Midrange compression driver development
Printed From: Speakerplans.com
Category: Plans
Forum Name: New Projects Forum
Forum Description: Forum for new speakerplans projects, in memory of Tony Wilkes, 1953 - 2014
URL: https://forum.speakerplans.com/forum_posts.asp?TID=108478
Printed Date: 26 March 2026 at 6:55pm
Software Version: Web Wiz Forums 12.08 - https://www.webwizforums.com
Topic: Midrange compression driver development
Posted By: JulianDA
Subject: Midrange compression driver development
Date Posted: 17 April 2024 at 10:49pm
|
Our latest project is is the development and construction of a 3-way fully hornloaded midtop. The goal is to reproduce frequencies from 200/250Hz and upwards with high sensitivity and a controlled directivity of 60 by 40 degrees. At the moment our low-mid driver is the 18S 12ND610 which will cover 200-800Hz. The Midrange is a B&C 6NSM51 which should be covering 800-3500Hz. The HF Driver is not specified at the moment. All we know so far is that it should be a 1" CD, but we are still searching for the right one (Any recommendations?). In this thread i want to talk about the midrange driver and horn. In my opinion it is the heart of the whole project and also the hardest part to get right. The plan is to design a midrange compression driver with the 6NSM51, which really just means to design a phaseplug with a pretty high compression-ratio analogous to your typical hf compression drivers. JBL did the same with their CMCDs (Cone Midrange Compression Drivers) as you can see in this Tech-Note: https://jblpro.com/en/site_elements/tech-note-cone-midrange-compression-drivers-cmcd" rel="nofollow - https://jblpro.com/en/site_elements/tech-note-cone-midrange-compression-drivers-cmcd I started this whole journey by reading and trying to understand the following paper from Mark Dodd and Jack Oclee-Brown: https://www.researchgate.net/publication/292662389_New_Methodology_for_the_Acoustic_Design_of_Compression_Driver_Phase_Plugs_with_Concentric_Annular_Channels" rel="nofollow - https://www.researchgate.net/publication/292662389_New_Methodology_for_the_Acoustic_Design_of_Compression_Driver_Phase_Plugs_with_Concentric_Annular_Channels The paper describes a way of generating a concentric phaseplug geometry which supresses the establishment of modes in the compression cavity by careful placement of the phaseplug channel entries. This work is for HF drivers, but the physics is the same for the midrange. So analog to Dodd and Oclee-Brown, i started with a modal analysis of the compression cavity of my -soon to be- compression driver in akabak. The following picture shows the result which is split in two halves, since i can only get akabak to show me the sound pressure over a straight line <- i managed to get two straight lines in the cavity:  For a real deep dive into all this you should read the paper, or even better the full dissertation. But in short: -The four lines you can see are four resonant modes in the cavity. -Only the first two are of importance, since the driver shall only play up to max. 4kHz. -By positioning the phaseplug channel entries at the nodes of one mode, all the modes below and including that mode are supressed. with that info it was clear that i needed two channels exactly at the nodes of the second mode. I set my desired compression ratio at 7:1 (its the value JBL used and called "moderate") and began construction of the plug in CAD. The following image shows the result:  The size of the channel entries is chosen according to the compression-ratio and the area of the section of membrane that "feeds" the channel. So the inner part of the membrane feeds the inner channel with a CR of 7:1 and so does the outer. The outer Channel is also curved to elongate it for pathlength correction (the difference is about 1mm. I could not get it better). Both Channels have the same expansion-rate and combine to an exit of 80mm diameter. As a whole the phaseplug starts with an area of ~20cm^2 at the channel entries and ends after a pathlength of 4,5cm with an area of ~50cm^2. Here it is. 3D-printed as three pieces that press together:   And here it is mounted to our test-horn:  One big problem with all of this is the simplicity of my akabak simulations. Since i can only work with TSP and the geometry of the membrane, all the stuff that can and will happen with a moving membrane (like breakup modes) is not part of my simulation. Because of this i know that my results will differ from reality. The most critical part that will differ is the position of the channel entries. But i aim to just produce more prototypes and go by trial and error from this stage on. But enough said. Here are two quick measurements i did. I also included the simulation-results. (Better measurements will follow eventually) First, here is the 6nsm51 cd on the horn (like in the last picture): Red is the measurement and blue the simulation  And this is just the Driver mounted flush to the horn without any phaseplug:  I think i am not to far off, but i always see a drop of 5dB starting at 2kHz that is not in the simulations. I did a lot more measurements and testing with other geometries etc. and it is a reoccuring "problem". Right now i dont have a clue why that is, so if anyone has some more insights or suggestions i would be more than happy to hear them.
