Revox G-36 open-reel tape recorder

The Swiss-made G-36 recorder had earned an enviable reputation among perfectionists during the few years that it has been available in the US, and our inability to test one (because of a backlog of other components for testing) became increasingly frustrating to us with each glowing report we heard from subscribers who owned them. Now that we have finally obtained one through the courtesy of ELPA (footnote 1), we can see what all the shouting was about, but we also have some reservations about it.

We must admit at this juncture that we have a certain, somewhat irrational tendency to be overly critical of minor weaknesses that we find in generally excellent components. If something is frankly mediocre, it is easy to accept with equanimity certain minor drawbacks which, in an otherwise top-notch component, would prompt us to become a bit hot under the collar. We can't help it; that's just the way we are, and we urge you to bear this in mind when we start, later in this report, making what seems to be an undue fuss about some of the things we weren't too happy about in the Revox.

We can state flatly at the outset that this is one of the most beautifully designed and constructed recorders we have had the pleasure to work with. Mechanically, it is a joy to use, with pushbutton-operated solenoids, gentle, positive tape handling, very quiet running, smooth and even high-speed winding, and safety features such as interlocked control modes, photoelectric tape sensing, and automatic braking in the event of an accidental AC supply failure or a tape breakage during fast wind. It is almost impossible to damage a tape on the G-36.

We were also happy to find that there are no head pressure pads in the G-36, which means more uniform head wear, reduced violining, and a lessened tendency for tape to squeal. Intimate head contact is obtained by simple tape wraparound, and the G-36's freedom from left-channel dropouts (footnote 2) is testimony to the effectiveness of this wrap-around arrangement.

Wow and flutter were so low, even at 3¾ips, that it was difficult to distinguish between the playback of a 3kHz tone and the direct output from an audio oscillator.

We had some misgivings, initially, about the Revox's rather low (70kHz) bias frequency, for experience had shown that even 100kHz bias will normally tend to beat against the highest harmonics in the program material, causing a slightly hazy sound and creating "birdies"3rapidly swooping whistles3when one sweeps through the upper range of an audio oscillator. To our amazement, the G-36 had no audible birdies at all, and produced beautifully transparent playback from complex musical program material. We could find nothing in the schematic that would account for this, but we suspect that somewhere, somehow, something is suppressing the spurious ultrasonic harmonics in the program signal during the record mode. Perhaps Revox will enlighten us?

The importer informs us that the G-36, as delivered, is now being factory-adjusted for use with 3M's low-noise 201, 202 or 203 tapes, so we used 203 (1-mil polyester) for our tests. First, though, we checked the machine's adjustments for accuracy and found them to be right on the nose. The measured response of both channels was so similar that we thought for a moment we were just measuring the first channel over again (fig.1).


After our initial response checks, we readjusted the bias in both channels according to the service manual (not normally supplied with the recorder, but available from the importer if you're strong enough on the technical end to cope with it), and repeated the tests. They matched the first set of measurements to within a fraction of dB.

Like most professional tape recorders, the Revox has what appear to be very modest frequency response specifications. And like most professional machines, it exceeded its specs by a respectable margin. The published signal/noise ratio figures are probably for regular tape coatings, which would explain why our unit exceeded these by a wide margin when using low-noise tape. We measured a hair under 59dB of S/N, which would translate to about 53dB when you subtract the 6dB advantage afforded by the low-noise oxide. Suffice it to say that, with its recommended tape, hiss was completely inaudible except when listening at very high levels to material with wide dynamic range. As a ready comparison, 59dB S/N is better than what comes off a typical Dolbyized disc.

Crosstalk interference between adjacent recorded channels was not as low as in some professional machines, but it exceeded its rated 40dB by 2dB, and proved to be acceptable under most listening conditions. Bear in mind, too, that the lower hiss of the recommended 3M tape enables one to hear soft adjacent-channel interference that would otherwise be swamped by the residual hiss of ordinary tape.

