This ringing is the consequence of impedance mismatch. If the transformer was loaded at the design impedance then the energy would have been consumed by the load instead of ringing with the parasitic winding capacitance. In effect the load resistance kills the Q of the parasitic resonance.
Never had distortion issues with a properly matched isolation transformer, we used 600:600 transformers where the secondary had a few more turns to give us a true 1 to 1 voltage and cased in a mu-metal can for screening.
@@MikinessAnalog Sort of, we used to have a 1V rms system level, so all inputs were designed to produce 1V rms at max volume. and just to make life more fun all outputs were in dBm. (0.775Vrms into 600 ohms).
I appreciate the folks that know to refer to the transformers as what they are, because telling me that a transformer is 1:1 tells me absolutely nothing! For all I know, I can be a 60 turn transformer with virtually no use outside of 60 AC modulator circuits.
@@RobertHallIV 1:1 just describes the turns ratio, meaning that the transformer is just providing an isolation barrier. With no changes to the voltage or current flowing. It doesn't describe any other properties, so it could be a small one for audio or a larger one for mains.
The frequency of that ringing is around 50kHz and would therefore be imperceptible to the human ear and therefore not have any practical implication to an audio amplifier with a typical frequency range of 20 - 20,000 Hz. And as other comments suggest, if the secondary winding was resistively loaded, the ringing would also be heavily attenuated anyway.
You _must_ match the impedances properly if you want to maintain signal integrity. A well made audio transformer in a well designed circuit will have negligible effect on audio quality. (That’s subjective, obviously). There’s loads of broadcast quality gear that has audio transformers in it. Mixers, microphone amps, wireless microphone receivers and so on. Transformers do appear to be falling out of favour, probably mainly due to cost as good ones are very expensive. They were ubiquitous in gear with XLR connectivity not so long ago.
I believe the answer is "Zobel network". If you crack open a simple audio isolator used for hum removal there'll probably be a couple of parts on one side to that effect.
These types of transformers are almost always used in XLR to "tip-ring-sleeve" 1/4th inch adapters, with the "ground" being the shield and the tip / ring being the inverting & non-inverting output. Not sure if they work proper in reverse though. (with signal coming from the 1/4 inch side)
In almost all of the mic preamp schematics that I am familiar with, they use a small capacitor across the output side of the transformer or a capacitor and resistor to dampen ringing and, as people have mentioned, they also will load it with a specific input impedance/resistor in the next stage. I suspect they may also have taken advantage of inherent miller capacitance to help with this in some tube designs in the past.
Precisely. IMSAI Guy would need to match the input and output impedance first off, to conduct a valid plot of the transformer's audio-range linearity (20 - 20K hertz). Adding a Zobel network (resistor and capacitor in series) on the output side of the transformer will help a lot in reducing/dampening any induced L/C distortion. This from my "somewhat limited" experience in working with audio transformer applications. Additionally, a Mu metal can at ground potential, to physically shield the transformer would also most likely help to reduce/eliminate any distortion from radiating into and out of the transformer. Just my $0.02. Fred
@@electronicengineer Yeah. I mean some of the distortion is also considered desirable. Not ringing, but harmonic distortion is one of the things people like in a transformer-coupled sound. There's also hysteresis that plays a role as well.
@@christopherventer6391 Agreed. A majority of tube amplifier aficionados seem to prefer the even harmonic frequencies produced by those tube designs, as they're commonly referred to as a "warm" sound. I was taught that a clean amplifier will not "color" the audio signal in any way, including even/odd order harmonics, IMD (intermodulation distortion), dampening, and usually, at the top of the list, is THD (total harmonic distortion). To each their own I say. I like to listen to "clean" audio, however I have a tendency to almost always end up with the bass (0 - 100 Hz) pushed up a bit and I suspect, due to my aging hearing, the treble/high end frequencies find themselves also pushed up quite a bit, as I chase the "perfect" sound... It is all subjective in the end. Have a great weekend Christopher. Fred
I'm guessing the ringing is from inter winding capacitance. Transformers can express parallel and series LC behaviour depending on the frequency of the input waveform. Other interesting distortion can be caused from DC biased drivers (single ended) which runs the transformer core with a fixed amount of magnetic flux. Modulating that causes distortion in the primary to secondary coupling, due to hysteresis in the core. Have you done a frequency sweep of the trans to find the parallel and series resonance points? Cheers,
My first thought is an impedance mismatch and my second is that an audio transformer is not designed for a 20 kHz square wave as any harmonics present in that wave are above the audible range. Who can distinguish a square wave at 20 kHz from a sine wave at 20 kHz? Nonetheless interesting discussion and thoughts shared in comments so thank you!
its just reflection due to iperdance inbalance. try adding a resistor parralell over the coils. depending on trafo in audio its often 600 ohm, so on input (generator) add a 550 ohm in serie with coil, and on output of trafo add a 600 ohm parallel with winding, ( or used 2 potmeter to get optimal values)
Try to drive input with band-limited square wave (upped frequency limit of 20-30kHz) and match impedance of transformer on input & output, then it would be a real test how it works with audio signals. Square wave that you put in is (because of fast rise time) having harmonics going to MHz range and input & output impedance missmatch causes ringing...
