Taken from Amateur Tape Recording, September 1965

Although this amplifier is not one recently released, it is definitely worth looking at in view of the expanding interest in transistorized audio equipment and the fact that the amplifier has passed through several stages of small internal modification since its inception at the beginning of last year.

In spite of its remarkably small dimensions of 11¼ in x 3¼ in x 8½ in, with a 12 in x 4 in front panel, this stereo amplifier is capable of delivering a healthy two times 15 watts into loads ranging from 15 ohms down to about 4 ohms at a total harmonic distortion below 1 per cent. Indeed, the amplifier is smaller than some stereo control units of the valve variety!

Small size, of course, is one of the advantages of transistor equipment, and just how much smaller solid-state equipment can be relative to comparable valve equipment can be realized by working out the dimensions of two 15 watt valve channels complete with stereo control unit, mains power pack and so forth and then setting them down by the side of the dimensions of the Pye equipment.

Transistors also have other advantages over valves. They are much more efficient than valves because they demand no heater power and because no power is lost in heating an anode. Hum problems in high gain circuits are very much reduced and microphony - a problem with valves running at high gain-is virtually eliminated.

Probably one of the biggest advantages hi-fi-wise is the elimination of transformers. Like all recent hi-fi transistor audio amplifiers, the Pye HFS30T has neither an output transformer nor a driver transformer. Hence, one of the major quality determining components - the output transformer - of the valve amplifier is removed.

Nevertheless, valves do have certain advantages over transistors. On the whole, valves are more robust. They can deal with overloads and peaky transients better than transistors. They give fewer design problems concerning circuit high-frequency performance. However, these things should, not be taken to imply transistor shortcomings. They are inherent characteristics, as compared with the valve, and they are taken fully into account in the design of transistor circuits and networks.

The HFS30T features a pair of printed circuit board amplifiers mounted flat one above the other. The bass, treble and volume controls are of the two-gang variety. These three controls, along with the selector switch and balance control are nicely arranged over the left-hand two-thirds of the front panel. The remaining third on the right-hand side accommodates the various switches. Here we have the on/off switch, a loudness switch and switches for rumble and scratch filters. Also located in this area is the pilot bulb, indicating whether or not the amplifier is energized.
Each amplifier channel incorporates nine transistors and one semiconductor diode, the circuit of a single channel being shown in Fig.2. It may be noticed that, like some other transistor amplifiers, the Pye is based on a Wireless World design by Tobey and Dinsdale (October 1961).

The first three transistors operate primarily as pre-amplifiers. The first gives RIAA pick-up correction by switched frequency-selective negative feedback between the collector and base circuits. The second transistor gives variable bass and treble control in a 'Baxendall' tone control network, while the third transistor is associated with the filters. The power amplifier section starts with a voltage and driver amplifier VT4 and VT5, along with the diode to assist with thermal stability, the output of which drives the low power complementary pair (VT7 p-n-p and VT6 n-p-n) which in turn drives the push/pull output transistors, VT8 and VT9, both p-n-p types. It will be noted that the six transistors of the power amplifier section are all directly coupled (no coupling capacitors). This technique leads to good thermal stability and very small phase shift.

The amplifier features three major feedback loops, one to balance the output impedance, another to correct transistor phase shift and the third is the ordinary loop which controls overall gain and distortion. Local frequency-selective feedback loops are also used in the equalizing and tone control circuits.

The main supply section is incredibly small when it is considered that the total audio power exceeds 30 watts at full drive. A small mains transformer works into a bridge-type rectifier system and a 2,000 µF electrolytic capacitor eliminates all signs of ripple.

The amplifier consumes only about 25 mA quiescent, but on peak drive signals the current may rise to around the half-ampere mark. Each channel is protected power-wise by a 1 A fuse. The ht line voltage is 48 volts negative.

As will be seen, the amplifier has provision for piezo and magnetic pick-ups, for radio and for an auxiliary signal input. With the test amplifier, the author's Deram pick-up was found to work best into the magnetic Input. Here the input impedance is sufficiently low to endow the Deram with velocity characteristics, the output of which is then approximately equal to a magnetic pick-up. Equalization under these conditions was handled admirably by the magnetic equalizing feedback loop, referred to earlier.

The input sensitivity of the earlier Mark I version was 7 mV on magnetic pick-up, but the latest models have various component value changes which put the overall gain of the amplifier up, resulting in a magnetic pick-up sensitivity of 2½ mV (for 15 watts output). This sort of sensitivity would almost allow the direct connection of a suitable impedance tape head, but then, of course, some modification would have to be made to the equalization for tape, as distinct from disc records.

