This review page is supported in part by the sponsors whose ad banners are displayed below

The modular internal architecture spreads out across the digital input board (which covers up the USB transceiver module beneath it and the wider analog i/o board below that); the two ICEpower output boards; a toroidal power transformer; the preamp/headfi board; the tube buffer; and the vertical input switching board behind the front cover.

The GI celebrates high part density and physical segregation of discrete modules for specific functions.

Flying leads of either computer-style ribbon cable or conventional hookup wiring connect these sub modules via push-on connectors for easy assembly and replacement.

Seven hex screws clamp the top cover to damped rails in the ribbed and round-cornered lateral extrusions. A square sheet of damping material is centered on this cover on the inside.

David Solomon. "We learnt an important lesson working with Anthony Michelson of Musical Fidelity when we still imported his gear to the US. We were shown first hand that experimenting with circuit topology and component layout is important to ensure the lowest noise and best performance for each circuit. Our design team had a similar philosophy so many hours were spent ahead of time optimizing each component, board and topology for best performance, then listening to each and adjusting it to the desired overall sound we were looking for. The amp challenge was not much of a challenge at all. With the huge improvements to the new ICEpower modules and the new sampling frequency of the power supply way outside the amp's bandwidth there clearly are no issues with bass deficiencies or a brittle weird top end. Running a common-mode rejection circuit with proper loading proved to be invaluable. We spent an enormous amount of time balancing the signal path to optimize the circuit. To be a little different, a Class A preamp stage was added feeding a single-ended to balanced converter dedicated to each channel. This worked well and helped us find the sonic sweet spot we were looking for.

"We started off by listening to the amp circuit. Jim Spainhour had found a 'killer amp' he was excited about. My first question was, A/B or D? I wanted nothing to do with any digital amp I'd heard. At his request the amp was sent to me for audition. After a week of it sitting there I decided to hook it up to a Decco2, call Jim and tell him how terrible it was. Much to my surprise I was blown away by the sheer bandwidth and the magic our team had included. My past experience left me with the impression that most digital amps had an anaemic low end and grain in the treble. With this amp however, I was convinced hands down and didn't want to send it back. Afterwards I hooked it up to a MF308 preamp and it really rocked. After approving the amp section, we tweaked and approved the 9018 DAC design. This is the best-sounding DAC we've ever done by a long shot. It's light years better than the designs we used in the Nova/iNova. I was stunned at the improvements even over the ESS Sabre 9016 we used in the iNova and iDAC. Next we moved to the preamp. We went through countless very expensive analog volume pots, all of which measured between 1-4dB of channel matching depending on position. We did not like their sound because it was flat with no snap or pace.

"At that point Jim suggested we change direction to a VCA. In the past I'd worked with those in older Krell and Accuphase gear but didn't understand how good they could be until I heard the direct comparison and immediate improvement in our own integrated. That was more like it. Now everything snapped into focus. Microdynamics and pinpoint imaging became sharper and more defined and snap and pace were there. Once we had our preamp sounding better than expected, we added Bascom's tube buffer. We all liked it already but wanted it a little mellower due to inherent problems we knew are present with many digital recordings.


"All musical material isn't created equal. More music than not is recorded poorly or compressed badly. When played through a purist circuit, the reproduction often leaves you uninvolved or even has you turn off music you otherwise really like. A lot of great Motown or hard compressed bands like Nine Inch Nails are like that. It's ear-fatiguing at best but not necessarily due to the equipment. Rather it's the poor recording practices so many recording engineers use these days. That's the real world. Running such a signal through a tube is almost mandatory if you don't want to get hit squarely in the ear canal with tweeter shrapnel. So we retuned the buffer to help with these types of recordings.


"On top of that, 30.000.000 people in the US alone use Pandora and that number is growing on all the streaming sites like MOG and Spotify. This is a huge market. We don't want to leave them out in the cold. There are millions that use nothing but streaming services. We've found that the tube buffer with streaming or MP3 sounds more engaging and musical. This is very similar to the iDac. Almost every reviewer who reviewed it used the settings which produce the worst measurements. As a result we've received many great awards for the iDac. With the ability to manipulate a DAC we all switch in and out the filters we like. I laugh because rarely do folks switch to or prefer the 'best' settings. We're all about the music. Numbers matter but we're also realists and know that most of the recordings people buy are not 'audiophile approved'. So does the tube circuit flavor the sound? You bet. But that's exactly what we wanted. We perform buffer in/out show demos all the time. Over half the listeners prefer it in. Just like the buffer on the iNova and previous units, we favor it when we find the need to soften harsh recordings. Good news for the purist? When you disengage the buffer, it's out of the signal path entirely. Poof. Either way is fine with us. Ditto for the DAC slope filter on the back.


"Each twin triode of the buffer is wired in parallel for dual mono operation, one per channel. This lowers the noise floor and improves channel separation. It is really mostly Bascom's original design, which we already liked a lot but tweaked further. We use a constant current source that biases the tubes equally so they never get imbalanced.

"The headphone circuit has a dedicated TI headphone driver chip that couples after the preamp stage and optional tube circuit. This is the first time we tried this and it worked much better than in previous units. The preamp stage is active and consists of a discrete Class A front end coupled to the voltage-controlled amplifier. During listening sessions we preferred this active circuit for its smooth response and ultra-high dynamics.


"Even though the ICEpower modules are permanently bridged for higher power, we've had no issues driving low-impedance speakers. The Martin Logan Montis at CES is a nominal 4Ω but dips to .52Ω at 20kHz. At RMAF the Sonos Elipsa Se started at 4Ω. The Martin Logan CLX we used descends to 0.7Ω. So far we haven't encountered one speaker yet which the Grand Integrated wouldn't properly drive.

"For USB we initially chose the Tenor TE8802L chip after again extensive listening tests versus other asynchronous USB receivers. Unfortunately Tenor didn't make it a driverless solution which we'd been expecting since January 2012. They also were late on a 64-bit Windows 7 driver. After considering this carefully we opted to abandon the Tenor platform in favor of the rightly popular XMOS solution which is proven, ultra stable, comprehensively supported and sonically excellent." High time now to break down the machine's individual features to see what performance class they each belong to.
Enlarge!