I had a super cheap DI box for a PA rig...(free gift with purchase of a box of cables). the transformer was microphonic. If you bumped the box, ringing came out of the audio. I'm sure that's a different issue. I put a high grade transformer in that DI box and then life was good. It's actually my favorite DI now, with one of the best grade transformers out there.
Every audio output transformer or intermediate stage audio coupling transformer I have encountered has the pins for the primary on one side of the core and the pins for the secondary on the other side of the core. Your transformer has six pins arranged as a 3x2 pattern. Typically, that means that one set of three is the center-tapped primary, the other set is a center-tapped secondary. Since you found that the two sides are connected together (low resistance across the core), and all your experimenting uses this connection, I am inclined to think this is a shorted output transformer or an impedance-matching autotransformer. In either case, or whatever the case, this is an oddball transformer, which may not be useful to provide representative results of audio transformers in general. As far as the goal of the experiment, the answer has to be "yes" - audio transformers add distortion, because every added component adds noise and distortion. It is only a question of how much and in what way. As other commenters have noted, you need to impedance match to transformers, or at least take the impedances into account, when doing such measurements. It is true that transformers have resonant points, which can be useful in many applications, and only the frequencies of interest matter, since the circuit design should ensure that parasitic oscillations are suppressed. Also, consider that audio output transformers that drive speakers see a motor coil (meaning it is a coil with a changing load on its inductance field), which also needs to be considered so that the nonlinearities of the speakers are canceled by nonlinearities in the audio output stage. With all this complexity, it is a wonder that it works at all! However, many of these distortions are small enough and the impedances are low, so just tuning circuits with designed-to-be audio transformer matched to resistive loads is often good enough for everyone except the very picky audiophiles among us. I spent a lot of time tweeking and tuning circuits and speakers to produce the best audio reproduction into reference microphones, only to have confirmation that the "experts" love the distortion they have gotten used to hearing. After dealing with descriptions like "warm", "soft", "bright", "tinny", "cool", "brassy", "heavy", "muddy", and many more, and trying to assign spectral and temporal distortions to each colorful term, I decided I must be deaf and went back to working on computers where 1s and 0s are always just 1s and 0s and answers are either right of wrong.
Transformers are sensitive to the load impedance, the source impedance and the current passing through the windings. Measuring the performances with another condition is somehow irrelevant
Not a very conclusive test. You took a random transformer and made a random measurement without proper load. Transformers, like pretty much any component must be well integrated into the curcuit with correct levels, correct load impedance etc. Even though the ringing technically is outside of the audible range here, no serious engineer would have accepted this result.
Help needed! How do I match impedance of inverter16v AC output SOURCE to 16vAC isolation transformer primary, that I have to hand WIND, for 4 ohms load of 32v AC 6 amps secondary ? Do I use the ohms of primary 16v + 12 amps division or the resistance .06 ohms of inverter output contacts terminals at 20khz frequency? Voltage and turns ratio same as 1:2. The online formula for inductance value of turns requires reactance value in ohms , at the frequency of source.
Use transformers all the time for tube amplifier output. 5k ohm to 8 ohms. I'm not sure what the application of the transformer was exactly, or tested with triangle waves before. But a 1k square wave has serious transients for an audio system. You don't have any load on the transformer either so it's going to misbehave. Match the source impedance with a load impedance and try again. Good transformers are expensive. Like the scope BTW, I have one.
assuming output from function generator is 50 ohm should you not have a 50 ohm times the transformer factor on the output of the transformer to match impedance? Then you can measure with 10 MOhm in parallel with that.
I'm testing an ART DTI (Dual Transformer Isolator) and non linear distortion is really bad. With a 100 Hz tone, the 3rd harmonic is only -60 dB and all the way out to the 9th harmonic is only about -85 dB.
