Page 1 of 4 Power Supply Choke Measurement 5 March 2020

Old power supply chokes typically don't come with a rating

plate stating their inductance and DC current rating.
Manufacturers like Rola were an exception, with labelling such
as 14H at 60mA measured at 10V 100Hz, with 560Ω DC
resistance for the choke type 14/60.
Within a valve amplifier, the high voltage (HT) DC supply with
“choke input” filter applies a very large 100Hz AC voltage
across the choke which is much greater than 10V, and
contains substantial higher harmonic levels (one end of the
choke cycles from 0V to the peak of the AC supply, and the
other end of the choke is pinned to the HT DC voltage).
Whereas a smoothing choke application, where the choke is
connected between two capacitors, experiences a much lower
AC voltage (likely to be less than 10V at 100Hz, with relatively ROLA 14/60 choke
low harmonic levels).
The inductance of an iron-cored choke can vary significantly with applied AC voltage (ie. ac current), and
with the level of DC current passing through the choke. Results below, from [1], show those characteristics.
So it is important to compare choke ratings only when similar operating conditions are being applied, and to
be aware that the choke inductance value by itself is only half the story for power supply use.

Choke inductance measurements

[1] showing variation with AC & DC
current.

Simple measurement scheme

Choke inductance can be measured using a
relatively simple method that passes DC plus AC
current through the choke. The test circuit uses the
choke to load the rectified output of a transformer
power supply. A low value sense resistor Rs is used
to make Idc and Iac current measurements. The
choke is shown as an inductance L and a DC
resistance Rdcr. The total DC resistance of the
loading circuit is Rdcr + Rs.
By using different AC supply voltage levels, Vac
(rms), the level of DC current can be varied. The AC
waveform applied to the choke is a rectified
sinewave with a DC level of 0.9*Vac, and so a DC
current of about Idc = 0.9*Vac/(Rdcr + Rs) flows
through the choke.

http://www.ozvalveamps.org/ Tim Robbins

Page 2 of 4 Power Supply Choke Measurement 5 March 2020

The AC voltage across the choke has a level of Vc=1.27*Vac/(n2-1),

where n=2,4,6.. (ie. the even harmonics of the mains frequency) [2]. The
harmonic levels drop off rapidly, so simply using just the n=2 (ie. 100Hz)
harmonic frequency indicates that the applied AC voltage on the choke is
approximately Vc = 0.42*Vac. The AC current can be approximated by
Iac = Vac/(1500*L), where L is in Henry - this approximation assumes
f=100Hz, the choke reactance (2.π.f.L) is much larger than Rdcr, and only
2nd harmonic current is significant. If we assume Rdcr>>Rs, then the
previous equations can be rejigged to show that DC current is larger than
the AC current by the ratio of about Idc/Iac = 1350*L/Rdcr, which is 38
times for the Rola choke example.
So this simple test method measures the small signal (incremental)
inductance of the choke, where a relatively large DC current is passing in
comparison to the AC current, and is similar to a smoothing choke
application. With respect to what the choke core experiences for this application, the graph on the right
shows the choke core magnetised with the DC magnetising force Hdc, and a smaller AC magnetisation level
∆H is superimposed which causes a cyclical magnetisation loop to be followed between A and Bm.
This measurement scheme does not inherently measure inductance at a given frequency and excitation
voltage, due to the measurement waveform including mains frequency harmonics and waveform distortion
from the mains voltage/transformer/diode rectifier. The influence of shunt capacitance is neglected, as it is
likely to be >> 100x the inductive impedance. In practise the scheme gives good inductance precision for
power supply choke design purposes, as the test frequency and waveform are typical of that application, and
especially given that choke inductance can vary so much with operating conditions.
The above measurement is made at twice the mains frequency, and a separate measurement is needed to
determine the general self-resonant frequency (SRF) of the choke. The self-capacitance in the choke
winding causes the rising choke impedance with frequency to level off at the SRF and then impedance falls
for ripple frequencies higher than SRF. Vintage power chokes of 10-14H, and DC current ratings of 60-
125mA are likely to have a self-resonant frequency of about 3-5kHz.
When the mains supply is turned off, or the supply disconnected in some manner, and twice during each
mains cycle, the DC current in the test circuit commutates through the diode bridge, which acts as a free-
wheeling diode (similar in action to the protection/suppression diode typically placed across a DC relay coil).

