Hi,
I hope the following information will be useful to members wishing to re-wind transformers. I can only write from personal experience from winding my own chokes and transformers as I’m completely self taught but it works for me.
Some form of rotational drive will be needed to turn the bobbin (Former/Winding Spool). I now own four coil winders but have previously used a variable speed wood turning lathe; the lathe was way over the top but I did wind a Murphy A30C field winding coil using it in fact I wound two and these take a tremendous amount of enameled copper wire. The first one I wound as a complete novice went well until almost completed then the lot collapsed resulting in the waste of a lot of wire as it all tangled badly. I’ll explain how to avoid this later.
There are many ways to mount a bobbin and get it to revolve; a simple hand crank will suffice for a one off but will be slow work; it would be better if it was geared up to rotate the bobbin in a ratio of say 10 turns of the bobbin to one turn of the crank. Simple shafting and a large and small wooden pulley would achieve this with a bit of imagination and effort. A variable speed electric drill mounted into a bench stand would be a simple way of revolving the bobbin and controlling the speed.
Take the electric drill as an example. Set it up in a bench stand either bought or home made. If the chuck will accept up to 10mm then buy a length of 10mm threaded rod with a few nuts and washers. The threaded rod can be cut to length with a hacksaw and rough edges filed smooth. This is a crude way but with care and patience should give a working coil winder.
The open circuit choke or transformer will need completely stripping right down to the bare bobbin (Former). If it is a transformer covered in black pitch which I hate with a passion; the pitch will have to be removed and a simple but messy way is to bake it in an oven until the pitch has melted and dropped clear; to aid this it pays to elevate the bobbin on something like a baking tin; set the oven around 160C as a start and see how it goes. I’ve even used a blowlamp working in the garage. Removing the pitch is one job I do not like doing.
Next job is to remove the laminations; if it is a largish transformer then the through bolts and nuts will need removing first. Now the fun begins because usually the laminations will be a tight fit inside the bobbin; this might be compounded by rust making them extremely difficult to remove. Grabbing outer laminations with pliers will ruin and distort them rendering them useless. “E” and “I” laminations are easiest to remove but the “U” and “T” can be a problem; if you come across one piece laminations as I did then these are a downright pain to remove.
LOPT laminations.
For “U” and “T” laminations; gently nip the transformer winding in a vice with the laminations lying horizontally. Locate the lamination joints and using a wide screwdriver and small hammer gently tap at the joints and try to separate them; work from side to side in order to drift a single lamination out; once it has moved and there is enough to grip with pliers then it should pull free; once the first three laminations are out the others will follow much easier. I find a craft knife helps to separate each lamination but care is needed to prevent a finger being cut. Stack the laminations neatly on the bench in the order they were removed. Now the bobbin will be free. If you have some old laminations to hand a piece cut from one of these can be used with as a drift.
Remove the outer insulation from the bobbin and this will reveal the top layer of the last coil usually the secondary. This will normally be thicker wire than the primary coil and there will be fewer turns of it. Two things are now very important and these are number of turns and thickness of wire (Wire Gauge).
Stripping insulation.
Reading books on transformer design is sure to scare a novice into not attempting trying to re-wind one but I found a way by trial and error of obtaining the information. I wish someone would have told me rather than me having to find out the hard way by unwinding what seems like miles of wire and counting the turns.
Simple output transformers with a primary and single secondary are dead easy to re-wind. Mount the bobbin in the drill leaving both hands free. Remove the outer insulation to reveal the outer layer of secondary turns and locate the end of the wire (Finish). Now be very careful and take a note of how many turns are on each layer either by counting directly across the layer if the wire is thick enough or by unwinding; carry on unwinding and noting the number of turns until the other end of the wire is found (Start). You now have the number of turns of wire for the secondary. Now for the thinner wire of the primary winding. This can be a real pain to count as there can be 3,000 turns of extremely fine wire on such a simple output transformer as for a DAC 90 if my memory is correct. Add to this sticky interleave insulation (insulation between each layer) and trying to count without losing the plot quickly becomes a nightmare. To complicate matters even further the wire can break at a blink and once the open circuit is located the wire could break many times causing a great deal of frustration.
There is a simpler way to determine the primary turns and I need to find the formula once again and I’ll add it at the end of the thread; it involves some very simple math’s that even I can understand; using this formula will give the turns ratio; I think the last output transformer I rewound had a ratio of 48/1 so by having already counted the secondary turns it is easy to determine the primary turns and for me this was the hardest part in re-winding as the actual winding is straightforward.
Establishing the wire gauge would sound to be difficult? In fact it is incredibly simple; measure the original wire accurately using a micrometer. I always measure over the enamel as this is the thickest wire that the lamination window will accept. Using very slightly thicker wire would give a more robust transformer but multiply this extra thickness many times and the bobbin won’t fit into the laminations. As modern enamel is much improved it is possible to wind whilst leaving out the inter-layer insulation this will allow very slightly thicker wire to be used and I’ve got away using this method in the past. Normally I will use the yellow transformer self adhesive insulation tape which can be bought through eBay adding a layer of tape to every third layer of wire. A bit of practice helps here.
Even a small transformer uses a lot of wire.
