01-10-2012, 11:31 AM
Power transformer stress
Power transformer stress.pdf (Size: 206.35 KB / Downloads: 29)
Many of us, restorers or users of old radios in particular, would like to extend the life of power transformers in old radios.
One of the ways to do this is to reduce transformer heat. The heat in a transformer comes from losses. While there are
several sources of loss, the predominant loss is normally in the resistance of windings. With only casual thought we might
assume removing 10 watts of load removes 10 watts of heat or stress on the transformer. This cannot be correct because the
transformer loss would always have to equal load power...an impossible situation.
Let's consider a typical old power transformer. The primary has to supply power for all the secondary windings. The power
loss in the primary is a combination of the perfect power factor load caused by the filaments and the sometimes very high
power factor of a capacitor input filtering system of the high voltage supply. The most common modification is to remove a
vacuum tube rectifier and substitute solid state rectifiers. Typical filament power is 2 amps at 5 volts. We are often told this
removes 10 watts of power from the transformer.
Since the primary is sized to handle all of the power load any heating by the rectifier filament is very low. The primary
current caused by the rectifier filament load is only around 92 milliamperes. There is very little change in line current by
removal of the filament load. We can reasonably estimate transformer heat by measuring the filament winding voltage
change as the rectifier tube is switched to an external 5 volt supply. In the case of a NC-303 I measured a change from 5.15
to 5.43 volts. This means the equivalent secondary resistance of the transformer filament winding is (5.43-5.15) / 2 = 0.14
ohms. The 0.14 ohm resistance would drop .28 volts at 2 amperes. The actual loss resistance is a little less than this...but it
is close enough.
We can now determine transformer internal heat caused by the rectifier filament. It is simple I^2 R heating, so the heat is
2^2*.14= 0.56 watts. Removing ten watts of filament load actually removed only .56 watts of internal transformer heat.
A second factor actually INCREASES power transformer heat when the rectifier is replaced with solid state rectifiers. This
occurs because the solid state rectifier is a much harder switch. A solid state rectifier either fully conducts with a minimal
voltage drop when forward biased or it is fully off when reverse biased.
A vacuum tube rectifier is rather soft when in conduction. It can drop 20-30 volts or more and does not supply the extremes
of peak current supplied by a solid state rectifier. Substituting a solid state rectifier for a tube rectifier without adding a
suitable series resistance can significantly decrease power factor of the load presented by a capacitor input supply. (It
doesn't do much to a choke input supply, since it already has a very good power factor.)