Futterman H-3A power amplifier

It is not at all unusual these days to find manufacturers producing "matched" speakers and amplifiers that are designed specifically for one another. But it is very unusual to find this being done by an amplifier manufacturer who doesn't make loudspeakers. The Futterman H3-A is one of these rarities—an amplifier designed primarily to complement one of the best, and one of the hardest-to-drive loudspeakers on the market: the KLH Model Nine.

The H-3A is unusual in another respect, too. It is probably the closest anyone has come to combining tube and transistor technology, for although it uses vacuum tubes, its output circuit and its lack of interstage capacitors and output transformer have more in common with transistor amplifiers than with tubes. Thus, although it runs at the typical high temperature of high-powered tube equipment, it can also have the ruggedness and dependability that are claimed for but rarely achieved from transistor amps, and it has the cardinal virtue of being repairable in the field, unlike transistorized units which must usually be returned to the factory or an authorized repair station.

Another difference between the H-3A and most transistor amplifiers is that, while the latter are designed to deliver maximum power to a load of between 4 and 8 ohms, the H-3A yields its maximum power (100 watts) into 30 to 40 ohms. Into 16 ohms, it gives 90 watts (continuous sinewave) per channel, and will put out about 45 watts per channel with 8-ohm loading. At 4 ohms, output drops off to around 19Wpc, which is a bit ridiculous for a $288 amplifier. So although the H-3A can be used with 4-ohm speakers, it would be rather foolishly wasteful.

The literature supplied with our H3-A also recommends it for use with Quad electrostatics, and this we must take very strong issue with. The Quad speakers are designed to handle about 15 watts per channel, and can be damaged by even momentary application of power in excess of 30 watts per channel. With the Quads' 16-ohm impedance, the Futterman could deliver almost 100 watts to one of them as the result of an inadvertent loss of a ground connection at one of the phono input leads. Although one might avoid such accidents for a long period of time, the consequences of such an occurrence could be a bit too serious to allow the possibility to exist at all. Better to use a lower-powered amplifier with the Quads.

As a matter of fact, the KLH Nine is the only 16-ohm speaker system we know of that can use more than about 25 watts of power. In general, 16-ohm designs tend to be horn-type systems, which can blow you out of the room with an honest 10 watts of input signal, and all of the less-efficient 16-ohm systems we know of have maximum power ratings of 50 watts or less. So in order to justify the use of the Futterman on anything but very inefficient 16-ohm speaker systems, it would have to be demonstrably better-sounding than competitive amplifiers at output levels of about 40 watts or less. And there's some very stiff competition in this power class.

The circuitry is very well thought out and the amplifier is beautifully put together. Indeed it looks more like a piece of laboratory equipment than a home high-fidelity component, with clean, very neat wiring and (among other things) an array of 14 machine screws fastening the bottom plate to the chassis. We couldn't find a single sheet-metal self-tapping screw in the whole amplifier, and we looked carefully.

In addition, the H-3A is more thoroughly fused than anything we've ever seen before, and in quite an unusual manner. Apart from the two standard front-panel fuses, for the B+ supply, there are 11 circuit fuses inside the chassis, each one consisting of a single strand of ordinary hookup wire—a thick strand for heater circuits, a thin strand for the lower-current plate circuits. And fastened to the underside of the chassis, via its own little holddown clamp, is a supply of the two different sizes of fuse wire. If this sounds like a slightly shoddy way of doing things, remember that many industrial and laboratory-type electronic instruments are protected by just such bare-wire fuses, soldered into the circuit.

Frankly, we very much like this idea, because it allows for comprehensive circuit protection at a cost far below that of an equal number of conventional fuses. The H3-A's extensive fusing, incidentally, is not because the unit is more likely to blow up than any other amplifier, but simply because without such protection, its high-current circuits could do serious damage to themselves (and possibly to the loudspeakers, too) if anything ever did let go.

The H-3A has eight internal setup adjustments—more than we've ever seen in a power amp, and detailed instructions are provided for making the adjustments "every thousand hours or so, or whenever tubes are changed." We checked the adjustments when we received our H-3A and again after about a couple of hundred hours' use. They were right on the button each time, so there's certainly nothing unstable about them. Some of them are fairly critical, though, so they should be maintained at the recommended settings if the amplifier is to give minimum distortion and maximum output without burning up the tubes.

The H-3A normally runs quite hot, so it should be well ventilated or, if it's located where you won't accidentally touch the uninsulated output tube caps, it could be used with its perforated cover removed. Despite the high operating temperature, though, none of the H-3A's tubes had the dull red plates or blue glow that are often a sign that tubes are being overworked. And the fact that it ran for about 200 hours without a measurable change in characteristics suggests that the tubes should last a long time.

Loudspeaker Choice
Because it seems (to us) obviously better suited for use with the KLH Nine than with any other speaker system, we did the bulk of our listening on that system, comparing it with the two best amplifiers we had found for use with the Nine: The Marantz Model 9 and the Dyna Stereo 70. We also tried it on two 8-ohm systems—the Janszen Z-600, with its electrostatic high end, and the Ampex 4010 all-dynamic system, and compared it again with the other amplifiers.

Sound Quality
With the KLH Nine, our initial reaction to the H3-A was that it was far and away the best thing we'd ever driven that speaker with. Highs were very crisp without being hard or brittle, and the overall sound was highly transparent and effortless. With prolonged listening, however, we became aware of a certain heaviness in the mid-bass range, which had the effect of covering up some of the output in the range below about 60Hz. Neither the Dyna Stereo 70 nor the Marantz Model 9 amplifier showed any tendency to do this, with the result that both of these gave the impression of tighter mid-bass and considerably more deep-bass output.

