Toroidal PT and OT Suppliers

[K O'Connor]

Hi Guys

My own toroidal transformer guru retired a while ago and that has caused me to think about what direction I want to take with my amp line. His rates for custom power and audio transformers were very much on the low side and he did not penalise you for prototypes or low quantities.

Building preamps is not a problem as there are many suppliers with appropriate off-the-shelf products that are suitable. It is really the support of power amps that is tricky. You can find good output transformers but the PT designs I have seen all leave something to be desired.

Note that when using a toroidal OT it MUST be FULL-POWER-BANDWIDTH. You CANNOT use the OT to restrict the frequency range as it will overload in an extremely non-musical manner.

The nearly universal problem with toroidal PT offerings is that there is no proper bias winding.

In a tube power amplifier the bias supply is the most important voltage in the entire circuit.

As readers of TUT (The Ultimate Tone) know, a separate bias WINDING on the PT is preferred in the method for achieving lowest-noise and to properly support a Power Scale solution. A bias TAP on a center-tapped plate winding is workable, but far from ideal.

Toroidal transformers are nearly ideal in their functioning and because of this we have to specify them and use them a bit differently than all other transformer types. This mainly manifests itself in how we configure rectification inasmuch as half-wave rectification should be avoided at all cost. This means that a bias TAP is not a good idea AND that the basic CTed plate winding is also not appropriate unless you are in fact using a tube rectifier - there are other ways to incorporate a tube rectifier which use a non-CTed winding. As TUTs point out, the CTed plate winding is really just two half-wave circuits over-lapped and each winding has to be able to support the full load - a real waste of wire and winding space. The discontinuous conduction of half-wave rectifiers injects noise into the PT, which couples into all the other windings and back to the mains. The efficiency of toroids makes this a much worse situation than with traditional transformer designs.

So, toroids need to be specified with windings that require a full-wave full-bridge rectifier, usually referred to simply as a "bridge rectifier" and often attained as an integrated 4-pin package.  The winding is used to support the load continuously over the full AC cycle of the mains to minimise EMI. The plate winding should be like every other secondary inasmuch as it is a single winding with no CT.

The plate winding may have a tap to allow different maximum output voltage using either a switch or hard-wiring. As TUT4 illustrated, this tap will not provide an alternate but lower voltage at the same time as the full winding is being used, nor the reverse, where the full winding cannot provide a higher voltage while the tap is in use.

The capacitive-coupled bias supply some manufacturers use with the proper plate winding is clever in all the wrong ways.  It has a high-ish impedance and imposes a discontinuous load on the winding, injecting nose. It has all of the wrong characteristics to supply what is the most important voltage in the entire chassis and will not support a bias regulator. It is truly a false economy.

In my survey of toroidal PT offerings for tube amps, I see the CTed plate winding BUT no 5V heater winding for the presumed tube rectifier?  The use of such windings continued well past the reign of tube rectifiers, and this can only be due to lethargy on the part of both the amp builders and the transformer suppliers - an unspoken agreement. In any case, many of the toroidal manufacturers today are newcomers to the industry and it must be assumed they are simply unwittingly copying the mistakes of legacy manufacturers without realising it. Or, they have some gaps in their understanding about tube power amps?

Heater windings tend to be plentiful to over-abundant in modern toroidal PTs for tube amps. Most have at least two 6v3 windings with many providing a CT to at least one of those. The CT is fairly useless since a faux-CT works as well, and in audio the use of a DC-stand-off for the heater winding is inexpensive and provides as much hum rejection as  DC heaters but without all the trouble and wasted heat.

You have to remember that with a toroid EVERY winding must fully cover the core to help shield it, which means that it must be made with enough wire to wrap around the full circle of the toroid and end right beside where it began. A CTed winding is in fact two separate windings that each must wrap all the way around the core, and then have the appropriate leads tied together and brought out as a center-tap. Doing this for low-voltage windings imposes some difficulty upon the manufacturer. The wire lead-outs impose unavoidable gaps in the shielding, so minimising the number of these is helpful in reducing EMI.

TUT4 presented London Power's 6/12 heater system. This uses a 12V CTed winding to supply a mix of 6V and 12V heaters. For lowest-noise from wiring, using 12V for preamp tubes that accommodate it allows the triode leads to never have to cross the heater wiring. We can do this in both hand-wired and PCB layouts. In a power amp, the 6V (octal) tube heaters can be distributed on either side of the CT (and with separate wire runs can use #22 wire). These loads do not have to be equal about the CT for the reasons cited above. With many of the tube amp toroidal PTs on offer, the two 6V windings can be wired this way and tied to a DC-stand-off. If one of the 6V windings has a CT we simply do not use the CT.

So, most of the available toroidal PTs for tube amps have no bias winding and have no 5V winding to support a tube rectifier. The latter is not really a problem for most builders. To have a proper bias supply, we end up adding an auxiliary supply of one form or another to create a bias supply of sufficient voltage and impedance.

