Ariel transmission line loudspeakers designed by Lynn Olson.
The Ariel is an efficient twin transmission line design which is well suited for use with low output 'single ended triode' valve amplifiers. Originally designed in 1993 it has since undergone a number of modification stages, primarily prompted by feedback from the large number of DIY enthusiasts who have constructed them.
The following discussion relates to the Mk6c version but previous versions are worth checking out, details are profuse and profound on the international builders club site :
The Ariel first came to my attention in 2002 when I was searching the web for an efficient speaker that might match a low powered SET amp. At that point Derek Walton , a man whose DIY ramblings I watched with interest, was proclaiming good results for them. So, trusting soul that I am, I thought I'd follow in his footsteps. I seem to be a bit slow off the mark as this design has been around since 1993 and its fans are legion, mostly I would add, not in the UK. However, the proclaimed 93db efficiency and the transmission line design seemed to be what I was looking for so I decided to have a go.
The design is basically a twin transmission line layout to support a D’Appolito format. The speakers are not small, standing 4 feet tall and 13 inches deep, but when it comes to sound and efficiency like this, size is not something one immediately wants to compromise. Or weight for that matter, each unit weighs in at 35 kg. The twin transmission lines dictate that this project requires a degree of application and it is therefore hoped that this article will not only give an insight into construction, but also help in deciding whether the necessary tools and skills are available at the outset.
I wont pull the wool here, this is a fairly complex cabinet, comparable to a folded horn. For this reason I would recommend the minimum set of tools to be a table saw(with accurate cross-cut fence), a good router, a vernier gauge and a clean work bench of at least 4 feet by 2 feet. If the table saw does not have a precise cross-cut fence then a large capacity mitre saw(8-12inches wide) is a must. Access to a spindle moulder is useful, albeit only for one simple but crucial task. The skill required to construct these speakers can be summed up simply as the ability to cut wood exactly square and to precise lengths. In addition to the woodworking skills there is a requirement for some very simple soldering.
Two minutes with the plan reveals the need for templates. To mirror match the cabinet sides would be very difficult without. I used 3/8 inch ply for templates.
Once satisfied that the templates are correct, the panels can be cut on the table saw and matched for length and square. The side panels are one thickness of MDF but the front and back are sandwich of MDF and ply to achieve the necessary acoustic damping. The front and back panels can be stuck together before or after routing for the internal partitioning, but it’s likely this choice will be driven by preference for manhandling light or heavy panels. Once joined, the front and back panels are pretty heavy. In the example speakers it was decided to locate the crossovers in the unused space at the top of the cabinets and the binding posts ¾ of the way down the back panel. If these locations are chosen then a cable channel can be routed up the back panel in the middle of the sandwich. Failure to remember this before laminating could be frustrating.
Having cut and checked all the panels, the next stage is to route all the rebates for the internal partitioning. The templates will obviously have been cut to suit whichever guide bush has been chosen for the task. If using a cutter with the same dimension as the internal partitioning ( 3/4 inch or 18mm) then this job is straightforward, albeit tedious. Depth of cut is not crucial here but I chose 6mm which gives a pretty strong build.
Now for the interesting bit, the internal partitioning. Method of construction will be entirely down to individual preference. The method adopted here (which I can guarantee works) was to lay one side panel on the workbench and glue the back panel to it, waiting for it to cure. All joints were made with PVA . Assembly was then a matter of inserting 1, 2 or 3 internal partition pieces at a time. Before gluing the partition pieces in place, they were positioned and the other side panel laid on top to ensure a good fit. Once lined up, the partition pieces were glued into the side that was lying on the workbench, with the unglued side replaced on top while the glue cured.
The internal partition pieces can be rebated or but-jointed to each other. If but-jointing (as in this case) the jointing obviously has to be precise. As the angle of the joints varies between 30 and 75 degrees this can obviously pose some problems because the standard mitre saw will only cut to 45 degrees. To address this problem an adjustable mitre jig was designed/constructed to allow the joints to be cut with the router. An annotated construction drawing of this jig can be found here.
When all internal pieces had been glued to the side/back lying on the work bench, the driver wiring, the felt damping, and the long haired wool were added(all as shown on Lynn's drawing). The wool was placed loosely in the transmission line as per recommendation, taking care not to place any in the bottom labyrinth. To prevent the wool from dropping down into the line a piece of nylon stocking can be glued into place, at the start of the labyrinth. At this point the remaining side can be glued onto the back and internal partitioning. It is recommended that the front panel is placed in position (unglued) while the side is drying.
Finally, the front panel can be glued in place, not forgetting to drill the bolt holes for the drivers and fitting the captive nuts behind. It has been said that drivers can be screwed into cabinets but it seems such a waste to do something that tacky after spending this amount of time on the cabinets.
The crossovers can be built to personal preference. In this case the circuit was built from 2.5mm mains wire on drilled veroboard as shown in the photograph. The units were securely located and wired into the empty space at the top of the cabinet before the top was glued into place.
I am no fan of MDF. Apart from the health issues I don't like the look of the finished product. I chose to lacquer mine for a piano-black finish. This was expertly done for me by a good friend, Dave Tuplin from T&G Motors in Alford Lincolnshire. Respect!
One point to note. With the mix of differing types of wood(mdf and ply) it has been found that movement over a period of time can cause joints to become visible under the lacquer.
(updated June 2006)
Hand on heart I must say that I was disappointed with the initial sound of the Ariels. Hooked up to my WAD 6550 PP valve amp the sound was subdued and foggy. First listen was done before they were lacquered, but on their return and subsequent re-assembly, they sounded a lot brighter and sharper. The passage of time has proved yet again that some speakers require a fair amount of running in, the Ariels seem to be in this category.
I will not go into detail about how I find the sound of the Ariels, there is a plethora of opinion which can be read, just follow the links from Lynn's Ariel pages. In addition, opinion seems to be polarised, some people love 'em and some people hate 'em so I won't stick my neck out.
I will say that I have been running them now for 3 years in my project recording studio where they are driven by my Single ended EL34 amp of about 6 watts. I have no complaints whatsoever, they tackle any type of program thrown at them, the imaging is razor sharp and presentation very neutral to my ears, which is ideal as I use them primarily for mix monitoring. In short they are one of the most invisible speakers I have had the pleasure of........
If you fancy the challenge, and I thoroughly recommend it, then good luck.
The above constitutes a precis from my original review in HiFiWorld from April 2005 which may still be available in the archives