Listening & Use
Subjectively, playback from commercial tapes was unusually clean and lucid, with a slight "foppishness" and a perceptible although not annoying deficiency of deep low end. Predictably, its own tapes (on 3M 203) were astoundingly transparent, but the ability to make direct A/B comparisons between the original and the playback made the high- and low-end differences somewhat more noticeable than from commercial tapes, where no such easy comparison is possible.

The G-36 has switching facilities to select stereo recording, mono recording on either track, echo effects, copying from either track to the other, or sound-on-sound re-recording with either track being mixed with an incoming signal. It all sounds very flexible, but here's where we started running into Things We Were Unhappy About.

The Revox has a built-in low-powered amplifier and a small loudspeaker (about 6" by 9") for monitoring and portable playback. This is a convenience in some rare instances, but in most applications stereo playback would be required, and headphones or a separate amplifier and speakers would provide this. In view of its limited usefulness, it seems to us that the built-in amp and speaker adds an unnecessary amount of weight to what is supposed to be a portable machine.

A selector switch allows the built-in amplifier and speaker to monitor the input signal or the tape playback from either channel or from both stereo channels combined (A+B), but there is no way of monitoring the input signals in stereo. There is only one, monophonic, output from the power amplifier, so you can't use low-impedance stereo headphones in place of the built-in speaker. The only source of stereo output connections is the high-impedance signal outlets, and there is no way of routing the input signals through these. They only yield output signals, from playback of the tape, so the machine must be recording before you can hear what's coming into it, and in order to balance a stereo signal by ear.

The VU meters, on the other hand, read only input signal, so while they can be used for setting maximum record levels before starting to tape, they can't be used for in-the-field spot checks of playback level and frequency response, as can the meters in most professional recorders. (Incidentally, unlike most small record-level meters, these are actually genuine VU types, with the standard NARTB ballistic characteristics.)

In addition to the lack of stereo input monitoring facilities, the output level from the high-impedance outputs (500mV) is a bit too low to provide adequate level for driving most headphones. It would be necessary in most cases to use an amplifying headphone adapter such as the one Koss makes for feeding Low-Z phones. And you'd still have to set stereo balance while recording, and would be unable to hear audible cues that would warn you that the program was about to start.

Of course, the foregoing criticisms are not likely to be of deep concern to you if you only intend to use the recorder in conjunction with a separate hi-fi system. You could always monitor the input signal through the main system, using the preamp's tape monitor switch for comparing input with tape output. But the G-36 strikes us as being poorly suited for live stereo recording in the field, via microphones. Indeed, the unit almost gives the impression that live-recording facilities were added as an afterthought, for the microphone input sockets are phono plugs, which have not been used on any microphone cable we ever saw. Suitable adapters could be made up for any Hi-Z mikes (or Low-Z types with matching transformers), but why should this be necessary?

Another thing that struck us as rather strange was the fact that although the Revox's designer included such sophisticated refinements as highly effective bias trap adjustments, a crosstalk null adjustment, and VU meter calibration adjustments, he neglected to include some other adjustments that are generally considered to be de rigeur on any precision recorder. There are bias current adjustments, but there is no means for coping with differences in sensitivity between different tape coatings. The bias adjustments are used to obtain the flattest possible high-end response from the tape being used, but in the case of the recommended 3M tape, optimum treble response made the tape playback (through the built-in monitor amp) about 3dB higher than the input signal. The resulting apparent treble increases (due to the increased loudness) tended to dilute the validity of the A/B comparisons.

When using the tape monitor switch on an external preamp, fed by the Revox's Hi-Z outputs, the A/B level difference (at normal record levels) depended on the level of the signal source from the tuner or phono preamp. (A preamp's tape monitor switch is typically located just ahead of its volume control, so equal volume between Source and Monitor positions of the switch demands equal signal voltages going to and coming from the recorder.) With a tuner having its own output level control, it was possible to match A and B levels by adjusting tuner output to match the recorder's output, but there was no way of making similar adjustments to phono preamp output levels, and no way of controlling the recorder's Hi-Z output levels to match the input signals to it. So, again, the A/B comparison is not likely to be very revealing, except to show that the recorder is indeed putting clean signal onto the tape.

