Home |
Search |
Today's Posts |
#1
|
|||
|
|||
S.E.X. amplifier review by Andre Jute from Glass Audio
This review of the Electric Tonalities S.E.X. amplifier appeared first
in Glass Audio. The S.E.X. kit has been revised since this article was last updated. It is now a parafeed design and a case is available. Details from http://www.bottlehead.com ************ How to Enrich Your S.E.X. Life for only $399 with the Single-Ended eXperimenter Kit by Andr=E9 Jute Highlights: Least expensive possible entry to qaulity single-ended tube amps; overcomes standard SE speaker problem with $30/set of 4 high-sensitivity speaker option; monobloc set; upgradable; expandable For: a proper dual monobloc amp to a clever design, simple construction, outstanding illustrated assembly instructions with phone and e-mail backup available, sensitive speakers a low-cost option, excellent sound from base model Against: typical low SE output requires sensitive speakers Prices: S.E.X. kit $399, with 4 speaker-drivers $430 Contact: Electric Tonalities, PO Box 2786, Poulsbo WA 98370, USA. Tel: +360 697 1936. Fax: +360 697 3348. http://www.bottlehead.com Even cheap single-ended amps off the shelf cost a lot, and most single-ended amps offer such modest output, and such steeply rising distortion (Reid Welch wrote here last year of distortion rising at one per cent per watt of output), that ultra-sensitive--read ultra-expensive--speakers must be factored in to the price of a usable system. With complete SE systems being so expensive, the would-be experimenter and tweaker's inclination to fiddle declines in direct proportion to the cost of the original item. Thesis, Antithesis, Synthesis Enter that fine old Yankee from Seattle, Daniel G Schmalle of Electronic Tonalities, who admits to being as cheap an electronics experimenter as ever lived. He correctly identified the four characteristics that make single-ended amps expensive: 1. Custom-trained labor on low unit numbers. 2. Custom casework in low numbers. 3. Custom transformers in low numbers. 4. Custom high-sensitivity speakers in low numbers. 'Custom' and 'low numbers' are the two price-elevators. Schmalle's S.E.X. or Single-Ended eXperimenter's Kit overcomes all these problems by packaging the amp as a kit, to which the purchaser adds the labour, by supplying a chassis plate which the purchaser encases, by using transformers made in their millions for other purposes, and by supplying inexpensive speakers to stack up until the correct sensitivity is achieved. This is the genius of simplicity. Schmalle, one is not surprised to learn, studied physics, the most reductionist of all the philosophies. The S.E.X. kit The kit, complete for two monoblocs, costs $399. With the four optional speakers it costs $430. I would suggest that, even if you already have horns or other 90dB/W or better speakers, you lash out the extra because these little spun aliminium speakers are worth having. For $150 extra the S.E.X. is delivered fully built; kit buyers who decide building it is too much for them can also return it at any stage for rectification and completion for the same price. The kit arrives in Ireland, after three airborne weeks that reflect very poorly on the US Post, in a cardboard box that has survived the experience but might not survive a return journey. Inside everything is wedged in according to a plan, with liberal application of foam chips, and nothing is broken. My packing checklist is signed by 'Dan', presumably the designer, and another parts count list is signed by 'Eileen'. The key element in the kit is actually the most unobtrusive: the two aluminium chassis plates on which the two monoblocs will be mounted. These plates are .040 thick, which is just thick enough not to bend under the weight of the iron if you handle the construction by the transformers rather than the plate. Every drop-through hole, valve mount, cable access and bolt-hole is beautifully laser-cut. There are no rough edges. The ali is even lacquered against oxidization. It is a shame to attach the well-made but still cheap wafer sockets to such nice plates--they cry out for nice ceramics--but Truth in Reviewing demands that first I build it the amp as the designer intended. The chassis plates are intended to screw onto a simple box that the purchaser provides. My chassis plates were not drilled at the corners for attachment because the original intention was that buyers would use the generally available Hammond boxes and drill their own holes to suit. In fact most will probably build simple wooden cases and screw the chassis plates to those. Since one must do simple woodwork or have it done for the speakers anyway, I am