Columns Retired Columns & Blogs |
..counts less than the specific implementation proper.
From the time the Luxman MQ-88uSE arrived at my house until I sat down to write these opening paragraphs, I knew only the following: It uses tubes for voltage and current gainthe latter accomplished with two KT88 pentodes per channelbut not for power-supply rectification; it's built on a spartan and decidedly old-school chassis, its vitals protected from people, and vice versa, by a removable tube cage; it was made in a limited edition of 100 units (the SE in the model designation stands for Special Edition; a similar amp, the MQ-88uC, will take its place in the line when all of these are sold); and its retail price is $5995 (USD). Also, because the shoebox-sized MQ-88uSE doesn't have room for output transformers of greater-than-moderate size, and because single-ended amps usually require much-larger-than-average transformers, it was a safe bet that it was a push-pull design.
I learned most of the above when I first saw the MQ-88uSE at the 2018 Montreal Audio Fest; not until today did I get around to inquiring about its US retail price, which is also $5995 (footnote 1).
Neither the design specifics that I ignored nor my reasons for doing so were trivial. In 331/SUB>3 years (footnote 2) of writing about domestic audio, I've gathered a great many opinionssome informed, others grounded in nothing more than whim, hunch, or emotionon the musical and sonic consequences of the various choices one makes in the design and construction of a tubed amplifier: Is it hardwired, or made using circuit boards? What's the output stage's class of operation? Is it auto-bias or fixed bias? Does the circuit include feedback, either local or global? Are the tubes run as triodes or pentodesand if as pentodes, what's the ratio between the amount of DC on the screen grid and the power-supply rail? Is the power supply built like a fretting hammer or a jackhammer?
An even bigger question, at least for me: Is the circuit designed along the lines of classic audio amplifiers, or does it involve such microprocessor-fueled refinements as sliding bias and switchable power-supply rails?
It was with open-eared ignorance of all of the above that I approached the Luxman MQ-88uSEwhich I describe, in the last page of my listening notes, as "one of the nicest-sounding and most engaging amps I've ever heard, bar none." Now for a look inside . . .
Historically grounded
Is it hardwired, or does it have circuit boards? The answer is a resounding Yes: All of the components in the MQ-88uSE's signal path, and the larger components in its power supply, are hand-wired, point to pointyet the MQ-88uSE's two solid-state full-wave rectifiers, plus a scattering of connectors and two relatively small capacitors, are mounted on a slender circuit board tucked away near the front of the chassis. Also mounted on small boards are the amp's left- and right-channel output-bias supplies, each fed by its own 60V secondary from the mains transformer. Both of those boards also contain the coupling capslabeled "LUXMAN OIL CAPACITOR"for their respective channels, plus a pair of miniature trim pots for adjusting bias current; the pots are tucked away on the hidden sides of the boards, accessible via small holes in the top of the chassis, just behind the output tubes themselves (footnote 3). From that, and the fact that the tubes' cathodes are at ground, we know that the MQ-88uSE is a fixed-bias design.
While poking around inside the MQ-88uSE, I took a few simple measurements with my Fluke multimeter. With the amp powered on and warmed up, and with shorting plugs on the inputs and loudspeakers connected to the outputs, I measured 485V DC on the plates of the output tubes and on the tubes' screen grids; indeed, on all four KT88 output tubes, those two elements were connected together through a 100 ohm resistor. With the same charge on their plate and screen grids, and with the output signal present on the former also appearing on the latter, we know the output tubes are run as triodes.
To know the class of operation of any fixed-bias tube amp requires a reliable means of detecting when that amp goes into clipping, usually determined with the help of a signal generator and an oscilloscope, both of which I lack. That said, given the amount of bias voltage used (footnote 4)I measured between 55.4 and 56.7V DC on the signal grids of the four KT88 tubesas well as the MQ-88uSE's highish plate voltage, I'm comfortable saying that the Luxman runs in class-AB. Connections between the output transformers' secondary windings and the cathodes of the ECC803 dual-triode input tube that comprises the amp's first stageeach channel has its own ECC803, the two halves of which are in parallelindicate the presence of global feedback.