There are, of course, several top-grade professional recorders whose playback volume is fixed (or, rather, is adjustable by relatively inaccessible screwdriver-slotted pots), but all of these have their own A/B monitoring switches as well as provision for accurately matching input and output levels.

Although the G-36 will accommodate 10½ reels, it will not accept the standard NARTB large-reel hubs, but requires instead the lO½ reels having the same small, keyed center hole as a 7" reel. A suitable 10½ reel is supplied with the G-36, but it would be nice if Revox included a couple of adapters for the NARTB hubs. Such adapters are available from the importer and probably also from most Revox dealers.

The G-36 is not designed for easy tape editing. In the Stop position, the tape lifters pull the tape away from the playback head, making it impossible to slow-cue the tape by hand. Fortunately, it is possible to cheat this arrangement by removing the head covers (they just lift straight up) and threading the tape behind the lifter rods.

Internal wiring in the Revox is very neat and orderly, but like most European machines we have seen, it does not look terribly easy to service. Some parts are buried under layers of other parts, and even some of the tube socket connections look a bit awkward to get test prods to. The service manual (available at extra cost) is comprehensive and detailed, but its format is rather confusing to anyone accustomed to American-style manuals. This is not recommended reading for the casual hobbyist who is weak on the technicalia.

Summing Up
All in all, while there are many truly excellent things to be said for the Revox G-36, we feel there are enough minor drawbacks to weigh the scales in favor of some of its competitors. The $600 price class has been dwindling of late, with the withdrawal of the Ampex F-44 and the impending discontinuance of the Dynaco Beocord 2000, but there are still some worthy contenders in the field, not the least of which are the TandBerg 64X and the Sony TC-777-4. Both of these can be set up for use with low-noise tape, both can provide somewhat smoother overall response than the Revox, and both have their share of minor shortcomings, too.

Of the three, our first choice would probably be the Sony, at about $150 more than the Revox. Between the Revox and the TandBerg (both around $550), we would bave to decide on the basis of specific requirements, for each machine has advantages and disadvantages that are not shared by the other.

On the other hand, we might even be tempted to sit back and wait until someone comes up with the ideal $600 recorder. But then, that's another story.

Footnote 1: As we went to press, we learmed that ELPA Marketing will no longer be carrying Revox recorders, as Revox is setting up its own sales office in the US. They will, however, continue to service the naw-discontinued G-36 units that were purchased from them, and will continue to honor warranties.

Footnote 2: Dropouts due to momentary loss of tape-to-head contact occur most frequently in the left channel of a quarter-track recording, because this channel is closest to the edge of the tape.


monetschemist's picture

A real stroll down memory lane. I remember, as a poor university student, lusting after a ReVox (I think a B77).

Also somewhat sobering to read the specs.

Any idea to review one of the "new" reel-to-reels out there?

Thanks for bringing great articles like this back to life.

JRT's picture

This is an interesting nostalgic article on a piece of equipment used in archaic recording schema, but (aside from narrow use special effect of using tape saturation to apply dynamic compression) it makes no sense to chase after it or anything like it now for general use in audio recording and playback as its performance has long since been eclipsed by the performance currently available on moderately priced AD/DA recording/playback hardware such as Lynx Studio Technology's E22 and E44 family of PCIe sound cards.

jmsent's picture

..were very different animals from the G36, and much better machines. They had addressed virtually all the complaints JGH expressed regarding the operation. And, they had made a big upgrade in the transport by doing a direct drive capstan system with servo control. They also made serviceability far better by using plug in cards. Of course, electronics were all solid state, not tubes, and this also improved on the noise specs. The G series is great for nostalgia buffs, and certainly built like tanks. But all the machines: Sony 777, Tandberg 64, and Revox G36 were pretty much obsoleted by the next generation of models from these manufacturers. I worked on all of them, and a lowly Sony TC 366 that sold for a couple hundred bucks could easily outperform a Revox G36. It was just a matter of newer technology filtering down market.

monetschemist's picture

Have you enjoyed any of these recordings?

All-tube recording chain (if you go the LP route, anyway).