Whilst I’m on turn’s ratio I’ll try to explain the absolute basics of transformer design. A transformer is actually very simple to understand for a novice. A transformer consists of a number of parts the main parts being the laminations and the winding wire; add to these the bobbin (former) insulation and lead out wires. Given the few parts involved what can be complicated? The turn’s ratio is very important and determined by the laminations; too few turns will result in a poor performance whilst too many will saturated the laminations.
I once put together a home turned wooden pulley with exactly 12” circumference at the bottom of its groove mounted between centers and used this to measure the length of wire removed from my first transformer. I didn’t understand that it was the number of turns and not the length of the wire that was important. I find it wonderful that I can fall into every black hole and make every mistake possible whilst learning something new.
Because we are discussing re-winding then all the necessary details are to hand if we know what to look for. I always thought winding radio mains transformers would be extremely difficult given the number of secondary windings and tappings. On the ones I’ve re-wound the primary has always been open circuit and as the radios are for use with a number of different voltages; tappings are added which at first would appear difficult for a novice to understand; I ignore the mains tappings and wind a single primary coil to work from 240V. That’s one big problem out of the way to start with.
Before removing the transformer from the chassis I mark every connection and pay particular attention to the heater connections and which valves they are connected to; I write this down and double check as a mistake could quickly lead to a lot of frustration. A clear sketch of the transformer is drawn; it’s mounting position and above all its connections; time is taken over this and it is done before breaking the connections. A digital camera is used to take close up pictures both above and below chassis.
Now for the important turns ratio. Having already established what the connections are I look for the valve heater connections which I have clearly identified on the transformer? The bobbin is then mounted into the coil winder and the outer insulation removed; taking careful note of the first wire end which could be marked panel lamps for instance the wire is unwound until the other end of the wire is found. This would then be the panel lamp secondary winding. The number of turns are not counted but a note is made of the wire gauge (Thickness) and also that this is the outer coil. The same procedure is followed for successive coils each coil only having two connections these being the finish and start.
Once the valve heater winding is found I unwind this very carefully accurately counting the number of turns; I can do this using the counter on the coil winder but it can be easily done counting each turn as it is removed; on a vintage radio there shouldn’t be many turns. This time I note the number of turns as well as the wire gauge. Hopefully the panel lamps and valve heaters will have their own separate windings and might be separated by the Rectifier windings. Rectifier windings could be center tapped so this too will have to be noted together with the wire gauge. Up to now all wire gauges will have been noted together with the position of each coil working from the outside. The only counting of turns has been done on the valve heater winding.
Winding on manual AVO winder.
Once all the secondary coils are removed the troublesome primary coil will be revealed. Again I note the wire gauge. A sketch will also have been drawn showing clearly the positions of the tails (connection wires) and which side of the transformer they emerge from.
All this sounds much more complicated than it actually is and it will have taken me longer to think and write these notes than to actually do the unwinding.
Once the primary is located and the details noted the wire is removed without counting the turns; this alone saves a lot of time and I can work all the winding details out from the valve heater turns.
With the old wire removed I can now work out the rewinding details with ease. For simplicity and using very basic calculations I’ll now take the mains transformer that I have just unwound as an example. Service sheets are a great help and I’ll look up the valve heater voltages of the winding I removed which I counted the number of turns from. Say there were 24 turns according to my notes. The service sheets (or valve data book) give the valve heaters as 6V. Divide the number of turns which is 24 by the heater volts of 6 and this gives the turn’s ratio of all the other windings which is 4/1.
To expand on this; the ratio of 4 to 1 means for every four turns on the primary winding will give 1 volt on the secondary winding whether this winding is for the valve heaters panel lamps or rectifier; it is a fixed ratio and by knowing this ratio virtually any voltages can be obtained within the limits of the laminations. The turn’s ratio is very important and I used to spend hours unwinding counting every turn of every winding until the penny dropped and I realized I could establish the ratio using the above method.
Now to work out the various coils that needs winding. First there will be the primary; unless authenticity is required on an expensive set then I simply wind for a single 240V supply (UK) eliminating any tappings thus saving a lot of extra work and joints. Taking the ratio of 4/1; the primary will need 960 turns of the wire gauge noted for “primary”. (240V x 4 = 960 turns).
The valve heaters will need 24 turns of the wire gauge noted for “Valve heaters”. (6V x 4 = 24 turns).
The panel lamps if 6V will need 24 turns of the wire gauge marked “Panel lamps” (6V x 4 = 24 turns).
The Rectifier winding will need a bit more care because it might be center tapped. If it is a simple single winding and the voltage needed is 350V then 1,400 turns will be needed of the wire gauge marked “Rectifier”. (350V x 4 = 1,400 turns). If this winding is center tapped then be careful because the number of turns need to be doubled giving 2,800 turns of the wire gauge marked “Rectifier”. This will give 700 volts across both outer ends but from the middle (Center tap) to each end will give the correct 350V. (700V x 4 = 2,800 turns with the tap at 1,400 turns). I found this out the hard way after rewinding my first radio mains transformer; I wound a single coil for 350 volts which the Rectifier valve needed adding the centre tap but I didn’t fully understand the action of the center tap until I wired the transformer into the chassis and powered up; the voltage reading on each of the Rectifier pins was exactly 175V. A very simple mistake to make but a mistake only made once. (350V x 4 = 1,400 turns; adding the centre tap dropped each side of the tap to 700 turns hence half the required voltage. The correct 2,800 turns with the centre tap added dropped each side of the tap to 1,400 turns giving the required 350V at each side of the tap).
