A Valve Pre-amplifier for 78s.
The musings of a newcomer (of mature years) to valve audio – part III.
WARNING: HIGH VOLTAGES, WHICH CAN BE FATAL, ARE USED IN THIS TYPE OF EQUIPMENT.
THE MAINS VOLTAGE IN THE U.K. IS OFTEN QUOTED AS 240 VOLTS – THIS CAN KILL YOU;
THE PEAK VALUE IS ACTUALLY 340 VOLTS. DO NOT WORK ON PROJECTS LIKE THIS UNLESS
YOU ALREADY HAVE PRACTICAL EXPERIENCE.
(I first built main-powered valve equipment when I was 12 years old,
under the supervision of an ex-Army professional Radio engineer.)
9th January 2017. There has been much water under the bridge since the end of November. We have acquired a mentor, who read the first two pages in this series, and kindly contacted us to get us on the right track. We often do things wrong, and sometimes eventually get them right, if we have not become discouraged first. Now, our friend (who was already very well known to us in other circles to do with 78 rpm records and gramophones) has generously given valuable advice – not least, in sending a circuit using two double triodes, rather similar to our Mk.I, but a more considered design, which he had worked on over a long period. I cannot give the circuit here, because I have not asked him yet for permission, though I intend to. But the component count is very low, and evidently nothing is included in the circuit that is not directly relevant to its primary function, viz., to raise the 2 or 3 millivolts from a moving magnet cartridge, to perhaps 100 mV or so, for input to a power amplifier.
Among the things we had got wrong, was: (a) The use of pads and strips of single-sided PCB for mounting components. Ironically, the added capacitance that these provide can have an effect on the performance of a phono pre-amp – while their use at far higher radio frequencies (say up to 10 or 15 MHz) is generally innocuous. (b) The buzz from our earlier power supply was due to the use of the common rectifier diodes 1N4007. They are relatively slow in switching over from the conducting state to the non-conducting, so a back-effect is produced in the secondary of the transformer to which they are connected. I am slightly hazy as to how it all occurs; but it is not necessary to understand it – only to obtain and fit ‘Ultra-fast, soft recovery’ diodes instead. Use of the UF4007 diode is a good step in the right direction. Of course, in a power amplifier, the ringing or buzzing effect is much less noticeable, because that gets a pretty good signal level as its input; but in a pre-amp, we’re only looking at a couple of millivolts, and the buzzing can be very invasive at that level. (c) We had fitted no less than two 10 Henry 50 mA chokes in association with three 47 µF smoothing and reservoir capacitors in our Power Supply Unit, and were very pleased with our sophisticated five-section filter. Our mentor tactfully pointed out that a smoothing choke will only become effective, when it is running relatively close to its current rating – in this case, say thirty mA or perhaps a bit more. Since our pre-amp only consumes about 8 mA, the chokes are essentially not doing very much at all! So out they came, to be replaced with 2.2K 3Watt resistors, as he advised.
In the meantime, we had decided that the hum problem could be further reduced by running the valve heaters on DC. Of course, the cathodes of the valves do not generate the hum; their thermal inertia is much too high to permit any trace of 50 Hz AC to get through. No; the hum is induced by the wiring carrying the AC heater voltage to the valve holders. That’s why they were traditionally twisted, in the hope of cancelling the alternating magnetic field they radiate. In addition, they were always routed round the sides of the steel chassis on which amplifiers (and pre-amplifiers) used always to be built, so the magnetism would mostly be taken up by the steel. Since metal-bashing – especially of sheet steel – is not exactly my bag, DC heaters were adopted with some relief.
The Mk.III pre-amp is pictured above, and the use of a 6X5 rectifier valve in the PSU got shut of the diode problem anyway. Since the pre-amp only consumes about 8 mA, the small 200 – 0 – 200 transformer is quite happy in that role. On the other hand, its 6.3 V heater winding was totally incapable of producing the 0.9 Amps of rectified DC we needed. We discovered this was because there is a power factor issue. You need a heater winding capable of at least double the desired current, and this tranny could not do that. So we just used a 6 Volt SLAB – ‘sealed lead-acid battery’. We already had a nice 12 Volt 7Ah battery, but the 6.3 V valve heaters could not be put in series because the 6SL7 only wants 300 mA, while the 6SN7 needs 600 mA. A cheap digital voltmeter we happened to have knocking about (Ł3 on ebay) was installed to remind us when to recharge the heater battery, which can be seen on the right of the image above.