not surprised that an attractive wooden case is preferred to a cold steel one. Before I heard how good this amp sounds, I was eyeing a plastic cake tin of just the right dimensions but have now decided that would be sacrilege. Future deliveries of the chassis plates will have screwholes cut into the corners for mounting onto the casework. The kit is utterly complete except for the case and few pieces of wire. The buyer must provide a heavyish gauge of copper wire for the ground line; two feet of television coax from any electrical store will do. He must also provide about a foot of signal wire. The instructions advise using the offcuts from the inline fuse fitting but this was too cheap for even the dumpster diver in me: I revolted, picked up a scrap of microphone wire from the floor and stripped out the two signal wires from that. And the owner must provide several 8-10in lengths of wire links, for which I bought a yard of six-core wire used by my favourite organ-installer so as to get multiple colours; stripped out mains cord will do fine. Suitable wiring was found in my own garage and in the garage of every (non-electronic) neighbour on every side. All the parts are of the cheapest possible supply, if not necessarily of the cheapest possible provenance. By using true mass-produced parts, now in surplus, a good quality is achieved without driving the price up. For instance, the output transformers appears to be from the age of tube television; they are beefy enough to handle 60mA or more. The mains transformers, also surplus, too are beefy items. Surprisingly in a kit at this price, the supply is choke rectified. Caps are electrolytic in the power supply filter, for supply decoupling and for cathode bypassing, but film for signal decoupling. On the whole, the parts quality came as a pleasant surprise. The circuit This is a KISS-KISS clever amp. The 6DN7 is a dissimilar double triode in a single envelope. Each channel has two 6DN7. The voltage amplification halves of the two tubes together form the first stage, which is configured as a mu-follower, a variety of constant plate-current circuit. (For an explanation of its operation, see the article by Alan Kimmel in Glass Audio, 2/93.) The remaining halves of each valve are arranged as a parallel single-ended output stage. The combined plate resistance is about 1000 ohm and the standard airgapped single-ended output transformer has a primary impedance of 2500 ohm. With a better optional output, no negative feedback is required. With the standard output there is a feedback loop from the secondary of the output transformer to the cathode of the lower tube in the mu-follower. About 10dB feedback is used. High frequency response is protected by a bypass capacitor in the feedback loop. The secondary of the standard OPT has a single 16 ohm tap but it will work with 8 ohm speakers. Resistors of two values are supplied in order to tailor the feedback loop for either 8 or 16 ohm speakers. This is a nice touch in so cheap an amp. Output impedance matching is important for best frequency response. The power supply is full-wave rectified by silicon diodes, followed by a two-stage capacitor input filter with the choke in the first stage and a resistor in the second. Power for each side is entirely independent of the other side; these are true monoblocs. Power supply is international for 110-240V mains. The voltage divider beyond the pi-filter raises the filament supply above ground by about 90Vdc to keep the cathode-heater voltage differential manageable. Construction What construction? It's a bit of fun and it is over too soon. Schmalle claims to put a pair of monoblocs together in 40m and with a little practice it should be possible. The parts count is low enough. But, most of all, there is a build-it-by-the numbers instruction scheme that makes this the best instruction set I have ever seen in a kit. The only thing missing is advice on which way round to put in the resistors... (Yes, I know. But good practice is to have all the tolerance bands pointing in the same direction.) The only tools required are two screwdrivers, adjustable spanner, wire cutter/stipper, soldering iron and a voltage/resistance meter. Needlenose pliers are handy for bending up leads neatly. Bring your own solder. Instructions are given for clearing the protective lacquer from the aluminum chassis plate and priming and painting it. I forwent this stage, intending later on to spray plate and transformers alike with Hammerite, which will stick to anything. The mains transformer, choke and output are first bolted to the chassis, the latter two loosely. No grommets for through-hole wires are supplied, so