Choke filter question

[catalin gramada]

Hey. Are the values of the filter chokes provided by comercial schematics rated to nominal inductance related in circuit please ? I asked as I have a bunch of Hammonds rated for 60 cps. I was able to modify some for my projects but I asked to be sure I may use it as direct replacements filters or in clones. Did the schematics the inductance value for 100- 120 cps or what ? Thanks.

[Chuck H]

I didn't look anything up, but,... I remember reading that Hammond makes a practice of designing power supply products to work with either 50hz or 60hz. So obviously they're aware of this and I'd bet it applies to all their chokes since they can't know who is using them and where.

[loudthud]

In general, the 60Hz note means the inductance was measured at 60Hz. There won't be a noticeable difference at 100 or 120hz. Some very high Henry chokes may have a self resonant frequency in the audio range, but you won't be using those in a power supply. DC current in a choke may alter the inductance by 10% or so, this is usually ignored.

[catalin gramada]

In general, the 60Hz note means the inductance was measured at 60Hz. There won't be a noticeable difference at 100 or 120hz. Some very high Henry chokes may have a self resonant frequency in the audio range, but you won't be using those in a power supply. DC current in a choke may alter the inductance by 10% or so, this is usually ignored.

That,s right. Couple of 5H Hammonds I have measured only 3H at 100cps. Thinking of worst scenario first but then reading the published datasheets saw it was rated at 60 cps.

[Helmholtz]

The magnetic properties of an iron core are highly non-linear.
In general the inductance value of an iron core choke varies with frequency, AC current and DC current.
To reduce such dependencies a filter choke must have an airgap.
An airgap works similar to NFB. It stabilizes inductance at a significantly lower value.

Still, inductance tends to
- drop with frequency,
- increase with AC current until beginning of saturation,
- decrease with DC current.

Means that a choke should be measured at operating conditions.
L-meter results typically are not very accurate.

Here's an excellent paper on chokes by trobbins: https://dalmura.com.au/static/Choke%20measurement.pdf

[catalin gramada]

The magnetic properties of an iron core are highly non-linear.
In general the inductance value of an iron core choke varies with frequency, AC current and DC current.
To reduce such dependencies a filter choke must have an airgap.
An airgap works similar to NFB. It stabilizes inductance at a significantly lower value.

Still, inductance tends to
- drop with frequency,
- increase with AC current until beginning of saturation,
- decease with DC current.

Means that a choke should be measured at operating conditions.
L-meter results typically are not very accurate.

Here's an excellent paper on chokes by trobbins: https://dalmura.com.au/static/Choke%20measurement.pdf

...I don't know... I did a jig test with a series resistor to measure one in dc current. The choke is 150mA dc rated and is pretty consistent: 7.8H at 10 mA DC 7.6H at 100 mA, but yes, it shows only 5H measured at 100cps on lcr meter. Thanks.
The calculation was done using this formula: L= (R / 2x3.14x100) x (VLac /VRac)

[Helmholtz]

...I don't know...

Why?
Don't your measurements confirm what I said above?

[catalin gramada]

Why?
Don't your measurements confirm what I said above?

A choke series with a resistor are not quite real condition. A pulse charging currents into a cap may change the game. This are. my doubts...

[Helmholtz]

A choke series with a resistor are not quite real condition. A pulse charging currents into a cap may change the game. This are. my doubts...

A choke with an iron core does not have one true inductance.
The pulse charging time method is likely to give unrealistic results as inductance changes with momentary DC current.

The method described by Tom Robbins gives a result closest to real filter application, where series impedance at ripple frequency in the presence of a DC bias matters. The external resistor doesn't influence the result.
As you know the DCR you could correct the impedance for pure inductance - if it matters.

The general dependencies I listed above exist independent of the measuring method.

[Helmholtz]

An example:
An ideal 5H choke has an impedance of roughly 3000 Ohm @100Hz.
A DCR of 100R increases impedance by only 1.7 Ohm.

[catalin gramada]

Thanks. What is the most important parameter relating saturation occur ? We get infos about dc rating, which is good but nothing about ac ratings regard max allowable conditions. I wonder how ac current limits in respect with a certain frequency may be deduced from specs please ?
I tested the choke with 50Vac rms ripple over what I deducted to be a 7.5H inductance , meant effective 4.7k reactance at 100 cps. I wonder where is the saturation limit when inductance collapse?
I asking as Hammonds chokes are easy "tunable" setting the gap but we have to determine ac current limits to saturation then.

[Helmholtz]

Core saturation shows as a significant drop of effective inductance.
It is caused by either excessive DC or AC current.
With a typical CLC filter DC current dominates.
Unfortunately choke datasheets don't state if the rated DC current is a thermal or a saturation limit or both.

Using the trobbins method you should be able to find the saturation limit for a given choke.

[trobbins]

I'd guess that nearly all chokes purchased or used or seen nowadays would be used in a CLC application, where L is unlikely to experience more than say 10-20Vrms across it (at twice the mains frequency, and with the application's dc current passing through it).

I've hardly ever come across manufacturer data for chokes specifically made for choke-input filter application for B+ supplies, where the choke likely experiences well above 100Vrms or 300Vpp along with the output load dc current. That makes it difficult to confirm that an unknown choke is ok for use in a particular choke input filter application. The test technique I typically use is relatively easy to align with testing chokes for CLC use, but maybe not so easy for choke-input application.

Determining when a choke core is reaching in to its saturation region is an interesting topic - which relates to the shape of the BH curve of the part. A test could aim to measure incremental inductance as DC current is increased, and note how the value of inductance falls as the DC current pushes the operating point further in to the saturation region of the core (that is the type of test I nominally do).

A different test could apply an ever increasing AC voltage, with no DC, and try and measure the large-signal inductance and note how that changes, however that may get a bit fuzzy to deduce what is happening as the waveform of the current will likely get more and more distorted. In its base form, that is the easiest test to measure inductance as it just requires an ac voltage supply and ac current meter, but the difficulty is then relating that to a power supply application.

Ronald Dekker (uTracer) recently tested a flouro choke using the step voltage method where the rise in current is measured and the inductance is then deduced by the slope - I added that test concept recently to the choke article after chatting with Ronald.

[catalin gramada]

...so, I.ll suppose will never know what did a certain choke in a certain circuit, except someone did true measurements and share the results. Due my last home experiments I found the screen supply as one of most important point regard amp presentation and replacing a choke with whatever have on hand selected by comercial specs is wrong.

[Helmholtz]

I don't understand your problem.

Do you really think you can hear a say 20% difference in choke inductance?

Are your filter caps selected to less than 20% tolerance?