Simulation accuracy with boxplot

[Kei]

Been having another look at the simulations from boxplot for my diy kef chorales. Drivers being used are Kef B200 SP1063 and SEAS 22TFF H1280. The cabinet is approximately 24L (265x255x480mm) and ported. (60x170mm)


This is the current crossover setup:


and it's simulation, with a large "hump" in the mid range:


Taking the 16uF capacitor out of the notch filter results in a much flatter plot (taking the whole notch filter out isn't as flat):



I'll have to admit that i cannot hear this large hump in the mid-range. I have also noticed that the bass response is much improved when it is simulated as a closed box vs ported. Is there any easy way to close the port effectively? Presently is is fully stuffed with dense foam matting, but that doesn't seal them. (and according to boxplot sims, a stuffed port has even worse LF than both open ported and closed box)

[dumptruck]

A ring of blue tak / rope caulk with a piece of cardboard or plastic on top.

[speaker dave]

Stuffing the port with foam is fully adequate to convert to sealed box. Marantz sold a series years ago that came with foam plugs for tuning. Even if it leaks a little you just have a leaky sealed box which will have lower Qm but no other consequences.

Your crossover phase summing could be better. You don't have cancelation at 3kHz but you get no gain, implying the phases are over 100 degrees apart. Try some simulations with higher or lower order on the woofer or tweeter. You could end up with a higher sum at the crossover point and a little flatter response.

Try a couple of ohms in series with the 12 uf. It will soften the woofer corner and change the phase shift. You may need to flip tweeter phase.

David S

[tvrgeek]

Your sim won't tell you about the diffraction problems you will get from this cabinet. That may really confuse what you hear vs the sim. You really need to measure.

[Dissi]

The simulation is not accurate unless you import real measurements of driver impedance and spl.

A heavily stuffed port results in a aperiodic closed box. That's the best you can do having a KEF B200 in only 24 litres.

[Kei]

Had the feeling the simulation was likely to be inaccurate. I know that it would be best to measure it, but i do not own any of the necessary tools. What would i need to measure it myself? (and approximately how much would it cost?) With the port dimensions (51mmx170mm) and box volume (24L), am i right that they are tuned to about 36Hz?

With the front baffle being inset and the B200 not being flush mounted like kef suggest it should be, i imagine things are quite different to the simulations. The grilles sit within about 3mm of the driver face and are cutout tightly around the woofer and the edges are chamfered. Teh tweeter and port cutouts are much larger. This is how they look with the grilles fitted...


This is the simulated phase response (also likely to be inaccurate)...


I enjoy listening to them, particularly in the small room they are used in. They do sound rather weedy compared to my celestion ditton 25's due to the lack off LF below 45Hz but they have smoother mid range and slightly crisper highs. (they are slightly bright)

[Kei]

What would i need to measure the response characteristics myself? (and approximately how much would it cost?)

[Kei]

Spent some time having a search about and it seems that a Behringer Ecm8000 with a suitable phantom power supply and an XLR to 1/4" jack matching transformer would be the right tools combined with some suitable software like JustMLS.

Have i missed anything out?

[mdocod]

You're (estimated to be) about 130 degrees out of phase at the crossover point on-axis.

is that simulation taking into account the differential in distance from listener to each voice coil or not? It's possible that you have much better summing through the crossover at a real world listening position than the simulation would suggest. A couple inches at 3000hz could swing that right back into phase but if the simulated phase includes the differential then I would suggest flipping the polarity of one of the drivers and have a listen.

[PeteMcK]

re:"Taking the 16uF capacitor out of the notch filter results in a much flatter plot " - if tried in the real world would also result in smoke from the amp....
the 22tff is an excellent tweeter, works well with a simple 2nd order xover...