Test Method
Use a true-rms meter to measure Vsac across sense resistor Rs (to derive Iac=Vsac/Rs), and to measure Vc
across the choke. Choke impedance Z is then Z= Vc / Iac. For most types of choke, the choke inductance
can be approximated by the impedance, such that choke inductance L = Z/(2.π.f), as |Z| = √((2.π.f.L)2+Rdcr2)
and (2.π.f.L)2 >>Rdcr2. A calculation spreadsheet is available [4], and accounts for the effect of Rdcr on the
choke inductance calculation. Measure the DC Voltage Vsdc across sense resistor Rs. DC current through
choke is then Idc=Vsdc/Rs.
Using a small value for Rs (ie. 10Ω) will require a meter with at least 1-10mV resolution, such as a cheap
Aneng AN8009. Some DVM’s like a Fluke 115 handheld won’t do (although it has a 600mV AC-DC range,
this can over-range due to the DC level exceeding 600mV even though the AC level being measured is low,
and so 6VAC range is only available). Raising Rs value to say 100Ω will help with poorer resolution meters,
and shouldn’t really affect accuracy or be significant compared to Rdcr for many chokes. Check your meter
performance specification as part of preparing to take measurements. Choke input filter parts are designed
for higher applied AC voltage – this requires a larger Rs (eg. 470Ω with higher power rating >20W) to apply
say 50Vrms and pass over 100mAdc, otherwise the measured inductance will be lower than rated.
A tapped transformer can be used to change the Vac level, to apply different DC current levels, and allow
inductance droop to be plotted. My first test setup used 12V, 20V, 32V and 52VAC secondaries, and two
multi-tapped 0-24VAC transformers would be quite practical, but I now use the heater supply of a vintage
valve tester with 0.6 to 117V in 19 steps. The supply needs sufficient current rating to suit the DC current
being applied to the choke (eg. a 1A secondary rating would suit many chokes used in valve amps). Any
diode bridge with 6A or more rating should be fine.
If needed, the value of the sense resistor Rs can be increased in order to lower the DC current level relative
to the applied inductor AC voltage level, as Idc/Vac ~ 0.9 /(Rdcr + Rs). If needed, the rectified waveform
could be RC filtered before the choke is connected, so as to attenuate higher ripple frequencies.

http://www.ozvalveamps.org/ Tim Robbins

Page 3 of 4 Power Supply Choke Measurement 5 March 2020

Choke Performance
The example 14H Rola choke was measured using the described test method. The nominal transformer
voltage Vac used for four measurement points is given in the following table. Rdcr was measured at 535Ω,
and SRF at 6kHz (C≈50pF).
Vac 12 V 20 V 32 V 52 V
Idc 17 mA 30 mA 50 mA 82 mA
Iac 0.6 mA 0.9 mA 1.4 mA 2.7 mA
Vc 5.85 V 9.17 V 14.1 V 23.1 V
L 15.5 H 16.2 H 16.1 H 13.6 H

The test results agree well with the 14H at 60mA DC part rating, noting that Vc is greater than the spec level
of 10Vrms for Idc=50mA, and hence measured L would be a bit lower at the Vc spec level. As Vc increases
when larger Vac is applied, the measured inductance can increase even though Idc has increased.
The measured drop in inductance with DC current, along with the DC resistance of the choke, can provide a
good estimate of the manufacturer’s DC current rating for the part. The following product curves of
inductance are from Chicago Transformer chokes at 10V 60 Hz excitation [3]. The power loss at rated max
DC current is about 5.5W for each choke (note that these are large chokes). The much smaller example
Rola choke has a max power loss of only 2W at rated 60mA DC. Hubelhank presented a consistent test
result using a similar measurement technique, circa 1956 [5].