AVO Douglas manual and Aumann electric winders.
Once the basics are grasped then any transformer can be wound. There is plenty of data around giving enameled wire gauges and maximum current. The output will need to be known in Watts and with a few simple calculations that don’t even tax me the wire gauges and lamination sizes can be worked out. It only took a little practice for me to wind transformers to my own specifications. It really is so simple.
Once the bare bobbin is exposed then I would never ever try to rewind if the bobbin was in the form of a tube without cheeks or flanges fitted. To do so is asking for the newly wound wire coil to fall apart at the ends. I now add a pair of tightly fitted glued cheeks. These cheeks can be thin Tufnol/Paxolin sheet or even cut from a cereal packet but during winding they must be supported from the outside of the bobbin. I have also used stout card and cereal box to make small bobbins with great success; once assembled and glued the bobbin is given a few coats of shellac which dries rapidly. The bobbin is then mounted into the winder using a threaded drive rod with two nuts and large wooden washers which are tightened sufficiently to nip the bobbin thereby affording it drive without slipping. The cheeks on the bobbin and the securing wooden washers will need holes for the tails to exit.
New home made former from Tufnol sheet.
I have designed my own transformers and wound them previously and have bought the new laminations from a local company. I simply did the basic math’s and took along the total watts and bought laminations to suit and whilst there asked the turns ratio of the laminations.
Inserting laminations into a LOPT.
I think knowing what I know about transformer winding that being charged £100 for a rewind is reasonable; the cost isn’t justified by the wire used although the wire is expensive but in the large amount of time involved in doing the work; it is the labour being paid for and I know I can spend a day re-winding a single output transformer so looking at it like this £100 these days isn’t a lot of money. With practice transformer winding becomes quicker especially if everything is set up with correct gauge wires to hand but the labour is still needed to strip and rebuild the transformer. I wouldn’t want to do this work as a full time job.
I’ve gone on at length in the hope of explaining in as much detail as I could how to wind a transformer at home. Mains transformers would appear to be extremely complicated to a novice but in reality all they have is more windings than a simple output transformer. Chokes are even easier to rewind being just a simple single winding.
http://www.youtube.com/watch?NR=1&feature=endscreen&v=3tkSrbQGv5Y
http://www.youtube.com/watch?v=Lt0uHyzr4C0&feature=endscreen&NR=1
There are many ways to make a simple coil winder and a counter is needed to accurately count the number of turns. These counters can be very complicated as in electronic counters or very simple which suits our needs in the form of manual counters.
One simple trick to get a working transformer whilst not having a counter is to unwind the original transformer noting the secondary turns which are usually few but for the primary with many turns to carefully count the turns from the outer layer only taking note. Now simply count the number of layers as the winding is removed from the bobbin. It is important not to lose count of the layers. To rewind using the same gauge enameled copper wire wind on the number of layers of wire previously counted taking care to do this neatly; the number of turns will be near enough for this kind of job.
Two rewound bobbins ready for laminations.
Perhaps my chum Gary can be persuaded to add his own excellent design of coil winder for the more ambitious experimenter.
Ed Dinning has very kindly given me permission to include his detailed notes on transformer design which goes into much greater technical detail than my notes. Ed is a top guy on transformers and I am sure Ed will have forgotten more about transformers than I’ll ever know. Thank you Ed and the notes are as follows in their entirety.
Kind regards, Col.
TRANSFORMER DESIGN, CONSTRUCTION & THEORY
Presented here is an overview of transformers for hobbyists that want to expand their knowledge of this essential electrical component.
Of the 3 basic passive electrical components, R, L, C, the transformer is simply a special case of the inductor with either taps or multiple windings that are coupled.
It will operate over the full range of frequencies and may be either iron or air cored; most of the following discussions will concentrate on low frequency, iron cored transformers.
The basic transformer to be discussed is the voltage (or power transformer), which operates at mains frequency and is step up or down. There are also special cases, such as current transformers, pulse transformers and audio (output and driver) types that cater for a wider band of frequency.
3 phase transformers are simply a special case of the single-phase transformer with the core arranged to give a balanced magnetic circuit.
Transformers consist of a core and windings; the core is usually an iron alloy to suit the application and the windings consist of coils of insulated copper or aluminium wire.
There is no absolutely correct design for any given transformer; as in all matters of good engineering there are many design and cost compromises to be made.
Having selected a core and decided on a flux density, the T/V figure is calculated and then it is a matter of seeing if the turns will fit on the bobbin available for the selected core.
THE CORE
All transformers follow the basic transformer equation and work by reason of the fact that the magnetic flux, produced by the primary applied voltage is constantly changing. This is basic magnetic theory. It appears confusing as it is taught in many systems of units. I will concentrate on the MKS system where the units are Metres. DO NOT USE mm!!!
This is a practical system where the unit of flux density is the TESLA. Normal transformers operate at levels around 1T depending on the materials used.
Transformer equation:
N = E / 4.44 B F Ae Where:
N = turns
E = applied voltage in volts
4.44 is a constant for sine waves
B = desired flux density in Tesla
F = frequency in Hz
Ae = transformer centre limb area in M²
Stalloy or silicon iron, which is widely used as a core material is normally operated at a flux density of 1T. Unisil, a grain orientated material, which is a little more expensive can be run at 1.5 T. Special alloys for Mil and aerospace applications can be run up to 2T where a compact design is a requirement.