Highs from both of these were somewhat softer than those from the Futterman, and although the Futterman's crisper highs did not seem to add any roughness to record groove breakup, we found that we preferred the slightly sweeter high end with most program material. There was not, however, a trace of the typical transistor hardness from the Futterman, even at very low listening levels.

At high levels, approximating a row-H seat in the concert hall, the Dyna was showing marked signs of breakup on power peaks, the Marantz was showing some roughness, and the Futterman was still as effortless-sounding as it was at low listening levels. Our efforts to overload the Futterman yielded nothing more than a succession of blown fuses in the KLH Nines. Normally, one would not expect to notice this much difference between a 70W (Marantz) and a 100W amplifier, but we've observed this seeming ability of transistor amps to put out more than their rated power, so it is not really so surprising that a transistor-oriented tube amplifier would behave the same way.

On the Janszen Z-600's and the Ampex 4010's, there was less tendency toward mid-bass heaviness, but the loss of deep bass seemed even more marked. The rest of the range, through the Ampex speakers, was superbly reproduced, with the same overall timbre and transparency as from the other top-grade amplifiers. Strangely, though, highs from the Janszen speakers were somewhat softened, as though there was a gradual rolloff above about 8kHz. Highs from the Janszen were in fact a shade softer than when using either the Dyna or the Marantz. We found the softening a help with some program material, but the majority of discs and tapes came through with a bit too much edge removed from them. The Ampex speakers, on the other hand, were not at all affected in this way. Their highs sounded almost exactly the same with the Futterman as with the other amplifiers.

With both of these 8-ohm speakers the Futterman's loss of power (due to the impedance mismatch) made it only slightly cleaner than the Dyna at output levels barely exceeding 35 watts on peaks. At the same level, of course, the Marantz was cleaner than either of the others, but again, not as much so as the power difference would suggest.

We found one other thing about the Futterman H-3A that may or may not be significant, depending on the circumstances. When using three-terminal pickups or a tonearm with a common grounding circuit—a bad arrangement to begin with-the Futterman exhibited fairly severe ground-loop hum. This could be reduced to a tolerable level by disconnecting the ground at one of its inputs, but was not entirely eliminated.

For instance, we had the trouble with a Decca pickup (3-terminal output) and a Dyna PAS-3 preamp or a pair of PAM-1 preamps with the stereo adapter. A 4-terminal pickup worked fine with the PAS-3, and hummed with the two PAM-l's until we opened up an input ground, which then eliminated the hum entirely. The PAM-1 combination, with its separate chassis, is especially prone to hum-loop problems anyway, but the PAS-3 is unusually good in this respect. With other preamps, it would be best to try the combination before buying.

This is one of those puzzling instances where an objectively superb amplifier just does not seem to deliver quite what one would expect in the way of subjective performance. Every instrument test we gave the Futterman H3-A (response checks, squarewaves, IM distortion measurements) showed it to be one of the best amplifiers we've ever encountered. Yet its low end performance on every speaker we tried it on was less satisfying to the ear than was that of some other amplifiers that don't measure as well. We wish we could at least volunteer some educated guesses as to why this might be the case, but we're stumped.

We can only say—with much regret, because the manufacturer clearly set out to produce an absolutely no-holds-barred unit—that we feel there are other amplifiers better suited for most loudspeakers, and that the H-3A does not sound quite as good to us on the KLH Nine as do some other amplifiers.

Harvard Electronics Co.
New York, NY 10025 (1967)
Company no longer in existence (2019)

John Atkinson's picture
I chose this review to republish on the website because this was the very first example of that extraordinary beast, the output-transformer-less tube amplifier, that I encountered!

John Atkinson
Technical Editor, Stereophile

Ortofan's picture

... an output transformer or a coupling capacitor?

As an experiment, buy a pair of "computer-grade" 800µF aluminum electrolytic capacitors, connect one in series with each channel's output of your amplifier and determine what effect, if any, they might have on the sound quality.

scottsol's picture

You should have realized, and I have verified this from the Acoustat .comments, that publishing old material is a colossal waste and that nobody could possibly be interested.

John Atkinson's picture
scottsol wrote:
You should have realized, and I have verified this from the Acoustat .comments, that publishing old material is a colossal waste and that nobody could possibly be interested.

With all due respect, the page views stats show that you are wrong.

John Atkinson
Technical Editor, Stereophile

scottsol's picture

My post was intended as a satirical play on that ridiculous comment in the Magnepan 20 thread. The Acoustat reference was a mistake as I had just posted about the 2+2.

michelesurdi's picture

how time does fly.nowadays it is quite common to suggest that tube amps apparently deli ver more than their rated power.

Bogolu Haranath's picture

Manufacturer's specs say ......... IM distortion <0.05% at 90W continuous into 16 Ohms ....... S/N ratio better than 93 db below 90 Watts ....... Damping factor 200 ........ SNR of 93 db is approx. 15 to 16 Bit resolution ....... Can an OTL tube amp achieve a damping factor of 200? ....... Beast indeed :-) ..........

Bogolu Haranath's picture

Certainly, Futterman was ahead of his time, prepared for future-proofing ......... He was thinking about hi-res audio in the 60s, with that kind of SNR :-) ........

T-NYC's picture

The amp was ideal with QUADs if you told Julius that's what you planned to use. He adjusted the output to lower wattage running pure class A. He also strongly recommended Rogers LS35A (15 ohm) for use with this amplifier.

Herb Reichert's picture

they were broken I fixed them (new caps and tubes) tested them and hooked them to both Quad 57s and LS3/5a. For about 48 hours I was completely taken with their midrange. Then I sold them quick while they were still working. lol


Ortofan's picture

... Hoovie's Garage approach to Hi-Fi.

Bogolu Haranath's picture

Wonder whether HR also has a photo-album diary? :-) ........