This review page is supported in part by the sponsor whose ad is displayed above

Reviewer: David Kan
Digital Source: Restek Radiant CD player, Assemblage D2D-1/DAC-3.1 Platinum, Deltec PDM Two DAC, Marantz SA6820, Philips DVP9000S, Pioneer DV-578A (home theater)
Preamp: KingRex Preamp, Dared SL2000A, NuForce P-9 [on loan]
Power Amp/integrated Amp: NuForce Reference 9 and 9 SE, NuForce Reference 9 V2 [on loan]
Speakers: JMlabs Micron, Dynaudio Facette, Klipsch Synergy F2, Apogee Centaur Minor, Mark & Daniel Maximus-Monitor w. Omni-Harmonizer, Maximus-Ruby, Maximus-Topaz, Maximus-Sapphire
Cables: Symphonic Line Reference interconnect, Clearaudio Silver Line interconnect, Deltec Black Slink interconnect, Luscombe LBR-35 interconnect, Unity Audio Solid Link interconnect [on loan], Symphonic Line The Fast speaker cables, OCOS speaker cables by Dynaudio, Unity Audio Solid Link single-wire/Shotgun bi-wire speaker cables [on loan], Orphee Audio digital cable, Aural Symphonic Digital Standard digital cable
Power Cords: Aural Symphonic Missing Link, Ensemble Powerflux, Unity Audio Link Precision Link [on loan]
Power Line Conditioning: Tice Power Block IIIC, Monster Power HTS-3500 Mk II (modified by Nu Force), Monster Power HTS-1000 Mk II
Room Size: 15' x 13.5' x 8 diagonal setup / 12' x 24' x 9' opens to 12' x 17' x 9' L-shape, short wall setup / 13' x 28' 8" x 9' with openings on one side to hallway and staircase, short wall setup / 15' x 15' x 8' home theater
Review component retail: $399 (fully assembled with Alps Blue Velvet volume control); DIY kits from $319

It all started with an email I received from 6moons reader and Tripath connoisseur David Dye back in April:

"Hello David, I had intended to give you a heads-up on having given Jay Hennigan of Winsome Lab your email address and suggested that you might be interested in reviewing his product. I take it by your response to Jay that you weren't too annoyed at me for doing so. There are a number of good low-powered Tripath amps out there but many people need more power than the Trends TA-10, Sonic Impact, Charlize and other similar amps provide. The buzz on the forums about the Winsome Mouse has been highly favorable, including on where folks are far less likely to get overly ecstatic about a new amp of the month. From a power perspective, the Mouse appears to be close to equivalent of a pair of Red Wine Audio Signature 70s without the batteries and PIO coupling caps. It should be able to drive most speakers in the 86-90dB range and I'm betting that it would be enough for some low efficiency but relatively easy to drive speakers, like my beloved 83.5dB LS3/5As. At less than $400, the price is also very attractive."

Outside the box
It never occurred to me that I'd ever need a bigger power Tripath amp. I am happy with my Trends TA-10 and KingRex T20/T20U. Rumor has it that higher power Tripath chips don't sound as good. Why would anybody need that - more power, inferior sound? In retrospect, I'm glad I didn't turn down the offer. As a rule, I always ask for bi-amp loaners. Jay made that happened within a week. I quickly set them up with my four Loth-X BS-1s. But there was no sound. I checked the cables for loose connections but all were intact and secure. I emailed Jay: "Both of them only give out very low volume distorted sound even with the volume knobs dialed to max." Jay knew what went wrong from miles away: "Do you have the speaker wires connected according to the silkscreen on the back? It is a bit unconventional as I wanted to avoid the idea of a "ground" or common terminal. The binding post colors match the input jacks. Each channel connects horizontally, speaker (–) on the left and (+) on the right. If you connect them vertically, you'll only hear the difference signal between left and right channels."

What silkscreen? I didn't even pay attention. I just plugged everything in the way I always do. So I did connect the speaker cables wrong! Why is this amp different? Like any Tripath amp, the Mouse has minimal input/output amenities: one pair of RCA line inputs, two pairs of speaker binding posts and a coaxial DC power input. But the binding post colors match the input jacks, meaning L signal input and L channel speaker (+ and -) are all white, R signal input and R channel speaker (+ and -) are red. And they are aligned horizontally. Actually, that makes a lot of sense. But so often, wrong is right. It takes a thinker to challenge conventions and stop habitual practice.