PS: getting work done on this will take some time, like all of my documented projects on this site. So please dont expect fast updates  |
Replies:
Posted By: VECTORDJ
Date Posted: 18 April 2024 at 12:58am
| Turbosound wrote an AES Paper that talked about Cone Mid Horns....Speaker Builder Mag had a paper on Cone Mid Horns....Checkout Funktion One designs....As You found getting above 2K flat is Black Magic....Good Luck with Your Project. |
Posted By: Elliot Thompson
Date Posted: 18 April 2024 at 3:17am
|
Have you measured the driver without the horn? The TS Parameters indicate an Le of 0.15 In free air you should be able to make it to around -3dB @ 6.1 kHz with a 0 dB starting point @ 1.03 kHz without the need of a Phase Plug. B&C TS Parameters in terms of 1 watt, 1 meter is lower than the advertised specifications once calculated. B&C states 100 dB sensitivity 1 watt, 1 meter whereas, calculation brings forth 94.66 dB @ 1 watt, 1 meter.
Best Regards,
------------- Elliot Thompson |
Posted By: snowflake
Date Posted: 18 April 2024 at 12:47pm
| you aren't trying to get very wide bandwidth out of this - can you not do better just using a front chamber to couple to the horn... |
Posted By: fudge22
Date Posted: 19 April 2024 at 12:59am
|
You have obviously spent a lot of time on the initial design and analysis of the phase plug, but then seem to have just given up and slapped it on to the back of a wooden horn, with no consideration as to how they couple. From the photos, there appears to be a relatively large discontinuity in cross sectional area from the phase plug exit and the throat of the main horn. Such discontinuities are not conducive of a smooth or optimum response. One of the benefits of using a phase plug is that it reduces the volume of the cavity between the diaphragm and the horn throat. This cavity behaves as an acoustic reactance acting as a shunt capacitance across the throat. However, it may be useful to measure the response of the driver horn combo without a phase plug to get a base line response. In his 1978 paper on phase plugs, Henricksen noted that circumferential phase plugs produced a notchy response compared to radial slit phase plugs. It might be worth experimenting with that type of design. If I remember correctly EAW used a form of radial slit phase plug to couple cone drive units to mid-range horns in some of their designs. |
Posted By: madboffin
Date Posted: 20 April 2024 at 4:43pm
|
One thing to check for, as you tidy up the design, is resonances in the component parts of the phase plug i.e. the cone and the ring. They can sometimes cause suckouts and an uneven frequency response. The technical way to do this is with an accelerometer (vibration pickup) but you can do a basic check for any serious resonances by holding a finger on the relevant part while doing a frequency sweep, which will indicate if further investigation is required. |
Posted By: KDW32
Date Posted: 21 April 2024 at 12:36am
Yes Zeus! Knowledge bomb right there.
|
madboffin wrote:Posted By: RoadRunnersDust
Date Posted: 21 April 2024 at 7:16pm
|
What cone-plug distance have you used? ------------- www.guildfordcablecompany.co.uk" rel="nofollow - www.guildfordcablecompany.co.uk |
Posted By: JulianDA
Date Posted: 22 April 2024 at 6:05pm
Do you have links to these papers if they are free to read? i could not find them. Surely i know about funktion one´s designs, but they work with a cone speaker "converted" to a ring radiator, so their designs are of little use to me.
Yes i did, but the driver was measured without any baffle. Here is the result. (red -> with horn)  to me it is clear that i loose a lot of spl in the higher frequencies because of the nearly constant directivity if my horn. The driver alone measures more hf, but its directivity also becomes very narrow, while my horn disperses the sound pressure over a larger radiation angle. That is also the reason for my whole Phaseplug Design. I need more output at the higher frequencies to then disperse it over a greater angle.
No i can't since i would not get the desired directivity with the hornthroat = SD. But by reducing the throat area till its small enough for the directivity and "hf"-SPL, i get problems because of the pathlength differences and throatchamber volume.
This prototype horn was always part of the design process. I know from previous experimentation that it works really well for how simple it is constructed. Even if it doesn't look like it, the horn is also a result of numerous simulations in akabak and the construction of two other prototypes before it.
I think the photos are misleading then. The only real discontinuity is the change from a round throat to a square horn. This change is made less drastic by the triangles in the corners. But i also did a measurement without those triangles and even that had no significant effect on the response.
