Redoing an old build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
After repainting and sealed around the vents a bit more, I'm getting the attached FR. Looks like the dip at the upper end of the passband has almost vanished. I added a little stuffing in the vented section, and that seems to have appreciably reduced the resonance at 400 Hz

The impedance curve still looks unusual however.
 

Attachments

  • 20180206-INF10-BP (stuffing)-FR.png
    20180206-INF10-BP (stuffing)-FR.png
    60.6 KB · Views: 176
  • 20180206-INF10-BP (stuffing)-IMP.png
    20180206-INF10-BP (stuffing)-IMP.png
    58.8 KB · Views: 182
#%(@Q#@$()&*!!

Apparently for the last test there was some leakage in the vented section through some of the screw holes in the access panel. Now that everything's buttoned up properly,. the hump in the response is back.

Grr...

I'm going to call this one "Enigma", because it's still a mystery to me why there's such a big difference between the sim and the actual result. And you'd think that a 6dB difference would make it rather insensitive, but it seems to be the other way around - it's like it's gained a few dB at low frequencies rather than losing a few at higher frequencies.

Oh well, at least the cutoff at the upper end is pretty smooth, and out of band noise is 20dB down from the passband. This 4th order BP box can actually be incorporated into a full range system without having to use a HP HP filter. And now that I know the "secret" about how to eliminate the lowest out of band resonance in a 4th order BP system, I'm tempted to try it out on another build.. :)
 

Attachments

  • 20180207-INF10-BP-FR.png
    20180207-INF10-BP-FR.png
    53 KB · Views: 171
Administrator
Joined 2004
Paid Member
Hi Brian,
Since the sub is normally rolled off well below 400 Hz, it might not be such a big deal anyway. That's probably what they were thinking when they used that woofer. It would be nice to be able to find out where it's coming from.

Do you have another different woofer you could place in there temporarily? It would just prove whether the 400 Hz hump is in the box or the driver.

-Chris
 
Hi Brian,
Since the sub is normally rolled off well below 400 Hz, it might not be such a big deal anyway. That's probably what they were thinking when they used that woofer. It would be nice to be able to find out where it's coming from.

Do you have another different woofer you could place in there temporarily? It would just prove whether the 400 Hz hump is in the box or the driver.

-Chris

That 400 Hz is definitely a by-product of the 4th order BP alignment. Its caused by a combination of vent "organ-pipe" resonance and the location of the vent's entrance in the vented chamber of the build.

Where this build is going "off the reservation" is the rolloff at the upper end of the passband, which is not predicted by the sim. The sim suggests that the passband will slope upward, and it's sloping downward.
 
Administrator
Joined 2004
Paid Member
Hi Brian,
Increasing inductance in the woofer VC would tend to roll the higher frequencies off. I don't know what the magnitude of the inductance and air load that looks like inductance to the amplifier is. But it seems that the 400 Hz peak is stronger than the mild effect I just mentioned. It really does look like some backwave is being reflected out the port. I think that was what you were saying here:
Its caused by a combination of vent "organ-pipe" resonance and the location of the vent's entrance in the vented chamber of the build.
The resonance of the port looks like a reactive mass to the system, it wouldn't resonate at 400 Hz because the box would damp that behavior. I was just seeing if there was a way to prove what it was for certain, and maybe some method to eliminate it. Sometimes aiming the inside entrance of the port in a different direction is all it takes. Or a simple deflector to break up the direct path to the port either from the back wall or the rear of the woofer.

I hate having unexplained deviations from the predicted response. That 400 Hz peak is crying to be solved as it is pointing to a problem.

-Chris
 
The resonance of the port looks like a reactive mass to the system, it wouldn't resonate at 400 Hz because the box would damp that behavior.

The vent itself will have resonant modes (pipe resonances) based on its effective length.