One other minor detail shows that the Mouse is not a follower: the power on/off button is located on the front panel. You can't argue that certain things were better the old-fashioned way. When a tiny Tripath amp is put inside the rack, reaching around for that little switch could require some yoga skills. When the Mouse's power button is depressed, four LEDs light up in sequence from top to bottom: red, red, yellow, blue, signifying peak, fault, temp and power. The top two light up for only a short duration during the initial power-up test protocol. Only the bottom blue LED illuminates throughout normal operation. I presume the other three will light up in case the circuit detects the onset of clipping, excessive output current or overvoltage/undervoltage and unhealthy internal heat exceeding 130°C. I'm happy to report that I haven't seen the yellow light during my hundreds of hours of auditions. The little box gets just mildly warm at all times. In fact, such protective circuits are already built into every Tripath chip. This is the first time I've seen them manifested in visual form. Unlike the other class T amps I've encountered, the LEDs on the Mouse never blind your eyes with Star Wars laser-gun intensity.

Cosmetically, the functional aluminum box is all Spartan simplicity. At first glance, it looks almost like the Diva control box of the WLM loudspeakers. That's befuddling but I can reassure myself that WLM and Winsome Labs Mouse are entirely unrelated. So why call it the Mouse? Because of size? Because it has the power to make elephants tremble? Whatever you want to read into that, it's a far more memorable and unique name than alphabet soups and numerals. Weighing 16 ounces and measuring 4.25" x 2.25" x 5" excluding connectors and control, the Mouse is one of the smallest, lightest class T amps yet yields the biggest output per cubic inch: 30/45wpc at 8/4 ohms to be exact. It is available in four versions: DIY kits with either Panasonic or Alps Blue Velvet volume control ($319/$349) or fully assembled amps with either Panasonic or Alps volume control ($369/$399). The pair Jay sent was the fully assembled Alps version. I am not qualified to report on the DIY kits but just to cover them briefly, I asked Jay to send me pictures of the kits and the assembly manual which is very detailed and looks like fun. (Actually, after going through the 9-page step-by-step manual, I'm confident that even I could handle the build.) Basically, the kit has all of the surface-mount components preassembled. The builder is required to solder the through-hole components, wind the output coils and perform the mechanical assembly. If you order a kit and later change your mind, you can send the unassembled kit back and have it assembled for a very reasonable fee. You'll find details on the Winsome Labs website.

Once I sorted out the speaker connections, my first mission was to check the output polarity with my handheld polarity checker and CD. Bless you, Jay, this Tripath amp is the only one I know with correct polarity. As all Tripath chips operate with an inverted input stage, Jay took the trouble to insure that if you wire up your speakers in the proper silk-screened polarity, the output related to the input is non-inverting. A lot of people don't bother with absolute polarity, claiming there's a 50/50 chance the recorded media is inverted and you couldn't hear the difference regardless. My experience tells me otherwise as mentioned in my Trends TA-10 review but I am not going to debate it here. Let me just ask you this: Would you invert polarity knowingly? For now, here's my first thumbs up for the Mouse. But the polarity story doesn't end yet. In a while, I'll let Jay further explain the how and why.

To really appreciate how much thinking outside the box Jay has done, you have to first peep inside the box. To understand why Jay did things the way he did, you have to know the man's background. I fired rounds and rounds of questions at Jay who responded in the most timely, professional and forthcoming manner, never dodging once. Without much ado, here are some excerpts from our marathon e-mail interview.

Can you tell me more about yourself and your company? Have you always been in the audio engineering business?

I've always been a gadget freak and experimenter and had built Heathkits and Dynaco kits as a kid. Winsome Labs has typically done projects along the lines of "build me a gadget that does 'X'". These have been custom projects, designed to order, none in the audio realm. My main love of audio gear has been the restoration of old jukeboxes as well as vintage components. I became fascinated with the Sonic Impact, explored the Tripath technology and found what I thought was the "sweet spot" between the low-power 12-volt all-in-one chips and the complex high-power T-class amps with discrete output devices and large power supplies. This is the first product without a specific single customer having it built-to-order - and the first audio product. Our company has been around for about ten years. It's a home business operated by my wife Winnie and me, hence the name.

How long did you work on the Mouse? How long to bring it to market?