The last measurement of my initial post shows exactly that :)
The paper that i linked also tested radial designs and found that a properly designed circumferential plug has superior performance. The "notchy response" comes from the resonant modes inside the compression cavity. The whole aim of this design is to suppress these modes from forming by carefull placement auf the channel entries. The Notch that i measured (around 2,5kHz) is exactly that. Its the first resonant mode inside the cavity that is still excited.
Thanks for the tip! I will try if i can feel anything...but with double ear-protection...this thing is LOUD with even the slightest input :D
The distance is 2mm. Hornresp showed 0.5mm excursion at 500W input and a highpass of 800Hz. I decided that my next step will be testing slightly different phaseplugs to see if i am in the right ballpark with the position of my channels. The printer is already printing two new variants. One with both channels moved 2mm more to the outside, and one with both channels moved 2mm more to the inside. I think trial and error is the only thing i can do at this moment. |
Posted By: Elliot Thompson
Date Posted: 23 April 2024 at 2:06am
Measuring without a baffle is the best way as the loudspeaker's frequency response is not influence by a boundary. It appears you are having issues (major dips) within the 4 kHz - 5 kHz range once attached to the horn. If you are looking for a wider dispersion at High Frequencies, try exploring the notes from Don Keele. He was responsible for Electro-voice (Hyperbolic) and JBL (Bi-Radial) wide dispersion horns in the 1970's - 1980's. Typically either design will offer no less than 100 degree horizontal dispersion at high frequencies. http://www.xlrtechs.com/dbkeele.com/" rel="nofollow - http://www.xlrtechs.com/dbkeele.com/ Best Regards,
------------- Elliot Thompson |
Posted By: Timebomb
Date Posted: 23 April 2024 at 4:27pm
|
I would try extending the central part of the plug into the the horn throat, try and smooth the transition. What does the impedance response look like around 2KHz? ------------- James Secker facebook.com/soundgearuk James@soundgear.co.uk www.soundgear.co.uk |
Posted By: fudge22
Date Posted: 23 April 2024 at 7:46pm
|
Ah! Sorry. From the photos, it looks like the front of the 3D printed phase plug is flat, and that there is a ring around the exit, perpendicular to the horn axis, where the printed part attaches to the wooden horn. If the taper of the wooden flare continues down at the same angle into the printed phase plug, so that there is no step, my comment was indeed wrong. |
Posted By: JulianDA
Date Posted: 24 April 2024 at 4:02pm
Thanks for the input, Elliot :) Its true that 4-5kHz is still a bit rough. But as one can see from the modal analysis of the compression chamber, most of it is probably just the second resonant mode <- So the same problem as in the 2,5kHz range. But i think i would be happy if there was some destructive interference around 5kHz in the final build. The FR and Impedance of the 6nsm51 shows severe cone breakup at that frequency and it would be nice to filter that out acoustically. Regarding the dispersion: I am more than happy with my construction. I just mentioned it as a reason for the reduced hf-output <- it will need the typical constant directivity eq correction when it is finished.
Could you explain why that should help me? I simulated and build quite a lot of prototypes before this and most had a plug reaching inside the horn. But to my knowledge this is only needed if the circumference of the horn-throat is too big for the desired hf-dispersion, or when the wavefront at the throat is not shaped in the desired way <- so the extended part of the plug acts more like a waveguide.
