SissySIT (bilo SufiSIT)

[Mikorist]

 Cinemag. 

[Woland]

@Mikorist za 7 dana su meni stigli

[ninja]

Da ne bude samo Miko što kači 

https://www.lennartlabs.se/#!/projects/articles/blnemesis

Spoiler

 

 

[Zen Mod]

Artikeln är endast tillgänglig på engelska

[mutayu]

Det er fullt av ål i luftputebåten min!

[Mikorist]

evo teksta

 

When transistorized audio amplifiers appeared in the late 1950s, most of them mimicked traditional tube amplifier circuits. Transistor radios of the time invariably used push-pull transistors between transformers in their output stages. Very soon, however, new transistor amplifier circuits without the transformers were developed and transistor circuitry took on a new route. This was only natural, since the much lower voltages and higher currents offered new opportunities.

Now, several decades later, Nelson Pass presents his PASS-SIT-1. I find this transistor very intriguing, since its electrical characteristics are very similar to those of power triodes. What if it also is able to deliver sonic characteristics similar to those of the power triode? Mr. Pass tried his PASS-SIT-1 in the classic, very simple, Nemesis circuit and presented some really encouraging results. He also suggested further exploration.

 

Introduction 

SITs (static induction transistors) have been around for quite some time. The SIT technology appears to have been developed in the 1950s at Tohoku University, Japan, and SIT FETs have been used in some commercial amplifiers. The SIT is a depletion mode FET with drain characteristics similar to those of the triode. High input impedance, very low drain resistance, yet high transconductance and power capacity, invite to the construction of very simple power amplifiers not depending on negative feedback.

Single-stage power amplifiers based on solid-state devices are rare, if not very rare. A few circuits have been published though, Mr. Pass' Zen circuit and subsequent variations possibly being among the most widespread and appreciated.

None of these circuits uses an output transformer, and for obvious reasons. Output transformers are bulky, heavy, and expensive. Secondly, solid state devices operate at significantly lower voltages and higher currents than most vacuum tubes. Consequently, the low frequency range is likely to be compromised by low transformer primary winding inductance. Besides, transformers appear not to be overly well recognized by the wider audiophile community.

In two previous Linear Audio articles, Mr. Pass briefly presents Jean Hiraga's one-transistor power amplifier design "Nemesis" [1], [2]. In the former article he explores the use of various MOSFET and JFET devices in this indeed very simple circuit. In the latter he introduces the use of the PASS-SIT-1 and presents some interesting findings. He also suggests further exploration using a larger output transformer with better frequency response than those he had available at the time.

This article will further explore the use of the extraordinary PASS-SIT-1 device in the Nemesis circuit. The output transformers used are most kindly provided by Jan Didden. A few design considerations and an evaluation circuit will be presented, along with some measurements to verify its performance. To conclude the evaluation, and to hopefully contribute to the experience with simple amplifier circuits, the results of a few listening tests might be helpful.

 

Why the SIT Nemesis? 

The traditional single-ended triode (SET) amplifier circuit, developed in the early 20th century and still in service, is well recognized by many for its ability to reproduce music with very pleasing subjective qualities. Generally, the SET amplifier merits have been subject to debate for many years, and it appears that subjective experience and objective performance are not always easy to match. It should be safe to assume that the last word on the matter has yet to be spoken.

Jean Hiraga presents, in his two Nemesis articles [3], [4], the merits of simple single-ended triode circuits and, in particular, refers to the classic Western Electric WE 25B amplifier. Based on the very same philosophy, he develops the Nemesis circuit to feature a common-source mode power FET in place of the vacuum tube. The Nemesis amplifier is reported to present a subjectively very high sound quality with highly 'natural' properties. Even in mono, on a single speaker, the sound is reported to give the impression of depth both in front of and behind the speaker. In stereo, especially the piano is presented with uncommon acuity, and very life-like impressions of the felt hammers and the pedals. The stereo presentation also reveals the presence of the room. Let me remind: this was back in 1985.

Figure 1: The Pass Labs PASS-SIT-1

 

The electrical characteristics of the PASS-SIT-1 (Figure 1) offer new opportunities. The current versus voltage properties are similar to those of the triode and include low drain resistance and high transconductance. I have designed SET amplifiers for many years and thus considers this development with extraordinary interest and curiosity. What if the PASS-SIT-1 can deliver listening experiences similar to those of the SET?

