Added To Basket 
1 x VT0000
Mains transformers for valve power supplies consist of at least two windings, or sets of windings, on an iron core to assist the coupling between them and thus improve their mutual inductance for a working frequency of 50 – 60 Hz.
Power from the AC mains is applied to one winding or windings (the primary) and the magnetic flux set up in the iron core and around the coil induces currents in the second winding or set of windings (the secondary), the voltages across these coils being either higher (step up) or lower (step down) than the voltage applied.
The size of each winding bears a very definite relationship to the power applied to or drawn from it, with the number of turns controlling the voltage and the resistance, expressed as the diameter of the wire, controlling the current.
The Core
The number of turns varies inversely to the size of the core. The core is built up of thin sheets of iron (latterly, 'electrical steel') in the form known as a laminated core, and this is a method used in practically all AC apparatus. Clearly the rapidly varying magnetic flux will induce currents in the core as well as in the windings around it, and if the core were one mass of metal with a very low resistance, the current so induced would be exceedingly high. It is necessary therefore to increase the electrical resistance of the core which can only be done as described, by splitting it into thin sheets and insulating each sheet from the next. Eddy currents will still flow but the total loss of power so caused will be far less than otherwise.
Core Laminations
Laminations are insulated in several ways – by chemical treatment of the metal surface, by varnish, by very thin cemented paper – and there are two main shapes of laminations, the E and I type (commonest) and the T and U type, both sets giving a three legged core (Figure 1a).
When the laminations are being inserted into the finished coils on their former they must be alternated, that is an E must go in from the left with an I from the right, then an I from the left and an E from the right and so on, the laminations being brought into tight contact with no air gaps.
The cross sectional area of the core, Figure 1b, is chosen from the formula given by The Radio Designers' Handbook, Iliffe, where:
A = √W / 5.58
where W is the Volt-amperes output (Va, or Watts), and A is the cross section area in square inches.
The formula connecting the number of turns in a winding with a given voltage, size of core, frequency and flux density is:
E = (4.44 x F x H x N x A) / 100,000,000
where E is the voltage supplied to or supplied by the winding, F is the mains frequency, H is the number of lines of magnetic flux per square inch in the iron and A is the cross sectional area of the core.
If E is allowed to equal 1 then the calculation will give the number of turns per volt for any winding on that core.
It is supposed that often home-made transformers will be rewound using materials to hand, and in this case the characteristics of the iron will not be known. The best compromise in such conditions is to let H equal 60,000 lines per square inch, a figure at which many power transformers are run, although if winding space and other conditions permit this may be reduced to 50,000 lines. A, it must be remembered, is built up of laminated sheets which have insulation on one side at least so that the actual magnetic area will be only 90% or so of the geometrical area. This measured area, then, should be reduced by 10% for the calculation.
The shape of the core must be well proportioned, each outer limb having half the width of the middle limb on which all the windings are placed in layers, thus occupying the window space "x x y" of Figure 1a. The general order of the windings is: primary inside*, nearest the limb; the HT secondary and the heater windings outside, of which there are usually at least two, one to supply the rectifier heater and one for the valve heaters of the apparatus.
* NOTE: Class 2 safety regulations require that the mains primary must be physically isolated from all secondaries through the use of a split bobbin, that is, one having effectively two formers separated by a centrally placed plastic wall. This is called 'double insulation' and may be subject to EC or US federal laws governing mains powered electrical apparatus. Alternatively it can be concentric wound (full-width layers) with an electrostatic screen between primary and first secondary, which is basically a sheet of copper foil with a lead-out wire that can be mains earthed.
Transformer Regulation
The regulation of the transformer is very important – that is, the virtue of its having only a small output voltage variation with varying current loads – and depends to a great extent on the iron of the core, the shape of the core and the filling of the window space with windings, there being no large gap between the last outside limbs. The core must be large enough and the wire diameter fully adequate to handle the loads expected.
(Often regulation of commercially made transformers are quoted in %, e.g. 10% being common. This means that the secondary turns ratio/Voltages are 10% 'overwound', i.e. 10% more. Measured off-load Voltages will then be 10% higher than expected, this is normal and not a 'fault'. Once loaded, the Voltage will drop because of internal resistance, which the overwinding compensates for.)
