Has anybody come across any actual measurements of distortion in ferrite core inductors? I've come across plenty of assertion, but only one guy who actually measured an iron core inductor:
<meta name="description" content="Klipsch audio systems provide the true audio/video lover a wide variety of high performance loudspeakers and loudspeaker systems for music and home theater entertainment centers, including iPod speakers, multimedia s,
It would be useful to have an idea of the actual magnitude of any hysteresis effects.
<meta name="description" content="Klipsch audio systems provide the true audio/video lover a wide variety of high performance loudspeakers and loudspeaker systems for music and home theater entertainment centers, including iPod speakers, multimedia s,
It would be useful to have an idea of the actual magnitude of any hysteresis effects.
Last edited:
interesting thread starter. Ive often wondered what ACTUAL measurement would reveal. Ive previously used ferrite chokes, with the provision that theyre high current types, and high enough that saturation is far less of a concern. Ive guess-timated around 5amp as a minimum.
I too would find measurements to be of interest, and to check my 'saturation safety margin' is adequate.
I too would find measurements to be of interest, and to check my 'saturation safety margin' is adequate.
just found this old (short) thread:
http://www.diyaudio.com/forums/multi-way/129352-articles-inductor-distortion-tests.html
http://www.diyaudio.com/forums/multi-way/129352-articles-inductor-distortion-tests.html
Pardon me for interjecting without having direct experience, but here is a way for measuring inductor distortion. I think this will give a more direct visual indication than the Klipsch IM plots. I haven't done this but I'm pretty sure it would work.
See attached diagram. You need to drive the inductor with a power amplifier and an LF oscillator. You can test it from 100 Hz down. The lower the frequency the earlier you will see the saturation.
Basically you need to put a resistor in series and then drive the combination hard. You will watch the voltage across the resistor to see inductor current. The voltage across the inductor is essentially the total input if the sense R is small.
You must have an oscilloscpe that floats. i.e. it doesn't ground the amplifier. Connect carefully or you will short the amp.
You must be in XY mode rather than a time/sweep mode.
I'm not sure of what the distorted waveform will look like, but it will develop bumps or flats on opposite sides and this will indicate the distortion. Different cores will develop visible distortion at different levels. Ampere turns determine flux level so number of turns on a given coil will be a factor.
Can someone try this?
David S.
See attached diagram. You need to drive the inductor with a power amplifier and an LF oscillator. You can test it from 100 Hz down. The lower the frequency the earlier you will see the saturation.
Basically you need to put a resistor in series and then drive the combination hard. You will watch the voltage across the resistor to see inductor current. The voltage across the inductor is essentially the total input if the sense R is small.
You must have an oscilloscpe that floats. i.e. it doesn't ground the amplifier. Connect carefully or you will short the amp.
You must be in XY mode rather than a time/sweep mode.
I'm not sure of what the distorted waveform will look like, but it will develop bumps or flats on opposite sides and this will indicate the distortion. Different cores will develop visible distortion at different levels. Ampere turns determine flux level so number of turns on a given coil will be a factor.
Can someone try this?
David S.
you can make sensitive measurement with the parts the other way around - gnd the inductor and measure the V across it - its already gnd referenced
Lissagous plot is often used for visualizing reactive part nonlinearity
with protection you can use a soundcard to see distortion at much lower levels than a 'scope
Lissagous plot is often used for visualizing reactive part nonlinearity
with protection you can use a soundcard to see distortion at much lower levels than a 'scope
I did see a comprehensive study once, but I don't remember where. My take away was that yes they are nonlinear, but only lower orders at very high power levels. By that time the SPL level is high enough that the ears masking would make these nonlinearities benign.
I no longer ever consider nonlinear artifacts that occur only at high power levels. Its the ones that occur at higher orders and low power levels that are the most audible. Basically, if the nonlinearity is audible at high power levels then you simply are not using high enough power capable devices. Fix that problem and the issue goes away.