No worries! I totally understand why you thought that....its because i never mentioned that there is a 3D-printed throat-adapter on the horn. The exit of my Compression Driver is 80mm wide and the throat of the horn is 96mm wide. So i printed a conical adapter from 80 to 96mm with a length of 12mm. So just from the information that i gave you it is clear that you interpreted the pictures in that way....there is just too much 3D-print visible at the horn throat :D |
Posted By: Elliot Thompson
Date Posted: 26 April 2024 at 3:08am
So I would imagine you are planning to use a compression driver to fill in the gap (4 kHz - 5 kHz) housed on a 60 x 40 horn to keep things uniform (seamless)? Best Regards, ------------- Elliot Thompson |
Posted By: snowflake
Date Posted: 26 April 2024 at 10:33am
| why have you chosen the crossover point to be 3.5kHz? plenty of 1" compression drivers can be crossed at 1.5kHz or 2khz... |
Posted By: smoore
Date Posted: 26 April 2024 at 11:38am
|
Hi Julian, Great to see someone doing some proper R&D! Well done for taking the step! I just wondered, have you got any pictures of your measurement set up and how you have analysed the data? The measurements of your driver on it's own are making me suspicious and you may be measuring reflections rather than your horn, and compression/phase bung. Both look great by the way! Cheers, Sam
|
Posted By: RoadRunnersDust
Date Posted: 26 April 2024 at 12:32pm
Have you missed the point of nearly all professional mid-range solutions for the past 40 years?  ------------- www.guildfordcablecompany.co.uk" rel="nofollow - www.guildfordcablecompany.co.uk |
Posted By: MarjanM
Date Posted: 26 April 2024 at 1:26pm
^^^^^THIS!!!^^^^^ ------------- Marjan Milosevic MM-Acoustics www.mm-acoustics.com https://www.facebook.com/pages/MM-Acoustics/608901282527713 |
Posted By: fudge22
Date Posted: 27 April 2024 at 7:53pm
|
Whilst I agree that a smooth transition between the “compression driver” exit, and the horn throat is essential, JulianDA, the OP, has clarified that this is the case. Given that there are no discontinuities, some explanation of why extending the central part of the plug is beneficial would be useful. |
Posted By: Teunos
Date Posted: 28 April 2024 at 12:52pm
|
The project looks great, kudos for the work done. I agree with TimeBomb and Marjan, extend the plug further into the throat. You write the throat is 80mm 50cm^2, consider that a standing wave in the throat would have its first null at ~2140Hz, which ligns up almost exactly with the first measured null. In the simulation, did you do a full BEM analysis of waveguide including horn?
------------- Best regards, Teun. |
Posted By: Timebomb
Date Posted: 28 April 2024 at 9:10pm
|
Its hard to tell from the photos but is the expansion rate smooth from the end of the phase plug smooth into to the throat of the horn? The geometry is quite different as you change from radial slits to the flat side walls of the horn, if the plug was extended the reflections across the throat would be reduced and diffused by the round plug. Did you aim for a curved wavefront or planar at the exit of the phase plug? ------------- James Secker facebook.com/soundgearuk James@soundgear.co.uk www.soundgear.co.uk |
Posted By: JulianDA
Date Posted: 28 April 2024 at 11:46pm
|
Thanks for all the replies! They sparked more thoughts on the matter for sure! In a few days i will post more pictures of the complete CAD model of the driver, the phaseplug and the horn. Hopefully this will clear thinks up a bit. After many suggestions of, in my eyes, knowledgeable persons to expand the centre of the plug, this will probably be my next try in getting a flatter FR. But first i will try the new plugs with slightly different channel entry positions. This test will probably be done tomorrow, so i can update you on that in a few days, too.
Thanks :) Yes, a standing wave at 80mm would have a null at that frequency. That could be a possibility. But i have to ask why there should be a standing wave there and not at all other parts of the horn with different dimensions? Would it just be because of the miniscule part of the throat were the outer walls change from a negative to a positive expansion and create a zero expansion segment by diong that? Would the only effect of the plug-extension be then to block the path for this particular standing wave until the outer walls of the horn start to expand? Lots of questions, but i am excited to learn more! For the simulation: i did a full bem analysis in AKABAK of the membrane (with the shape as close to the real one as a pair of calipers and some tricky measuring could get me) driving the compression-chamber (just the shape of the membrane and surrounds shifted by 2mm to the front) then the phaseplug (with interfaces at each end) connected to a model of the horn in use, which is then coupled to the outside via another interface. All that in 4pi with 180° polar measuring points at 1,5m distance. I did some pre-meshing of my models in gmsh and remeshed certain parts again in AKABAK if they needed a finer mesh as indicated by the standard deviations calculated by the solver. I didn't use the model of the horn in the simulations for the initial design of the phase plug. For that i used a model of a plane wave tube (with a 100% absorbant boundary at one end) and later a conical horn shape with the same boundary at the end. <- this was just to get a more accurate look at the achievable FR. But after my phase plug was designed, i simulated it with the horn that i had from previous tests. <- just as with hf compression drivers, the horn shouldnt be critical for the performance as long as there is no large impedance missmatch at the throat |
Posted By: snowflake
