´I have remade the ports for my DIY subwoofer with a 15TBW100 driver. The ports were previously undersized (two round 103mm ports, 164mm² port area), the tuning frequency was about 32.5 Hz with a net internal volume of 100 liters (~3.5 cu ft). The undersized ports had a 14mm radius flare on both ends.
The new rectangular ports have a total 296mm² area and a 55mm flare radius on both ends, but they sit directly against the bottom and walls, so only 2 of 4 sides of the ports are flared. The tuning frequency increased to 38.5 Hz, and the net internal volume is now 81L, due to the larger ports.
I've made frequency response measurements of the subwoofer at different power levels before and after the changes using UMIK-1 (6db gain) and REW, and put together some port compression data in a spreadsheet.
All measurements are ground plane 2m. Velocity in m/s and compression in db.
The spreadsheet shows input voltage, WinISD simulated port velocity and measured compression compared to the 5.66V sweep for Fb - 2.5 Hz, Fb, and Fb + 2.5 Hz. It also shows compression at 130 Hz (low impedance) in order to get some indication of possible power compression, depending on sweep level.
Some things I found interesting:
Port compression mostly affects frequencies below Fb. The undersized ports had only 0.7 db port compression at 77,9V and 35Hz, despite a (very high) simulated port velocity of 36,3 m/s.
Power compression at 130hz was considerable, although later in the sweep (more time for the voice coilt to heat up), it means that power compression likley affects the high power sweeps. For instance, there is more compression at 41Hz, 112V after the mod, than at 35Hz, 77,9V before, despite lower port velocity (28,3 vs 36,3 m/s).
The new rectangular ports have a total 296mm² area and a 55mm flare radius on both ends, but they sit directly against the bottom and walls, so only 2 of 4 sides of the ports are flared. The tuning frequency increased to 38.5 Hz, and the net internal volume is now 81L, due to the larger ports.
I've made frequency response measurements of the subwoofer at different power levels before and after the changes using UMIK-1 (6db gain) and REW, and put together some port compression data in a spreadsheet.
All measurements are ground plane 2m. Velocity in m/s and compression in db.
The spreadsheet shows input voltage, WinISD simulated port velocity and measured compression compared to the 5.66V sweep for Fb - 2.5 Hz, Fb, and Fb + 2.5 Hz. It also shows compression at 130 Hz (low impedance) in order to get some indication of possible power compression, depending on sweep level.
Some things I found interesting:
Port compression mostly affects frequencies below Fb. The undersized ports had only 0.7 db port compression at 77,9V and 35Hz, despite a (very high) simulated port velocity of 36,3 m/s.
Power compression at 130hz was considerable, although later in the sweep (more time for the voice coilt to heat up), it means that power compression likley affects the high power sweeps. For instance, there is more compression at 41Hz, 112V after the mod, than at 35Hz, 77,9V before, despite lower port velocity (28,3 vs 36,3 m/s).
Of course you would find more compression with more voltage at or very close to tuning. This is a region dominated by the impedance minima, where you are dumping some real power into the coil. Driver movement is near stall, and the coil is rapidly heating.
This has nothing to do with the larger ports.
This has nothing to do with the larger ports.
“Power compression at 130hz was considerable, although later in the sweep (more time for the voice coilt to heat up), it means that power compression likley affects the high power sweeps. For instance, there is more compression at 41Hz, 112V after the mod, than at 35Hz, 77,9V before, despite lower port velocity (28,3 vs 36,3 m/s)”
What i meant by this is exactly what you are saying, higher compression despite larger ports means that thermal compression significantly affects the results at the high power sweeps.
I may not have been clear enough.
What i meant by this is exactly what you are saying, higher compression despite larger ports means that thermal compression significantly affects the results at the high power sweeps.
I may not have been clear enough.
I have some difficulty to replicate your speaker and port specifications.
the two ports are 103 mm long, correct?
the port cross section surface is 164 mm2? that seems waaay to small if I understand it correctly.
If I try to define a 100 liter enclosure with 32,5 Hz tuning that needs one port with about 60 cm2 and 10,3 cm length (or two with 30 cm2 each).
I did some research regarding port compression and according to my findings the relation of port radius and air displacement is the relevant one.
if you like have a look here:
https://www.diyaudio.com/community/...rbers-and-port-geometries.388264/post-7755603
This is brilliant. I have been heavily considering rebuilding my cabinets for my two Dayton Epique 7’. I used two 2” pipes which ended up being a tune of ~27hz. There’s a rap song with a stepped bass line with a bottom note of 27hz and I aim my designs to be able to reproduce this. Juggling velocity, 1st resonance, length and such with the small subs I like to employ generally ends with my fighting to get below my magic 27hz tuning and I tend to end up right at 27hz.
I have found that the lowest note tends to be attenuated to my dismay. The idea that I may be dealing with power compression as opposed to velocity compression is extremely interesting. Might it be better to use smaller diameter ports that push the stalling of the cone lower? Essentially, this should push the group delay spike down lower as well?
I have found that the lowest note tends to be attenuated to my dismay. The idea that I may be dealing with power compression as opposed to velocity compression is extremely interesting. Might it be better to use smaller diameter ports that push the stalling of the cone lower? Essentially, this should push the group delay spike down lower as well?
What levels are you listening at with the small subs? Approximate power input?
At lower volumes, our hearing becomes less sensitive to low frequencies, so that could be causing it.
https://en.wikipedia.org/wiki/Equal-loudness_contour
I made some new tests in an attempt to separate thermal and port compression.
Instead of using sine sweeps, I used short CEA-2010 bursts to minimize voice coil heating.
The subwoofer in the OP with 296 mm2 ports would compress 1 dB at 120V input and simulated port velocity of 35.6 m/s, with CEA-2010 bursts. I measured 1.6 dB compression at the same voltage and frequency when i used sweeps instead. The difference is most likely attributable to thermal compression.