The first prototype was built a little over a year ago, as a one-off more or less for fun. I have a pair of AR-6 speakers in my office, so I built it for use with iTunes on my Mac laptop. I built a second one with the parallel mono design for more oomph. A friend who is into high-end audio came over for dinner and heard it driving the Infinity Kappa 7s that he had earlier sold me. The source was an unremarkable Sony CD player. He was looking for the amp in the system, insisting it had to be something bigger. I showed it to him and he begged me to go into production. I did some more tweaks, substituted some higher quality parts in the input stages and optimized the mechanical design and packaging. I crunched the numbers, looked at what was out there and decided to offer both a kit and assembled version. We went live with the product in February of 2007.

Inside the box
At this point we have to go inside the box before moving on with the interview. After removing eight screws and the volume knob and sliding off the aluminum cover, what comes into sight is a very clean circuit layout optimized on a four-layer PCB utilizing both through-hole as well as surface-mount components. Signal paths are extremely short and there's almost no internal wiring except for the short silver plated Teflon insulated leads to the speaker binding posts. The Alps Blue Velvet volume control takes up a lot of space, proportionately speaking. It looks like there's not enough room for upgrading the Bennic caps to Auricaps or other boutique parts. In the middle of the PCB is your Tripath chip. But this one is merely the TC2000 controller. Unlike other smaller-output Tripath chips like the 2040 or the 2020 where the controller and power stage all built into one module, the TP2050 comes as a separate chip for more application flexibility and sonic integrity. With the Mouse, you have to unscrew three more screws to remove the bottom panel to see it - er, them. The Mouse is using two TP2050s snuggly pressed against the bottom panel with heat-dissipating gel and a metal plate as heat sink in-between. Ready now for another round of Q&A? Bear with my technical ignorance. What's that saying? Clever answers for stupid questions...

Can you call this a dual-mono design, bearing in mind the TC2000 is stereo?

I guess you could but it wouldn't really be accurate. True monoblocks use completely separate power supplies and input stages for absolute maximum separation and elimination of sag from one channel to the other. The Mouse of course has a common input stage and power supply so it wouldn't be fair to call it dual mono any more than a stereo amplifier with two separate pairs of 6L6s in push-pull is automatically dual mono.

Even the air-coil inductors are for monaural circuits? Well, that might be normal even for stereo. Is there a reason for mounting them at 90 degrees to each other? On Tripath's own evaluation board, they are mounted parallel.

They're toroid cores, not air-coil, large enough however that saturation isn't an issue. Even though toroids are essentially self-shielding, turning the adjacent parts 90 degrees reduces the magnetic coupling and thus crosstalk. Two coils per channel are normal.

What are the mini potentiometers for? DC bias? Where do you probe to measure, at the speaker terminals?

DC offset across the speaker terminals. This is typical of Tripath designs. Adjust for minimum DC with no signal.

You do not use the usual made-in-China gold-plated speaker binding posts. What are these?

Made in the USA, EF Johnson, an old-school connector manufacturer, and a silver-plated part.

The line-input is just ordinary or gold-plated?

Gold plated, although not a super-premium type part.

The plastic LED reflecting the channel: Is that specially made or an existing item? That is so clever. All other T amps have LEDs that blind the eyes.

It's called a light pipe. I'm not 100% pleased with it as there is some light leakage to the adjacent indicators. Not specially made, it's a stock part available in a number of configurations. I think it's nicer than using a bunch of individual LEDs and wiring. One design goal was to come close to a "wireless" package. A lot of audio gear seems to have a maze of internal wiring. I wanted to eliminate that and the antenna effect that random wiring can have on noise. The light pipe eliminates eight wires for four LEDs. The originals have a green power LED (I'm kind of an old-school guy) but my beta-testers wanted blue for the bling factor. I caved in but used a fairly high series resistor to tone it down. The light pipe diffuses it as well.

Why did you opt for a push button instead of toggle switch? Was it an aesthetic or functional decision?

A little of both. It eliminates exposed hardware on the front panel and allows the use of a high-current part.

The overload protections, do you use the built-in circuit from the ST L7805 or Tripath chips?

The 7805 is a 5-volt regulator. Overload protection is on the Tripath chips.