I protected the choke and output wires by slipping pieces of the stripped-out co-ax outer case (see below) over them. Next the wafer tube sockets, RCA phono input, binding post outputs, and on/off switch are bolted in. Be careful to handle the assembly only by the mains transformer or the beautiful chassis plate will eventually bend. Four strips of four tags each are bolted onto the feet of the choke and the output transformer and their bolts are then tightened up, an admirable piece of value engineering. Next, all tag points are numbered in order according to the plan supplied as the first instruction. Each instruction has space for a tick to be made once it is executed. Every instruction is complete in itself and polarity instructions are repeated as required. It would be impossible to go wrong. Most of the main links are made with the leads from the transformers. The filament current is carried between the tubes by the offcuts. Most of these wires are just long enough to reach the tag to which they must be attached. This is actually a good thing, because otherwise one cuts them a little too long for fear of cutting them too short, and the construction looks amateurish. A ground bus is made from a piece of thickish single strand copper wire the user must supply. It is threaded through four tags and grounds the chassis plane at the input RCA's uninsulated ground. I had an offcut of television installation co-ax which I stripped out for the wire, reusing the sheath as makeshift grommets where the choke and output transformer wires pass through the chassis. The resistors, diodes and capacitors are then fitted between tags, or between tags and valve bases, or from valve bases to the ground bus. Each component is in turn described as to appearance and markings, which struck me as a bit superfluous as a multimeter is anyway required for the impressive test schedule and can therefore be used to identify parts. Still, every single part could be identified from the description without any external aid, and those requiring polarization polarized. An idiot could build this amp! Though the S.E.X. is a simple device with the minimum possible parts count for the specified performance, it is also a sophisticated amplifier, so that the parts counts is quite a bit larger than the theoretical text-book minimum. Some of the sixteen tags become heavily soldered with so many junctions to be made. This is not helped by the fact that, in true American style, the tinning on the tags is third-rate. (I noted American plating problems too when I was in automobiles: everything imported chromed from the States into Europe had to be reworked by professionals.) The instructions warn not to let the holes in the tags fill up with solder but the holes are too small. My solution was to make the excess wire on the first component to be soldered to each tag into an extended 'ear' standing straight up, around which further component leads may be looped and soldered. When everything is finished, one must carefully trim away the surplus of the ears because otherwise they act as RF antennae. Finally, several links must be made, for which the user must supply his own wire, and for which distinct routing instructions are given. With all this point-to-point wiring, one is forced to think in three dimensions! I made mine from multicolored wire for aesthetic reasons. It is clear that the placement of the components and the wire-dress have received much attention. The instructions provide drawings of the top and the bottom of the chassis, photographs at several stages of construction, and a full-size wiring diagram. Descriptions for using ready-bought boxes or making a wooden case are offered. In the end I cased mine as monoblocks for ease of experimenting. The short endcap at the bottom right in the photo is to route the leads through to direct connection. The socketry is in the chassis plate just to fill the holes... Test and Measurement A full three stage test sequence is described. First, every junction has a resistance measurement given and taken. Then the tubes are installed and amp is switched on for the first time. If the filaments light up and nothing else does, the voltage is checked at every junction. If the voltages check out, plug in a speaker and a pre-amp and listen up! With the negative feedback loop optimized for 16 ohms, the gain of the amp is about 12.5dB. Frequency response with 10dB of feedback looks good, with 20Hz only 1.2dB down, reference 1kHz, and -3db at nearly 16Hz. THD with a 1kHz input is under 2% at 1W output, and under 4% at 2W, the rated power. That's pretty much what is claimed and, with third harmonic distortion at least 