Losses
The main losses in a transformer are "iron" and "copper" losses; those watts lost due to eddy currents and the purely magnetizing effect on the core, and the watts lost due to the currents flowing in the resistances of the windings. Theoretical transformer design requires these losses to be equal when the transformer will be at its most efficient working level, but for the purposes of small transformer design it will be sufficient to base all calculations on a theoretical efficiency of 80% instead of 90% or so which, with care, will be obtained. These losses will be dissipated as heat and any transformer which heats up in working to anything but a small degree is inefficient and wasteful. Power is being lost, regulation will be poor and insulation will be subjected to the most undesirable strains. A good transformer will work for hours with a temperature rise which can scarcely be observed by touch.
Example
A transformer is to be made with the specification: Primary to be tapped to 210, 230, 250 volts (traditionally these taps provide the option for 'trimming' the output); Secondaries, 350-0-350 volts (for a full-wave rectifier) @ 120 mA, 6.3 volts @ 3A (valve heaters) and 5 volts @ 2A (separate isolated rectifier heater).
The watts ratings, therefore, are: -
350 x 120 mA (only half the HT winding supplies current at one time) = 42 watts.
6.3 x 3 = 18.9 watts.
5 x 2 = 10 watts.
giving an output total wattage of 70.9 watts.
or, say, 71 watts.
The cross sectional area of the core should be at least:
A = √71 / 5.58
or 1.5 square inches, and assuming an efficiency of 80%, which should certainly be bettered in practice, the input wattage is therefore:
71 x 100/80 or 88.7 watts.
At a working input voltage of 230, therefore (the usual mains voltage), the primary will take 88.7/230 amps or 0.4 amps, nearly, and the wire must be chosen to carry this current safely. The question of insulation enters here.
Construction
Commercial transformers, as inspection will show, are most often wound with enamelled wire, but conditions are different from those pertaining to home construction. The commercial transformer is machine wound so that the wire can be, and generally is, slightly spaced between turns so that there is no rubbing of the enamel, whilst the wire tension can be more accurately controlled.
For amateur construction enamelled wire can be (and is most conveniently) used but on no account should it be wire taken from old coils or transformers. It must be new and every precaution must be taken to ensure the covering is not cracked, kinked or rubbed, for a breakdown in insulation in any winding renders the whole transformer useless.
A suitable core is now chosen, one with an area of 2 square inches (reducing to an electrical area of 1.8 sq. ins.).
The turns per volt formula becomes, then,
(4.44 x 50 x 60,000 x N x 1.8) / 100,000,000
but if desired a factor can be produced relating to all transformers where H is taken as 60,000 by leaving out the terms N and A.
This factor, obviously, for 50 Hz mains, is:
[(4.44 x 50 x 60,000) / 100,000,000] x AN
= 0.1 332 AN
so that the formula for this transformer becomes:
0.1332 x 1.8 x N
= 0.24N
and N = 1/0.24 or 4.2 turns per volt.
The windings can all be calculated, then, the primary having 250 x 4.2 = 1,050 turns tapped at 966 and 882 turns, the secondary has 700 x 4.2 = 2,940 turns, centre tapped, the valve heater secondary has 6.3 x 4.2 = 26.5 turns* and the rectifier secondary has 5 x 4.2 = 21 turns.
(* Ideally the valve heater secondary should be centre-tapped to provide a balanced heater feed; the centre-tap connecting to 0V ground so as to arrange equal and opposite voltage fields in twisted pairs of heater wires that hence cancel each other out, thus minimising hum pick-up in sensitive valve amplifier circuits.)
The size of wire, as already shown, affects the current flowing in the winding, and for this type of transformer the gauge may be chosen on the basis of a current flow of 2,000 amps per square inch.
The primary draws 0.4 amps so from a wire table it will be seen that SWG 26 enamelled copper wire (ecw) will be suitable; for the HT secondary enamelled wire with an interleaving of thin waxed paper between each layer should be used, and to carry the 129 mA SWG 34 ecw will be suitable.
SWG 18, enamelled, will suit both heater windings, and to make up losses one extra turn is usually added to the calculated figures for these two coils.
Former Dimensions
It is now necessary to pay some attention to mechanical details and to check over the dimensions of the former. The size of the window space, "x x y", as shown in Figure 1a, is 11/8" x 17/8" and the former may be supposed to be made of one-eighth material, card or paxolin. This will reduce the available space in three directions, leaving the depth of the window one inch and the length one and five-eighths inches. The space taken by each winding must now be calculated.
The Primary
SWG 26 ecw 48 turns to the inch, so that the former will take 48 x 15/8 turns per layer, or 78 turns. The number of layers will be 1,050/78 or 14 layers and the height will therefore be 1/3".
