Primare A35.2 power amplifier

My writing desk looks out over a large garden with chickens, bees, and feral cats. My chair sits only six feet from loudspeakers, playing softly on my left. Between the speakers sits whatever painting I am working on. That painting hangs no more than 10 feet from the oscilloscope and drill press in my kitchen. Best of all, my desk is only six feet from squadrons of ravenous sparrows attacking the suet cage on the fence outside my window. These real-world attractions keep my right and left brain in balance. Similarly, reviewing both analog and digital sources, as well as tube and solid-state amplifiers, keeps my review practice in balance. But not always. Sometimes my Apollonian self struggles to balance my Dionysian self. Especially when reviewing class-D amplifiers.

You see . . .

Class-D amplifiers are about quantities: like high power with high damping factors delivered at low temperatures in small, lightweight boxes at low dollars-per-watt prices. Class-D is a left-brain pleasure that declares, "Look what a smart shopper I am."

Class-D lovers venerate Bruno Putzeys, the Belgian engineer who kick-started class-D's current popularity, first with his UcD power amp module for Philips, and subsequently with his NCore amplifier modules and various switch-mode power supplies for Hypex.

I've been a forever fan of Bruno's class-D modules because they deliver exceptionally clean, dynamic power at very reasonable prices. Best of all, Hypex makes these modules available, not only to other manufacturers (like Bel Canto, NAD, and MBL) but also as kits made available to DIY people. Looking at the bigger picture, it's fair to say: Putzeys's inventions are currently reshaping the landscape of high-fidelity audio.

To me, the chief beauty of class-D is: It brings the specter of upper-class sound to the parlors of the merchant class.

Class-A, on the other hand, is about qualities. Long favored by gentry, class-A appeals to right-brain audiophiles with a penchant for sensuous excitements and leisure-class pleasures. Think stone fireplaces, linen sheets, and wines from Châteaux. Like aged beverages, class-A warms the listener's blood. Class-A amplifiers are typically low power, reside in heavy chassis, and operate at high temperatures.

But times are changing. Young Royals have started to abdicate. And class-D amp manufacturers are seeking Royal Warrants.

Class-D amplifiers are tiptoeing into the Lord and Lady's listening chamber. They are sneaking in because they are being presented in understated, fashionably sculpted chassis, at Oxford Street prices. The subject of this review, the $3495 Primare A35.2 stereo amplifier, is one of these new breeds of fashionably sculpted, "gentrified" class-D amplifiers. My plan is to see how it sounds while sipping Châteaux Margaux.

Sequence is everything
The only way I know to review a basic amplifier is to take out the amp that's driving whichever speaker is in the system and replace it with the amp I'm scheduled to review. My review deadlines are always the 10th day of each month, so usually, on the 11th or 12th day, I connect the new amplifier to whichever loudspeaker was in the system at the end of the review period. My first hope is that it works—that some form of relatively undistorted sound emerges in equal portions from both channels. If that happens, I can relax: As long as the amp doesn't catch fire or blow fuses, I will likely earn my shilling and three pence for the month.

On December 11, 2019, I connected the just-arrived (from Sweden) Primare A35.2 amplifier to the been-here-forever Magnepan .7 quasi-ribbon panel speakers. The .7s have a current-hungry 3–4 ohm nominal impedance, which I assumed would work very well being driven by the A35.2's 400Wpc into 4 ohms. And I was right. Immediately, the modest Maggies made clear, undistorted sound, from both channels, causing me to relax and listen closely to Alexandre Tharaud's faux-cabaret album Barbara (24/44.1 FLAC Erato/Qobuz). This recording, an homage to a chanteuse Française known only as "Barbara," is a 2019 favorite of mine. It's nicely recorded and, through the A35.2, sounded especially well-articulated.

The tone was good, and the A35.2's transparency was conspicuous, but the Primare amplifier sounded a lot different than the 20Wpc "class-A-like" Schiit Audio Aegir amplifier it replaced. With the Primare amp and Maggie .7s, this well-made recording presented Alexandre Tharaud with a very distinct outline. The lower register of his piano became more solid but less resonant. Tharaud's voice acquired a well-articulated but slightly insentient quality. More obviously, the Primare made the Aegir sound grainy and slightly blurry.

