When constructing a MLTL is it ok to mount the driver and port on the narrow side of the enclosure?

It is OK to mount the driver and port on either side, the important feature is the length and the placement of the driver and port along the length. Within reasonable limits, the dimensions of the cross-section area and the selection of the front of the enclosure are not so important.

also, I may be missing it, but I can't find any articles on the mass loaded lines on Martin's site, just the worksheet.

There is no formal documentation or derivation. The design evolved fomr the ML TQWT described in Project #2. Then in Project #4 under the enclosure design section I described how a ML TL works. I have copied that significant paragraphs below.

"I have been asked many times about the difference between a ML TL design and a simple bass reflex enclosure. From the outside the two look very similar and performance wise there is not a large difference. I think that the principle difference is the way the air volume in the cabinet is used to provide the spring that interacts with the mass of air in the port to form a resonant system.

In a bass reflex cabinet, the air in the box is compressed to a uniform pressure to form an air spring. Typically no damping material is added to the inside of the box so that the Q of the box remains high and the effective volume of air is predictable from the internal dimensions of the box. The shape of the bass reflex box is not that critical, only the internal volume matters. A bass reflex enclosure can be represented as a lumped mass hanging on a spring. If you displace the mass the entire spring stretches. When you let go, the mass oscillates at a predictable frequency that is a function of the springrate and the mass of air in the port. The key point is that the entire spring stretches linearly. This is a simple one degree of freedom mechanical system.

In my opinion, one of the negative attributes of a bass reflex enclosure is that any strong standing wave resonances in the enclosure will not be sufficiently damped. The lack of fiber in the center of the air volume allows energy from the back of the driver to potentially excite resonances and produce unwanted acoustic output that escapes through the port opening. Some people try and mitigate this problem by placing the port on the back of the enclosure. Placing the port on the back of the bass reflex enclosure may require more standoff from the rear wall and lead to room placement problems. The ML TL enclosure design requires stuffing in the internal volume of the enclosure. The presence of this stuffing is part of the design cycle and the amount and location is accounted for in the design process.

The ML TL enclosure can be thought of as a form of transmission line where quarter wavelength standing waves are used to provide the spring for the mass of air in the port. To physically model a straight uniform TL, clamp a yardstick to the edge of a counter or desk and pluck the free end so that it starts to vibrate. The vibration pattern is analogous to the air velocity in a TL. The TL's air velocity is zero at the closed end as is the yardstick's motion at the clamped end. The TL's air velocity is a maximum at the open end as is the yardstick's velocity at the free end.

There are two ways of changing the frequency of vibration for the yardstick. If you shorten the length cantilevered off the counter, the frequency of vibration will increase. Make the length longer and the frequency decreases. This is how straight TL's have traditionally been tuned by adjusting the length. The second way of tuning the frequency of the yardstick is to add a lump of mass to the free end. Put a piece of modeling clay on the free end and watch the frequency decrease. What I have done to the classic TL is put a lump of mass at the terminus end using a restrictive port. For a given frequency, I can shorten the TL (make it stiffer) increasing the tuning frequency and then add mass (air in a port) to pull the frequency back down and get a similar tuned result. One other benefit of having a lump of mass at the terminus is a rolled off port output above the first quarter wavelength resonance. This result is similar to a bass reflex port's response. I did this first with the ML TQWT and then with a straight TL. If you try the yardstick analogy, I think by changing the length and adding mass to the end you can demonstrate to yourself exactly what I am doing in my MathCad computer models."

If you want to accurately model a ML TL, the best worksheet is the "Ported_Box" worksheet where you can optimize the placement of the driver and the port.

Hope that helps,