The simulation doesn't "lie." Your port size is capable
of causing significant audible noise. If you can't hear it the way the speakers are being used, then it doesn't matter so never mind. (appears to be a rear facing port? the close positioning to the wall probably helps tune down the box a bit lower and eliminate some noise.) Keep in mind though, that your TV may have a high pass filter in place so you may get much better performance from a line-out to a nice little chip amp (they are cheap, don't worry). Also, Those little "3"" drivers have a lot of excursion capability. Just because they a "small" does not mean they can't overwhelm a 1" port.
I took the liberty to show that guesswork builds are not generally going to be very flat (something i myself have only learned very recently, so please don't feel bad)
The attached is a simulation of your speaker build. There is no baffle step loss (google it) simulated into this, but you probably don't have too much with the placement of your speakers against the wall and one of them even being somewhat "corner loaded" so it's not a big deal....
Now.... The interesting thing (as I have learned recently) is that through proper simulation, this relatively un-flat speaker, could have been made much flatter at (likely) no significant change in cost.
As you can see in the simulation above, the x-over is not effecting the woofer the way the calculated x-over values were expected to. This is because the component values chosen were based on trying to get a ~3800hz x-over on 8 ohm drivers. The problem is, the woofer is not 8ohm at 3800hz, The woofer voice coil does not act like a normal resistor, but instead, like an inductor. The result is rising impedance with frequency.
Here's the trace of the impedance curve (ignore the large peak at Fs, it would be different with box loading anyways and it does not effect the X-over)
With that in mind, I suggest the following change (lowest cost change with greatest benefit):
You can use a cheap low inductance 10W 8Ohm resistor and an electrolytic non-polar cap for the zobel network. Nothing fancy needed here since it does not carry any signal to the speaker, it just corrects the behavior of the x-over network by correcting the impedance that it "sees." The value of "zobel C1" is not super critical, something close to 10uF will work well. (5-15uf range, I like the way ~10uF sims out best personally).
EricEdited by mdocod - 1/16/12 at 2:16am