Running at a higher flux density allows the turns/ volt to be reduced and the cost of copper and copper losses; however this increases the losses in the iron and increases the magnetising (idling) current. Design curves are available showing watts loss per Kg at various flux densities.
Traditional design theory would make copper and iron losses equal at normal load levels (say 80% of full load)
An important fact of life with iron cores is Saturation. When we increase the voltage across an iron cored coil the current will initially increase in a linear manner. When we reach a certain level the current will start to increase much more rapidly than the voltage. This is the “Knee point” where a lot of transformers are designed to operate.
Note that Stalloy has a fairly rapid turn into saturation, Unisil much less so. Hence lower distortion when used in the output transformer of a Hi-Fi amplifier.
To determine the operating level of an unknown core it is simply a matter if winding a know number of turns onto the core, and then plotting a V/A curve. At the point at where it starts to turn into saturation will be the operating level, and it is then easy to read off the Turns/Volt from the plot.
Other core losses are determined by the circulation of eddy currents in the laminations. Thinner lams and lower frequencies give lower losses, but increase the cost of the lams. A solid core would have very high losses, hence the lams are lightly oxidised to give insulation between them. These losses are frequency dependant.
Most laminations available these days are “lossless” types.
This means that they are “E” & “I” shapes stamped out of sheet so as to leave no waste. Many other forms will be encountered in vintage equipment.
Transformers should normally be laminated to leave as small a gap between lams as possible, interleaving and tapping them together during assembly. The exception to this is transformers carrying DC, as in Class A valve output stages, where the standing DC current could cause saturation and a small air gap, usually made from thin paper is inserted between the “E” & “I” lams that are not now interleaved.
Clamping bolts, when inserted through the lams should have an insulation washer between the lam face and the nuts. If this is not done there can be high currents circulating through the bolt.
Toroidal transformer cores can be run at higher flux densities and have the advantage of a much lower stray field, useful in audio applications
THE WINDINGS
These are usually of enamelled copper wire but various other insulations are available. The insulation should be as thin as possible consistent with a suitable electrical withstand level. Windings often had a thin layer of paper between layers of winding. With present wire insulation this is not necessary every layer, and is often replaced with Type 56 Polyester tape every few layers, again depending on application.
The normal wire current density is typically 3A per mm² and this figure is often shown in wire tables.
Note that wire sizes, in SWG, AWG or mm are based on bare wire, not coated diameter.
The current density figure is based on a max internal temperature for the transformer of 120°C; the limit for normal materials, other design factors may dictate different values.
Wire tables will also give the resistance/ Mtr of wire. This is used to calculate the winding resistance from the mean length of turn. Note that copper resistivity increases by 0.3% per °C rise.
The transformer equation gives the primary turns; secondaries can then be calculated by the turns ratio. If very exact ratios are required it may be necessary to increase the primary turns so as to get an integral number of turns on the secondary of importance.
If an exact voltage is required it is often necessary to add compensating turns to a secondary to allow for the volt drops (calculated from mean length of turn and resistance per mtr) in both primary and secondary.
It can also be necessary to allow for the increase of resistance due to temperature rise as well.
Further compensation is sometimes necessary to allow for the imperfect coupling between primary and secondary which appears as a parasitic loss.
Primaries and secondaries are normally wound on top of each other for good coupling; where additional safety separation is required, they may be wound side by side with a centre insulated barrier (some transformer kits with pre-wound primaries are like this).
This gives inferior coupling and additional compensating turns will be needed to compensate for this.
For very close coupling the primary and secondary will be wound together with Bi-filar wire. Where insulation levels make this impossible, the primary and secondary are wound in several sections, one on top of the other. For balance this would usually be one more primary section than secondary section.
Wire is commonly insulated with synthetic enamels in either 1 or 2 coats. Most modern enamels can be vaporised with solder and are known as solderable types.
Where high levels of insulation are required in a small space (such as a switch mode transformer), then triple insulated wire can be used. This is certified for supply to low voltage windings without additional insulation.
An earthed copper screen is often placed between the primary and secondary; this is for both safety and as an interference screen against noise impulses.
Windings are not normally carried to the outside of the bobbin on multi-layer windings due to the danger of a turn slipping down the side of the bobbin and “seeing” a higher voltage, possibly causing breakdown. Paper margin tapes, a few mm wide are often used here.
With EHT transformers this margin is often tapered inwards (wider) as the outer layers are added and the winding voltage to earth increases.
Copper is the important part of the winding so thin insulation is used, as there is only a relatively low voltage between turns and layers. This is usually capable of withstanding a minimum of 120ºC.
The windings are often vacuum impregnated to seal against moisture and prevent chafing movement under operating or fault conditions. It is possible to spray varnish them while building, or they can be paraffin wax impregnated in a container of molten wax.
For voltages above 6KV it is usual to seal the transformer in a can of specially refined mineral oil.
Tapping points on windings can be “brought out” of the body of the winding by looping the wire out and back again. Lead-out wires are often soldered on to the winding at an appropriate point. These must be well insulated.
The forgoing should allow a reasonably competent person to design and build their own transformers when size and cost are not the driving factors as in commercial applications. It should always be remembered that these devices are connected to the mains and can easily kill or start serious fires; all possible safety precautions should be taken when using them.
