Page 4 of 4
BUILDING THE 300B The 300B amplifier is available as a kit, but we will be offering a transformer set for those intrepid souls who would like to do the rest themselves. It comprises a mains transformer, two output transformers, two chokes (mounted underneath the chassis), and two inter-stage phase-splitter transformers, which come encapsulated in a metal case to protect their secrets; Tim says they are very special.
The set of seven transformers weighs no less than 16kgs (35lbs), so be aware of carriage charges, as well as the need for a strong chassis. In addition to a good knowledge of electronics, you'll need civil engineering skills to build this monster. No wonder the feckless fled to transistors all those years ago!
Ideally, a strong steel chassis should be used, with a protective bottom plate, to prevent accidental contact with lethal voltages (550V H.T.). The valves can cause burns, so a well ventilated top cover is sensible too. These items are part of the kit, but a cover is an optional extra.
Anyone attempting to build this amplifier from the transformer set I imagine will have a good electronics knowledge, plus some practical experience with valves. If you don't have this, then for safety's sake, I suggest you take up knitting instead. This is a highly specialised amplifier and although it is easy to build, having few components (not including transformers!) it has a lethal 550V H.T. line inside, which makes testing or fault finding potentially hazardous. A skilled electrician or radio/TV repair man should be able to handle the job though.
PARTS There are some unusual parts in this new design. The power supply uses high quality 'motor run' 25µF polypropylene smoothing and reservoir capacitors, instead of polarised electrolytics. They are rated at 630V d.c., do not retain a charge after switch off and are neat. The slightly cheaper alternative is a pair of 47µF electrolytics wired in series, with d.c. balancing resistors across them, and a I µF or so bypass polypropylene to keep their impedance low at high frequencies. That amounts to quite a few components just to do one job, so the 'motor run' caps., from Electromail, seem an elegant and sensible solution. They keep high frequency noise off the power lines too, contributing to sound quality.
The chassis layout is very important; spacings and orientations are chosen to minimise hum. We suggest the mains transformer is best placed in the centre of the chassis, either at front or rear. For DIY, at the front makes most sense, since the mains runs across the chassis only once and the input valves can be placed close to the input sockets, whilst also staying away from the mains transformer, which should of course be kept away from all low level circuits and transformers to avoid hum induction.
If a valve starts to fail and you'd prefer to run away than towards the mains switch, then the properties of the fuse need to be relied upon. The amplifier draws 0.9A from the mains (240V), but switch on/off surges go way above 10A, so a slow blow 1.6A fuse is recommended in the transformer primary circuit. The transformer and rectifier valves handle current well enough to make a secondary (H.T.) fuse unnecessary.
Put the mains transformer at the rear, unless you don't mind the unusual appearance of a centre-front mounted unit. Put the output transformers either side, spaced away from the mains transformer by at least 50mm to lessen hum induction. The cores must be pointing in different directions too. Similarly, the chokes go under-chassis and away from the mains transformer; put the power supply components in-between.
The transformer cores should be electrically isolated from the chassis, that's why the screws have fibre washers underneath their heads. Ideally, all transfonmers should be spaced off the chassis by an insulating gasket. The mains transformer is best spaced up above the chassis by 3-5mm, using spacing nuts on each of the comer bolts. This lessens hum induction and allows convective cooling currents to pass freely up past the transformer. It runs very cool, as is to be expected with a well designed amplifier with high quality parts.
On the subject of cooling, large vent holes should be put into the chassis to ensure heat from the cathode resistors is vented. It is good policy generally to heavily vent an amplifier such as this, for its heat output is substantial. Our own amps have vents around the valves and a spaced-off mains transformer, plus a vented bottom plate and cover of course, to ensure the unit runs cool. Even though the heating effect of the amplifier is noticeable in a room, it is only equivalent to two 100W light bulbs.
WIRING When building the amp, my advice is to wire in the 'services' first. By this I mean power supply, earths, heater leads and what have you. After that I wire the stages in succession, from input through to output, using red cable for all dangerous H.T. lines, black for earth and yellow for signal. Remember to use heavy cable for the heaters, since they draw quite a lot of current. Twist them tight and keep them close to the chassis.
It is very important to split the earths and run them back to one earthing point on the chassis, by the input sockets; this avoids instability. In particular, make certain the earthy end of the output valves (R 15,16) runs back to the single (star) earth point using heavy cable and without any other connection being made to it. Having suffered instability more times than I prefer to remember and it's a brute to clear - avoiding it should be a priority. In my experience, good earthing and keeping outputs away from inputs results in a very stable unit, one where even grid stopper resistors are often unnecessary.
To lessen hum pickup from induction loops, run signals alongside their earth return lines. They can often be lightly entwined, but if you do this watch out for the effects of distributed capacitance, or 'stray coupling'.
