November 6th. "Sorry for the long wait but as the saying goes (cough), great things take time. This is a completely different icOn than the previous review sample was in July. I hope you will like it. The basic functions, crossover, output selection etc. all work. Special A/B comparisons & testing features will need some more time. I'll have some question about the details. I don't want you to wait longer so after some more testing and measuring, I want to send this to you by Monday. The extra features and final customizations based on your feedback will be added by firmware upgrade. With rotary switches that would be just a bit more difficult." Pál was ribbing. With rotary switches, this flexibility would be clunky to the extreme; if some of it was even possible.

"The standard Model 3 has 7 different active output configurations selectable by remote. Any of those can be the default start-up. 1 = RCA 1; 2 =  RCA 2 ; 3 = XLR 3; 4 = XLR 4 ; 5 =  RCA 1+2; 6 = XLR 3+4; 7 = 1-4. The LP/HP outputs are always active. I can remove any of these options or or add any other new combination together with the HP/LP outputs. Choose the combinations sensible for you. We can change them anytime with a firmware update. For A/B testing here are the current extra functions of the big remote. F1 toggles outputs RCA1+2, F2 XLR1+2, F3 RCA1/XLR1. 'Back' toggles the sub output, 'set' is still unassigned. That's the beauty of microprocessor control. Modern gadgets, including audiophile versions, should be smart, future-proof and easily customizable by the end user." Here is the Model 3 manual, here how to upload new firmware.

My new sound|kaos Vox 3awf had already shipped a few days prior. Our upstairs system would thus welcome two new additions to dial in further. Hifistay's amazingly effective Mythology X-Frame rack reviewed in the interim had been left on an extended manufacturer's loan. This had driven up the installation's overall resolution very demonstrably. That bed was very tidily made. Now fun and games in it were ahead. The breakdown of that bedding was Soundaware D100Pro SD card transport ⇒ Chris Sommovigo Tombo Trøn digital coax ⇒ Denafrips Terminator ⇒ icOn 4Pro SE ⇒ Zu Submission sub + Crayon CFA-1.2 or Bakoon AMP-13R ⇒ sound|kaos Vox 3awf ⇐ Furutech GTO-2D NCF passive power delivery, Hifistay X-Frame resonance control for equipment, Stella60 isolation footers beneath speaker stands, Crystal Cable UltraConnect interconnects, Black Cat Cable Lupo speaker cable. Simple but carefully selected. I asked Pál about balanced sources whose XLR output is twice the voltage of their RCA version. "Yes, that gain advantage passes intact through the icOn. Its XLR inputs use a special balanced line-receiver chip not op amp."

Whilst the Acelec Model One still were in play, the icOn slid onto its allotted shelf. I dimmed down its big numbers to not distract. Now the Crayon amp's own volume was bypassed. A quick 'mute' on its remote could kill the monitors to hear just the Zu sub. That was useful to arrive at the correct bass volume. Once set and using the LP/HP paths with Zu's filter bypassed, I had a perfectly seamless blend. It also became clear in a hurry that Pál's execution of an active 4th-order low-pass fixed at 40Hz was far superior to the broadly variable 4th-order filter of Zu's Hypex plate amp. But more surprising than perfectly integrated infrabass because I'd not considered it at all were dead-obvious gains in dynamic range for the monitors. On the face of it one objects. With Pál's filter, they were -6dB down at 40Hz, so perhaps three decibels more attenuated than unfiltered. Why should that make much if any difference to the dynamic range in the midrange well above 40Hz? Hello adverse square functions.

Let's quote my old boss Pat McGinty of reborn speaker house Meadowlark Audio. "Each time you double current flow through a voice coil, you generate four times the heat. Impedance rises linearly with heat. As current increases, output falls off fast. Complicating things, the time constant is long. Think of the wires in your toaster. They take a while to heat up, a while to cool. Also, the crossover's transfer function modulates with impedance changes. For a low-pass filter, corner frequency rises with impedance. So heat equals glare. And filter Q and total system Q both wobble with impedance. The trick is to not heat up your coils. That's a tall order when the typical woofer only turns 2% of power into sound. The rest goes to heat. About mechanical compression, the woofer motor works by moving a coil against a static magnetic field. Since the amount of coil inside the sweet spot of that field diminishes with excursion, motor force falls off fore and aft of center. As the moving system approaches its mechanical limits, the elastic components are stretched tight. So the motor force is quasi linear for the first few millimeters, then falls off quickly. Small excursions generate pretty much twice the output for twice the input power. But ask a woofer to really stroke and you'll see a falling relationship. The entire band's dynamic linearity modulates with coil position. Thermal compression sums with mechanical compression. All of the shortfalls add up to reduced dynamic range."