Why can't you just look at the voltage or current waveform (depends on the setup) with a spectrum analyzer and note the harmonics? Thats certaily hos I would do it.
I no longer ever consider nonlinear artifacts that occur only at high power levels. Its the ones that occur at higher orders and low power levels that are the most audible. Basically, if the nonlinearity is audible at high power levels then you simply are not using high enough power capable devices. Fix that problem and the issue goes away.
Why can't you just look at the voltage or current waveform (depends on the setup) with a spectrum analyzer and note the harmonics? Thats certaily hos I would do it.
Actually the best results are when you drive with a current source and look at the spectra. Power handling is quite frequency dependent. So you usually need a power amplifier and high value resistor or feedback to make a current source.
The link in the referred to thread had 2 tone IM measurements. The IM products were in the -60dB region (and so pretty inconsequential). I thought a visual approach would show the effect only when it got to a serious level, probably in the 1 to 3 % region, and also give an intuitive feel for what was going on. Not everyone has a distortion analyzer.
Yes, many ways to skin the cat.
David
actually I like both methods. Gedlee has a point but the test method you(dave) describe may be more accessable to the more amateur audio amateur
If I had thought to bring them along to work, I have a Voltech I could be testing them with right now, in stead of twiddling my thumbs on a ZPF load run.......
If I get around to it this week Ill be sure to post any results I get, IF they look valid
If I had thought to bring them along to work, I have a Voltech I could be testing them with right now, in stead of twiddling my thumbs on a ZPF load run.......
If I get around to it this week Ill be sure to post any results I get, IF they look valid
Yes and no. Looking at THD alone is not significant since we do not know if this is high order or low order, THD just lumps it all together. I would want to see the actual spectrum of the nonlinearity once it has actually reached saturation. I would say that the film does agree with my assesment that it takles a lot of power to drive thesae coils into nonlinearity. If, as I suspect, the nonlinearity is low order, then 1% would not be audible. Even 3% of pure second order nonlinearity is not audible.
Dave - a laptop and some free software will do spectrums. There is no excuse these days for not using these tools. The LazyJew figures (no offence intended, I can't spell it either) figures will only show rather large amounts of nonlinearity. I don't think that 3% would even be visible.
Dave - a laptop and some free software will do spectrums. There is no excuse these days for not using these tools. The LazyJew figures (no offence intended, I can't spell it either) figures will only show rather large amounts of nonlinearity. I don't think that 3% would even be visible.
The secret method to quickly find how to spell just about anything is to hit your second browser tab, go to google and start typing. It will quickly get you to the correct spelling and more.... lissajous! (Did take 5 letters!)
Slowly saturates it! really! A few percent in my last one week test, cut short by a power outage, so I don't know the long term limit.
@gedlee ive used VPI resin aka micafil. Epoxy based resin, thinned to a light varnish viscosity. Then wrapped in glasstape or polyester tape, and varnished again. Then the whole lot heatshrinked with raychem HT tubing.
Overkill i admit, but a couple of layers of masking tape is all you get on some ive used. Generally the Murata hi current types are ok since they are at least heat shrinked.
These inductors will see a lot more than 30mv in other apps without effect.
@simon: over a week 30mv MAY weakly magnetise the ferrite, but not saturate it. AC current will demag so its a non issue. Just imagine all those ships with PWM propulsion converters, with their 2000amp harmonic filters...working continuous for years...id imagine that would be a problem that wouldnt be tolerated.
Overkill i admit, but a couple of layers of masking tape is all you get on some ive used. Generally the Murata hi current types are ok since they are at least heat shrinked.
These inductors will see a lot more than 30mv in other apps without effect.
@simon: over a week 30mv MAY weakly magnetise the ferrite, but not saturate it. AC current will demag so its a non issue. Just imagine all those ships with PWM propulsion converters, with their 2000amp harmonic filters...working continuous for years...id imagine that would be a problem that wouldnt be tolerated.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.