The high voltages and low currents of the SET circuits are not very useful for driving loudspeaker voice coils. This is why an output transformer is almost universally used. Transistor circuits, on the other hand, are fully capable of driving loudspeakers without the output transformer. To keep this evaluation in line with the intentions of Messrs. Pass and Hiraga, however, and to hopefully contribute some experience in relation to other PASS-SIT-1 circuits (published elsewhere), the output transformer is retained.

 

Design considerations 

This section describes in brief the attempted operational design and performance of the evaluation amplifier. The treatment is similar to most textbooks on SET design, although the more detailed design procedure is outside the scope of this article.

The PASS-SIT-1 static operating point should be located in the most linear portion of its operating area. Ideally, the operating point should be chosen to give the highest output power possible, at the lowest distortion possible, with the highest efficiency possible, and all this without exceeding any of the maximum rate specifications. These are of course conflicting requirements, and, as always in this business, compromise is required.

Single-ended triode amplifiers are capable of producing a very good perceived sound quality. One of the most important, if not the most important, components is the output transformer. Interfacing the output power device to the loudspeaker without degrading amplifier performance or sound quality is a tall order. Thus, the demands on the output transformer are high and the transformer should be selected carefully. For this evaluation, however, the transformer choice is already made; the Nemesis transformers most kindly provided by Mr. Didden.

One sometimes overlooked fact with single-ended amplifiers, is that the power supply is an integrated part of the signal path. The current loop in the output power device and the transformer primary winding inevitably has its return path via the power supply. Thus, power supply properties are as important as those of the output power device and the transformer.

 

Transistor data 

This is where we need the data sheets. Since the PASS-SIT-1 is not yet in production, and data sheets are not available, we have to find our own data. Figure 2 depicts the operating area, the ID vs. VDS diagram, for one of the individual sample transistors.

Figure 2: ID vs. VDS for VGS between -7 and -10 V as measured in one of the individual samples

 

It should be noted that these measurements are not taken under controlled conditions, such as constant temperature, rather at the bench top at reasonable thermal equilibrium on the heat sink. Thus, this diagram should not be considered representative for the PASS-SIT-1 transistor in general.

As already pointed out in [2] this diagram appears very similar to the plate diagram of power triodes such as the 300B. If these properties, in a circuit similar to the SET circuit, can also produce a similar sound quality remains to be determined.

 

Output transformer data 

The output transformers are shown in Figure 3. Their weight exceeds 6.6 kg each. The exact origin of the transformers is unknown to me, although it has been suggested they were wound by Tango. They are intended for use in single-ended power amplifier circuits, has a center-tapped primary winding and two secondary windings.

Figure 3: The output transformer

 

Many parameters affect transformer performance, the most obvious being losses (core losses and winding resistance) and frequency range determining factors (inductance, leakage inductance and winding capacitance). The low frequency range of an output transformer is primarily governed by primary winding inductance and the internal resistance of the output power device. The high frequency performance is governed mainly by the leakage inductance, but also (to a lesser degree) winding capacitance.

Since the primary winding in a single-ended output circuit will carry direct current, the core will be pre-magnetized. To avoid core saturation by signal currents, a gap is introduced in the core. Unfortunately, this gap also reduces winding inductance. Some measured transformer data are listed in Table 1.

Table 1: Some output transformer design data. Inductance was measured at 50 Hz.

Circuit diagram 

Since the PASS-SIT-1 is a depletion mode device, its idle current should be controlled by a gate voltage which is negative with respect to the source. This voltage can be provided by source resistor voltage drop (self bias) or by a separate voltage supply (fixed bias). Both schemes will be evaluated and are shown in Figures 4a and 4b, respectively.

Figure 4b: Amplifier circuit using self bias

 

Figure 4b: Amplifier circuit using fixed bias

 

The simplicity of the original Nemesis circuit is striking, and this circuit is even simpler. Besides the transistor and the transformer very few components are required. Alas, since this is a single-ended circuit, power supply quality is of major importance.

 

Operating point 

The selection of a suitable operating point requires careful consideration. As mentioned above, we are facing conflicting requirements. The most important compromise would be the tradeoff between output power and distortion. Since the transformer is already provided, its primary load 64 ohms will be used in this evaluation.

Although FETs in the TO243 case normally are good for higher power levels Mr. Pass recommends we keep PASS-SIT-1 dissipation below 40 watts (just like the 300B, as it happens). FETs are normally more linear in the higher drain current range, but since I didn't want to run one or more Amperes through the transformer primary winding I turned in the higher voltage, lower current direction. Measurements also confirm that the device under test appears more linear in the higher voltage region.