© Ed Dinning 2009
I hope the following information will be useful to members wishing to re-wind transformers. I can only write from personal experience from winding my own chokes and transformers as I’m completely self taught but it works for me.
Some form of rotational drive will be needed to turn the bobbin (Former/Winding Spool). I now own four coil winders but have previously used a variable speed wood turning lathe; the lathe was way over the top but I did wind a Murphy A30C field winding coil using it in fact I wound two and these take a tremendous amount of enameled copper wire. The first one I wound as a complete novice went well until almost completed then the lot collapsed resulting in the waste of a lot of wire as it all tangled badly. I’ll explain how to avoid this later.
There are many ways to mount a bobbin and get it to revolve; a simple hand crank will suffice for a one off but will be slow work; it would be better if it was geared up to rotate the bobbin in a ratio of say 10 turns of the bobbin to one turn of the crank. Simple shafting and a large and small wooden pulley would achieve this with a bit of imagination and effort. A variable speed electric drill mounted into a bench stand would be a simple way of revolving the bobbin and controlling the speed.
Take the electric drill as an example. Set it up in a bench stand either bought or home made. If the chuck will accept up to 10mm then buy a length of 10mm threaded rod with a few nuts and washers. The threaded rod can be cut to length with a hacksaw and rough edges filed smooth. This is a crude way but with care and patience should give a working coil winder.
The open circuit choke or transformer will need completely stripping right down to the bare bobbin (Former). If it is a transformer covered in black pitch which I hate with a passion; the pitch will have to be removed and a simple but messy way is to bake it in an oven until the pitch has melted and dropped clear; to aid this it pays to elevate the bobbin on something like a baking tin; set the oven around 160C as a start and see how it goes. I’ve even used a blowlamp working in the garage. Removing the pitch is one job I do not like doing.
Next job is to remove the laminations; if it is a largish transformer then the through bolts and nuts will need removing first. Now the fun begins because usually the laminations will be a tight fit inside the bobbin; this might be compounded by rust making them extremely difficult to remove. Grabbing outer laminations with pliers will ruin and distort them rendering them useless. “E” and “I” laminations are easiest to remove but the “U” and “T” can be a problem; if you come across one piece laminations as I did then these are a downright pain to remove.
LOPT laminations.
For “U” and “T” laminations; gently nip the transformer winding in a vice with the laminations lying horizontally. Locate the lamination joints and using a wide screwdriver and small hammer gently tap at the joints and try to separate them; work from side to side in order to drift a single lamination out; once it has moved and there is enough to grip with pliers then it should pull free; once the first three laminations are out the others will follow much easier. I find a craft knife helps to separate each lamination but care is needed to prevent a finger being cut. Stack the laminations neatly on the bench in the order they were removed. Now the bobbin will be free. If you have some old laminations to hand a piece cut from one of these can be used with as a drift.
Remove the outer insulation from the bobbin and this will reveal the top layer of the last coil usually the secondary. This will normally be thicker wire than the primary coil and there will be fewer turns of it. Two things are now very important and these are number of turns and thickness of wire (Wire Gauge).
Stripping insulation.
Reading books on transformer design is sure to scare a novice into not attempting trying to re-wind one but I found a way by trial and error of obtaining the information. I wish someone would have told me rather than me having to find out the hard way by unwinding what seems like miles of wire and counting the turns.
Simple output transformers with a primary and single secondary are dead easy to re-wind. Mount the bobbin in the drill leaving both hands free. Remove the outer insulation to reveal the outer layer of secondary turns and locate the end of the wire (Finish). Now be very careful and take a note of how many turns are on each layer either by counting directly across the layer if the wire is thick enough or by unwinding; carry on unwinding and noting the number of turns until the other end of the wire is found (Start). You now have the number of turns of wire for the secondary. Now for the thinner wire of the primary winding. This can be a real pain to count as there can be 3,000 turns of extremely fine wire on such a simple output transformer as for a DAC 90 if my memory is correct. Add to this sticky interleave insulation (insulation between each layer) and trying to count without losing the plot quickly becomes a nightmare. To complicate matters even further the wire can break at a blink and once the open circuit is located the wire could break many times causing a great deal of frustration.
There is a simpler way to determine the primary turns and I need to find the formula once again and I’ll add it at the end of the thread; it involves some very simple math’s that even I can understand; using this formula will give the turns ratio; I think the last output transformer I rewound had a ratio of 48/1 so by having already counted the secondary turns it is easy to determine the primary turns and for me this was the hardest part in re-winding as the actual winding is straightforward.
Establishing the wire gauge would sound to be difficult? In fact it is incredibly simple; measure the original wire accurately using a micrometer. I always measure over the enamel as this is the thickest wire that the lamination window will accept. Using very slightly thicker wire would give a more robust transformer but multiply this extra thickness many times and the bobbin won’t fit into the laminations. As modern enamel is much improved it is possible to wind whilst leaving out the inter-layer insulation this will allow very slightly thicker wire to be used and I’ve got away using this method in the past. Normally I will use the yellow transformer self adhesive insulation tape which can be bought through eBay adding a layer of tape to every third layer of wire. A bit of practice helps here.
Even a small transformer uses a lot of wire.
Whilst I’m on turn’s ratio I’ll try to explain the absolute basics of transformer design. A transformer is actually very simple to understand for a novice. A transformer consists of a number of parts the main parts being the laminations and the winding wire; add to these the bobbin (former) insulation and lead out wires. Given the few parts involved what can be complicated? The turn’s ratio is very important and determined by the laminations; too few turns will result in a poor performance whilst too many will saturated the laminations.
