Actually, you have disputed it via a gut feeling. As you have stated, you are not now capable of deriving the relevant equations to prove your point one way or the other.
I on the other hand, do this for a living presently. Soon enough, the production hardware I am testing will be complete, so my test setup will go away. I've only 180 more 10 Khz magnets to go. They required 150 plus strand litz as a consequence of the proximity caused 2nd orderresistive component. Single 15awg was a joke, double #17 was little better, so we bagged the solids and went with the real PITA stuff. . In addition, I work with several world leading magnetic physicists who understand my point exactly, are in total agreement, and one has actually used OPERA (IIRC.or was it Roxio) to validate what I have stated. Claiming that the em "textbooks do not support this" is first.. not suprising to me, and second, of little value. You would be better off finding either textbooks which deal with the issue, or perhaps e-mail other active players in this field. Switchmode power supply technology is the field where this becomes an issue.
That is why I asked if you used a zero crossing switch. It heavily depends on where in the cycle the turn on occurs. If all one did was flip a switch, the results will by design, be random..
Nice link, thanks.
As the materials advance, the high value for flux will climb.. We limit to about 1.5 tesla in our steel, but mainly to limit saturation at pole tips to below 2 or 3 tesla.
Indeed. One only has to go to "products" to read about one of their materials. While the verbage will be obscure to those not familiar with magnetics, this paragraphs says it all:
Peformance Garantee
Non-Impregnated toroidal cores are normally offered with the following standard guarantee:
0.3mm Grade M4 50Hz 1.0 Tesla limit: 19.0 A/M
Of note is the 1.0 tesla and 19.0 A/M... This is a statement of the coercivity of the material. 19 amps per meter is the field in the center of a loop of current 1 meter in diameter carrying 19 amps. The vendor is indicating the maximum limit of residual induction after the 1 tesla is taken away. It is by no means, close to the saturation of the material, it is more than 2 orders of magnitude below.
We specify steel by saying after an induction of 10,000 A/M, there can be no more than 80 A/M coercivity. Meaning, the field required to produce zero induction. And with over half a million pounds purchased, it has proven to be an easy specification for steel vendors to meet.
Of course they did. The purpose was to use an x/y matrix of wires to set or reset a core. The third wire was the pickup.
I'd be interested to hear jn's view on the remnance stuff.
Are we saying this is after the core had been disconnected from the supply for some time, and then reapplying power gives the issue? I would expect the remnance is related to the magnetizing current, and on a torroid ( especially a big one) this would be very low. Not enough to pop fuses the way some people are suggesting.
Ahh the joy of . . . Magnetics
What you expect it consistent with my understandings. Where in the cycle the switch closes it the important factor, not where it opens.
We certainly agree on that point at least..
jn
Last edited:
This issue of extreme inrush current of large toroidal transformers clearly separates designers who are rather theoreticians (or just theoreticians) from those who have own experience with hundreds of units (and starter circuits) running successfully in real world praxis, and who also have experience with real problems of some different suggestions. I do not wonder that JC sometimes has used the therm "amateurs".
Working the metal alters it's magnetic characteristics.
The flux will also take the easiest path, which the air in the hole will not be.
The worked edges of the hole caused by drilling will smear, and the result will be eddies which confound the measurement. Even fly cutting, taking off a few hundred microinches at a time is problematic. It is necessary to lap the surface using 400 -600 grit emery to guarantee no continuity.
It would be easier to use a DC secondary to magnetize the daylights out of the core, then record turn on currents without a load first, then try a load second.
Toroids typically have higher transfer bandwidth, so I believe the lack of control over turn on timing is what is being seen. Even though the toriod is rated at frequency, the elevated dI/dt of an abrupt input transition will eddy the living daylights out of the laminations. Especially if the secondary is capable of grabbing the fast transient into the cap.
jn
The issue is how much the material remains magnetized under the worst case conditions.
The normal magnetic swing goes from not quite saturated one way to not quite saturated the other way.
If there is enough remaining field that under very rare conditions combines with the turn on magnetization causes the core to reach saturation then excess current will flow. (Close to saturation a core often buzzes.)
Now a transformer designed for 50/60 hertz use will have more safe area operated at 60 hertz.
So the required test is how much magnetic energy remains in the core after a worst case switch off.
J.N. In making core memory they did find the toroid stored the energy better than other shapes. A bit (pun intended) of common sense (pun also intended.)
So in my opinion I would have to side with J.N. that the remnant field should not be enough to push into saturation particularly at 60 hertz. It may however contribute to a moderate increase in start-up current at 50 hertz in cores that come very close to the design limits. Particularly with the normal manufacturing tolerance for steel. (The stuff I buy is +/- 5% on the important parameters.)
ES
Magnetostriction or loose laminations causes buzzes. Saturation per se does not, but can do so if the nonlinearity causes higher harmonics in the windings, and they are allowed to move.
I just gave a friend a toriod made of 1 mil thick 1.5 inch wide lamination ribbon (I had the toriod, I didn't make that), wound 40 turns primary, 228 turns secondary, for some kind of experiment. It transfers 1.5 kW at 20 Khz, the max output of the RMX1450 I loaned him. Cool to the touch. Unfortunately, I didn't stop to think about noise...the thing sings, they foamed the daylights out of it to quiet it down..
Hence the specification coercivity.
Of course. The fewer jogs and discontinuities in the reluctance path, the better.