Sensitivities at the other inputs are now piezo pick-up 70 mV, radio 35 mV and auxiliary 35 mV, compared with the early model sensitivities which were respectively 120 mV, 100 mV and 100 mV. The modifications leading to the increase in sensitivity have resulted in an improvement in the overall performance of the amplifier and a slight extension of the pre-amplifier frequency response.

Readers with the early Mark I amplifier can have the current modifications incorporated by a Pye dealer. These consist mainly of a change in value of five resistors.

Tape facilities
For tape enthusiasts, the amplifier has a tape record output delivering 100 mV at the sensitivities given above. This socket provides a 'record' programme signal, while the 'auxiliary' input socket will accept the signal present at the 'monitor' socket of a tape recorder for replay at hi-fi.

One outstanding feature of the amplifier is the very low background level. Hum is virtually non-existent, and hiss is many times below the level of the programme signals. Indeed, the power amplifier section has a signal/noise ratio in the order of -90 dB, while under the worse pre-amplifier condition, namely magnetic pick-up input, the ratio is still -60dB.

There is virtually no breakthrough from one channel to the other from a full-gain listening test, and the crosstalk figure was found to be around 400 times down - measured at 1,000 c/s.

The rumble and scratch filters are preset to take over at about 100 c/s and 3,500 c/s respectively (at 3dB points), the characteristics of these relative to the normal amplifier response being shown in Flg.3.

Another interesting feature of the amplifier is the loudness compensation filter. With this switched in, a maximum boost in the order of 15dB occurs at around 30 c/s as shown In Fig.4. This bass boost effect increases as the volume control is retarded, and its purpose is to maintain the apparent bass loss which occurs at lower output levels.

The amplifier power output is rated at nominally 15 watts per channel. But this is a 'music rating' and not a sine wave one. If the amplifier was driven to sustain a power of 15 watts from a sine wave signal source, then two things would happen. One, the power requirements would soon draw all the power from the supply reservoir capacitor, resulting in a fall of voltage and a consequent fall in audio power and rise in distortion, and two, the output transistors would dissipate more heat than could quickly be removed by the heat sinks resulting in probable failure of the output transistors.

Music-power rating
Fortunately, music may incite a peak power of, say, 15 watts but only for a very short period of time, the rest of the time the power being very much below this. From these factors has evolved the music-power rating, which is common to many transistor audio amplifiers - not only the Pye. After all, we are interested only in programme sounds; it is only when we test that we may wind an amplifier up to full output on a sine wave. But we should never do that on transistor equipment.

The power output is somewhat dependent on the impedance of the speaker load. At a load of 15 ohms, the output power can be taken up to 15 watts for just a little below 1 per cent total harmonic distortion, while at a load of 8 ohms, the power can be taken up to about 18 watts for the same amount of distortion.

The impedance of speakers tends to rise with increase in high frequency signal, which means that the amplifier is presented with a load that rises in value towards the treble end of the spectrum. Thus, it would seem likely that the output power of the amplifier falls as the result of this increasing load impedance. But much here depends on just how a particular speaker behaves over the audio spectrum. This problem is brought up by the authors in the Wireless World articles referred to earlier.

The proof of the pudding ... on test under ordinary domestic conditions the HSF30T is a joy to use and listen to. The author has now had the opportunity to try the amplifier under a great diversity of conditions both mono and stereo, radio, disc, tape and with quite a selection of speaker systems.

Probably the best speaker system tried so far employs a pair of Wharfedale Super 8/RS/DD units loaded into a pair of somewhat special corner horn cabinets. The amplifier has been driven at high levels on all sorts of music, and that music with a wide dynamic range is adequately handled. Even on sustained peaks there was literally no apparent distortion. This would indicate a very long power-unit time-constant and good transistor sink action.

The amplifier accepted the stereo outputs from an Akai 44S tape recorder with no trouble at all and complete freedom from hum-loops over which no special precautions were made. It operated equally well from programme signals provided on the one hand by a simple transistor radio and on the other by a considerably more elaborate valve FM tuner. Both low and high output piezo and magnetic pick-ups, both for mono and stereo, were used with the amplifier and no undue problems were at any time encountered.

The Pye HFS30T is a worthy hi-fi amplifier and at the new price of £55 13s 0d (walnut case, if required, £4 14s 6d extra) can be thoroughly recommended to those readers who are contemplating going solid-state. G. J. K.