I once put together a home turned wooden pulley with exactly 12” circumference at the bottom of its groove mounted between centers and used this to measure the length of wire removed from my first transformer. I didn’t understand that it was the number of turns and not the length of the wire that was important. I find it wonderful that I can fall into every black hole and make every mistake possible whilst learning something new.
Because we are discussing re-winding then all the necessary details are to hand if we know what to look for. I always thought winding radio mains transformers would be extremely difficult given the number of secondary windings and tappings. On the ones I’ve re-wound the primary has always been open circuit and as the radios are for use with a number of different voltages; tappings are added which at first would appear difficult for a novice to understand; I ignore the mains tappings and wind a single primary coil to work from 240V. That’s one big problem out of the way to start with.
Before removing the transformer from the chassis I mark every connection and pay particular attention to the heater connections and which valves they are connected to; I write this down and double check as a mistake could quickly lead to a lot of frustration. A clear sketch of the transformer is drawn; it’s mounting position and above all its connections; time is taken over this and it is done before breaking the connections. A digital camera is used to take close up pictures both above and below chassis.
Now for the important turns ratio. Having already established what the connections are I look for the valve heater connections which I have clearly identified on the transformer? The bobbin is then mounted into the coil winder and the outer insulation removed; taking careful note of the first wire end which could be marked panel lamps for instance the wire is unwound until the other end of the wire is found. This would then be the panel lamp secondary winding. The number of turns are not counted but a note is made of the wire gauge (Thickness) and also that this is the outer coil. The same procedure is followed for successive coils each coil only having two connections these being the finish and start.
Once the valve heater winding is found I unwind this very carefully accurately counting the number of turns; I can do this using the counter on the coil winder but it can be easily done counting each turn as it is removed; on a vintage radio there shouldn’t be many turns. This time I note the number of turns as well as the wire gauge. Hopefully the panel lamps and valve heaters will have their own separate windings and might be separated by the Rectifier windings. Rectifier windings could be center tapped so this too will have to be noted together with the wire gauge. Up to now all wire gauges will have been noted together with the position of each coil working from the outside. The only counting of turns has been done on the valve heater winding.
Winding on manual AVO winder.
Once all the secondary coils are removed the troublesome primary coil will be revealed. Again I note the wire gauge. A sketch will also have been drawn showing clearly the positions of the tails (connection wires) and which side of the transformer they emerge from.
All this sounds much more complicated than it actually is and it will have taken me longer to think and write these notes than to actually do the unwinding.
Once the primary is located and the details noted the wire is removed without counting the turns; this alone saves a lot of time and I can work all the winding details out from the valve heater turns.
With the old wire removed I can now work out the rewinding details with ease. For simplicity and using very basic calculations I’ll now take the mains transformer that I have just unwound as an example. Service sheets are a great help and I’ll look up the valve heater voltages of the winding I removed which I counted the number of turns from. Say there were 24 turns according to my notes. The service sheets (or valve data book) give the valve heaters as 6V. Divide the number of turns which is 24 by the heater volts of 6 and this gives the turn’s ratio of all the other windings which is 4/1.
To expand on this; the ratio of 4 to 1 means for every four turns on the primary winding will give 1 volt on the secondary winding whether this winding is for the valve heaters panel lamps or rectifier; it is a fixed ratio and by knowing this ratio virtually any voltages can be obtained within the limits of the laminations. The turn’s ratio is very important and I used to spend hours unwinding counting every turn of every winding until the penny dropped and I realized I could establish the ratio using the above method.
Now to work out the various coils that needs winding. First there will be the primary; unless authenticity is required on an expensive set then I simply wind for a single 240V supply (UK) eliminating any tappings thus saving a lot of extra work and joints. Taking the ratio of 4/1; the primary will need 960 turns of the wire gauge noted for “primary”. (240V x 4 = 960 turns).
The valve heaters will need 24 turns of the wire gauge noted for “Valve heaters”. (6V x 4 = 24 turns).
The panel lamps if 6V will need 24 turns of the wire gauge marked “Panel lamps” (6V x 4 = 24 turns).
The Rectifier winding will need a bit more care because it might be center tapped. If it is a simple single winding and the voltage needed is 350V then 1,400 turns will be needed of the wire gauge marked “Rectifier”. (350V x 4 = 1,400 turns). If this winding is center tapped then be careful because the number of turns need to be doubled giving 2,800 turns of the wire gauge marked “Rectifier”. This will give 700 volts across both outer ends but from the middle (Center tap) to each end will give the correct 350V. (700V x 4 = 2,800 turns with the tap at 1,400 turns). I found this out the hard way after rewinding my first radio mains transformer; I wound a single coil for 350 volts which the Rectifier valve needed adding the centre tap but I didn’t fully understand the action of the center tap until I wired the transformer into the chassis and powered up; the voltage reading on each of the Rectifier pins was exactly 175V. A very simple mistake to make but a mistake only made once. (350V x 4 = 1,400 turns; adding the centre tap dropped each side of the tap to 700 turns hence half the required voltage. The correct 2,800 turns with the centre tap added dropped each side of the tap to 1,400 turns giving the required 350V at each side of the tap).