The HT Secondary
SWG 34 ecw winds 100 turns per inch so that each layer will contain 100 x 15/8 or 162 turns. The number of layers will be 2,940/162 or 19 layers, and these will be one-fifth inch high.
Heater Secondaries
SWG 18 ecw winds 19.7 turns per inch so that one layer will contain 19.7 x 15/8 or 32 turns so that each heater winding will fit into a layer comfortably, and the whole wire height of the two windings together will be under 1/8".
The total height of the wire alone, then, is 1/3 + 1/5 + 1/8 or 2/3 inch, leaving 1/3 inch space for insulation.
Assembling
When the former is made, shellaced and perfectly hard, the cheeks may be drilled for the leads using the figures above as guides or the holes may be made as the work progresses providing there is no chance whatever of damaging the wire insulating in any way. The Primary is wound first, the wire being cleaned properly with spirit, not by scraping, and having a flexible lead soldered to it.
The soldered joint must be perfectly smooth with no sharp points or projecting wire ends, and it is then covered with insulating sleeving which carries the flex lead through the cheek. The wire is then wound either by hand or by a simple winder, which is much to be preferred. All that is needed is a spindle turning in end plates or bearings, a handle at one end.
Two adjustable cheeks are then mounted on the spindle to grip the former tightly, the spindle (which might well be a long screw threaded rod) passing through the centre hole of the former. The former is then rotated with the right hand, the wire being fed off its reel and tensioned evenly with the left. The turns should be laid evenly side by side and counted as they are put on, in the absence, as is likely, of a mechanical counter it is convenient to mark every twenty turns on a sheet of paper.
The primary winding is not interleaved so that when the end of one layer is reached the wire is wound straight back on itself and tension must not be over tight for each corner of the former presents a sharp right angle bend to the wire whilst the lower turns have to sustain the considerable strain of all those windings above them.
It is necessary to understand the effect of one short-circuiting turn in any winding. It would consist of a very low resistance loop in which, therefore, a very high current would be induced, this causing heating and consequent burning of the insulation on adjoining turns of wire, whilst the extra load reflected into the primary might cause that winding to be overloaded to the fusing point. It must be realised that the current flowing in the primary depends entirely on the load being drawn from the secondaries, with the secondaries disconnected the only current flowing in the primary is the small core magnetizing current and the winding acts as a choke.
The taps for the various primary voltages can be taken out in the same manner as the taps on coils, by drawing out a loop of wire and returning the wire to the next turn without any breaks or joins, or a flex lead may be soldered to the winding at the correct turn and well insulated. Whenever possible taps should be arranged to fall at the end of a layer so that they may be passed straight through the former cheek. If, however, they have to pass over several turns the insulation must be perfect and on no account must unevenness of winding be allowed in the next layers. Any bump in the centre of the coil will be magnified in the later layers with a corresponding strain on wire and insulation.
When the primary is finished, and a flex lead soldered to the last turn, the winding must be insulated from the following coils. The best material is Empire Cloth interwoven with glass fibres and known under such names as Glassite, but plain Empire Cloth may be used. Every part of the primary must be covered, the insulation being carried up snugly to the former cheeks.
Electrostatic Screen
Many transformers have an electrostatic screen* wound over the primary to prevent interference from the mains being induced into the secondaries. It consists simply of one layer of fine insulated wire – SWG 34 ecw, for example, one end of the wire being anchored internally and the other brought out through insulating sleeving. The end brought out is earthed to the apparatus' chassis. Naturally just as much attention must be paid to the insulation of the screen as of any other winding; no load is taken from it as only one end has a connection but shorting turns would give rise to the same heavy overloads mentioned above.
* Electrostatic screens have fallen into general disuse with the advent of split bobbins, the close capacitive coupling that resulted twixt primary and secondaries without one is no longer much of an issue due to the physical separation. Still advisable to have one in the case of concentric wound.
Study of any existing valve power supply transformer will show that the full HT voltage is established between the HT and rectifier heater windings, and so the insulation between them must be perfect. Any breakdown here will immediately ruin both transformer and rectifier valves.
When the former is wound it is given a last covering of cloth and the laminations are inserted into the centre aperture in order as already explained. The E and I laminations of the core must be inserted carefully for it may be possible to run a sharp edge or corner into and through the former material, cutting or scraping the primary winding.
The laminations must be clamped into a solid mass with metal clamps (now called 'frames') which also provide fixing holes for bolting the transformer to its chassis.