You have z-banding in your print back there. It's a 25 dollar upgrade and then 3d prints look smoother, even at standard rough settings. It's a 10 minute install. Don't let yourself be gouged by official brand name z axis upgrades. As long as there's a motor, coupler, rod and a few other parts buy it. It's better than auto bed leveling like a bltouch.
A well-designed speaker looks like 8-20 Ohm resistor, which I presume is not included? This would damp out the oscillation. Also, the frequency is a bit beyond 20 kHz.
What is the output load? What is the frequency of the ringing? All the world is comprised of sine-waves in various combinations and durations. Step functions are pretty short duration, like a kick drum. Nor are they symmetrical.
Could you make a video about transformer. What are power transformer What are audio transformer What are output transformer What are line matching transformer What are directbox transformer Are they all same? Can they be subsituted to eachother? Thankyou
This is useless video! Transformer needs to be impedance matched on both primary and secondary side in order to have signal integrity and meaningful square wave signal handling. Second reason why I call this video useless: I was fully expecting a read distortion measurements! It is a well known fact that unless a special circuit is implemented to drive a transformer, it would distort low frequency signals - say below 100 Hz. Holding the poor transformer on vice would not improve things. So poor planning, poor execution.
This ringing is the consequence of impedance mismatch. If the transformer was loaded at the design impedance then the energy would have been consumed by the load instead of ringing with the parasitic winding capacitance. In effect the load resistance kills the Q of the parasitic resonance.
Never had distortion issues with a properly matched isolation transformer, we used 600:600 transformers where the secondary had a few more turns to give us a true 1 to 1 voltage and cased in a mu-metal can for screening.
"Line level", correct?
@@MikinessAnalog Sort of, we used to have a 1V rms system level, so all inputs were designed to produce 1V rms at max volume. and just to make life more fun all outputs were in dBm. (0.775Vrms into 600 ohms).
I appreciate the folks that know to refer to the transformers as what they are, because telling me that a transformer is 1:1 tells me absolutely nothing! For all I know, I can be a 60 turn transformer with virtually no use outside of 60 AC modulator circuits.
@@RobertHallIV 1:1 just describes the turns ratio, meaning that the transformer is just providing an isolation barrier. With no changes to the voltage or current flowing. It doesn't describe any other properties, so it could be a small one for audio or a larger one for mains.
The frequency of that ringing is around 50kHz and would therefore be imperceptible to the human ear and therefore not have any practical implication to an audio amplifier with a typical frequency range of 20 - 20,000 Hz.
And as other comments suggest, if the secondary winding was resistively loaded, the ringing would also be heavily attenuated anyway.
You _must_ match the impedances properly if you want to maintain signal integrity. A well made audio transformer in a well designed circuit will have negligible effect on audio quality. (That’s subjective, obviously). There’s loads of broadcast quality gear that has audio transformers in it. Mixers, microphone amps, wireless microphone receivers and so on. Transformers do appear to be falling out of favour, probably mainly due to cost as good ones are very expensive. They were ubiquitous in gear with XLR connectivity not so long ago.
well when you load it with 10M ohm probe that's what happens....this is not designed/intended to have 10M ohm loading...
I believe the answer is "Zobel network". If you crack open a simple audio isolator used for hum removal there'll probably be a couple of parts on one side to that effect.
A transformer is not only an inductor, but also a capacitor due to it's windings next to each other.
These types of transformers are almost always used in XLR to "tip-ring-sleeve" 1/4th inch adapters,
with the "ground" being the shield and the tip / ring being the inverting & non-inverting output.
Not sure if they work proper in reverse though. (with signal coming from the 1/4 inch side)
I was going to say 'but your scope is not terminating the transformer correctly, so it will ring', but I see I've been beaten to it.
In almost all of the mic preamp schematics that I am familiar with, they use a small capacitor across the output side of the transformer or a capacitor and resistor to dampen ringing and, as people have mentioned, they also will load it with a specific input impedance/resistor in the next stage. I suspect they may also have taken advantage of inherent miller capacitance to help with this in some tube designs in the past.
Precisely. IMSAI Guy would need to match the input and output impedance first off, to conduct a valid plot of the transformer's audio-range linearity (20 - 20K hertz). Adding a Zobel network (resistor and capacitor in series) on the output side of the transformer will help a lot in reducing/dampening any induced L/C distortion. This from my "somewhat limited" experience in working with audio transformer applications. Additionally, a Mu metal can at ground potential, to physically shield the transformer would also most likely help to reduce/eliminate any distortion from radiating into and out of the transformer. Just my $0.02. Fred
@@electronicengineer Yeah. I mean some of the distortion is also considered desirable. Not ringing, but harmonic distortion is one of the things people like in a transformer-coupled sound. There's also hysteresis that plays a role as well.