In this case, I think the 400 Hz peak is related to the dimensions of the box. It used to be huge when the vent was longer, but decreased significantly when the vent was shortened. I suspect that if I shorten the vent so its entrance in the box is halfway between the front and rear panels, that 400 Hz peak would likely disappear :). At the moment it's 20dB below the passband and basically inaudible when I use the subwoofer as part of a three-way system with no x-over.

The woofer's inductance is taken into consideration in the Hornresp sim. However the upper part of the passband, and impedance around that area of the response is not a good match for the sim, and I haven't figured out why yet.
 
I thought this might of interest - it's a THD measurement of my Enigma BP subwoofer (the rebuilt version). Notice how the THD peaks at 50 Hz and decreases on either side? I think this is a (beneficial) side effect of BP designs, where the vented chamber is acting as an acoustic filter. In this case, the vented chamber is acting like an acoustic HP filter starting from around 117 Hz, so any distortion components at or above this frequency will be subject to the filtering actions of the vent. Driver displacement in this design is actually supposed to peak around 40 Hz, but the build sounds pretty clean at higher volumes, because of the distortion-filtering feature of the design.
 

Attachments

  • 20180315-distortion (enigma).png
    20180315-distortion (enigma).png
    77.3 KB · Views: 151
Where this build is going "off the reservation" is the rolloff at the upper end of the passband, which is not predicted by the sim. The sim suggests that the passband will slope upward, and it's sloping downward.

Indicating the vent is acoustically a lot longer and the front chamber smaller, so I cut the vent area in half and got close to the measurement and some reduction in the front chamber Vb got it closer still, so my SWAG is that between the bracing/whatever over the driver combined with the vents being close to boundaries and maybe further preloading the driver a bit due to the layout, close quarters........ ?

GM
 
Sigh, I found a minor error in the model, see corrected information below and attached....

Brian,

I still don't think your input parameters are accurate.

Use Vtc to set the ported-side chamber volume (it's cubic centimetres, so adjust your numbers accordingly) and then you want:

S1 = port flare entrance
S2 = port tube diameter
S3 = port tube diameter
S4 = port flare exit

I'd set L12 and L34 to be equal, and probably an exponential expansion. L23 doesn't matter, since it's basically a straight pipe.

Properly modelling a cabinet can take quite a few steps.

Chris
 
If I do that, I lose the opportunity to model the offset of the driver in the vented section, and this influences the size of the peak around 400 Hz.

The wavelength of 400Hz is 86cm, so half-wavelength is 43cm. My bet is you've either got a standing wave or a pipe resonance.

Since it's designed to be a subwoofer, I'd prioritise simulating that part correctly first.

Chris
 
If I don't include the driver offset, I lose the opportunity to address one of the more troublesome aspects of a 4th order BP build - the out of band response peaks. The offset deals with the most important of those peaks - the first one. The others are higher up in the passband and can be reduced with a bit of lining in the vented section.

In its current form, the "Enigma" can be run without a LP filter, as the out of band noise is significantly below the passband level. Even then, there's some room for even more improvement, as an FR measurement suggests that the first out of band peak can be reduced even further if the dip in the response caused by the offset driver lines up directly with it.

I have another possible explanation for why the passband response seems a bit off from the model - semi-inductance. When I have some time, I'll pop the driver from the enclosure and measure it to determine the semi-inductance parameters.
 
If I don't include the driver offset, I lose the opportunity to address one of the more troublesome aspects of a 4th order BP build - the out of band response peaks. The offset deals with the most important of those peaks - the first one. The others are higher up in the passband and can be reduced with a bit of lining in the vented section.

Brian,
I understand that.
I was simply offering a reason as to why the pass-band measurements and sims don't line up, and a potential solution to that.

Chris
 
I think this is a (beneficial) side effect of BP designs, where the vented chamber is acting as an acoustic filter.

so any distortion components at or above this frequency will be subject to the filtering actions of the vent.

but the build sounds pretty clean at higher volumes, because of the distortion-filtering feature of the design.

These are the EXACT reasons why I prefer the sound of BP4/BP6's over Sealed and BR's!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.