[2004-09-18]

My Fostex FE166E speakers are a little bit too forward for my tastes. The tradeoff for thier very nice resolution and detail is a bit of shoutiness and in your faceness that I am getting a little tired of. I thought they would get better with breakin, but a couple hundred hours hasn't done much. I had the opportunity to hear a pair of Matt Wiebe's Fostex FX200s in open baffles and the sound was almost the exact opposite - very laid back and smooth. If the Fe166Es are trying to jump out and smack you in the face with the music, the FX200s are just sitting back and playing the music for you. I hope the ff225K will have a similar presentation, but with a bit more efficiency.

The thing I really like about my Martin King TQWTs is the bass. I can't believe how deep it is for fairly puny fullrange drivers. Hoping to get results as good, I used his Mathcad worksheets to design these enclosures as well. This time I copied his Lowther DX enclosures so that I would be able to try out several fullrange drivers in the same boxes, including the FF225Ks, JBL LE8Ts, FX200s, and possibly even lowthers if I can beg, borrow, or steal a pair :) The mass loaded transmission line designs appear to be incredibly driver-independent, and look like they give similar response with just about any driver you can fit in the baffle cutout.

Be sure to visit Martin King's fantastic page on transmission line speaker modeling, quarter-wave.com.

I added about 40% to the various Q values of the driver, because I have heard it suggested that this would help to model the effect of using the speakers with low damping factor tube amplifiers.

The SPL response is pretty good. I would like about 10Hz lower frequency extension, but oh well. My FE166E speakers do go a bit lower, but I think the added cone area will give these 8" drivers a little bit more oomph.

Thorsten Loesch recommends using this cool eave-filler stuff to scatter the reflections directly behind the driver because these lightweigt fullrange drivers are particularly susceptible to high frequencies being reflected back to the speaker. For $3 at Home Depot I figured it was worth the expense :) The gray rectangle below the driver cutout is an adhesive vinyl floor tile, also from Home Depot. I wanted to double the thickness of the front baffle, but I screwed up cutting the MDF and ran out of material, so this was my token effort to damp it a bit more than the rest of the box.

After listening to the drivers sitting on the workbench for a couple hours while assembling the cabinets, it was nice to finally hear some bass out of them once they went into the unfinished enclosures. Bass at this point is quite boomy and ragged because of the untuned port opening, and lack of stuffig in the speakers. The sound is promising though, and the nasal midrange appears to be going away with breakin.

A quick note about construction techniques. I cut the panels for these using a circular saw and a straight edge. I find this to be much easier, and less dangerous, than trying to feed a hoge sheet of MDF through a tablesaw. It is also plenty accurate as long as you use identical measurement procedures for every cut. The panels are withing 1/16, and very square. I use either a piece of unistrut channel or a factory edge of a strip of MDF as a straight edge, and clamp it down using a pair of those spring clamps that look like big metal clothespins. The holes are cut using a router and circle cutter to make the first hole, then for the second hole, just clamp the panel with the hole on top of the panel that needs the hole cut in it and use a router bit with a bearing on top of a flush cutting blade to transfer the hole to the uncut panel. Works like a champ.

I find that a baffle step correction circuit is essential for getting the tonal balance right with speakers like this, but the efficiency loss of a speaker-level BSC circuit is an unfortunate tradeoff. Paul Joppa offers some insight into line level baffle step correction on the new Valve Wiki. He uses this same enclosure designed for lowthers as a baselind for his line level discussion. I built a version of his circuit designed for the 50k imput impedance of my Monkey amp, and it sounds really great. You still lose some gain between your preamp and power amp, but there is no power loss between the amp and speaker, where it really matters. And you don't need a zobel network to get a flat impedance response from the speaker. Read Paul Joppa's notes for the potential gotchas with this design though. You need a low source impedance to drive it, and low cable capacitance behind the BSC. I now plug these BSC modules directly into the input jacks of the amp using male-male adapters to minimize the cable capacitance between them and the amp.