Uncommon iron-cored chokes for valve amp power supplies

The typical flouro ballast choke is compact, double insulated, uniformly gapped by two rows of C laminations
butting to a square lamination central core, and appears fine to sit at 600VDC or more and be used for a
choke-input filter. A 240VAC 18/20W choke is very common, measures about 1.5-2H, and would be suitable
for up to 300mA DC (power dissipation up to 5W). A 240VAC 9/13W choke measures 3H at 100mAdc and
2.7H at 200mAdc, with Rdc=140Ω and SRF=5kHz, and would be fine for up to about 200mAdc (power
dissipation about 5W).
An ATCO EC18/20 240VAC 50Hz choke gave measurement
levels of Rdcr=54Ω, SRF=10kHz (150pF shunt):
Vac 12 V 20 V 32 V 52 V
Idc 147 mA 260 mA 420 mA 517 mA
L 1.88 H 1.69 H 1.45 H 1.26 H

Some Wurlitzer organs included a large number of note inductors (iron-cored with variable gap setting), of
which a few provide up to 4H inductance and are suitable for up to 30mA DC (eg. suitable for screen and
preamp filtering).

http://www.ozvalveamps.org/ Tim Robbins

Page 4 of 4 Power Supply Choke Measurement 5 March 2020

A Wurlitzer 500407 inductor from a 4100B organ (set with minimum

gap) gave measurement levels of Rdcr=266Ω:

Vac 12 V 20 V 32 V
Idc 30 mA 50 mA 90 mA
L 4.1 H 2.3 H 0.9 H

Ripple trap assessment

The test circuit can be used to measure the changing AC current harmonics passing through a choke with a
parallel capacitor (and series dampening resistance), sometimes referred to as a ripple trap (a technique
used in power supplies to enhance the attenuation of the dominant 2f ripple component – see [6]).
Displaying the ripple current (voltage across sense resistor) on a spectrum analyser shows the increasing
attenuation of the 2f harmonic as the parallel capacitance value is increased towards LC resonance, but also
shows a corresponding increase in the magnitude of higher ripple frequencies being passed through. The
filter capacitor following the choke bypasses the ripple currents, with the net result of a lower rms ripple
voltage across that capacitor. Given the likelihood of choke inductance being higher than its rated value
when DC current is below the rated level, and given the increase in higher order harmonics with increasing
capacitance, it is recommended that a lower capacitor value is used than what would be expected to tune
the rated inductance at 2f – perhaps at least 20% lower. The dampening resistance RC in series with the
capacitor is typically twice the choke RL, although the Q can be calculated as √(LC) / (RC+RL).
Output Transformer Primary Inductance
OT primary inductance influences the low frequency roll off of an amplifier, and forms an RL high pass filter
with the output stage valve’s internal resistance Ra. The choke test circuit can be used to measure primary
winding inductance at a given DC current level for an SE output transformer, and to indicate the change in
inductance at zero, idle and twice idle current. In a PP amp, apart from the OT design itself, a DC imbalance
from valve mismatch and/or bias mismatch can significantly lower primary inductance L.
Plate-Plate winding inductance of a PP output transformer can be measured with an unbalanced DC current
level. When testing the P-P inductance of an output transformer, the test circuit applies DC current in both
half-primary windings, but as the DC current is not being cancelled by the CT feed location, then the DC
current level equates to an imbalance level twice as large. It is likely that the test circuit Rs will need to be
made fairly large in order to suppress Idc down to a level typical of an unbalanced output stage.
The plot on the right includes a 68H P-P
inductance value from a simple transformer fed
test circuit (ie. no DC bias), as well as 3 test
values taken in the choke test circuit when using
high values of Rs (up to 6k8Ω).
The transformer is a Red Line AF5/20 with 5kΩ
PP, and 15W hi-fi ratings. If idle bias current was
50mA nominal, then the 4mA test value would be
equivalent to an 8mA imbalance between valves
(eg. 46mA + 54mA). In that situation, the primary
inductance would be down about 25%, compared
to a balanced output stage.
Similarly, a common-mode choke can be
measured for differential inductance with an
unbalanced DC current level.
References
[1] ‘The Measurement of Choke Coil Inductance’, C.A.Wright & F.T.Bowditch, 1927.
[2] Fourier Analysis, Lucas Illing, 2008
[3] Chicago, Transformers and filter reactors, CTC-58.
[4] http://www.dalmura.com.au/projects/OT%20calcs.xls
[5] 'Use those "junk-box" chokes' by S.H. Hubelhank, 1956. Reprinted in Sound Practices, Fall 1994.
[6] "Smoothing Circuits: (2) Inductance-capacitance", ‘Cathode Ray’, WW Nov, 1949

http://www.ozvalveamps.org/ Tim Robbins