Date Posted: 29 April 2024 at 12:14am
there's several reasons you might do it - I just wondered what the thinking is in this design. thanks for your insightful reply 
|
Posted By: Earplug
Date Posted: 29 April 2024 at 10:43am
|
An interesting project - but really shows that the physics is difficult and simulations obviously have some severe limitations. Measurement and experimentation is the way - but also, in the end, using a decent DSP unit with several parametric corrections may be the only solution. And generally, the simplest. You can really drive yourself mad trying to tweak and tweak and tweak. Ultimately a bit of a waste of time. There has to be a point where you say enough is enough, and just use some electronics to get the thing sounding good. All the big manufacturers do that, even going back decades. I just picked up some EV S200's that a friend found while clearing out his shed! I really love those cabs, but they also need their processor to really sound good. ------------- Earplugs Are For Wimps! |
Posted By: Elliot Thompson
Date Posted: 29 April 2024 at 1:30pm
At some point in time, you must put down the simulators and, focus primarily on measurements. It is inevitable if you want to increase your knowledge fundamentally on loudspeaker design. Best Regards, ------------- Elliot Thompson |
Posted By: Earplug
Date Posted: 29 April 2024 at 2:51pm
Yes, definitely, but what I meant was that on top of that, you should also expect to have to do some correction/s with electronics! That's been part of the process since day 1!  ------------- Earplugs Are For Wimps! |
Posted By: Elliot Thompson
Date Posted: 29 April 2024 at 3:20pm
Yes! We are in full agreement. I was just reinforcing what you said. Too many live off of Simulations and Presets. They don't realize such methods are just estimates not guaranteed. Best Regards, ------------- Elliot Thompson |
Posted By: Earplug
Date Posted: 29 April 2024 at 3:44pm
Yup. Only things guaranteed are death - & taxes!!  ------------- Earplugs Are For Wimps! |
Posted By: fudge22
Date Posted: 01 May 2024 at 9:41pm
I’m guessing that you never designed a horn whose length was an integral number of half wavelengths at the crossover frequency. That said, I agree that electronic correction is a valid method to get the response required.
The original poster did put down the simulators, made his design, and took some measurements. Now he is stuck and remeasuring is unlikely to solve the problem or increase his knowledge. Whilst measurements are an important part of any design process, at some point you need to understand the theory behind the simulations so that you can have any hope of understanding the measurements. With regards the OP’s problem, it might be worth him revisiting the JBL paper that he linked to in the first post, and try to figure out the difference between the CMCD-81H and the CMCD-61H. The former exhibits the same dip in response, that his own design does, at just over 2KHz, but the latter does not. Other things to consider, which may or may not have been accounted for in the simulation (so feel free to disregard any or all of the following), and may have a large or negligible effect on the response. Does the model assume that all the sound passes through the nearest port/slot in the phase plug? If so, sound passing across the chamber and exiting the multiple slots will cause dips in response not predicted. Take into consideration the mass of the air in the ports, which will act as a series inertance with some added acoustic resistance. Treat the diaphragm as a simple rigid piston. Consider the speed of wave propagation through the diaphragm material. Does the frequency/depth of the notch change when you measure the response off axis? |
Posted By: Elliot Thompson
Date Posted: 02 May 2024 at 3:16am
Understanding the limitations of a Simulator will lead you to wards real-world measurements. Simulators are based on the perfect scenario. Real-World Measurements are based under real-world conditions.
Best Regards, ------------- Elliot Thompson |
Posted By: Earplug
Date Posted: 02 May 2024 at 9:57am
Haha, no - nothing that complex, but I did design a small system based roughly on the Martin F2 (1 x 18" subs & 2 x 12" + 2" mid tops). My idea being to make something lighter & easier to move. My main system uses Martin Philis. Lovely cabs, but very heavy. I have a good carpenter friend here, now unfortunately retired, that literally had years & years of experience building cabs that helped me - and his (practical) experience was essential. Along with SMAART, of course! That system actually turned out very good and I still use it all the time. Amazing SPL (and sound quality) for it's size. ------------- Earplugs Are For Wimps! |
Posted By: JulianDA
Date Posted: 18 July 2024 at 4:17pm
|
I finally have time to post more of this project. I printed 2 more phaseplugs with slightly different entry points of the outer ring (shifted +2mm and -2mm) The simulations showed a massive difference, but in practice it was almost nothing. I think the reason ist the real behaviour of the membrane in comparison to the assumed stiff piston movement in the sim. The resulting responses are the following:  the green curve is the one with the entry shifted +2mm to the outside. I deemed it the best. This is the same response of that pp but with added phase:  there is clearly something wonky happening at ~2kHz but with only 3 eq points i get the following response and phase:   is it fine to use it as it is now? to me it looked like destructive interference inside of the phaseplug that caused the 2kHz dip. But is it really a problem, if it is easily solvable with eq and the corresponding phase response is also good?
