Hi guys,
The modelling is specifically with slot ports. The tuning of the ports of the finished box was not as calculated and the ports seem to be resonating at 80hz (the recommendation for this volume seem to be 80hz tuning anyway.) Since I was pushing the limits of the port area, it may well have become over damped hence this result. Also, after converting the inches to mm plus the cutting accuracy of the panels has resulted in each port being 5mm taller than planned. I suspect this to be very critical in a design like this. The response from 60-200hz is jagged.( Dips at 63hz and 125hz and peaks at80 and 160hz)
I don't have the proper equipment to measure things and just use apps really. This build and evaluating is my first speaker project. I will try again with a bigger version of this enclosure and see what happens.
The modelling is specifically with slot ports. The tuning of the ports of the finished box was not as calculated and the ports seem to be resonating at 80hz (the recommendation for this volume seem to be 80hz tuning anyway.) Since I was pushing the limits of the port area, it may well have become over damped hence this result. Also, after converting the inches to mm plus the cutting accuracy of the panels has resulted in each port being 5mm taller than planned. I suspect this to be very critical in a design like this. The response from 60-200hz is jagged.( Dips at 63hz and 125hz and peaks at80 and 160hz)
I don't have the proper equipment to measure things and just use apps really. This build and evaluating is my first speaker project. I will try again with a bigger version of this enclosure and see what happens.
Here's an in-situ (on desktop, backs to the wall) measurement, mic placed close to where I listen:
The 400hz dip is desktop bounce cancellation.
IG
An externally hosted image should be here but it was not working when we last tested it.
The 400hz dip is desktop bounce cancellation.
IG
Sim of IG's W4-1320SIF DBR
Using the dimensions provided by IG, here is the sim in AkAbak with the speakers far from a back wall (60 in) and measurements at 1 m and 1 watt. The bass extension and location of the 150 Hz is quite close. The oscillations would be damped in reality but I have no way to model the damping.
Using the dimensions provided by IG, here is the sim in AkAbak with the speakers far from a back wall (60 in) and measurements at 1 m and 1 watt. The bass extension and location of the 150 Hz is quite close. The oscillations would be damped in reality but I have no way to model the damping.
Attachments
Not bad! Thanks for doing this.
Here are very nearfield measurements to try and isolate cone output, vent1 output and vent2 output.
Blue: cone
Red: vent1 (top)
Green: vent2 (bottom)
An externally hosted image should be here but it was not working when we last tested it.
Very close indeed, right down to the 750Hz peak on the red trace, which I assume is the 9" half-wave resonant mode of the main chamber.
IG
Last edited:
A notable difference is the tuning of the main chamber; I measure ~60Hz compared to your model's ~70Hz.
IG
IG
Very compelling comparison of simulation v measurement.
How difficult would simulation of CBT array be?
Keel's Constant Beamwidth Trasnducer line array.
http://www.diyaudio.com/forums/mult...idth-transducers-line-arrays.html#post2162871
http://www.diyaudio.com/forums/mult...idth-transducers-line-arrays.html#post2162871
Barleywater,
A curved line array... The cuved nature requires some geometry to define the relative x, y, z coordinates and the angle of each driver's diaphragm. Once you have that, it is easy as you define the location of each driver and how they are connected (parallel series arrangement). I am assuming the back just goes into a sealed common or separate stuffed chamber or aperiodic enclosure. If you provide the geometry (x,y,z, angle) for each driver and the driver T/S params I can model this. I don't feel like working out the geometry as this is not something I am particularly interested in. Akabak will provide spl polar plots but not color coded contour xy plots.
A curved line array... The cuved nature requires some geometry to define the relative x, y, z coordinates and the angle of each driver's diaphragm. Once you have that, it is easy as you define the location of each driver and how they are connected (parallel series arrangement). I am assuming the back just goes into a sealed common or separate stuffed chamber or aperiodic enclosure. If you provide the geometry (x,y,z, angle) for each driver and the driver T/S params I can model this. I don't feel like working out the geometry as this is not something I am particularly interested in. Akabak will provide spl polar plots but not color coded contour xy plots.
One more thing about the DBR sim in akabak: it uses real box dimensions (not just volume) and port dimensions with locations based on actual placement on baffle. The length of the box will have an impact on making it behave as a MLTL if stretched out long. If one were to do a hybrid MLTL and DBR what parameters should be tuned or lengthened?
- Status
- Not open for further replies.