EPIC CF4 info thread

[Ricci]

I've been messing with my 2 pairs of CF4's for a while and figured I would document some of it here. []. So far I haven't done much, but here's what I've got. One of my speakers (right rear surround) had a set of K31E drivers in it when I got it. I wasn't told and didn't find out until I removed the drivers from it. I've since obtained a set of the original neo 12's from a fellow forum member and replaced the K31e's which were also sold to another forum member. The K31E's sounded ok in there, but it did sound slightly different from the other speakers. The midrange was a bit different. Very obvious during pink noise. Now all 4 speakers sound the same as they should.

I'm kind of a bass junky and I always play around with ideas and modeling programs. To me the CF4's had a nice bass response with a lot of punch, but it seemed a bit boosted in the 50-80hz area, which is more than likely a room issue. I used a WFT3 to measure the K30N's (also K31'E's) and calculated the volume of the CF4 enclosure about 6.3cu ft and probably right at 6cu ft after accounting for displacement caused by the drivers, bracing ,ports, etc. I then plugged this info in and modeled it. The modeling showed that the tune of the stock alignment should be 45hz[:^)]? Should be 30-35hz going by the Klipsch FR spec and what I've heard. This would definitly cause a peak in the bass and seems kind of high to me. After doing this I decided that I wouldn't mind trying them out with the tune lowered a bit. I've since replaced the 2.5" ports with 10" ports. This should've dropped the tune down to about 30hz and would help flatten out the upper bass according to the modeling. I ran this for awhile and I do like the sound better with the lower tune. The extension is improved and the upper bass slightly reduced. I've since added a large and powerful SW system to this, so I may return them to the stock tuning as the extra extension isn't needed.

I'll be posting a bunch of graphs shortly that document things , or show the changes.

[Ricci]

This is a set of graphs showing the vent tuning of the CF4 in various configurations. This is a close mic low volume sweep measurement of the top driver to show the minimum motion notch.

PINK trace is sealed.

PURPLE is the stock config with both 2.5" ports showing a 31hz tune.

GREEN is with one 2.5" port and one 10" port indicating a 28hz tune. I had to try this as I've always wondered about how this would work. Apparently it appears like an average of the 2 unequal ports length to the system. In this case a pair of 6.25" ports. Not sure what other issues this may cause, but Polk used unequal ports in some of their designs.

RED is with one 2.5" stock port and the other plugged. 25hz is the tune.

BLUE is the new tuning with both 10" ports. The tune is 23.5hz.

YELLOW is with 1 10" port and the other plugged. 19hz tune.

[Ricci]

Here is a measurement showing only the stock tune and the new tune with double 10" ports.

This graph shows the FR effects at the listening position of the 2 different port tunings. The lower tune has a more extended response.

Here is another one from a different listening position.

[Ricci]

This is the FR of the mains pair of CF4's at the listening position with the lowered tuning. No EQ and no room treatments. 1/6 octave smoothing applied.

Here is a waterfall of the bass range below 200hz of the same response.

Finally here is the combined response of the front CF4's with the dual SDX 15 subwoofer, which is run about 3 or 4 db hot.

[Ricci]

The most interesting things are the fact that the unequal port lengths appear to equalize each other in regards to tuning the enclosure. I wonder what detrimental effects a set-up like this would cause and perhaps whether there may be a few advantages? I'll have to think on that one.

The other is that none of the enclosure modeling programs accurately predicted the low tunes that were measured. The 31hz tune with the stock 2.5" long ports just does not jive with the available enclosure volume. Neither do the 10" long port measurements. [^o)]. What is predicted is 45hz vs 31hz actual and 30hz versus 23.5hz actual. The modeling programs indicate that an enclosure volume of 9.4cu ft is needed to get the measured tunes with the ports used and that much volume is just not there. Intrigueing.

[bodcaw boy]

what is your measurement system again? the graphs look impressive.

[Ricci]

The measurement program is called REW. It is a free software download available at The Home Theater Shack. Many DIYer's use it. It's not as full featured, powerful, or probably as accurate as the true professional stuff, but it's great for people like me.

[chili bone]

The most interesting things are the fact that the unequal port lengths appear to equalize each other in regards to tuning the enclosure. I wonder what detrimental effects a set-up like this would cause and perhaps whether there may be a few advantages? I'll have to think on that one.

The other is that none of the enclosure modeling programs accurately predicted the low tunes that were measured. The 31hz tune with the stock 2.5" long ports just does not jive with the available enclosure volume. Neither do the 10" long port measurements. . What is predicted is 45hz vs 31hz actual and 30hz versus 23.5hz actual. The modeling programs indicate that an enclosure volume of 9.4cu ft is needed to get the measured tunes with the ports used and that much volume is just not there. Intrigueing.

I have a pair of CF3s and they have a 6" port. In doing some research on them, I found out these speakers were first introduced with the 6" port but then later changed to the 2.5" port. If you do a forum search, there is a member that discusses this in length.

Carl