After looking at these fancy CD waveguides for about 10 years, i finally found some time to wrap my head around mabat´s ath waveguide generator and made my own. Thank´s to sheer endluss information on this forum, i feel pretty confident that i am on the "right" track so far. After simulating probably 100s of iterations between christmas and newyears, i think i found a compromise that i am going to build.
Below you can see the current state CAD design and ABEC SPL response with enclosure. It´s supposed to be mated to an 8" fullrange, used as a Midwoofer up until 1,5kHz in this case, in a 60l BR enclosure. I want to use basotect on the inside walls to reduce mid reflections. Main design goal is to get rid of room reflections and keep as much midrange dynamics as possible. If general frequency response is within a reasonable target its a welcome bonus.
As much as i´d love to just get myself a Hypex FA (you can see it semitransparent in my drawing), i am not quite confident enough to sell my beautiful NS1000s to be able to afford those.
As of now i plan on building said enclosures, measure both drivers and use VituixCAD to design a simple crossover and take care of frequency response with DSP. I could go forward and spent another 2 weeks to setup a working Akabak simulation, but i am not sure it´ll be precise enough to get me in the ballpark, working with a CD horn in this case.
I´ve been looking into a Le Cléac`h crossover to compensate for HF driver position and increased center to center distance.
https://www.melaudia.net/zdoc/jml_crossovers_etf04.pdf
Honestly i have no idea how it is going to work out, but i might need some directions for crossover design.
Looks like i need to use a 10kHz first order lowpass to get somewhere close to a linear frequence response and attenuate efficiency to match with my woofers 90db? Is it possible to combine a first order lowpass with a 3rd order butterworth for a Le Cléac`h crossover concept? I might get some strange tweeter response below my crossover frequency i guess...
I will document further development here. As somebody building his first proper pair of diy speakers, i might need input at some point. 🙂
Below you can see the current state CAD design and ABEC SPL response with enclosure. It´s supposed to be mated to an 8" fullrange, used as a Midwoofer up until 1,5kHz in this case, in a 60l BR enclosure. I want to use basotect on the inside walls to reduce mid reflections. Main design goal is to get rid of room reflections and keep as much midrange dynamics as possible. If general frequency response is within a reasonable target its a welcome bonus.
As much as i´d love to just get myself a Hypex FA (you can see it semitransparent in my drawing), i am not quite confident enough to sell my beautiful NS1000s to be able to afford those.
As of now i plan on building said enclosures, measure both drivers and use VituixCAD to design a simple crossover and take care of frequency response with DSP. I could go forward and spent another 2 weeks to setup a working Akabak simulation, but i am not sure it´ll be precise enough to get me in the ballpark, working with a CD horn in this case.
I´ve been looking into a Le Cléac`h crossover to compensate for HF driver position and increased center to center distance.
https://www.melaudia.net/zdoc/jml_crossovers_etf04.pdf
Honestly i have no idea how it is going to work out, but i might need some directions for crossover design.
Looks like i need to use a 10kHz first order lowpass to get somewhere close to a linear frequence response and attenuate efficiency to match with my woofers 90db? Is it possible to combine a first order lowpass with a 3rd order butterworth for a Le Cléac`h crossover concept? I might get some strange tweeter response below my crossover frequency i guess...
I will document further development here. As somebody building his first proper pair of diy speakers, i might need input at some point. 🙂
This is one of the benefits to simulating as you can define that target and do whatever is necessary.
How would you generate a viable SPL curve of a compression driver in a horn? There are no TSP Data or something similar as far as i know...
Best would be to use the mearsured response as an input into VituixCAD, or am i missing something?
Best would be to use the mearsured response as an input into VituixCAD, or am i missing something?
You can consider using your PC as your DSP during speaker development. Especially if your PC has HDMI output. Connect your PC to a multi-channel home theater receiver using HDMI. Use Equalizer APO to listen to the different crossovers you want to compare. Independently amplify each driver using the different surround channel outputs on the HT receiver.
You can also do it with a USB audio interface if you have one, however, I personally think a PC with GPU/HDMI to HT receiver is the easiest setup because the receiver has matching amps plus a volume remote. No need to worry about the amplifiers that way.
After you work out your crossover you can decide if you want to convert to passive circuits or buy Hypex.
You can also do it with a USB audio interface if you have one, however, I personally think a PC with GPU/HDMI to HT receiver is the easiest setup because the receiver has matching amps plus a volume remote. No need to worry about the amplifiers that way.
After you work out your crossover you can decide if you want to convert to passive circuits or buy Hypex.
Funny you mention a AVR Reciever. I have just sold a Pioneers Reciever because i wanted to get rid of that gigantic thing....
But 4 channel USB audio interfaces are reasonably priced here in Berlin, i might just pick one up...good idea!
I have tinkered with Camilla DSP on my raspberry streamer, but its supposed to work as on PC aswell.
But 4 channel USB audio interfaces are reasonably priced here in Berlin, i might just pick one up...good idea!
I have tinkered with Camilla DSP on my raspberry streamer, but its supposed to work as on PC aswell.
That didn´t go so well..
At some point the nozzle hit something and moved the whole build plattform. Z-hop is activated, but i have been printing with filament i never used before. Today i did a set of calibration prints (k-factor and extrusion multiplier was way of. I´ll try again with standart support structure..
At some point the nozzle hit something and moved the whole build plattform. Z-hop is activated, but i have been printing with filament i never used before. Today i did a set of calibration prints (k-factor and extrusion multiplier was way of. I´ll try again with standart support structure..
Still nothing on the printing front. I dove into multichannel DSPing with a HDMI reciever though, but to no avail. Win11 only lists 2channel audio as available audio streams. The reciever is capable of 8channel PCM streaming with a lot higher bit rates than listed. Is there a way to force this somehow? According to Intel it should be available..