Each channel also has its own ECC802 dual-triode tube, used as a cathode-coupled phase splitter, a means of doing the job that also adds voltage gainand so the MQ-88uSE is regarded as a true three-stage amplifier. If I understand correctly, this type of phase splitter was first described by the British tube manufacturer Mullard, and used in their famous "5-10" amplifier circuit of 1954, so named because it used five tubes to produce 10W of output power.
As for the power supply, it's as simple as can be: the aforementioned solid-state rectifiers, three big 600V filter caps aboveboard, one or two smaller ones belowdecks, and a hefty chokewhich, like the EI-style mains transformer and Luxman's limited-edition OY-series push-pull output transformers, is designed and wound in-house. Other parts are equally well pedigreed, including small-signal and output tubes from JJ Electronic, of Slovakia. As Jeff Sigmund, president of Luxman America, told me via e-mail, the tube-selection regimen for the limited-edition MQ-88uSE is referred to by Luxman as their "high confidence" process, meaning that the tubes undergo a more exacting than usual battery of tests and more precise matching to ensure the closest match and the best ultimate specs.
The construction quality is excellent throughout. The chassis is steel painted semigloss black, with a brushed-aluminum top plate and a painted steel bottom cover fitted with four cushioned feet. Inside, the wiring layout was sensible and clean, with no stray strands or shreds of insulation, and the soldering was neatly done. The inside of the chassis is painted, and I saw no stray droplets of solder or resin.
The MQ-88uSE's only controls are a pushbutton power switch and a level control on the front and, around back, a rotary switch that allows the user to remove that level control from the circuit. Also in back are two pairs of RCA input jacks: one pair affected by the level control, and a Direct pair, which is not. I had to strain to hear the difference between the two: a dulling of the trebles with the control in the circuit, so minuscule I might have imagined it. Specs include a power output of 25Wpc into all loads, 416 ohms; an input impedance of 32k ohms; and a signal/noise ratio of 105dB.
Swinging volts
Experience suggests that my DeVore Fidelity Orangutan O/93 loudspeakers love low-powered, push-pull amplifiersand the Luxman MQ-88uSE seemed awfully fond of the DeVores. Indeed, the 25Wpc Luxman amp had much in common with my 20Wpc true-pentode, zero-global feedback, fixed-bias, class-AB Shindo Laboratory Haut-Brion amp ($12,500). Each had its own distinct voiceput me in the dark and I don't think I'd have much trouble telling one from the otheryet both coaxed from the DeVores a much higher-than-average level of musical involvement.
Footnote 2: My first day on the job at The Abso!ute Sound was Monday, January 7, 1985. I began writing about the Luxman MQ-88uSE in early May 2018.
Footnote 3: The owner's manual doesn't mention those pots, let alone instructions on how to adjust the bias of the output tubes. The Luxman amp is shippedin a rigid double carton with well-designed packingwith its tubes preinstalled, presumably with their bias set at the factory.
Footnote 4: Not to be confused with bias current, of course.
..counts less than the specific implementation proper.
It is a beam tetrode; the KT means "Kinkless Tetrode".
With all the "technical detail" you go into, I'd think you could get this basic detail right.
Hi
When KT-88 is hooked as a pentode with standard push-pull (non-ultra linear) output mode, its Ip-Vp transfer curves are KINKED.
It got to be hook up as pentode mode at high Vp to enable to use such very high fixed bias voltage -45V or higher.
So Art was correct to say KT-88 is a pentode in this situation.
Jack L.
Canada
Thank you for the review. I'll never own one of these (kids in college) but it's still great to read about the products from such a classic brand. Their amps are always just so smooth and musical. I have a circa 1982 L-580 integrated and it still sounds beautiful. Maybe a little dark and rolled off but it just makes music.
I was happy even looking at the pics.
I agree!
Over the past couple months, I've read loads of articles about audio and audio equipment. Almost all ended up being contentiously commented on. OK, nobody says that's not allowed. But, it's not exactly cheery.