Transformer Kits
To save a lot of hassle it is worth noting that mains 'transformer kits' are available in 50VA and 100VA sizes from this site. These kits take a good deal of trouble out of creating a customised mains transformer for whatever purpose; each is provided with a split-bobbin former with primary windings already wound, E and I laminations, covers, solder tags, and mounting/clamping frames. The existing primary pre-determines the volts ratio so it is only necessary to calculate secondary turns based on this, then the wire gauge necessary to fill the former.
The operation of the power supply as a whole may here be considered, with reference to Figure 02, where the transformer just described is shown in its circuit. The HT secondary has been wound to give a RMS voltage of 350 which means that the peak voltage will be 350 x 1.414 (peak value of a sinusoidal wave).
Thus the rectifier anodes will have peak voltages of 495 volts, the whole winding having a peak voltage across it of 990 volts and even after the voltage drop due to the rectifier is allowed for the capacitor A has a voltage across it well in excess of 350 volts – probably 450 volts. This explains why the voltage rating for this capacitor is necessarily high; a 350 volt working component would soon fail in this position.
Figure 03 
shows a full-wave rectifier using the separate 5 V heater winding (this example is a GZ34 in a PSU for an audio power amplifier). This is necessary since the DC potential between the output and ground is too excessive for heater to cathode insulation to withstand for very long. Hence, no attempt is made to insulate this valve's heater from its cathode nor share it with the main heater chain. Instead it has its own heater winding, the DC potential of which is as per the HT level.
The heater voltage for such rectifiers is typically 5 volts, previously 4 volts, hence by this very specification you can tell that the heater is not meant for the usual 6.3 V supply. There are exceptions where the HT DC is not expected to be high, and the rectifier heater may then be able to share the common grounded heater supply – reference to the relevant specification data is required to make sure.
Choosing The Reservoir Condenser Value
The actual value of the 'condenser' (old-fashioned name for capacitor) in microfarads is more or less of a compromise, for the final output voltage of the power supply depends to a great extent on the size of the reservoir. If it were to be omitted, the output voltage would be very low, and as it rises in capacity so the output voltage rises towards the peak value. Before the peak voltage is reached, however, the condenser is excessively large (and expensive), but moreover, it would be drawing very heavy currents from the rectifier valve on each surge or peak of the cycle and the valve would soon lose its emission (read as 'fail').
The earliest electrolytic capacitors were restricted in value, being typically 8 μ F – 16 μ F, then later, as much as 32 μ F. Nowadays values of say 47 μ F at 450 V working can be easily sourced, but, as was the case for the smaller values, there is still an advantage to be had by adding a series choke to reduce the HT ripple, see Design and Construction of Low Frequency Chokes and Single-Ended Output Transformers.
Circuit Protection
Valuable protection to the rectifier and transformer can be given by inserting simple fuses in the circuit, shown as 'F' in Figure 2. These can be conventional types, or they can be of the anti-surge control 'flash lamp' bulb type, viz, actual light bulbs with a current rating to suit the load to be taken from the power supply with extra provision for any surges that might occur as the condenser charges up.
High Voltage Transformers
It is unlikely that the amateur will attempt the task of winding a High Voltage Transformer such as would be used to supply a large cathode ray tube, but a few points of High Voltage practice might be touched upon.
Firstly, the peak inverse voltage across a typical television transformer might reach as high as 10,000 volts, so that great care is essential during testing to see that no risk of touching any live circuit is taken.
Secondly, the positive side of such a power pack is usually earthed, so that strain is placed on insulation in many ways. For example the primary of the transformer might easily be earthed via the mains; in such a case the end of the secondary nearest the primary would be the earthed end, thus preventing a large potential difference directly across the insulation separating the windings.
Thirdly air insulation is often relied upon. At high voltages a trace of moisture upon an insulating surface might give rise to sparking or arcing which, while slight at first would rapidly become something approaching a short circuit. For this reason the layers of the secondary are not carried to the end cheeks of the former and as the winding grows outward from the centre the layers are made shorter, giving a pyramidical or stepped effect. In this way, as the potential above earth rises through the winding so does the distance between any earthed object and the winding increase.
Fourthly, the potential difference between the rectifier heater winding and the H.T. winding makes it necessary to have perfect insulation between the windings, a separate heater transformer helping in this respect. Metal rectifiers give very good results for cathode ray tube power supplies.
Copyright © B. Babani (with modifications)
explore ··· 