Having wired the amp, check and re-check all connections, looking for connection errors and dry joints in the soldering and electrolytic capacitors the wrong way around. Ideally, the amp should be powered up in stages. First run up the power supply alone, disconnected from the H.T. line at C I 1/ L 4 junction, using a variac. If the variac hums furiously, you're drawing too much current, so shut down and find out why. Otherwise, power up and watch the H.T. rise with a d.c. voltmeter, making sure it settles at around 575V with no load.
Then with all power disconnected and the H.T. line disconnected from C I I /L 4, measure its resistance to ground. If this is low or a dead short, there's a fault that must be cleared. Expect to see I MOhm here, but wait for the capacitors to charge.
If all is well, connect H.T. line to power supply and switch on. Although this is for most people the nail biting bit, very little can in fact go seriously wrong, since the fuse will blow immediately if there's a fault, so there won't be much of a flash or a bang!
I suggest that the feedback connections are left off until proper working has been verified. If the amp goes unstable when feedback is connected, then the feedback polarity is wrong. Before connecting up the feedback, check all d.c. levels; they should be within 10% of our published values assuming 550V H.T. Be aware that mains voltages fluctuate from place to place and over the length of a day and that this will affect all internal voltages.
APPLYING OVERALL FEEDBACK As I explained in the introduction, the design rationale behind this amplifier was to produce a unit that was basically linear (distortion free) without feedback The 300B valve is prime contender for this role, because it was designed in 1928 to be an inherently linear audio amplifying device. Feedback wasn't used then, so valve manufacturers were obliged to consider this requirement.
The introduction of feedback freed valve designers from such constraints, allowing them to optimise other properties, like gain. So the innately linear amplifier became a thing of the past, by force of events. That's why the 300B is so popular today.
But feedback has many attractions, reduction of output impedance being one of them. Without feedback an amplifier such as this has a high output impedance; this one measures 50hms. It would be no problem if loudspeakers had a flat impedance, but most don't. Impedance usually varies wildly and this affects frequency response, the amplifier develops an actual frequency response that mimics the loudspeaker's impedance curve, although far less pronounced. So if, as in most two-way speakers, there's high impedance across the midband, then an amplifier such as this will have a lift in its response here as well.
All the same, I've used the 300B with Heybrook Quartets and our own large KLS-I and have enjoyed superb results. At the end of the day, the response changes aren't so great as to be obstructive; the properties of the 300B, which I'd describe as a beautifully open and clear sound, sweetly transparent across the midband and treble regions, yet with strong deep bass, come across all the same.
Overall feedback can be applied to reduce output impedance and lessen 'speaker sensitivity though (see circuit). The amp won't take more than around I OdB of overall feedback, with R8 increased to 470 0hms to reduce tertiary feedback, before going unstable. Experiment showed that around 6dB of overall feedback is ideal, set by making R8b 3.6kQ. If more overall feedback is to be applied, R8b set at 2.7k being about maximum, tertiary feedback will need to be kept low by setting R8 high. With no overall feedback, R8 can be reduced to 180 Ohms.
For best square wave response (minimum ringing) when R8b is set to 3.6k, 2000pF (C6b) needs to be shunted across it for phase compensation. With 6dB of feedback applied, output impedance fell to 2 0hms and distortion halved from 0.2% to 0.09% at I W / I kHz. Frequency response and sensitivity don't alter significantly.
Readers might like to try balancing tertiary feedback against overall feedback. Note that if the amp goes unstable whenever overall feedback is applied, then the feedback is positive, not negative. Reverse either the primary or secondary connections on the output transformer.
FEEDBACK SETTINGS
R8 R8b
Tertiary only 180 none
Overall & tertiary 270 3.6k
Max. overall+t 470 2.7k
OBTAINING THE PARTS The 300B valves are commonly available from the various valve suppliers around Britain, most of whom can be found advertising their surprisingly wide and esoteric stocks in this Supplement. This amplifier needs four, which new will cost around £300. Valve bases are available from the same suppliers. The transformers are available as a set (7 off) from us, price £430 inclusive of VAT and carriage in the U.K. These are available now. The other components are widely available from suppliers around Britain. We will be offering a kit of all parts, including a high quality welded steel chassis with top cover and bottom plate, transformers and all parts of best quality, except valves, for £750 inclusive of VAT and carriage.
SAFETY WARNING
Lethal voltages exist in this amplifier. We do not suggest you attempt to build it unless you are conversant with valve circuits and safety precautions. For safety, never hold earthed metal· work when testing. Make sure your body is isolated by rubber soled shoes. You should posses a voltmeter capable of reading up to 1000volts. The final power amplifier should have a protective underplate and a wire mesh top cover, since the 300B valves run very hot and will bum.
END
Go to Developing 300B for the updated final design.
|