On the other hand, the couple I've read that were written by Art Dudley about this kind of product have always left me happy. *That* is cheery.
... 100w power rating, and it typically does closer to 140W.
It also has an excellent phono amp, with a separate MC pre-preamp.
If you haven't done so already, it would be worthwhile to have the 50, or so, electrolytic capacitors replaced. Also, several of the transistors in the power amp section run too hot, so that should be addressed, as well.
http://www.tapeheads.net/showthread.php?t=28573
Yes, that's the one. The large power caps were upgraded when I purchased it 10 years ago or so. It's in my secondary system now so probably won't do further upgrades but agree with you - it's very powerful and has a nice, rich, warm sound. I know that "Tapeheads" article detailing the restoration of that model and it's great. The restorer loved the sound of the amp.
... dynakitparts for a mere $600:
https://www.dynakitparts.com/shop/st-35-kit-120-vac/
They are charging for the looks and the name of the company :-) .........
With all the proprietary parts and hand assembly...not to mention the thought of the company turning a profit to pay it's workers. That's a very nice looking piece that seems like a very solid performer...something the company should be proud of. I was rather shocked to see the US corporate office almost in my back yard, of all places.
Dear Art, here in Italy Luxman MQ-88uSE is sold only in a bundle with its pre-amplifier companion Luxman CL-38uSE that also has an interesting phono MM/MC. It could be interesting to know your opinion about the couple, maybe with a little listening-review, and, in this case to know differences between you reference pre and the Luxman pre. Any possibility?
"Luxman MQ-88uSE runs its output tubes as triodes, not pentodes. These tubes must be stupid: They don't know they're supposed to be slow and temporally imprecise when used as triodes." quoted Art Dudley.
This is the tonal characteristic of any triode-converted pentode using a resistor: relatively slow, soft, lacking punch vs its pentode configuration.
Decades long established audio amp. manufacturers, like Lux, Japan still sticks to the old-old-school design of triode strapping with a resistor. What a shame !
That said, if a pentode is properly converted to a triode using some new school type topology, it will sound fast, punchy like a pentode. I design-built a few such conversions with good success many years since.
Instead of using a resistor strapping the screen grid of the pentode directly to its plate like many amp manufacturers, including Lux, did in the past, which sounds slow & punchless, I have used a new topology, which I called it "Plate-screen split potential topology".
Instead of using a resistor so simple, the topology requires strapping the screen grid of the pentode to its plate with a circuitry of high quality capacitors, resistors & diodes, providing a STABILIZED potential difference between the plate & the screen grid. With the right split potential, the trioded pentode will sound superb:- faster, more detailed, more transparent than & as powerful as its pentode mode !
The 'trick' is to choose what right potential split ! I experimented quite a while until I found the right plate-screen voltage split. The superb sound of such triode conversion is very rewarding.
My first successful triode conversion was done on my vintage Dynaco ST-70 power amp many years back. I installed the pentode-triode split-potential conversion of EL-34 push-pull output power tubes. With a pair of toggle switches, I can switch back & forth pentode-triode modes.
Such split-potential triode mode sounds so much better than its original pentode output mode in term of speed, transparency & delicacy. I would never want to go back to the pentode mode again.
Listening to believing
Jack L.
Canada
I'm a little late to this conversation, but I could not pass up a great deal a few months ago on a new one of these amps, and it's just now broken in. Art didn't get into this much in his review, but the Variable mode can be used to connect a source component directly into the amp using the attenuator on the front panel (so it seems like this should be the Direct mode, but it's just the opposite). I've done this with my CD player (Naim), analog front end (Rega 10, Benz Ruby Z and Musical Surroundings phono pre/Auditorium step-up), and my oldie but goodie Yamaha tuner. IMO, they are all more satisfying this way as opposed to going through my Shindo preamp. Now don't get me wrong, my Monbrison (the original, not the newer one) is wonderful, and it's more convenient using the preamp (you can only connect one source at a time using Variable mode.) But used with this option, I think this amp might be even better than Art says it is, and he really liked it. And thanks Art, you have never led me astray.