Prior to mounting the drivers, I had to break in the new Chrystal drivers. I was familiar with their more than average requirements so dutifully gave them about 225 hours connected to my vintage Advent 300 receiver tuned to an FM station. For the woofers, Mark had wobbled them for 48 hours with a 20Hz signal. I was supposedly good to go.  Both of these break-in regimes are mandatory. Otherwise the speakers will sound forward and thin. I can't emphasize this enough. Additional improvement still occur afterwards. My woodworker and I mounted the drivers, then I did the wiring/soldering using the hookup wire supplied with every kit. Robert advises to bypass the terminals on the back of the widebanders and instead solder directly to the Litz leads leading to the voice coils for that last measure of purity by avoiding the metal terminals. Under the conditions in the wood shop, I was not comfortable taking this on and so soldered the wires to the terminals. May the audio gods have mercy on me.

Placing the Mandala in my upstairs listening room was not difficult. Unlike the sealed box woofers for the Prometheus, the open baffle woofers cannot be moved separately from the front baffle. By design, the front-facing slot of the woofer cab sits flush with an identical slot in the front baffle. This precludes experimentation with placement of the dipole woofer at other room locations as per this article. The HiFi Zine writer notes that dipole woofers excite fewer room modes when placed parallel with the room axis so at least this placement is consistent with this recommendation. Robert explained that his dipole woofer is an N-frame design and does not suffer the 6dB/octave roll-off of simpler dipole woofers. It also behaves differently than the W-frame dipole woofer used by the author of this article. All I needed was a solution for a crossover and amplifier to power my Ultimate woofers. And so the odyssey began! For the crossover, Robert suggested a MiniDSP 2X4 digital processor with variable frequency and slope settings. This is an inexpensive $99 but powerful tool to design and/or optimize speakers. It converts the analog input to digital, applies the selected settings through a USB connection with the control software program on your computer in the digital domain, then converts back to analog.  There are two inputs (right and left) and four outputs. I only used two of the outputs to the woofer amp(s).

At first I followed the leads of Mark Loewen, the Canadian distributor, and his US colleague Bill Demars. Bill uses Dayton A500 plate amps at ~90Hz. The Dayton xover uses a 4th-order slope. Mark too crossed at 90Hz with a MiniDSP and Canadian 50wpc SAR labs MOS 100 amp with 0.75V input sensitivity and 100KΩ input impedance. However, he runs reproductions of vintage Shearer bass horns, not Ultimate OB woofers. To get going, I tried a humble Adcom GFA-535 I had around. Robert Bastanis was doubtful that it'd do the job. What I first ran into was difficulty in gain matching the widebanders/tweeter with the woofer. The baffle output was much higher even with the MiniDSP set to max. Whilst the sensitivity offset between upper and lower amp was my first suspicion, further reading on the MiniDSP site revealed that the MiniDSP only delivers 0.9V. It couldn't drive the woofer amp to full output.

One solution was to try another amp with higher input sensitivity. To accommodate the MiniDSP, I borrowed a vintage 50wpc McIntosh MC502 amp set to 0.75V input sensitivity. With this combo, I could balance the output level with the trim pots on the front of the McIntosh. But it still couldn't support the large excursions needed for even moderate high-amplitude bass. I ended up abandoning the MiniDSP and borrowed both a vintage Dahlquist DQ-LP1 active xover and a pair of Dayton A500 plate amps (500wpc into 4Ω) with their built-in gain and xover controls. I mated the Dahlquist with a 440wpc/8Ω Crown XLS 2500 stereo amp, a 100wpc/8Ω Hafler DH-200 and finally a 400wpc/8Ω Crest CA6 pro amplifier. None of them proved satisfactory. The Crown amp sounded blurred, the Hafler was unremarkable. The Crest had enough power but its bass had a very thuddy cardboard quality whose tasteless signature imposed itself on every bass instrument. Also, I found its cooling fan most annoying.

There was also quite a bit of upper bass resonance. Both Robert and Mark felt that I needed a more powerful amp to properly control the woofers. [Here the reader remembers that these very large drivers don't see the usual back loading from a closed or even ported box; and suffer out-of-phase cancellation between the woofer's front/rear emissions to enforce the need for high power and high damping – Ed]. A visit by an electronics engineer friend who also does speaker design got to the root of the problem. The 90Hz crossover setting was simply too high for the Ultimate woofers. We moved down to 65Hz and the upper bass resonance vanished. We then nudged the filter back up and settled on 75-80Hz. Next I put the Dayton A500 plate amps on the woofers, sitting them upright on the floor just to the outside of the woofer cabinet's rear opening like shown below. I also briefly tried the 150wpc/8 Keiga 5230 plate amps but they sounded somewhat anaemic so I stuck to the Daytons.