|
Posted By: Teunos
Date Posted: 19 July 2024 at 9:02am
|
At some point theory and practice start to deviate as models (i.e. simplified versions of reality) become sufficiently flawed. That should show any possible resonances/standing waves as being the cause of your frequency response. Impedance and distortion measurements would be valuable as well. Regardless, as long as you don't heavily gain up the null at 2kHz, using 3 (minimum phase IIR) EQ points that simultaneously flatten the phase response i.m.o. is perfectly acceptable and something you should always do prior to reverting to FIR filters. However, if the cause of the null is physical and you can solve it by changing structural features instead of compensating for it electrically, you probably should. Guessing your knowledge level from your handiwork i guess you already knew that. ------------- Best regards, Teun. |
Posted By: Teunos
Date Posted: 19 July 2024 at 9:12am
Your horn acts as acoustic impedance matching device. The throat will show a large discontinuity as expansion suddenly increases tenfold. Therefore the throat is by far the most sensitive to generate diffraction or excite modal resonances, regardless of their directional preference (down or across the throat).
My, and probably others' who replied the same (TimeBomb or Marjan for example) is that extending the phase plug into the horn might ease the transition and smear the effects over a bigger area. This won't alleviate the problem fully but at least lower the Q of the effects. ------------- Best regards, Teun. |
Posted By: JulianDA
Date Posted: 19 July 2024 at 11:13am
|
Thanks for your input Teunos!
Surely i can post those graphs :) what would be "sufficient spacing" for your liking? I am still kind of a novice regarding decay and waterfall plots. I mean i can tell when things are looking really bad, but thats the extend of my knowledge :D
I think i have to row back on my statement regarding the eq´s and their effect on the phase... all i did yesterday was playing around with my raw measurements in REW and using the build-in eq function to simulate what can be done.
In my opinion it is a very valuable function that is surprisingly accurate. But if my issues in frequency/phase response are from resonant effects inside of the phaseplug things will not be as easy as REW makes it seem. I will have to test my eq points again with real measurements, but i guess the result will be less than ideal.
I think i will just do a rather quick and dirty change to my phaseplug design to extend the central part while keeping the flare rates of both channels more or less equal and test that in the next measuring session :) My concern with expanding the middle of the plug was always that the flare rate of the plug will be lower. As far as i know a high flare rate is necessary to achieve low throat distortion. Making the exit diameter of the whole thing bigger to account for that would also be less than optimal for the achievable high frequency dispersion characteristics. Accoustics is complicated stuff |
Posted By: JulianDA
Date Posted: 19 July 2024 at 4:51pm
I think i will just test a different centerpiece of the plug that is extended 2cm into the horn at an diameter of 2cm. This is the best i can do without changing the entire plug geometry which is a real PITA   Do you think this will be enough to see if it has any effect? |
Posted By: JulianDA
Date Posted: 22 July 2024 at 1:45pm
|
Yesterday i measured the setup with the extended plug. Sadly the weather was so bad, that i could do just the one measurement. To show it in relation to the same pp without the extension i had to match the levels manually in REW. This is the Frequency response. The green trace is the original Plug whilst the orange one shows the extended one (with a reflection/resonance around 500Hz that could not be gated).  this is the distortion:  for this i also had to match the levels manually. so the absolute numbers are not included and i just look at the shape of the curves, i.e. where the most distortion occurs. |
Posted By: Elliot Thompson
Date Posted: 24 July 2024 at 2:40pm
|
So how does the two compare in terms of sound? It appears each one offers their own trade offs. Best Regards, ------------- Elliot Thompson |
Posted By: JulianDA
Date Posted: 26 July 2024 at 3:58pm
I thought a lot about this and it made a lot of sense to me that this is the root of my problem. But then i remembered that i made a measurement of the CD mounted to the horn but without the corners at the throat -> so it should exaggerate the problem of the assumed throat-reflection even more because of the greater discontinuity. (The reason for that measurement was to see how smooth the transition at the throat really has to be) This is the measurement without corners (blue) compared to one with the corners in place (red) cropped at 200Hz for better visibility. Sadly it had to be done indoors, so there are some reflections in the lower frequency range that could not be windowed out from the response.  As this change presents a much bigger discontinuity at the throat and has little to no effect on the ~2,2kHz dip i am now even more confused of its origin. The following pictures are for better understanding of the involved geometries. The first one is a cut through the vertical phaseplug and horn profile and the second a cut through the horizontal profile. The highlighted part is said "corner" that was added to smooth the transition from round to square.   and for the complete picture here is the view from the front with all 4 corners at the throat in place: 