AVO Douglas manual and Aumann electric winders.
Once the basics are grasped then any transformer can be wound. There is plenty of data around giving enameled wire gauges and maximum current. The output will need to be known in Watts and with a few simple calculations that don’t even tax me the wire gauges and lamination sizes can be worked out. It only took a little practice for me to wind transformers to my own specifications. It really is so simple.
Once the bare bobbin is exposed then I would never ever try to rewind if the bobbin was in the form of a tube without cheeks or flanges fitted. To do so is asking for the newly wound wire coil to fall apart at the ends. I now add a pair of tightly fitted glued cheeks. These cheeks can be thin Tufnol/Paxolin sheet or even cut from a cereal packet but during winding they must be supported from the outside of the bobbin. I have also used stout card and cereal box to make small bobbins with great success; once assembled and glued the bobbin is given a few coats of shellac which dries rapidly. The bobbin is then mounted into the winder using a threaded drive rod with two nuts and large wooden washers which are tightened sufficiently to nip the bobbin thereby affording it drive without slipping. The cheeks on the bobbin and the securing wooden washers will need holes for the tails to exit.
New home made former from Tufnol sheet.
I have designed my own transformers and wound them previously and have bought the new laminations from a local company. I simply did the basic math’s and took along the total watts and bought laminations to suit and whilst there asked the turns ratio of the laminations.
Inserting laminations into a LOPT.
I think knowing what I know about transformer winding that being charged £100 for a rewind is reasonable; the cost isn’t justified by the wire used although the wire is expensive but in the large amount of time involved in doing the work; it is the labour being paid for and I know I can spend a day re-winding a single output transformer so looking at it like this £100 these days isn’t a lot of money. With practice transformer winding becomes quicker especially if everything is set up with correct gauge wires to hand but the labour is still needed to strip and rebuild the transformer. I wouldn’t want to do this work as a full time job.
I’ve gone on at length in the hope of explaining in as much detail as I could how to wind a transformer at home. Mains transformers would appear to be extremely complicated to a novice but in reality all they have is more windings than a simple output transformer. Chokes are even easier to rewind being just a simple single winding.
http://www.youtube.com/watch?NR=1&feature=endscreen&v=3tkSrbQGv5Y
http://www.youtube.com/watch?v=Lt0uHyzr4C0&feature=endscreen&NR=1
There are many ways to make a simple coil winder and a counter is needed to accurately count the number of turns. These counters can be very complicated as in electronic counters or very simple which suits our needs in the form of manual counters.
One simple trick to get a working transformer whilst not having a counter is to unwind the original transformer noting the secondary turns which are usually few but for the primary with many turns to carefully count the turns from the outer layer only taking note. Now simply count the number of layers as the winding is removed from the bobbin. It is important not to lose count of the layers. To rewind using the same gauge enameled copper wire wind on the number of layers of wire previously counted taking care to do this neatly; the number of turns will be near enough for this kind of job.
Two rewound bobbins ready for laminations.
Perhaps my chum Gary can be persuaded to add his own excellent design of coil winder for the more ambitious experimenter.
Ed Dinning has very kindly given me permission to include his detailed notes on transformer design which goes into much greater technical detail than my notes. Ed is a top guy on transformers and I am sure Ed will have forgotten more about transformers than I’ll ever know. Thank you Ed and the notes are as follows in their entirety.
Kind regards, Col.
TRANSFORMER DESIGN, CONSTRUCTION & THEORY
Presented here is an overview of transformers for hobbyists that want to expand their knowledge of this essential electrical component.
Of the 3 basic passive electrical components, R, L, C, the transformer is simply a special case of the inductor with either taps or multiple windings that are coupled.
It will operate over the full range of frequencies and may be either iron or air cored; most of the following discussions will concentrate on low frequency, iron cored transformers.
The basic transformer to be discussed is the voltage (or power transformer), which operates at mains frequency and is step up or down. There are also special cases, such as current transformers, pulse transformers and audio (output and driver) types that cater for a wider band of frequency.
3 phase transformers are simply a special case of the single-phase transformer with the core arranged to give a balanced magnetic circuit.
Transformers consist of a core and windings; the core is usually an iron alloy to suit the application and the windings consist of coils of insulated copper or aluminium wire.
There is no absolutely correct design for any given transformer; as in all matters of good engineering there are many design and cost compromises to be made.
Having selected a core and decided on a flux density, the T/V figure is calculated and then it is a matter of seeing if the turns will fit on the bobbin available for the selected core.
THE CORE
All transformers follow the basic transformer equation and work by reason of the fact that the magnetic flux, produced by the primary applied voltage is constantly changing. This is basic magnetic theory. It appears confusing as it is taught in many systems of units. I will concentrate on the MKS system where the units are Metres. DO NOT USE mm!!!
This is a practical system where the unit of flux density is the TESLA. Normal transformers operate at levels around 1T depending on the materials used.
Transformer equation:
N = E / 4.44 B F Ae Where:
N = turns
E = applied voltage in volts
4.44 is a constant for sine waves
B = desired flux density in Tesla
F = frequency in Hz
Ae = transformer centre limb area in M²
Stalloy or silicon iron, which is widely used as a core material is normally operated at a flux density of 1T. Unisil, a grain orientated material, which is a little more expensive can be run at 1.5 T. Special alloys for Mil and aerospace applications can be run up to 2T where a compact design is a requirement.