Robert provides these setup instructions. "Baffle placement - first connect the open baffle amp like you connect any other speaker. You can find the right placement by listening to solo vocals. With as much space between the baffles as possible, move them away from the front wall until the voice rises in height. Voices should be centred between the speakers. Check that the speakers are exactly parallel. For the woofers, right phase is essential for natural sound although changes are relatively subtle. Switch one woofer amp off. Raise the crossover frequency for the woofer to max and raise the volume a bit. Turn the phase knob while listening to music until the bass is maximum fat and boomy at the listening position. Reduce the level and crossover to normal. For the xover adjustments, switch the woofer amps off and listen to the imaging of the baffles only. Choose a choir or a big orchestra. Then switch the woofer amps on and compare the imaging. Reduce or raise the crossover knobs (of both woofer amps, both always at the same frequency settings) to the highest possible frequency which doesn't affect the imaging of the baffles. My own method is to raise the crossover frequency in about 10Hz increments from lowest position and compare the imaging of each step to the imaging of just the baffles. The right bass volume depends on the room. In most cases the volume of the woofers is set too high in the beginning. The phase and crossover knobs of the two woofer amps should always be set identical. The volume may differ a bit because most rooms and speaker placements are not 100% symmetrical. The quality of records differs a lot too and it is impossible to find the right volume for all your records with just one volume setting. Choose a position which is fine for about 90% of your records and reduce or raise the crossover frequency a bit when you listen to a record which shows a different tonal balance. You will probably end with three positions. The middle position is right for most records."

With the Daytons on the woofers, their phase and xover settings roughed in, my final speaker placement had the outer edges of the baffles sitting 82" from the front wall and the inner edges 81", giving a very slight toe-in. Do not assume that this speaker always needs this much space behind it. My room has eaves behind the speakers and this may require more space than a room without eaves. I asked Robert and learnt that "the ideal distance from the front wall for 99% of installations is 75-150cm."

Robert recommends pointing the baffles straight ahead but he also allows that a customer might prefer the imaging with a slight toe-in and that it depends on the rest of the stereo and personal taste. I would like to add a few comments about my own setup procedure. I found that the xover and phase settings can confound each other. If your xover is set too high, the Ultimate woofer's upper bass resonance can mask your ability to hear the phase differences. I would set the xover not above 65Hz when determining correct phase. I could hear effects in the lower midrange that allowed me to set the phase correctly. With the Dayton A500 plate amps, the phase setting is just 0°/180°. Other plate amps have a rotating knob that allows continuous settings between 0° and 360°.

With the switch at normal, lower midrange instruments and vocals sounded fuller and more natural. Inverted, they sounded somewhat pinched. After setting phase, come back to the xover point and raise it to where you begin to hear upper-bass muddiness. Then back off just enough to eliminate the muddiness. You can fine-tune more as you play additional recordings.

In another email exchange, Robert had some additional comments. "The phase relationship between woofers and open baffles is determined by the relation between the amp which powers the baffles, the active xover and the amp which powers the woofers. The power amps can have differences in phase. The active xover also changes phase and finally the woofers are a few centimetres behind the widebanders. In most cases, phase between woofers and baffles in the xover region is not at exactly 0° or 180°. For best possible phase, most xovers offer variable phase shift and digital xovers can set variable time delay. When your active xover doesn't offer phase shift, the only possibility is to change polarity at the woofer binding posts. This may be difficult if the correct value is between 0° and 180°. I recommend to follow the short setup manual I sent you and to check phase with raised xover frequency for the woofers. A bigger overlap between baffles and woofers makes it easier to determine which position gives the fattest most powerful bass. That is the right phase." While I am hesitant to disagree with Robert, I found the phase easier to set when the xover point was 70Hz or lower. Another factor that can interfere is whether the woofer cabs sit directly on the floor or on some sort of isolation devices. I would recommend placing the woofer cabinets on cones when making these adjustments. Otherwise, the cabinets may couple to the floor and muddy the bass. Bastanis advises that on some installations, the woofers may sound best directly on the floor. He encourages trying both ways. In my room (Berber rug on the floor), the bass was definitely cleaner when the cabinets were supported by cones. I used some vintage Audio Solutions metal cones I had for years. Probably further improvements can be made with more modern isolation devices like Nordost Sort Kones.