Figure 5 shows a sample operating point and a static load line in the ID vs. VDS diagram.

Figure 5: Sample operating point in the ID vs. VDS diagram and a static load line. Quiescent current is 600 mA at 60 V drain to source (circle). The static load impedance is 64 ohms.

 

To investigate amplifier performance at a few different operating points, three operating points were selected and tried. Important transistor characteristics at these operating points are detailed in Table 2.

Table 2: Operating point data and transistor characteristics.

 

The very low internal drain resistance is of great interest, since it helps create an acceptable low frequency range in a transformer with somewhat limited primary inductance. The high voltage amplification factor is in itself almost a prerequisite for a one-stage power amplifier.

Output power and distortion, self bias 

A test circuit was set up for electrical evaluation and listening tests. The circuit was wired point-to-point on a couple of bookshelf boards. This route is fast and easy, and allows for quick changes.

Figure 6 shows harmonic distortion (THD) of a 1 kHz signal versus output power.

Figure 6: THD of 1 kHz versus output power in the three different operating points

 

Theoretically, the quiescent 500 mA will limit current swing to an output power of 8 watts. 600 mA and 700 mA will limit output power to 11 and 16 watts, respectively. These limitations are apparent in the diagram. Onset of clipping is asymmetrical and very gentle, and as with any good SET amp, second order THD thus generated is benign to our hearing.

The diagram confirms that the PASS-SIT-1 appears more linear at higher VDS. As usual in SET design, however, the tradeoff between low THD and high output power is also quite obvious.

 

Output power and distortion, fixed bias 

The above measurements were repeated with the test circuit using fixed bias. Compared to the auto-bias scheme there is only a very minor reduction in THD and no increase in output power.

 

Choice of operating point 

Increasing VDS above 70 volts with a 64 ohm load will result in dissipation above 40 watts during positive signal half cycles. While this may be entirely acceptable (lower dissipation during negative half cycles will keep average dissipation slightly below 40 watts) I'd recommend against setting reliability at risk.

If load line slope can be made smaller (lighter load) it will keep dissipation below 40 watts also when moving VDS to even higher voltages. This may induce other problems, however. I repeated the above measurements with a 128 ohm load and the outcome was immediate; maximum output power doubles and THD quite more than doubles - the cut off region is reached way too quickly.

For further evaluation an operating point close to 60 volts, 600 mA and a 64 ohm primary load (4 ohm secondary windings) was chosen. This choice gives a 10 watt amplifier with promising data.

 

Frequency response 

The frequency response of the SIT Nemesis is strongly governed by the output transformer. A transformer wound for currents in the Ampere range will use the corresponding wire gauges, effectively lowering the number of turns that can be practically fitted on the bobbin. This obviously limits winding inductance.

As mentioned above, the low frequency range is primarily governed by the primary winding inductance. Despite the low inductance of this output transformer the resulting low frequency response is quite respectable. The frequency response at 100 mV input signal is shown in Figure 7.

The low drain resistance of the PASS-SIT-1 is indeed very helpful in this case, but a transformer with some more turns of wire will further extend the low end. The high frequency response is very good thanks to the low number of winding sections and layers.

Figure 7: Frequency response at 100 mV input. The - 3 dB points are 27.1 Hz and 60.6 kHz.

Listening impressions 

The SIT Nemesis amplifier is listened to by myself and two friends, both experienced with SET amplifiers. Various music is played and special attention is given to the presentation of voices and instruments, the depth and spaciousness of the sound stage and, of course, to the musicality and ability to engage. During evaluation a pair of relatively large, high-sensitivity (96 dB/W/m), 3-way, loudspeakers are used.

The first thing that really strikes is about resolution, contour and dynamics. We unanimously comment on the immediate attack, the liveliness of transients, and the extraordinary dynamics. The transients of string instruments such as the guitar, the cello, and the piano are presented with stunning realism. Energy distribution is excellent. If a singer in a band suddenly increases the strength of her voice, only her voice releases that energy, nothing else.

We can clearly see the individual instruments. A voice and a piano are presented as just that, a voice and a piano - not a sum of the two. The bass and the kick-drum are physically separated also on the sound stage (provided they are separated on the recording, of course). The separation of voices and instruments is very convincing and easily conveys the impression of individual musicians performing together.

Many of the recordings we listened to have many small sounds generated by the artists, the instruments, and the background; they are reproduced with surprising clarity. The concept of Downward Dynamic Range, as introduced by late Allen Wright, captures the requirement of an amplifier to allow us to hear low level sounds while louder ones are playing. The SIT Nemesis really does this and the sense of 'live' is fascinating.