30dB below the second, it is no surprise that a charming, clean sound results on appropriately sensitive speakers. Sensitivity for the full 2W output is 2V input. High frequency response with the standard output transformer depends on the negative feedback loop, so the instructions should be followed carefully. Experiments with zero negative feedback operation via the standard output transformer were not encouraging for those who take measurements seriously, with response tailing off under 10kHz. Load is less sensitive and we found that attaching 8 and 4 ohm speakers to S.E.X amps set up for 16ohm operation and negative feedback operation created a very good sound despite restricting the high frequency to about 13kHz. With none of the setups we tried did we see any instability like motor-boating. SE amps are not really about measurements. They are an acquired taste in sound, like the acquired taste for malt whisky, both of which require practice. First, one needs suitable speakers. The S.E.X. Kit Speaker Ultra-fi operates on a series of exclusionary acronyms. They come down to all-triode, no negative feedback, low-power amps, with the hidden proviso: driving horns. Horns are huge, expensive and have zero spouse approval. But even lesser speakers of enough sensitivity to multiply 2W to realistic volumes are few and far between and very expensive. This speaker limitation puts single-ended amplifier kits of any price out of bounds for many. The S.E.X. amps were themselves designed to power Lowther drivers in horn enclosures. The cheapest Lowther drivers cost precisely twice as much per pair as a pair of S.E.X. monoblocs and the woodwork for the horns will at least square that difference. Thus the availability of the cheap speaker option to the S.E.X. kit by itself makes it unfair competition for all other triode single-ended amp kits, regardless of the fact that the whole thing, including speakers, is one third the price of the next cheapest competitor. The speaker option costs $30 and provides four 5in MCM spun aluminium drivers with Fs=3D115Hz. These speakers are definitely worth having, and worthwhile in their own right. They extend the usability of the S.E.X. monoblocs tremendously for everyone but especially for experimenters who can wire them in series, parallel and series parallel for a variety of sensitivities and efficiencies. Plans are provided to mount the four speakers in pairs in the centre of a flat baffle 2x4ft. I was not impressed by this scheme, as Olson showed a long time ago that any driver on a flat baffle is subject to less frequency cancellation, and therefore offers greater output, if mounted off-centre. With two drivers, it becomes more difficult to mount the drivers so that no distance to the edge is a multiple of any other, but the difficulty is aesthetic rather than geometric. Start with a position one driver width from the side and two driver widths from the top, and the second driver in vertical line down and as close as possible to the first driver. If it looks okay by eye, do it. If it is not pleasing, move the drivers and measure to the edges: distances must not be multiples of each other. Splits into elevenths, as in 4/11 and 7/11, are good. Having decided on a wooden case for the amp and intending to have the baffle made at the same later time, I needed some way of listening to the supplied speakers in a hurry. From the garage I fetched several carpet tubes awaiting the trashmen's attentions. The speakers fitted nearly perfectly into the top and were taped in with common parcel tape. The tubes are 78in long and 5in diameter. They are placed with the speaker upwards, firing towards the roof, and leaning against the wall at an angle to leave the bottom open. No stuffing is used: the tubes are completely unobstructed. Listening When I first heard about single-ended amps at $399 for a pair of monoblocs, $430 with speaker drivers, I thought it was a scam. From experience designing tube amps for others to manufacture, I know what parts can cost. You should be aware that the S.E.X. in this test enters a hostile environment. The nearest cheap kit amp we have is 250% the S.E.X. Kit's price, and 50 times its output; the nearest in type of sound is even more expensive and has near enough 20W of triode output. Our cheapest CD player is a Quad. Our cheapest speakers cost $1600 in 1979 and our standard reference is Quad ESL-63. We conducted the S.E.X. listening test with a Quad 34 pre-amp which costs more than a fully built S.E.X. It has in the past happened that makers of an amp, intimidated by the company they were invited to keep, declined to have it reviewed here. Now, after listening, I