I was also surprised because every time I muted the Primare amp, to talk on the phone or answer the door, the damn thing shut off! It would start up again when I pushed the Standby button, but the ordeal was annoying. When I checked the A35.2's online owner's manual, it said, "By default, if no signal is present for 20 minutes the A35.2 will automatically go into standby to save power." The manual also said, "to disable automatic standby: press and hold down the front panel button for a few seconds, and release once the indicator light has flashed two times and become steadily illuminated." This was easier said than done. Later (see below) I discovered a switch on the back panel that turned this 20-minute-to-standby-scheme into a genuinely pleasant feature.

The Primare A35.2 is a stereo amplifier that weighs 26lb, measures 5.7" high × 16.9" deep × 15.7" wide, and features what Primare calls their "proprietary UFPD2 analogue class-D amplification technology" to deliver 200Wpc into 8 ohms, and the ability to be easily bridged into mono for 800W (!) into 8 ohms.


On the back panel are two 3.5mm jacks for input and output of a 12V trigger; an RS232 input; one pair each of RCA and XLR (balanced) inputs; and of course, right and left speaker cable binding posts. There is also a row of four small toggle switches. The first switch turns on and off the input sensor that allows the amp to automatically come out of standby if an input is sensed. (Her Ladyship liked this.) The second switch toggles between balanced and single-ended inputs. The third switches between two-channel and bridged-for-mono operation. And the fourth allows for normal or +6dB gain (the latter in bridged mode only).

Class-D Primare style
According to the design brief I received via email from Primare rep Terry Madalen, "The main reason that class-D was invented was for efficiency, and many of the performance benefits derived from class-D amplification are the result of the significantly greater ability of the circuit to deliver more of the power fed into it as musical signal output. Some manufacturers of class-D based amplification do not embrace the full potential this technology can provide, but our amplifiers do so in combining state of the art efficiency together with state of the art audio performance." (When I read these SOTA claims, my urge to insert a snarky remark was high, but I restrained myself.)

The design brief went on to explain why Primare thinks the A35.2 is superior to other class-D amplifiers:

"A major source of distortion in class-D designs is the demodulation filter on the output that becomes unstable with variations in loudspeaker impedance unless it's controlled by sufficient feedback. The failure to provide enough feedback to cope with dynamic changes in loudspeaker impedance across the audio band results in rising THD with frequency. While sounding very dynamic and vivid initially, it can become tiring and uncontrolled in the long term, especially when driving complex loads.

"UFPD provides a consistent 26dB feedback loop gain across the entire audio range and is stable way beyond the audio band. Rather than have the amplifier and then the filter as discrete stages, the UFPD design integrates the two, making control with feedback much more immediate and accurate."

I asked Terry who designed the A35.2:

"We design our products by hiring consulting designers with specific talents in various building block components for each product's design. So, in the case of A35.2 it is a product of the particular expertise in the design of Class-D power output circuitry of Patrick Bostrom (Sweden) and in the case of power supply, Bjorn Sandberg (Denmark). Bent Nielsen (Sweden), our chief engineer since the company's founding, then works with their fundamental designs to allow them to work most sympathetically together within a produce-able, and elegant, design."

According to Terry, the UFPD2 (amplifier) and AFPC (power supply) modules, as well as the chassis, are manufactured by "Our OEM manufacturer in China." When I asked where the A35.2 is assembled, he said, "In one of the largest and most sophisticated OEM manufacturing facilities in China, with the final quality-control testing being done in our facilities in Sweden."

More listening via Magnepan .7s
After a week of casual listening, I began to get a good feel for the Magnepan .7s driven by the Primare A35.2. The beauty of Magnepan's quasi-ribbon panel speakers is that they are not wood boxes with dome tweeters, and they generate an enormous soundfield. Best of all, they disappear when I close my eyes. The silence of the A35.2 amplifier let the quiet Maggies disappear more than usual.

Primare AB
US distributor: MoFi Distribution
1811 W. Bryn Mawr Ave.
Chicago, IL 60660
(312) 738-5025

JRT's picture

Herb (Cc: JA2 & JA1), Within this review you mentioned your Magnepan 0.7 planar loudspeakers and linked to a partial review (no measurements) from 06 August 2015. You have had a long time with these, and I think a follow up review addendum to that article could be an interesting read, and very much more so if JA1 includes a set of measurements.

Bogolu Haranath's picture

If you deliver free pizza to JA1's home for next couple of months, JA1 may consider doing the measurements of Magnepan .7 speakers ....... Just kidding :-) .......