The sound stage is very three-dimensional. Its width easily extends outside the speakers with no tendency of splitting up. Depth is presented seemingly limitless. Interestingly, it's also easy to see instruments at different heights; a guitar at waist height and a trumpet at mouth height. Not all recordings contain this height information, I guess it depends on the production process. A contrabass is difficult to record due of its sheer size, but a recording done right presents the contrabass as a two-meter tall instrument.

Airiness of the sound stage and (recorded) reverberation are impressive. The sound stage is perceived as 'big' and the presence of the venue very realistic. I know from previous experience that these properties are quite difficult to preserve in amplifier design, but the SIT Nemesis does a very good job here.

There is no doubt during listening that the music reproduction is engaging. None of us is completely still during listening. Although a strictly subjective experience, it's easy to engage in the music to the extent that the setup actually "disappears".

 

Concluding remarks 

The PASS-SIT-1 is nothing less than an extraordinary device. The triode-like characteristics, low internal resistance, high amplification factor, and distortion signature offered by the SIT Nemesis circuit makes it a very competent performer for anyone interested in a solid-state version of single-ended amplifier qualities.

Despite the electrical similarities between the PASS-SIT-1 and the power triode, the evaluated circuit does not sound like a triode amplifier. There's a slight lack of body and warmth when compared to a serious SE 300B amplifier. The transient and dynamic abilities, on the other hand, appear to be more pronounced than with the SE 300B.

Following the above listening impressions, however, it brings most of the SET qualities. I tend to believe that the output transformer, in combination with the distortion signature of the PASS-SIT-1, plays a vital role in achieving this. Much has been said about transformer performance in audio amplification, but its contribution to subjectively pleasing listening should not be underestimated.

Nor does it sound like a modern (push-pull, high power, NFB) transistor amplifier. Its character appears to constitute a blend of single-ended triode qualities with qualities of modern solid-state amplifiers. Considering the classic circuit and its musical potentials, however, the SIT Nemesis performs more closely to the SET than any of the very few one-stage, solid-state, non-feedback amplifiers I have previously come across.

No intermodulation, clipping, frequency range, or other distortion effects were detected during listening sessions.

Unfortunately, the PASS-SIT-1 cannot yet be bought in the local store. But apart from this, the major concern will be the sourcing of a suitable output transformer. Its size, weight, and cost is likely to limit the number of future owners, but anyone lucky enough to listen to the SIT Nemesis will find it truly amazing!

 

Acknowledgements 

I owe special debts to Nelson Pass for providing the PASS-SIT-1 transistors, Jan Didden for providing the output transformers, Anders Björklund and Eive Skoog for invaluable support and criticism in listening evaluations and discussions.

 

References 

[1]  Nelson Pass; "The Arch Nemesis"; Linear Audio, Vol. 0, September 2010; pp. 20-30.

[2]  Nelson Pass; "SIT Nemesis!"; Linear Audio, Vol. 1, September 2011; pp.141-148.

[3]  Jean Hiraga; "L'amplificateur Némésis ou l'hommage au WE 25 B"; L'Audiophile, January 1985.

[4]  Jean Hiraga; "L'amplificateur Némésis, 2. Montage et mise au point"; L'Audiophile, spring 1985.

[vladd]

To je jedna faza u razumevanju sit-a, i jednostavno mora da se luta tom sumom. Trziste je narezano sa SET-om, i treba se nasloniti na vec postojece duboreze.

Ono sto je moja primedba, principijelno, za SIT je trafo potreba od 0 do beskonacno.

Druga primedba bi bila da ne treba SIT da je "zamena" za lampe, nego da od karakteristika lampe uzme ono najbolje, a moze, a to je paleta harmonika, tip klipinga, a da izbaci previsoke napone, jaka polja, ponasanje elektroda u pregrejanoj zoni plazme, i naravno, izbacivanje igrarija oko grejanja, variranja i geometrija i desavanja usled visokih temperatura, koje je tesko odvesti izvan...

"Zamena" za lampe u vidu fetrona je davno napravljena, i neinteresantna..prevazilaze se neki drugi problemi lampi...

Covek spomenu "ozbiljan sistem sa 300B", a to je retka ptica u svetu hife, veoma retka, posto je i 300B izuzetno neobicna i unikatna zivuljka, daleko drugacija od drugih lampi, sa ozbiljnim zahtevnostima i komplikacijama u drajvovanju, koje, ako nije akurentno, moze da ubrlja sve njene inicijalne prednosti.