@@christopherventer6391 Agreed. A majority of tube amplifier aficionados seem to prefer the even harmonic frequencies produced by those tube designs, as they're commonly referred to as a "warm" sound. I was taught that a clean amplifier will not "color" the audio signal in any way, including even/odd order harmonics, IMD (intermodulation distortion), dampening, and usually, at the top of the list, is THD (total harmonic distortion). To each their own I say. I like to listen to "clean" audio, however I have a tendency to almost always end up with the bass (0 - 100 Hz) pushed up a bit and I suspect, due to my aging hearing, the treble/high end frequencies find themselves also pushed up quite a bit, as I chase the "perfect" sound... It is all subjective in the end. Have a great weekend Christopher. Fred
I'm guessing the ringing is from inter winding capacitance. Transformers can express parallel and series LC behaviour depending on the frequency of the input waveform.
Other interesting distortion can be caused from DC biased drivers (single ended) which runs the transformer core with a fixed amount of magnetic flux. Modulating that causes distortion in the primary to secondary coupling, due to hysteresis in the core.
Have you done a frequency sweep of the trans to find the parallel and series resonance points?
Cheers,
My first thought is an impedance mismatch and my second is that an audio transformer is not designed for a 20 kHz square wave as any harmonics present in that wave are above the audible range. Who can distinguish a square wave at 20 kHz from a sine wave at 20 kHz? Nonetheless interesting discussion and thoughts shared in comments so thank you!
If it rings, answer it! But seriously, it will ring a lot less when it is terminated in the appropriate impedances on the input and output.
Fascinating. I wonder how much the iron in the vice is influincing the workings of the transformer.
its just reflection due to iperdance inbalance. try adding a resistor parralell over the coils. depending on trafo in audio its often 600 ohm, so on input (generator) add a 550 ohm in serie with coil, and on output of trafo add a 600 ohm parallel with winding, ( or used 2 potmeter to get optimal values)
Try to drive input with band-limited square wave (upped frequency limit of 20-30kHz) and match impedance of transformer on input & output, then it would be a real test how it works with audio signals. Square wave that you put in is (because of fast rise time) having harmonics going to MHz range and input & output impedance missmatch causes ringing...
Did you load the transformer with a resistor?
This forces current through the primary and minimises the influince of the iron....
I had a super cheap DI box for a PA rig...(free gift with purchase of a box of cables). the transformer was microphonic. If you bumped the box, ringing came out of the audio. I'm sure that's a different issue. I put a high grade transformer in that DI box and then life was good. It's actually my favorite DI now, with one of the best grade transformers out there.
Every audio output transformer or intermediate stage audio coupling transformer I have encountered has the pins for the primary on one side of the core and the pins for the secondary on the other side of the core. Your transformer has six pins arranged as a 3x2 pattern. Typically, that means that one set of three is the center-tapped primary, the other set is a center-tapped secondary. Since you found that the two sides are connected together (low resistance across the core), and all your experimenting uses this connection, I am inclined to think this is a shorted output transformer or an impedance-matching autotransformer. In either case, or whatever the case, this is an oddball transformer, which may not be useful to provide representative results of audio transformers in general.
As far as the goal of the experiment, the answer has to be "yes" - audio transformers add distortion, because every added component adds noise and distortion. It is only a question of how much and in what way. As other commenters have noted, you need to impedance match to transformers, or at least take the impedances into account, when doing such measurements. It is true that transformers have resonant points, which can be useful in many applications, and only the frequencies of interest matter, since the circuit design should ensure that parasitic oscillations are suppressed. Also, consider that audio output transformers that drive speakers see a motor coil (meaning it is a coil with a changing load on its inductance field), which also needs to be considered so that the nonlinearities of the speakers are canceled by nonlinearities in the audio output stage. With all this complexity, it is a wonder that it works at all! However, many of these distortions are small enough and the impedances are low, so just tuning circuits with designed-to-be audio transformer matched to resistive loads is often good enough for everyone except the very picky audiophiles among us.
I spent a lot of time tweeking and tuning circuits and speakers to produce the best audio reproduction into reference microphones, only to have confirmation that the "experts" love the distortion they have gotten used to hearing. After dealing with descriptions like "warm", "soft", "bright", "tinny", "cool", "brassy", "heavy", "muddy", and many more, and trying to assign spectral and temporal distortions to each colorful term, I decided I must be deaf and went back to working on computers where 1s and 0s are always just 1s and 0s and answers are either right of wrong.