Here’s the underneath. Following the advice of our mentor, we used tag board to mount the chain of five components that constitute the first part of the RIAA equalisation network – more on that in a moment. Most of the other small components go from the valve bases to earth, ground, deck – call it what you will. I call it deck, actually. Therefore, there is a piece of thick copper wire which passes over the two valve-holders – a bus-bar, they call it, from the Latin ‘omnibus’, meaning ‘for all’ – I bet you’ve always wondered where the term bus-bar came from? I always did, until three minutes ago, when I looked it up on Google. But dear reader (if there ever are any), I digress… As we were saying, all the other components are earthed to the bus-bar. The bus-bar itself is not connected to the ‘chassis’ or copper ground of the PCB chassis; instead, it returns to the negative side of the power supply. There is a very good reason for this; it just escapes me at the moment.
THE RIAA PROBLEM. (What is RIAA? You could look at my page phonopreamplifier.htm , but probably best not; you’d soon get terminally bored, and since you’ve read this far anyway, I really don’t want to let you down.) Suffice it to say, that the RIAA equalisation – a gradual reduction of volume across the audio frequency spectrum as a record plays – is only required for vinyl LPs and EPs. Indeed, it is mandatory for playing those properly. The trouble is, I don’t play LPs and EPs – only 78 rpm records. And they require different equalisations. Therefore, I have had to meddle, and indeed tamper with my mentor’s circuit. Specifically, I have removed just one capacitor – a small item, costing perhaps 10p (12˘) – the function of which was to reduce the upper frequency response to suit LPs, EPs and singles. But nearly all 78s require the upper frequencies to be left ‘as is’ – a ‘flat response’ as it is called. The benefit of this omitted component to the sound of most 78s can be spectacular.
Airy vistas of sound open up; the Broad Sunlit Uplands of the sound of well-recorded 1920s 78s beckon urgently to us. Just listen to a 45 second extract of Joe Venuti’s Blue Four playing ‘Ragging The Scale’. It was recorded in New York in May 1930. (You do need to listen to this example on a good quality sound system to get the full benefit of it. For example, my PC sound card goes into a NAD 3020A amplifier driving KEF Uni-Q speakers; using a pair of disposable ear-buds won’t give you much idea, sorry!)
No processing whatever was done on the sample. We have just simply played a very nice copy of the U.K. Parlophone 78 rpm issue (R-778) using an appropriate size stylus – in this case, an .0028" truncated elliptical. All the presence, immediacy & upper-frequency response of the violin was always present on the master from the day it was recorded. Of course, the upper frequency of these Western Electric system recordings was limited to about 4 KHz, or at the most, 5 KHz. It seems laughable today, when most equipment can handle everything from Zero Hz to 20 KHz. But even a 4 KHz limit can make a pretty sound, can it not? And most people can’t hear anything near 20 KHz anyway, except children, dogs & bats. My own hearing finishes at 6 KHz.
Had this record been played with the wrong RIAA EQ, most of that precious top end would have been lost – a classic case of throwing the baby out with the bath water.
Naturally I picked an excellent quality recording for this demonstration. This disc is so well recorded it would have sounded superb on a Dansette *.
We have for the time being, retained the RIAA ‘bottom end’ boost, because this approximates fairly well to the bass lift required by the Western Electric and Blumlein recording systems. However, some adjustment here would be useful. We’re working on it.
One innovation we must mention, concerns the block-like thing you see at bottom right of the above image. At the top right is a seven-position switch, which can ‘tone down’ the upper frequencies if we so desire. Occasionally, some recordings were a bit ‘toppy’. Each position of the switch brings in a capacitor which progressively reduces the top response. But the most critical portion is right at the start of this range. Therefore, we thought of having a variable capacitor, in order to micro-manipulate the top end. The block is a twin-gang variable capacitor, with a total of about 750 pF. It has a slow-motion drive, requiring three turns of the knob to cover the whole range. It has been found very useful indeed for just trimming off any stridency in a recording. This idea cannot possibly be a ‘first’; but you don’t often see variable capacitors in phono pre-amplifiers!
* Dansette. A make of portable mains-powered three-speed record player, popular in the UK in 1950s & 1960s. The term has probably become generic by now for any of the scores of similar record players of that era. Carefully & efficiently refurbished Dansettes are always to be found on ebay, at startlingly high prices, until one considers the powerful nostalgic attraction of those who would re-live their heady youth again. Good luck to them all, say I!