Butikasenje i neke konfiguracije sitnih lampi preslikano na specificnu 300B su prosto promasaj, njoj treba 80-90V dobrog, cistog i brzog svinga za drajvovanje, pa ako se to obezbedi kako treba postignut je pun pogodak.

Trafo, za ovu situaciju iz teksta bi trebao da bude minimalno 4x veci, u vecem zasicenju, sto znaci znacajno vise gvojza(i to hibridnog, i niklenih zona i gvojzenih paketa), znacajno vise bakra i nesto vise procepa, cak uz mogucnost varijabilnog procepa, kao ideja for the record.

Ili izbaciti trafo, kao najlakse, ne udvarati se razmazenom trzistu(sic! , od kojego se zivi ) i uraditi SIT da radi kako moze... 

[ninja]

Evo, ako hoće neko origi, sa zapaprenom, preprodavackom cijenom, a može ustediti slanje

https://theartofsound.net/forum/showthread.php?61272-FS-First-Watt-SIT-3-amp

[vladd]

Uf, koliko je tu bazicne frivolnosti, da moras i da pljunes po spravi koju prodajes.  

[Mikorist]

Pa dobro. Nije mnogo kalirao. Cirka 1500 EUR 

[Zen Mod]

on da kalira ?

jok sine , nego je 570 podigao cenu , ako je kao sto mislim da ga je uzeo ovde

[Zen Mod]

uostalom ,nije ni bitno

nisam nikad nikome gledao u budjelar , necu ni njemu

svaka mu sreca pjevala , al' Graal-a nema

 

pa nema

[neskor]

Covek samo shvatio/cuo da zvuk lampe ipak nije tako lako zameniti

 

[Zen Mod]

@neskor

ne zalazim ni u to , svak ima svoj ukus

jedino - grese oni koji imaju preferencu na cevi , iz bilo kog razloga , ako misle da ista moze zameniti cev

i obrnuto - ko ima preferencu za SS (iz bilo kog razloga) , ne postoji lampas koji to moze da zameni

 

[Mikorist]

Lampe su za mazohiste

[vladd]

Ne treba ga menjati, doticni rad lampe, vec prevazici, nadmasiti.  

Prevazici ga po realnim i objektivnim kriterijumima. Od neophodnog damping faktora, do izbegavanja izmene interne geometrije lampe visokim temperaturama.

A zadrzati paletu harmonika, radnu krivu i ponasanje ampa prema kompleksnom opterecenju.

Ali, tako je to kada se postave pogresni postulati, i nisko brdo za penjanje.  

Nije on ni prvi ni poslednji koji zbog neznanja, ili pogresnog saznanja, napravi gresku, ali mu realno nije bitna.

Mada, praksa je pokazala, i silna osluskivanja, da ljudi ipak vole, da nazovemo ofarbani, ili intrigantan zvuk, svako na svoj nacin, apsolutno legitimno, a jedino je nelegitimno postavljati sebe u centar sveta objektivnosti i daviti ljude nametanjem licnog ukusa i subjektivnog kriterijuma.

[Zen Mod]

Miko , budi ozbiljan 3 minute , dnevno , ako je ikako moguce

[Mikorist]

2 minutes ago, Zen Mod said:

Miko , budi ozbiljan 3 minute , dnevno , ako je ikako moguce

pa brate. još mi fali da glancam lampe 

[vladd]

Sto bi ih glancao, postavi ih pod haubu, ne moraju da budu kao svecnjaci...  

Lampe su poseban i lep svet.

[Mikorist]

27 minutes ago, vladd said:

poseban i lep svet.

Možda si u pravu, a možda i nisi.

Mnogo puta mi se dogodilo da pošlusam veoma hvaljene i ponekada ludački skupe lampaše i zvučnike i ostanem mrtav hladan posle.

Recimo kao posle ovog fi fi sajma.  Ništa posebno. Sećam se samo slušalica....

[neskor]

taj gore pomenuti SIT-3 je svirao/gostovao u par vrhunskih lampaskih sistema i nazalost ni u jednom nije bio bolji

@vladd ljude ne zanima ni kriva, ni radna tacka, ni lucidnost dizajnera, nego krajni rezultat - zvuk

 

 

[Mikorist]

2 minutes ago, neskor said:

krajni rezultat - zvuk

Najbolje se čula juče ona harfa uživo.