have changed my mind. The S.E.X. is not a scam, it is a bargain. And it is not embarrassed to stand up in the very best company. The sound of the S.E.X. on the electrostats is good on voice and music. On 2W into a 85dB/2.83V speaker one must not expect a symphonic crescendo to be realistic, especially not on test records with real explosions over the open Thames! The voice test is especially difficult for low-powered amps but the S.E.X passes with flying colors for articulation, clarity and timbre. With elderly Bang & Olufsen speakers of about 90dB sensitivity but 4 ohm impedance, the sound is rivetting, including the explosions. So far, we have merely compared familiar reference points. We now know that with known quality speakers, the S.E.X. monoblocs will give fine sound. The surprise came when I hooked them up to the piped optional speakers, wired in series to give 16 ohm impedance for the best match to the standard output transformer. Of course, the nitpickers will complain that I have enormously enhanced the bass by using a pipe so much longer than the quarterwave. But on my kind of music--early music, single voices and choirs, small groups--these pipes are just right. For instance, on gregorian chant the ambience is awesome. The S.E.X. monoblocs recover the detail, and the optional speakers present it. All right, I am not going to swap my courteous, correct ESL for cheap little aluminium speakers, but their raw power is attractive. In tests with John Gielgud and Kenneth Brannagh's voices, and some cat recordings of my own, the actors fooled my family, and the cat my cats, who arrived hackled for battle. These little speakers were originally intended merely to give people a taste of single-ended reproduction while they decide whether to lash out for a horn to use with their S.E.X. amps. Personally, I see no need to move on to the masochism of horns... I'd rather double up on a baffled or piped array of the little spun speakers. Four in series-parallel for instance would offer sensitivity of 96dB. There is an option to monobloc two monoblocs, so that with two pairs of S.E.X. amps one could put out 4W into 8 or 16 ohms for use with less sensitive speakers. I don't see the point, unless one intends buying the Superwhamodyne speakers readymade or in kit form from Electronic Tonalities for use in quite large rooms or for party-time rock volumes. They are essentially the cheap baffle idea extended and enclosed and with a woofer added for greater SPL. On measurement we found that the S.E.X. ran out of steam, depending on feedback and to a lesser extent on output load, somewhere between 13kHz and 16.3kHz, but my fifty-year-old ears do not detect a treble deficiency when the amps were used with any of the available speakers. The bass on the pipes was so much more impressive than on any of the other speakers that I am investigating ways of making the piped installation permanent! The ear has an oddity, in that if the HF is cut down, then to make the reproduction sound right the bass must also be restricted to balance the sound. I never felt that this point had arrived even with zero negative feedback arrangements. The optional speakers are thus a perfect match for the amp. Tweaking Of course, any single-ended amp is near the top of the envy-class, but those with feedback are not quite in the top 1%. Thus the major upgrade to the S.E.X. that any owner can perform is to fit a better output transformer, which permits less negative feedback or even zero negative feedback operation. Almost any purpose-built single-ended audio output transformer with a primary around 3Kohm and capable of handling 60mA will provide an improvement. If you want to upgrade the perfectly good standard transformers, look into the Tango 808 or any of the smaller Audio Note SE transformers intended for 300B amps. Another transformer which will fit and is well spoken of, though I have not heard it myself, is the UBT-1. Any of these transformers should permit running with zero negative feedback, thus putting the S.E.X. amp firmly into the prestige stakes. Without the negative feedback the amp has 21dB of gain into 16 ohm, so be careful of sensitive speakers and even more sensitive ears when first firing it up! Given a suitable transformer, the S.E.X. with zero feedback measures -3dB at 28Hz, -1dB at well over 16kHz, and 5%THD at full power of 2.9W; figures taken at 1kHz into 16 ohm. 3rd harmonic is 26dB down from 2nd. With 2.5dB feedback, 4.5W at 5%THD and 22Hz-25kHz is available. You can thus spend extra money on fancier transformers for the audiophile street cred of zero feedback, or use less feedback and still enjoy extended bandwidth. Either