Running at a higher flux density allows the turns/ volt to be reduced and the cost of copper and copper losses; however this increases the losses in the iron and increases the magnetising (idling) current. Design curves are available showing watts loss per Kg at various flux densities.
Traditional design theory would make copper and iron losses equal at normal load levels (say 80% of full load)
An important fact of life with iron cores is Saturation. When we increase the voltage across an iron cored coil the current will initially increase in a linear manner. When we reach a certain level the current will start to increase much more rapidly than the voltage. This is the “Knee point” where a lot of transformers are designed to operate.
Note that Stalloy has a fairly rapid turn into saturation, Unisil much less so. Hence lower distortion when used in the output transformer of a Hi-Fi amplifier.
To determine the operating level of an unknown core it is simply a matter if winding a know number of turns onto the core, and then plotting a V/A curve. At the point at where it starts to turn into saturation will be the operating level, and it is then easy to read off the Turns/Volt from the plot.
Other core losses are determined by the circulation of eddy currents in the laminations. Thinner lams and lower frequencies give lower losses, but increase the cost of the lams. A solid core would have very high losses, hence the lams are lightly oxidised to give insulation between them. These losses are frequency dependant.
Most laminations available these days are “lossless” types.
This means that they are “E” & “I” shapes stamped out of sheet so as to leave no waste. Many other forms will be encountered in vintage equipment.
Transformers should normally be laminated to leave as small a gap between lams as possible, interleaving and tapping them together during assembly. The exception to this is transformers carrying DC, as in Class A valve output stages, where the standing DC current could cause saturation and a small air gap, usually made from thin paper is inserted between the “E” & “I” lams that are not now interleaved.
Clamping bolts, when inserted through the lams should have an insulation washer between the lam face and the nuts. If this is not done there can be high currents circulating through the bolt.
Toroidal transformer cores can be run at higher flux densities and have the advantage of a much lower stray field, useful in audio applications
THE WINDINGS
These are usually of enamelled copper wire but various other insulations are available. The insulation should be as thin as possible consistent with a suitable electrical withstand level. Windings often had a thin layer of paper between layers of winding. With present wire insulation this is not necessary every layer, and is often replaced with Type 56 Polyester tape every few layers, again depending on application.
The normal wire current density is typically 3A per mm² and this figure is often shown in wire tables.
Note that wire sizes, in SWG, AWG or mm are based on bare wire, not coated diameter.
The current density figure is based on a max internal temperature for the transformer of 120°C; the limit for normal materials, other design factors may dictate different values.
Wire tables will also give the resistance/ Mtr of wire. This is used to calculate the winding resistance from the mean length of turn. Note that copper resistivity increases by 0.3% per °C rise.
The transformer equation gives the primary turns; secondaries can then be calculated by the turns ratio. If very exact ratios are required it may be necessary to increase the primary turns so as to get an integral number of turns on the secondary of importance.
If an exact voltage is required it is often necessary to add compensating turns to a secondary to allow for the volt drops (calculated from mean length of turn and resistance per mtr) in both primary and secondary.
It can also be necessary to allow for the increase of resistance due to temperature rise as well.
Further compensation is sometimes necessary to allow for the imperfect coupling between primary and secondary which appears as a parasitic loss.
Primaries and secondaries are normally wound on top of each other for good coupling; where additional safety separation is required, they may be wound side by side with a centre insulated barrier (some transformer kits with pre-wound primaries are like this).
This gives inferior coupling and additional compensating turns will be needed to compensate for this.
For very close coupling the primary and secondary will be wound together with Bi-filar wire. Where insulation levels make this impossible, the primary and secondary are wound in several sections, one on top of the other. For balance this would usually be one more primary section than secondary section.
Wire is commonly insulated with synthetic enamels in either 1 or 2 coats. Most modern enamels can be vaporised with solder and are known as solderable types.
Where high levels of insulation are required in a small space (such as a switch mode transformer), then triple insulated wire can be used. This is certified for supply to low voltage windings without additional insulation.
An earthed copper screen is often placed between the primary and secondary; this is for both safety and as an interference screen against noise impulses.
Windings are not normally carried to the outside of the bobbin on multi-layer windings due to the danger of a turn slipping down the side of the bobbin and “seeing” a higher voltage, possibly causing breakdown. Paper margin tapes, a few mm wide are often used here.
With EHT transformers this margin is often tapered inwards (wider) as the outer layers are added and the winding voltage to earth increases.
Copper is the important part of the winding so thin insulation is used, as there is only a relatively low voltage between turns and layers. This is usually capable of withstanding a minimum of 120ºC.
The windings are often vacuum impregnated to seal against moisture and prevent chafing movement under operating or fault conditions. It is possible to spray varnish them while building, or they can be paraffin wax impregnated in a container of molten wax.
For voltages above 6KV it is usual to seal the transformer in a can of specially refined mineral oil.
Tapping points on windings can be “brought out” of the body of the winding by looping the wire out and back again. Lead-out wires are often soldered on to the winding at an appropriate point. These must be well insulated.
The forgoing should allow a reasonably competent person to design and build their own transformers when size and cost are not the driving factors as in commercial applications. It should always be remembered that these devices are connected to the mains and can easily kill or start serious fires; all possible safety precautions should be taken when using them.
© Ed Dinning 2009
Happiness is a wreck of a cabinet to restore.