In the US, Antek has a range of PTs and OTs
https://www.antekinc.com/transformers/

In Poland, Toroidy has PTs and OTs, serving all of Europe
https://sklep.toroidy.pl/en_US/index

There are certainly many other companies providing off-the-shelf devices for hobbyists and builders.

Note that the information for each device is often scanty and there is no "match-up" suggestions of PT + OT.

There are often a range of finishes from standard tape-wrapped, to available cans, and then potted centers or potted in a can.

[K O'Connor]

Hi Guys

With some of the available toroidal PTs - and maybe some traditional types - it may be necessary to verify the windings and leadouts as the documentation is sometimes minimal.

The first step is to use an ohmmeter to "ring out" the windings. It is handy to use an alligator clip test lead to connect one meter lead to a given winding lead, then both hands are free to probe other leads to see if they connect to the reference wire.

When you find leads that are related, twist them together and flare out the stripped ends so they do not short to each other.

Once you have found all of the related lead sets, make sure none of them are touching each other.

Use a Power Limiting Safety Socket (Mandatory Project in TOT Tonnes of Tone) with the lowest-wattage bulb you can find to power the PT. The primary or primaries of the PT should be specified and if it is a dual wire them in series. This will cut all the secondary voltages in half if you happen to be in North America or Japan. For the rest of the world, high-voltage plate windings will be near their full voltage rating so BE VERY CAREFUL.

With many transformers that have dual secondaries that are identical, the windings also have identical colours. This is fine if you simply wish to parallel them for maximum current, but quite useless for series-connection. You have to ring out the leads to find the related pairs THEN the colours are useful in that it should be easy to connect them for series-aiding (highest voltage from end to end of the series pair).

The above procedure "phases" the windings, where you can see if the voltages add together or cancel out. For example, suppose you have two 6V windings that each have two green wires. You ring them out to see which pairs of wires are for each winding, twisting the related wires together. Initially you simply take a random wire from each pair and link these together. Measure the voltage across the free ends and you will see either 12V or 0V (zero). If it is 12V, mark the free end of the first winding with some tape, or a marker. Mark the connected end of the second winding similarly.

Just to clarify, the phase between windings does not matter unless you need to interconnect them.

[K O'Connor]

Hi Guys

From our writings, the ideal power transformer for a tube power amp has these windings:

PRI-1: BLK (120) - WHT (0)
PRI-2: BRN (120) - ORG (0)
Electrostatic shield - GRN
SEC-1: RED (hi-V) - RED (0), appropriate current, plate
SEC-2: VIO (hi-V x 0.2 to 0.15) - VIO (0), 100mA bias
SEC-3: YLW (6) - WHT (0) - YLW (6), appropriate current

Note that the bias winding is not specified as a set voltage as it is on so many PTs for tube amps; instead it is proportioned to the high-voltage plate winding.

For example, say we need 500Vdc, we would want the raw bias supply to be between -100Vdc (hi-V x 0.2) and -75Vdc (hi-V x 0.15) to accommodate every tube sample from low-gain to high-gain. This would also accommodate using a bias regulator if the amp is to be Power Scaled. In absolute voltage terms, when the bias supply is 20% of the plate supply there should be no tube that is uncontrollable, where at 15% very few tubes remain uncontrollable.

Note that in the 6/12 heater system, the power tube heaters are distributed over each side of the CT and it does not matter if these loads are equal. Maximum current is based on the sum of individual heater loads.

For example, say there are 2x EL-34 and 3x 12AX7.
One EL-34 socket is wired to each side of the CT.
The three 9-pin socket are wired across the full 12V (pin-9 not used).
The load per side is equal to one EL34=1A6, plus 3 x 150mA = 450mA
The entire winding is then rated as 6-0-6V or 12Vcr at 2A05.

Some octals need 1A8 each for their heater, so we might spec the winding at 3A or even 4A. The hidden "magic" of this approach is that the heater winding is only rated for half the power tube complement plus the small-tube complement AND with fan-out wiring only needs small-gauge wire to each octal socket.

If there is a need to power relays or switching circuits, or other things, we can add an auxiliary winding having BLU wires. This may be as follows:
SEC-4: BLU (12) - BLU (0), 100mA, or whatever current is needed for the auxiliaries. 12Vac accommodates the use of a 7812 3-pin voltage regulator to give 12Vdc for relay coils, etc. The AUX winding and related DC output has no "polarity" until you ground one end or the other.

If you need split-rails for the auxiliaries, then a 24Vct winding is needed, as:
BLU(12) - WH (0) - BLU(12) at the needed current.

Of course, the aux windings can be other voltages as appropriate to your needs.

If the PT is for other power tube types, such as 300B, 2A5, etc, then appropriate heater windings are needed, where each winding is given its own colour