Not all T-amps are created equal
I still had further questions. But Jay was going away on a short trip. Before I ceased my fire and let him fly off, I had one more question: "I'm planning to put the Mouse to the ultimate test after I finish all the comparison auditions. That is to use it to drive my Mark & Daniel giant killers (83dB, 4-ohm, requiring a high-current amp). But I need your permission first. Do you think that would be too risky?" Jay responded: "Go for it. It won't hurt the amplifier. Worst case, if you push it really hard and the impedance drops below about 2 ohms for long, the fault light will come on and the amplifier will mute for about a second. This will repeat if the condition continues. The same thing happens if an output is accidentally shorted."

Jay returned in a week and the email marathon resumed.

Welcome back, Jay. Everything is still in one piece. Congratulations, you passed the ultimate test! It sounded excellent!

Glad to hear it. The paralleled configuration of the 2050 chip does very well with low-impedance and reactive loads.

Also, the Mouse is the only Tripath amp that's not inverting polarity at the speaker posts. From what I know, all Tripath chips use an inverted input. How did you put that right at the output?

As the output is a half-bridge, it is a differential balanced output. Neither of the speaker terminals is 'ground' or common. This is one reason I didn't go with the conventional black-red binding post colors, to reduce the risk of customers assuming that the negative speaker terminal is grounded. As both speaker posts are active and swing in opposite directions, it is trivial to wire and label the outputs to mirror the input polarity. As you note, the input stage to Tripath controllers is inverting. I just verified this with an oscilloscope and wired and labeled the outputs to undo the inverted input stage.

Now I understand why it's called the Mouse. It'll bring the giant elephants to their knees. I just quickly glanced through the Tripath datasheets (TK2050 or TP2050?) and began to understand a bit more. But I still don't understand the half-bridge part. I thought all Tripath chips were full-bridge when in stereo mode.

TK2050 is just a part number for a set of two chips, a TC2000 controller and a TP2050 power stage. The TC2000 is a stereo modulator/controller. The TP2050 is a power chip that can be configured either as stereo or with both channels paralleled for monaural operation at twice the current. If you look at the data sheet, you'll see that the four-ohm performance of the TP2050 in stereo mode leaves something to be desired. The Mouse uses a TC2000 controller to drive two TP2050s, each in monaural mode. You can think of it as a TK2050 with an extra TP2050 power stage for the second channel. This gives the performance boost of the parallel TP2050 configuration in stereo with only a single controller necessary.

The main difference between the Mouse amplifier and the SI, Trends and KingRex designs is that they use a nominal 12-volt Tripath chip with the controller and power stage in the same package. With these chips there isn't a practical way to use the Tripath controller with a power stage having a higher rail voltage or current capability. The TK2050 uses a separate modulator and power stage. The clipping level is now based on the higher rail voltage of the TP2050, in the case of the Mouse 30 volts. I've taken this further by using a separate TP2050 for each channel. This gives better bass control and higher current capability for lower impedance loads. My initial prototype was pretty much a stock TK2050 stereo design. I liked the sound and performance but I found it inadequate to drive "difficult" speakers such as Infinity Kappa 7s. Going to a parallel monaural configuration on the output stage changed that and improved the sound overall.

The "secret sauce" of T-class is the modulator/controller, not the output stage. In fact, the TP2050 is really a Tripath-branded STI Microsystems STA-505. And you are of course correct. It is configured as a full bridge design. Each side uses two half-bridge amplifiers. Regarding the previous discussion, getting correct absolute polarity is just a matter of which half-bridge input connects to which of the two modulator outputs from the controller.

The last answer really explained everything. Someone had been thinking outside the box. That's why when I used the Mouse or Mice to drive the Mark & Daniel Topaz and Sapphire, the results were so astounding. The Mouse is unique and should not be perceived as a Red Wine Audio Signature 30 minus the SLA batteries, which isn't and cannot be bridged to begin with. Perhaps the Mouse is more like the Signature 70 as a one-box stereo amp without the SLA batteries? But let's talk more about power supplies. The SMPS (switched-mode power supply) of the Mouse, what most people would simply call the AC/DC adapter, doubles the size of most other class T amps. This is because the Tripath TA2040 and TA2020 for instance can operate on 12V SMPS whereas the TP2050 requires a parameter of 10 to 36V. The SMPS of the Mouse is rated at 30V/3.3A. After studying the Tripath data sheet, I had another question: "ST L7805 output max 24V and TP2050 input max 36, am I missing something here?"