Actually i did not listen to the new modification since we had really limited time and the weather was beyond shitty
 |
Posted By: Isitone
Date Posted: 16 October 2024 at 11:58am
| any progress in this interesting project? |
Posted By: JulianDA
Date Posted: 19 October 2024 at 8:07pm
|
Yes there is progress :) we did some groundplane measurements to get a better look at the "real" response of the plug and horn combination without all reflections. Maybe i will post those on another day. In the last weeks and months i designed a much better horn of which the last parts just finished 3D-printing. The next step will be to make a negative mould from it and then build it with glasfiber and resin. Hopefully this process will be done next weekend. With the better Horn it will be much easier to optimise the Phaseplug in the later stages of this project. |
Posted By: JulianDA
Date Posted: 09 November 2024 at 2:20am
Sooo we put together the 3D-printed Horn that i designed. It came out really nice if i say so myself:  We used this for quick measurements and to make a plaster-mould for later fibreglas fun.   The mould looks pretty rough right now and is split in two parts, but it will be fine after glueing it and smoothing it with some filler and sandpaper. Sometime in the next weeks we will try to build the first "real" horn from glasfibre. I will keep you updated :)
|
Posted By: Contour
Date Posted: 09 November 2024 at 8:14am
| Nice work! |
Posted By: Keen
Date Posted: 12 November 2024 at 3:55am
| Awesome! |
Posted By: Bams
Date Posted: 12 November 2024 at 8:10am
| this i a wholesome thread to follow! keep up the nice work and story of this build. |
Posted By: Teunos
Date Posted: 12 November 2024 at 5:25pm
|
This is awesome! I can't see any warping of the 3D print, looks like a really decent printer you have and some skills to get it this good. The mold also came out great.
Did you eventually solve the dip, or just accept it? ------------- Best regards, Teun. |
Posted By: JulianDA
Date Posted: 12 November 2024 at 10:59pm
|
The printer is in fact a cheap one straight out of the box without any modifications (It is an Ankermake M5C). We are also really happy with its performance. 3D-printing definitely improved a lot in the last years. Regarding the dip i did nothing new since my last post with measurements. I will address that "problem" when the Horn ist finished and the dip is still there. My thinking is as follows: Since the horn has no abrupt changes in area/flare, no straight walls and no diffraction apparatus of any sort, it itself should behave fairly good when fed with a appropiate wavefront. With this in mind, it should be way easier to work on the phaseplug when its prototypes are mounted to this new horn. And just for completion: The horn has a size of 500x360x240mm (WxHxD), a constant directivity of 60x50° (HxV) and it loads the driver to 400Hz (will probably be used from 600-800Hz up) Here is a measurement of the printed horn with the extended phaseplug in the near field (20cm) with 1dB increments, no crossover, no eq, no smoothing (but windowing at 14ms after peak)  Here is a measurement at 60cm with some eq to get it flat. (We could not measure further away this time without reflections) No crossover, no smoothing (but windowing at 14ms after peak)  Here are measurements of the eq'ed response in 10° increments from 0° to 60° in one direction of the horizontal plane (so 0-120° absolute coverage angle). Here i used 1/6dB smoothing so one can actually somehow read the graphs  And the same for the vertical. This time from 0° to 40° (so 0-80° total).  We really need to get these measurements done outside....but i think one can see a tendency of how this horn performs. There will be better measurements when we finished the one in glasfibre :) |
Posted By: Teunos
Date Posted: 07 December 2024 at 9:14am
|
Measurements look great, but doing them outside would be a great idea. The null at 550Hz looks like it might come from a reflection. Is this a gated measurements? What's the vertical scale on those measurements? Difficult to really assess directivity without that :-) Looks like usable range is up to about a crossover around 3kHz consdering directivity, should be high enough to cross into a sweet 1" exit, 2"driphragm driver. ------------- Best regards, Teun. |
Posted By: JulianDA
Date Posted: 09 December 2024 at 1:12pm
-We will do "the real measurements" outside when we have constructed a horn from glasfiber.