We used +/- 15% from the mean. Our vendors would try to purchase the full quantity, and if a melt was not large enough, we had them shuffle the laminations to guarantee consistency
jn
J.N.
We ain't arguing. But how does magnetostriction correlate with saturation?
There was the acoustical consultant who specified the amplifiers be flown with the loudspeakers in a theatre. As he was US based and specified US made products and the theatre was in London, the amplifiers made an annoying buzz (pun also intended.)
The acoustician's cure was to specify a sound isolation enclosure. I would have used a buck transformer on the power supply.
ES
We ain't arguing. But how does magnetostriction correlate with saturation?
There was the acoustical consultant who specified the amplifiers be flown with the loudspeakers in a theatre. As he was US based and specified US made products and the theatre was in London, the amplifiers made an annoying buzz (pun also intended.)
The acoustician's cure was to specify a sound isolation enclosure. I would have used a buck transformer on the power supply.
ES
Both can cause noise.J.N.
We ain't arguing. But how does magnetostriction correlate with saturation?
Magnetostriction will be at 2f.
Loose laminations will pull in at 2f, and can buzz as well.
External metal will move as 2f.
An external magnet will move as f.
Loose windings as 2f.
Saturation will cause a change in the input waveform because the reluctance will drop at the magnetization extremes. That additional current at the peaks will be more audible than the primary f.
Sounds like a spec concern. Seems to me that perhaps there was a difference in powerline frequency, no?
I probably would have spec'd the correct line frequency.
Course, maybe the cover plate to the amps resonated at 50 hz, so were quiet on this side of the pond..
jn
J.N.
When the core is saturated more flux leaks. This can vibrate other parts, but what the buzz is about is the linear to nonlinear shift in the movement of the actual transformer parts (mostly the core laminations.) Now most laminations when moved at 50, 60, 100, or even 120 hertz don't make as much noise as say 480 or 960. The there is the Fletcher Munson effect coming into play. So that is why you get a buzz when you saturate.
The other cause is of course "DC" on the AC line preventing symetrical movement on the B-H loop.
Now of course you can't have theoretical DC on an AC line due to the infinite time requirement. So lets just leave it at a voltage offset you can measure with a DC current meter in series with a resistor.
Now as to amplifier noise, there is no spec for the acoustic noise produced by an amplifier. So one rated for 50 or 60 cycle use is in spec when it is slightly noisy. Keep in mind a theatre should be NC25 or lower, thus very sensitive to noise levels. I have a theatre in Alfred N.Y. that has a similar problem. One of the power amplifiers makes more noise when the lights are dimmed. That leads me to believe the dimmer system uses pairs of SCRs and is not perfectly symmetric. Of course the solution in that case would be to add a wall to the amplifier room, as it is open to the stage.
When the core is saturated more flux leaks. This can vibrate other parts, but what the buzz is about is the linear to nonlinear shift in the movement of the actual transformer parts (mostly the core laminations.) Now most laminations when moved at 50, 60, 100, or even 120 hertz don't make as much noise as say 480 or 960. The there is the Fletcher Munson effect coming into play. So that is why you get a buzz when you saturate.
The other cause is of course "DC" on the AC line preventing symetrical movement on the B-H loop.
Now of course you can't have theoretical DC on an AC line due to the infinite time requirement. So lets just leave it at a voltage offset you can measure with a DC current meter in series with a resistor.
Now as to amplifier noise, there is no spec for the acoustic noise produced by an amplifier. So one rated for 50 or 60 cycle use is in spec when it is slightly noisy. Keep in mind a theatre should be NC25 or lower, thus very sensitive to noise levels. I have a theatre in Alfred N.Y. that has a similar problem. One of the power amplifiers makes more noise when the lights are dimmed. That leads me to believe the dimmer system uses pairs of SCRs and is not perfectly symmetric. Of course the solution in that case would be to add a wall to the amplifier room, as it is open to the stage.
Agreed.
Wow..I haven't seen a Saturable Core Reactor in decades..Now that's a really old theater..
jn
Agreed.
Wow..I haven't seen a Saturable Core Reactor in decades..Now that's a really old theater..
jn
No, Not Saturable Core Reactors, Silicon Controlled Rectifiers (for the hard of Humor.) SCRs pair instead of triacs for higher power handling. A slight difference in trigger level or on state voltage would cause the problem.
I have at my bench a linear slide transformer removed from a very old theatre dimmer!
I'm sure I told the story somewhere of the Chicago orchestra pit I played in one night on a tour in 1972 that had d.c. supplied to the stand lights, tubular incandescents which, when supplied with a.c., generated, collectively, an unacceptable level of acoustic noise.
The hapless guitar player, who plugged in to one of the convenient outlets just before the performance, wasn't warned by the staff. The line fuse in his amplifier was not equipped for interrupting d.c. and the unmistakable and acrid smell of burning power transformer wafted out of the pit. Fortunately no one shouted FIRE, but needless to say the guitar parts were absent that night.
The hapless guitar player, who plugged in to one of the convenient outlets just before the performance, wasn't warned by the staff. The line fuse in his amplifier was not equipped for interrupting d.c. and the unmistakable and acrid smell of burning power transformer wafted out of the pit. Fortunately no one shouted FIRE, but needless to say the guitar parts were absent that night.
Your example switches on at peak..... but that means real work.
Example of switching on at the wrong moment.
jan
Which has been agreed produces higher inrush.
It does not however, indicate that it was a result of core magnetization.
jn
- Status
- Not open for further replies.