm BCA resonant absorbers in my back wall, 10-20cm of spare stuff in front of that and my 20cm Absorber stands in front of it. And the main room resonance doesn't care, it's still there. Not long ringing and broadband but with all the peaks and dips. As this is a resonant phenomen it doesn't need a lot of reflected energy to be excited. So I'm really interested what you did with your backwall (Newell style angled hanging panels or just filled) and how good you can get with that.
Looking forward to your progress!
- Yes, I know waveguides can't be brickwall 🙂 I was just saying that ideally. Now if I could get same frequency response on the mid / high from -30 to +30 horizontally, with at the same time be down 3 or 4 dB vertically, that would already be awesome.
- On the KH420, the vertical response, 30° down is already between -3 and -6 depending on the frequency (not uniform but not bad), that's already quite an improvement. For the standing desk position, I never master standing, it's more for producing, it forces your to be active on your legs, you focus more, you're in the flow easier, so desk reflections are less of a problem in that position. And with the ASC48F, I can as well control the vertical directivity in the low mids with the help of the Purifi woofers. That's why I want to keep it a 3-way, because with the 4 way of my cross-over, I can split horizontally or vertically the woofer array for pattern control. With a few simulations of good waveguides candidates, everything can then be simulated easily in VituixCAD 🙂 The fact that I'm in the infinite baffle scenario also simplifies a lot of things.
- Right now I don't have a HP filter on my "club" setting (which is basically IIRC a lowshelf with +6 dB under 100 Hz), but I already quite a few time engaged the excursion limiter on my Volt 12" (which are playing quasi sealed right now, not the best for them). And I put my excursion limiter at linear xmax, not mechanical xmax, so I have room, but I didn't listen at club levels TBH.
- Also, maybe I will stay with the ATC midrange (in a custom WG) instead of the Bliesma, who knows. But If I go Bliesma, I can recoup probably ~5k from the sell of the PMC, which ~ pays for the 8 Purifis. I can't do that if I stay on the ATC (and I've no money raining on my head).
- Maybe It will be in two steps, with first putting a better tweeter and waveguiding the mid/high section of the PMC to test, then go full in a second step.
- My room is 22 square meters of big membranes absorbers on back and side wall (tuned between 26 and 37 Hz depending on the position), with 1m25 Newell waveguides in front of them. If you are interested a bit more, there is a studio construction diary on Gearspace called "Storm Mastering - a mastering NE room in a military fort" that you can find easily 🙂
- Anyway, feel free to PM me to continue this conversation if needed.
Back on waveguides, I tried to model with Akabak's axisymmetric mode the KH420 horizontal profile (it takes 20 sec to compute on my laptop with high treble resolution vs the 7 hour the 3D model took on my work computer), and I got more pronounced peaks and dips. It seems mouth diffraction is bigger. Maybe the elliptical shape helps on that part, or maybe the simulation isn't as good in the axisymmetric mode. Keep in mind the rounding is quite small as well. Here are the polars for a circular waveguide with exactly the same profile as the horizontal one in the KH waveguide, same scales.
Tonight just before leaving work, I'll launch the tweeter simulation on the same KH file.
I have a weird request for help with a wave guide.
I have Scan Speak D3004/6040-00 Tweeters.
I want the smallest diameter wave guide I can get. Biggest goal is to reduce off axis reflections due to small environment.
I'd love if I could attach via the 3 set screws that I already have for the Scan Tweeter. ideally just bolt on the front. I have a friend with a 3d printer, but I have zero 3d CAD design experience. Would love some help getting a small wave guide to bolt on.
I have Scan Speak D3004/6040-00 Tweeters.
I want the smallest diameter wave guide I can get. Biggest goal is to reduce off axis reflections due to small environment.
I'd love if I could attach via the 3 set screws that I already have for the Scan Tweeter. ideally just bolt on the front. I have a friend with a 3d printer, but I have zero 3d CAD design experience. Would love some help getting a small wave guide to bolt on.
Following up, tweeter simulations :
And a last one, normalized, midrange & tweeter, same dB scale and 3 dB steps instead of 1 :
Midrange simulation seems to looks like the measurements. Tweeter simulation has way more peaks & dips than the measurements. I guess the 3D model isn't accurate, and it matters more on the treble than in the mids.
Or maybe it's because of normalizing at 0°, exactly dead center, on a simulation with no loss at all vs a real measure with some loss and never exactly dead center ?
I made a gif of augerpro's measurements, 4,5,6 and 8" guides. Changes are affected by other factors too, like relation of radiator/wg diameters and depth of the wg.
Last edited: