A Valve Pre-amplifier for 78s.

Fourth steps…


The musings of a newcomer (of mature years) to valve audio – part IV.


board top






(I first built main-powered valve equipment when I was 12 years old,

under the supervision of an ex-Army professional Radio engineer.)


26th January 2017. While being very pleased with the 6SL7/6SNT pre-amp described on the previous web page, we were still aware of some failings. There was a lot of hum. When I say a lot, that is a relative term. When you were playing a 78, you couldn’t hear any hum. But when the system was idling, if you listened carefully, you could just hear a little hum. If you went up to one of the loudspeakers and bent down near it, you could definitely hear hum. And if you turned up the gain to maximum in idling mode (which would never occur when playing any record!), then there was indeed a lot of hum. Now I don’t intend to get neurotic about these little projects; but this was obviously unacceptable. After a lot of thinking, we decided that our use of the old-time 6SL7 and 6SN7 valves was maybe responsible. For example, if you moved your hand anywhere near the 6SL7, the hum got much worse. This is a very familiar phenomenon that used to be called ‘hand capacitance’, and if, like me, you have built simple short-wave receivers, you will be familiar with it. When you move your hand near the knob to adjust the receiver, your body-mass acts as a capacitor which you are coupling into the variable capacitor you’re trying to adjust, so tuning goes awry.


What could/should we do? Put our big, old valves into steel screening cans in order to shield them from external influences? I thought not. Doctors do not treat illnesses solely by supressing their symptoms. Well, our prototype Mk.III pre-amp with large octal valves was built on an 8" (20cm) x 5" (12.7cm) sheet of single-sided printed circuit board. This, in retrospect, is quite a large area, and we had spread the bottles apart, in order to leave plenty of space for fiddling with the intervening replay EQ. All these long leads had a chance of propagating hum – of whatever origin – in spite of our use of a 6V SLAB (sealed lead-acid battery) to supply the valve heaters. It was possible that the ground-plane of copper foil was picking up hum from the mains transformer of our NAD power amplifier – on which, in the absence of a better place, we had always sited our various valve efforts. (But we had also set them up elsewhere, in case transformer hum was coming through the steel casing of the NAD. This proved not to be the case. Or at least, the NAD was but a very trivial hum source.)


Therefore the board shown above was made. It’s 1/8" (3mm) fibreglass, 5.25" (13mm) x 2.5" (6.4mm). The ceramic B9A bases – which are extraordinarily cheap on ebay – fit into 22mm holes, easily bored by step-cone drills – again, a set of three almost absurdly cheap on ebay. The three brass pillars are our own; 1/8" (3mm) brass rod, 1/2" long, drilled 1/16" (1.5mm) each end, interference fit. I should say, that they were supposed to be 1/2" long; but having been made by yours truly, they were all slightly different… 8^)


board bottom


Underneath, an earth bus-bar of 1/16" (1.5mm) brass rod links all the phono negatives, and connects to the centre pillar. It may be a little long; we had thought of making it a ‘Z’ shape; but still, this is a prototype – hopefully the last!


final top


The finished item is mounted on four wooden dowel legs, making it easy to test.


final bottom


Point-to-point wiring was used. There’s not a lot to this circuit, is there? 8^)  You’ll notice the pairs of phono sockets are in parallel. We do not play stereo. discs; but we do need both channels from the pick-up – while the pair of sockets on the output (right-hand) side is provided so that we don’t have one spare phono plug flopping about & accidentally shorting something out! You will also see a heterogenous collection of components – such as ultra-cheap polyester capacitors, small 1/4 W carbon film (aiee!) resistors, some of them in high voltage positions. Also, three instances of required values being made up from two components in series, or parallel. Please don’t worry; as I keep on saying, this is a prototype, and I already have to hand a complete set of much better-quality components of optimum values, obtained from The Hi-fi Collective, as recommended by my mentor. These will be employed on the final version. Happily, I was able to return to my old habit of using small pads of single-sided PCB, superglued down, to form anchorage points. As the board is just fibreglass, there is no ground plane of copper as hitherto, which might form undesired mini-capacitors with the pads.


Mk IV in place


So: here it is. The board sits just below a new top, made of 1/16" (1.5mm) fibreglass. It’s secured there by two 1/2" (12.7mm) brass distance pieces. There are 24mm holes so that the valves can be ventilated – though there is no pressing need for that, as the modest 230 V HT doesn’t worry the ECC83 & ECC82 very much. No, I am not going to install blue LEDs underneath them. 8^) The heaters are run from a 9 V transformer inside the PSU on the left. When rectified by a bridge of UF4007s and smoothed by a 4700 µF electrolytic, the valves receive 11.2 V with the twin heaters in series/parallel. This is rather below the permitted tolerance, but it seems to work OK. 


What does it sound like? I’m pretty happy with it. The hum is almost non-existent. The compactness of this new board seems to have been effective. With the experience gained in this exercise, the final board could be smaller still, and even let into a die-cast box – hang the expense. After all, playing 78s is a very serious business, isn’t it?


Here's a short mp3 of part of ‘Cornfed’ by Red Nichols’ Five Pennies. New York, 20th June 1927. It’s the chorus that mostly features Adrian Rollini on bass saxophone, with Vic Berton on tympani and cymbal behind him. Click on the label below to hear it; or probably better, right-click and open it in a new tab, then come back to this page. Of course, to get a good idea of the sound, you’ll need to listen to it on a reasonable quality set-up. A pair of disposable 99p ear-buds are unlikely to do it justice.  8^)




Original issue was U.S. Brunswick 3597. This sample was taken from the equivalent U.K. Brunswick in pretty good condition, played with a 0.0032" truncated elliptical stylus in a Shure SC35 cartridge. The (binaural) output from the cartridge was put into mono. through the above pre-amp, and then into the auxiliary (i.e. flat) input of a NAD 3021 amplifier. My loudspeakers are very modest – Monitor Audio MA6 Budget Version. Also, my living room is small, but the sound-stage of this – and most other – 78s comes over very well. I hope you’ll think so, too.


Before closing, I would like to stress that this extract has had NOTHING done to it whatsoever. Perhaps that might more accurately be stated as: I have done nothing to what came out of the pre-amp. No de-clicking, no de-crackling, no re-equalisation, and above all, NO DIGITAL DE-NOISING. Of course, it has passed through umpteen transductions since Nichols, Mole, Dorsey & Co. performed this music 86 years ago. Most of those, we can’t do anything about, as far as I know. But what we can do, is to retrieve the maximum musical information from a disc, (a) as long as it is in reasonable condition; (b) we use an optimum size stylus, and above all (c) give the information on the disc every help in coming across to us. At very least, the sound of the sample above, I contend, would make an excellent basis for sensitive and thoughtful conservation.


Brunswicks are often written off as ‘tubby – no top end’. Not true – at least in this case. Which is why I chose a Brunswick for this page. Most Western Electrics are far better! Another mentor of mine, the late John R.T. Davies, once wrote, as I remember it: ‘The top end, given a chance, will unfold into airy vistas of sound.’ Indeed it will – when gently cossetted!


It occurs to me that I did not check the musical pitch of the record; alas, my attention span has drawn to its close, sorry. But I think my main intention has been set out fair and square:


Get the best sound off the record in the first place. You can’t make a silk purse out of a sow’s ear. (Ancient proverb.)