There is a solution to the problem. You can use this tool:
https://custom-resolution-utility.en.lo4d.com/windows
to manually add features to your output.
https://custom-resolution-utility.en.lo4d.com/windows
to manually add features to your output.
Now wait a minute: is this just a matter of selecting the right submenu in audio output settings and or EqAPO, or is Win11 in its standard incarnation not capable of multichannel audio via HDMI, so we need the above patching app?
In my case i had problems using HDMI multichannel PCM with my 12th gen i7 Intel Laptop. It was simply no option in Windows audio output settings, as pictured above.
Thanks to CRU i was able to add a lot more output options, that are available as a setting now. Wasn´t able to check if my AVR does actually recieve multichannel PCM now.
Thanks to CRU i was able to add a lot more output options, that are available as a setting now. Wasn´t able to check if my AVR does actually recieve multichannel PCM now.
Finally on a good way to get nice prints. Printing in Segments and fit them later delivers much nicer results. Still need to improve proper support structures to avoid damaging surfaces during removal and they seem to have a bigger influence on resulting surfaces on the opposite side. Somehow i had quite a few collisions printhead vs. printed object.
Has anyone tried printing with low infill percentages and fill with plaster of paris or similar?
Has anyone tried printing with low infill percentages and fill with plaster of paris or similar?
@Psytron
Thanks for the offer of sharing some of your hard earned 3D printing experience.
Mine is limited to printing 100s of small fixturing parts in ABS, which has been pretty uneventfull, so moving on to a waveguide that i originally wanted to print in one part was challenging to say the least.
I am trying to document my route to a functional design here in this thread.
Currently i still haven`t resolved my printing issues. Coming from Amazon Basic PLA (worst) to Sunlu PLA+ my results improved significantly. When the printer reaches the waveguides mouth (shallow overhang), it still starts curling up, hitting the printhead, clogging up the nozzle or just move the whole builtplate.
I am printing with an 90° intereference layer between object and support structure and 0.2mm distance (i think).
Are there any slicer setting i need to check for you to get an impression of what might be going wrong?
Thanks for the offer of sharing some of your hard earned 3D printing experience.
Mine is limited to printing 100s of small fixturing parts in ABS, which has been pretty uneventfull, so moving on to a waveguide that i originally wanted to print in one part was challenging to say the least.
I am trying to document my route to a functional design here in this thread.
Currently i still haven`t resolved my printing issues. Coming from Amazon Basic PLA (worst) to Sunlu PLA+ my results improved significantly. When the printer reaches the waveguides mouth (shallow overhang), it still starts curling up, hitting the printhead, clogging up the nozzle or just move the whole builtplate.
I am printing with an 90° intereference layer between object and support structure and 0.2mm distance (i think).
Are there any slicer setting i need to check for you to get an impression of what might be going wrong?
Last edited:
It does depend on filament however not everything can be solved in the slicer. Sometimes geometry is the best way to solve the problem. This can be done in a number of ways. One way is the orientation of the part, another is to add/subtract material(alter the shape) where the issue arises, splitting the part allows you to re-orient where the issue occurs. Another approach is to design your own support in a non-critical area, this can be bonded strongly to the curling area and cut off post print.
I usually split and have the critical areas in a good orientation to print or expect to spend time sanding a prototype. For a finished product I always expect to have to finish all surfaces and/or 3d print a mould and cast/composite layup the horn.
Yeah i can see where you are coming from.... design isn´t exactly perfect for printing. I was hoping my stiffeners would limit movement enough.
Do you think connecting those to the built plate would be enough. I have sketched a rough line where distortion usually becomes a problem.
Or should i add even more connections to the build plate?
All in addition to normal support structure as seen in this image:
Do you think connecting those to the built plate would be enough. I have sketched a rough line where distortion usually becomes a problem.
Or should i add even more connections to the build plate?
All in addition to normal support structure as seen in this image:
What you want to do is remove high degree overhangs that are not firmly and rigidly pulled down. You could do this in many ways. More, strongly attached support is one way. Stepping the outside of the horn is another, splitting the model in half would also work.
Here is the finished prototype.
Waveguide is printed in one piece, 5% infill and filled with sand. Final measurements will be posted, once im finished with DSP insanity. Still using EqualizerAPO through a AVR for the meantime.
I am already looking into adding some FRS5Xs for MEH foolery, to be able to extend the waveguide to 500 Hz. Right now crossover is about 1300 Hz, but i really don´t like the Visaton BG20. Just no way to make it behave in any shape or form. Any 8" aluminium woofer with enough bass extension is what i am looking for. Maybe Dayton RS 225.
I tried simulating some variant of that waveguide with bandpass mids and enclosure in akabak, but it is pretty much impossible to solve with my PC. I will just have to try it out. I plan on printing a waveguide with modular "inserts" to try different phaseplug/enclosure parts for the mids.
Waveguide is printed in one piece, 5% infill and filled with sand. Final measurements will be posted, once im finished with DSP insanity. Still using EqualizerAPO through a AVR for the meantime.
I am already looking into adding some FRS5Xs for MEH foolery, to be able to extend the waveguide to 500 Hz. Right now crossover is about 1300 Hz, but i really don´t like the Visaton BG20. Just no way to make it behave in any shape or form. Any 8" aluminium woofer with enough bass extension is what i am looking for. Maybe Dayton RS 225.
I tried simulating some variant of that waveguide with bandpass mids and enclosure in akabak, but it is pretty much impossible to solve with my PC. I will just have to try it out. I plan on printing a waveguide with modular "inserts" to try different phaseplug/enclosure parts for the mids.