So first off I should come clean and admit the cardinal sin of building someone else's design. That is trying to use an alternate driver, but first hear me out.
I am building this http://projectgallery.parts-express.com/speaker-projects/synergy-horn/
And having listened to the bass and lower midrange of the morel kw-1 and dayton dcs165-4 in sealed boxes I much prefer the sound of the morel
I know drivers can sound better in horn, but in a synergy horn or least this design the woofer don't see much loading until above 200hz.
Therefore I'm trying to use the morel kw-1 in the design and use the same mid and tweeter.
So far I'm trying to model the original woofer in the horn and compare this to the measured response from the designer. So far I've tried modelling the driver at the mouth and found this to have too big of a peak at 500hz, however modelling it as an offset horn I get hardly any peak as if their is barely any horn loading.
Anyone have suggestions? I've attached my hornresponse inputs and a .pdf of the project here. See post 153 here https://techtalk.parts-express.com/forum/tech-talk-forum/1331723-synergy-horn-build/page11 for the response I'm trying to replicate. All three drivers raw responses are in post 63.
I am building this http://projectgallery.parts-express.com/speaker-projects/synergy-horn/
And having listened to the bass and lower midrange of the morel kw-1 and dayton dcs165-4 in sealed boxes I much prefer the sound of the morel
I know drivers can sound better in horn, but in a synergy horn or least this design the woofer don't see much loading until above 200hz.
Therefore I'm trying to use the morel kw-1 in the design and use the same mid and tweeter.
So far I'm trying to model the original woofer in the horn and compare this to the measured response from the designer. So far I've tried modelling the driver at the mouth and found this to have too big of a peak at 500hz, however modelling it as an offset horn I get hardly any peak as if their is barely any horn loading.
Anyone have suggestions? I've attached my hornresponse inputs and a .pdf of the project here. See post 153 here https://techtalk.parts-express.com/forum/tech-talk-forum/1331723-synergy-horn-build/page11 for the response I'm trying to replicate. All three drivers raw responses are in post 63.
You do need to model it as an offset and Hornresp has a MEH wizard for this purpose. Modelling it the other way is just treating it like a bandpass which is why there is a pronounced peak before rolloff.
In the parts express thread Hornresp inputs were given for the horn which look completely different to the ones you are using.
Different woofers will react differently and it not a simple matter of choosing the one you like the most in another setting, they may not work well in this one.
The Vtc is the volume in front of the cone and you would need to estimate that somehow, the morel might be similar to the dayton or not. The area Atc is probably close enough if the drivers are the same size but there is a difference in Sd.
The sharp dip is the reflection notch from the tweeter, that should change with the right horn parameters and look like the graph in the other pdf.
Try using the values from the Parts express forum pdf for Hornresp in a MEH wizard and estimate if there is a likely Vtc difference, you have already changed the TS parameters so there isn't much else to do if you plan to copy the design as is.
That's because the horn isn't at least a 1/4 wavelength in its center axis direction to support much below 550 Hz without significant contributions from the room's boundaries (boundary gain). You need a much bigger horn to support horn-only loading below 550 Hz.
In the K-402-MEH design, Hornresp modeling didn't do much other than tell me how much sensitivity there was from dual 12" woofers vs. dual 15" woofers (I chose dual 15"), and where to put the "off-axis ports", i.e., the ports not on the central axis of the horn...those ports other than the throat itself. In this regard, Hornresp was useful for MEH work.
However, the power response of the Hornresp was clearly not very close to what I measured after I put everything together. This was no problem in my K-402-MEH design because I already knew the capabilities of the K-402 horn with a throat-mounted 8" driver only--i.e., the "KPT-305-MB" module--and knew that what Hornresp was telling me wasn't accurate. In fact, Hornresp was way off in being able to predict on-axis SPL of the woofers (inferred from the power response that Hornresp tells you).
I wouldn't worry too much about Hornresp predictions in on-axis SPL or even power response--otherwise known as integrated polar SPL response.
Above 1 kHz, I also found that Hornresp wasn't that good in predicting actual on-axis SPL--which otherwise should be pretty close to the power response at frequencies above ~1 kHz due to the horn's ability to contain all the high frequency driver's output without loss of polar control.
Chris
Thanks for your input fluid, If you read the fine print in the PE forum it states that this is the wrong hornresp model. He is using a different driver and a different flare rate.
When modelling an offset horn do I find the area of the horn in the center of the woofer port hole? or on the edge?
When modelling an offset horn do I find the area of the horn in the center of the woofer port hole? or on the edge?
Although they don't match the actual response In all the hornresp models I've made the morel and dayton have fairly similar freq response to each other. With the morel having a bit more hf extension. It has a lower MMS and Le.
In xsim I have a model of the current crossover with the dayton dcs165 with the free air frd. Should I just design a crossover for the morel that matches this curve? For example the current crossover starts to roll of the dayton at 200hz to combat the rising response of the horn for an ultimate 400hz crossover to the mid.
I just don't know how to get a good phase response and a good null if I flip polarity of the midrange. I'm fairly new to crossover design.
In xsim I have a model of the current crossover with the dayton dcs165 with the free air frd. Should I just design a crossover for the morel that matches this curve? For example the current crossover starts to roll of the dayton at 200hz to combat the rising response of the horn for an ultimate 400hz crossover to the mid.
I just don't know how to get a good phase response and a good null if I flip polarity of the midrange. I'm fairly new to crossover design.
The best way to know what to so with a passive crossover design (I recommend something like first order to retain good phase response) is to build the horn, place the drivers on it, and measure each set of drivers acoustically and electrically, then use that in some sort of passive crossover simulator to design the passive network with actual horn-driver response--not simulated. Using XSim only isn't what I'd call a success-oriented approach.
Chris
Chris
Last edited:
Chris, That's what the original designer John H did he measured all drivers frequency response in the horn and then made a xsim model with all the raw responses. I'm basically just trying to figure out a new low pass section of the crossover but leaving the mid and tweeter as is.
I'm trying to take a first stab at the low pass crossover with the morel. If it needs modifying after it's assembled that's ok, I just thought with hornresp and xsim I could figure something out.
I'm trying to take a first stab at the low pass crossover with the morel. If it needs modifying after it's assembled that's ok, I just thought with hornresp and xsim I could figure something out.
Nothing. But you need real data to design a passive crossover for MEH design (and other horn-loaded designs)--not Hornresp simulations...in my experience.
Chris
Any approach not using actual acoustic/electrical measurements of the horn/driver sections individually and in combination (your pick of software package to do that...I'd use REW, etc.) isn't what I'd call a viable approach to having a real hi-fi result--in my experience. You need the measurements of the actual hardware and how it actually performs to get the passive crossover network right--assuming that you're also trying to get relatively flat SPL and phase response (i.e., plus or minus 1.5-3 dB using psychoacoustic smoothing). It's like the difference between the living and the dead...again, in my experience.
If it were me, I'd invest in a miniDSP 2x4 HD and get that right first with crossover and EQ filters to achieve a good overall transfer function response (SPL and phase responses), then use the active crossover settings to help you arrive at a usable passive network design--just like commercial loudspeaker engineers do it nowadays.
Chris
I did not read the fine print, it seemed odd that the entrance would be 4cm but it is.
I'm not sure which area you mean. bwaslo's spreadsheet and accompanying pdf is a good way to understand all the options.
I hadn't realised what you were trying to do. Measure what you end up with and work out the crossover from there.