Transformers are sensitive to the load impedance, the source impedance and the current passing through the windings. Measuring the performances with another condition is somehow irrelevant
Not a very conclusive test.
You took a random transformer and made a random measurement without proper load.
Transformers, like pretty much any component must be well integrated into the curcuit with correct levels, correct load impedance etc.
Even though the ringing technically is outside of the audible range here, no serious engineer would have accepted this result.
Ringing at 120kHz isn't a problem if the signal is filtered above 20kHz
Help needed! How do I match impedance of inverter16v AC output SOURCE
to 16vAC isolation transformer primary, that I have to hand WIND, for 4
ohms load of 32v AC 6 amps secondary ? Do I use the ohms of primary 16v +
12 amps division or the resistance .06 ohms of inverter output contacts
terminals at 20khz frequency? Voltage and turns ratio same as 1:2. The
online formula for inductance value of turns requires reactance value in
ohms , at the frequency of source.
Use transformers all the time for tube amplifier output. 5k ohm to 8 ohms. I'm not sure what the application of the transformer was exactly, or tested with triangle waves before. But a 1k square wave has serious transients for an audio system. You don't have any load on the transformer either so it's going to misbehave. Match the source impedance with a load impedance and try again. Good transformers are expensive. Like the scope BTW, I have one.
I saw 1827 k ohms for the second dc resistance of the coil. Inductors are like people... they resist change.
assuming output from function generator is 50 ohm should you not have a 50 ohm times the transformer factor on the output of the transformer to match impedance? Then you can measure with 10 MOhm in parallel with that.
Great video but why does it ring?
I'm testing an ART DTI (Dual Transformer Isolator) and non linear distortion is really bad. With a 100 Hz tone, the 3rd harmonic is only -60 dB and all the way out to the 9th harmonic is only about -85 dB.
I wonder what the reading would have been if he'd flipped the primary and secondary windings?
Just wondering how the FFT would compare, input to output, of a sinewave (~10k Hz)?? Nice demo!!!!
You may need to do impedance calculations then terminate both sides in its proper impedance then test again
I never knew that, thank you.
You have z-banding in your print back there. It's a 25 dollar upgrade and then 3d prints look smoother, even at standard rough settings. It's a 10 minute install. Don't let yourself be gouged by official brand name z axis upgrades. As long as there's a motor, coupler, rod and a few other parts buy it. It's better than auto bed leveling like a bltouch.
A well-designed speaker looks like 8-20 Ohm resistor, which I presume is not included? This would damp out the oscillation. Also, the frequency is a bit beyond 20 kHz.
What is the output load? What is the frequency of the ringing? All the world is comprised of sine-waves in various combinations and durations. Step functions are pretty short duration, like a kick drum. Nor are they symmetrical.
AUDIO TRANSFORMERS REDUCE DISTORTION VS. BARE WIRE CONNECTIONS.
Yeah you made a resonant filter. You need to load the transformer correctly
Yeah, say a 600:600 unity matched transformer does not do that. Gimme a day of so, and I'll throw my sig gen and scope on it.
Could you make a video about transformer.
What are power transformer
What are audio transformer
What are output transformer
What are line matching transformer
What are directbox transformer
Are they all same?
Can they be subsituted to eachother?
Thankyou
interesting subject, so many types, so many applications. I'll add it to my list
@@IMSAIGuy i have been trying to make a direct box for a piezo device.. and i dont know much about transformer.
Will be waiting for the video 😊😊😊
Why would you need a transformer
@@IMSAIGuy to convert impedance.. without it piezo sounds very tinny. To match impedance
hmm, not familiar, I've only seen active amplifiers on piezo to both amplify and convert impedance.
In audio, a 'pure' signal is not always the best sounding signal. Short 4 min well done vid here: th-cam.com/video/1n-lXFTuMfI/w-d-xo.html
Did I read the meter wrong or was that 120 ohms/1800 ohms, not 120 ohms/180 ohms?
1800 Kohms is what I saw @ 1:06.
@@robjordan63 it reads .1826 kOhm = 182.6 Ohm
@@BaldurNorddahl Ah, thanks, I see now.
This is useless video! Transformer needs to be impedance matched on both primary and secondary side in order to have signal integrity and meaningful square wave signal handling.
Second reason why I call this video useless: I was fully expecting a read distortion measurements! It is a well known fact that unless a special circuit is implemented to drive a transformer, it would distort low frequency signals - say below 100 Hz.
Holding the poor transformer on vice would not improve things. So poor planning, poor execution.
I found it useful as many viewers helped educate me and possibly others