Bogolu Haranath's picture

Take a look at the measurements of Magnepan LRS ...... .7 measurements could be similar :-) ......

avanti1960's picture

with the 4-ohm terminals on the Harbeths I suggest you again try the 8-ohm terminals.
Doing so should greatly improve the dynamics and impact of the bottom six octaves.

georgehifi's picture

"All measurements were taken with Audio Precision's auxiliary AUX-0025 passive low-pass filter, which eliminates noise above 200kHz. Without the filter, there was 433mV of ultrasonic noise present at the loudspeaker terminals."

Of course they had to be other wise it would have looked like this.

When is Stereophile going to bench test a Class-D that get rid of this because of the use of 1.5mhz switching used with the appropriate output filter like on the Technics SE-R1 Class-D that won't need the Audio Precision's auxiliary AUX-0025 passive low-pass filter, which eliminates noise above 200kHz to be used to see it 1 or 10k square wave.

Cheers George

SNI's picture

There really is no reason to enclude a carrier wave into the measurements. If you could hear the switch frequency the yes, but i guess you can´t.
Typically the switch frequency is placed arround 5-600 KHz where there is a sweet spot, providing the best compromise between, losses, EMC and distortion in higher order modulated amps.
The only reason for Technics to include filtering is EMC approval.
Ripple in the output of an amp, cannot be accepted almost anywhere in the world. That is nogo.
Therefore you often se inductors applied to the output of class D amps, especially those from OEM manufacturers.
This because often OEM modules are used in active speakers, where no such EMC regulations apply.
No speaker in the world can replay those frequencies, and they are way to inductive, to cause any problem at all.
The same goes for passive speakers, but EMC rules goes for anything. Even your hairdryer has to comply with the same rules.
No ripple allowed.
Because of this you can see a variety of solutions to this problem in class D amps. Some use a ferrite core with output wires twisted arround. Some even uses transformer coupling and so on.
Anyways the best audiosolution is no filtering. No speaker reacts on HF ripple at all, and if doesn´t make your mobilephone unusable, one shouldn´t really care about it.

Bogolu Haranath's picture

May be HR could review the comparably priced ($3,995), Rogue audio DragoN tube/class-D hybrid amp, 300/500 WPC 8/4 Ohms ........ DragoN may provide that high frequency 'bloom' HR is looking for :-) ........

Briandrumzilla's picture

No thanks.

tonykaz's picture

Mr.HR is certainly describing & revealing High performance levels for the Reviled Class D.

The top Pro guys like Bob Katz have already embraced the enhanced capabilities of this new design, what increase in perfection is offered by Amp Companies presenting similar product at the $150,000 price level ?

Schiit make ( in California of all places ) far more complex products at far lower prices. Why do these Primare people feel compelled to have unknown sub-contractors building their gear ( our gear ) ?

I ask for fair priced, quality gear designed and built by Proud & loyal employees. Shopping out Asian Labor to make a fast-buck is Taboo.

Put a : Proudly made in Sweeden sticker on each unit and watch sales double.

Tony in Venice

ps. what is the point of three chassis feet?

David Harper's picture

instead of listening to your equipment with music why don't you guys try listening to music with your equipment?

zalimci's picture

Out of sheer wanton extravagance and curiosity I bought a slightly used A35.2 contrary to my gut feeling. The key lever was the reference to the way this review praised its ability to deal with the Magneplanars due to the possibly unprecedented high damping factor, I have both .7s and 20.7s. I have amps which quote damping factors as low as 10 and as high as 250, but I'm led to believe that the A35.2 value is greatly higher, inherent to class D technology.

Well it made a very good initial impression, sounding as clean and fast as expected. As with the .7s in the review the lower end of the sound picture was very quick and tight. Some familiar recordings sounded different, a couple possibly slightly better than previously. Some less pleasing...

Then back to ones that pleased, and they didn't please as well second or third time around. I had possibly been overkind to this amp in that for the week or so prior to arrival I had been using an elderly solid state amp by the name of Dawn Audio Command 2, a small Britsh enterprise based in Sussex. Plenty of power and grip etc, not overly refined and a gift from an older friend whose hearing has become impaired to the extent that listening is no longer a pleasure. So, this morning it was time to try a few of the discs that I keep for the performance not the recording quality, and they were the killers!

Class D may be the future, but for me that future is yet to happen. The A35.2 is definitely worlds ahead of my previous Class D experience, also Primare - their CD10 all in one system of a few years ago. The A35.2 is certainly a good amp, I would guess a huge leap forward for class D, but for all the grip it may have on the likes of my Maggies, it has little of the depth and body and presence of my valve (tube) amps. Looks like I shall continue to line the pockets of my electricity supplier for the forseeable....