-All measurements that i posted in my latest post were gated at 14ms. I am also pretty sure that the dip is caused by reflections. -The vertical scale of all measurements is 1dB per increment as stated in my post. The microphone was shifted by 10° for each new measurement in the "directivity-graphs" starting with 0° as the first measurement. -> This means the desired 60° directivity in the horizontal plane is reached by the 4th measurement (purple) and 50° in the vertical is between the 3rd and 4th (orange and green) -We are eyeing on the B&C DE360 as the HF Driver over this Horn. The desired Crossover is somewhere between 3000-3500Hz but it can be higher if thats a possibility. But i would like the midrange horn to be well behaved a good bit higher than crossover. We will see how this turns out with the glasfiber horn and optimised phaseplug. |
Posted By: JulianDA
Date Posted: 03 February 2025 at 6:19pm
|
We build the first horn from GRP. Its not the prettiest and still needs a lot of cleaning and finishing, but there will be some measurements -hopefully outdoors- that follow in a few weeks. This is the current state of it:   |
Posted By: smoore
Date Posted: 21 March 2025 at 3:11pm
| Good work! Looks solid! Any developments? |
Posted By: JulianDA
Date Posted: 26 March 2025 at 2:53pm
|
Yes, there are developments. I still did not get to measure the horn/driver combination outdoors, but the work still continued. For the phaseplug development i now use mostly impedance measurements to spot resonances. This worked with great success! The following graph shows a comparison in impedance and frequency response of 3 different phaseplugs measured on the GRP-Horn.  -The first one (green) is the original PP constructed from the results of my simulation. -The second one ("purple") has shifted PP entrances by 2mm <- seemed like the right direction -The third one (red) is the newest PP which has shifted PP entrances by 4mm This is the newest PP (in the pictured version with a twist in the outer channel for path-length-matching)  The first frontchamber cavity resonance at ~2,6kHz is almost completely suppressed. It seems like shifted entrances by 3,5mm will be perfect. The second frontchamber cavity resonance at ~4-5kHz is a bit more complicated. It should also be suppressed by now. The problem at this point is that the driver itself has strong resonances (You can see them in the impedance and response measurements in the drivers datasheet: https://www.bcspeakers.com/en/print/lf-driver/6-5/8/6nsm51-8" rel="nofollow - https://www.bcspeakers.com/en/print/lf-driver/6-5/8/6nsm51-8 ) Things i learned: - Impedance measurements are a perfect tool for spotting resonances and working on them! especially when good fr-measurements are difficult to do - the hump between 1-2kHz comes from the driver itself and has to be EQed or dealt with in the crossover - extending the center of the plug, like previously suggested, makes no difference now that the horn is perfectly adapted to the PP - the resonances of the diaphragm around 4-5kHz will probably be a problem in the future
- matching path-lengths makes things worse. My reasoning is, that matched path-lengths are only valid if both PP entrances are driven in phase with each other. This condition may not be met with the realities of a real diaphragm that does not move piston-like at those frequencies The next development will be changing the areas of the PP-channels. Right now the outer channel has a bigger area than the inner one. I will test same area channels and channels where the inner one has the bigger area. |
Posted By: Timebomb
Date Posted: 29 March 2025 at 9:27pm
|
Looks nice :) Some good work gone into it! You making me think about doing one for the Celestion CN0617M or the CN0515M, i had a play with them last year. Are you matching the path length so that the two path lengths are the same? Or are you accounting for the time it takes for the wave to travel through the paper diaphragm to the outer ring? ------------- James Secker facebook.com/soundgearuk James@soundgear.co.uk www.soundgear.co.uk |
Posted By: Line Array
Date Posted: 30 March 2025 at 10:15pm
|
fun project but i think what QSC did in their 4-way Cinema speakers with a phase plug on the 10" midrange is more interesting ... it appears they used a 10" B&C driver that has a nasty resonance in its response and they designed a phase plug that tuned that resonance out ... they have a replacement parts page where you can find out the models of drivers they use ... or you can call them ... i don't see much value in taking a midrange driver that is already optimized to work without a phase plug and then adding a phase plug to it but if you can do what QSC did and actually use a phase plug to equalize response of an otherwise flawed driver then that is pretty cool most dedicated midranges probably will use some kind of soft felt paper that damps the resonance but if you can suppress a resonance with a phase plug you could use a driver made out of harder paper, say carbon fiber impregnated paper that was maybe on its own quite peaky in response and create something unique ... maybe that is already what you're doing in which case i apologize ... you can't really horn load a cone midrange because the horn would get impractically large - you can only really waveguide load it in which case the function of a phase plug is slightly different than in a horn ... in a horn you're padding out the volume of the cavity that would otherwise short out HF like an acoustical capacitor ... in a waveguide you can use a phase plug like QSC did to EQ the nasty out of the driver, or maybe to slightly narrow the effective diameter for better dispersion ... or both ... anyway good luck
|
Posted By: JulianDA
Date Posted: 18 April 2025 at 2:00pm
Thanks James :) I matched the lengths just geometrically and started my iterative prototyping from there. Everything else was to much guesswork for my liking. Keep us updated if you go through with your project! I will be following it with great interest! Its always nice to inspire people to start new projects :) |