way you also get much more usable power, which has a beneficial influence on the reproduction of dynamics and transients, the exact level and nature of the benefit depending on your preferred music. For early and vocal music, I would lose the feedback. For symphonic music and rock I would keep the feedback and enjoy the extra power. To test the expandability and likely ultimate quality of the amp, and to discover the natural limits of the 6DN7, I replaced all the iron with very upmarket Lundahl double chokes and mains, output and interstage transformers. All of this had to be outboarded because the Lundahl outputs, for example, are more than four times the size of the standard transformers. This arrangement was tested side by side with a choke-coupled Real McCoy Type 33 300B. Progressive optimization of this S.E.X. setup for the greater capability of the Lundahl iron, with tube rectification, and uprated polypropylene capacitors, Holco resistors and gold connectors substituted where appropriate, enabled us eventually to listen to the gods from some little distance, whereas the expensive 300B (with an even higher component spec) put us among them on equal terms. Since no one in his right mind will add over a thousand dollars' worth of transformers to a $399 amp, I didn't bother measuring what was obviously a vastly superior performance on listening tests. A fairer test but still not quite fair, would be against the Audio Note Kit One at about five times the price of the little Electronic Tonalities amp. Against my recent memory of the dark tone of the ANK1, the S.E.X. with its 6DN7 sounded infinitely more cheerful. I was impressed with the versatility of the S.E.X. as a workbench for experimentation and for a quick test of components destined for a more expensive amplifier. Conclusion? The experience with the Lundahls convinced me that after adding a better output transformer than the standard one, the biggest single improvement the experimenter can make to a S.E.X. is to add an interstage transformer between the driver and power stages. Remember when first testing this modification that the interstage transformer adds another a third to the gain of the amp. Take care! All other components are replaceable. In order of decreasing marginal value, including aesthetic value, I would order the lesser improvements as: ceramic sockets, gold input and output socketry, a consistent paint job, a lacquered case--and only then would I start increasing the added value of the sound beyond the replacement OPT and interstage transformer coupling. Again in order: polypropylene caps, preferably uprated, to replace the electrolytics, fancy signal caps, trick resistors (carbon composites may have synergy), silver wire and solder certainly for the signal chain and perhaps elsewhere as well. The instructions conclude with six pages of suggested tweaks the user can perform, in which not even all the above is included, but many, many others find a place, from the trivial to the possibly earthshaking. There is also half a page of possible tweaks on the S=2EE.X. Kit Speakers, not including my pet carpet tube pipes. Recommendation If you have ever built an amp from a kit, or even thought about it, or wondered about single-ended tube sound, you should build this one. Construction is easy, backup is available, it works well; the speaker option makes it instantly usable. For this price you can't go wrong. It is a good amp at any price, even if you just build it to listen to. As an amp for budget-conscious single-ended experimenters, it is without peer or competition. Highly recommended. October 1998: The S.E.X. amp kit has been upgraded to parafeed and many of the components changed, though the price remains the same. Information from http://www.bottlehead.com =A9 Andr=E9 Jute 1997, 1998, 1999 Andr=E9 Jute was educated at universities in South Africa, Australia and the United States in psychology, economics and business management. He has worked in advertising, management consulting, and as a political and military advisor, and is now chairman of the graphic design house Communication Jute and the industrial design firm Real McCoy. He is a distinguished novelist and writes a column on classical music read every week by 9.2m music lovers. Before taking up tube amp design, for his hobby he designed and built complete automobiles.=20 ************ |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Re KISS 123 by Andre Jute: Why the KISS 300B is ZNFB | Vacuum Tubes | |||
Re KISS 123 by Andre Jute | Vacuum Tubes | |||
KISS 114 by Andre Jute | Vacuum Tubes | |||
KISS 113 by Andre Jute | Vacuum Tubes | |||
Is Andre Jute real ?? | Vacuum Tubes |