You are correct PMA. JN's simple circuit is just a voltage controlled current source, as I used for motor drive amps while working at Ampex, 50 years ago. The only 'difference' is the value of Rs, which is usually less resistance, just to keep the circuit efficient, and not waste too much heat (or voltage swing) across the sense resistor.
I suspect a complete discussion of the Ls/Rs model of an inductor is warranted.
When voltage driven, the in phase current is due to the resistive part, the 90 degree out part of the current is due to the reactance.
When you multiply the voltage times the current at any instant, the positive value part is energy that is heading to the source. The negative value means energy that is returning to the source. The integral is the net transfer of energy.
When the V and I are in phase, that is dissipation. When they are 90 out, that is reactance.
The second coil can only see the magnetic field, it cannot see any dissipation. Dissipation is a local effect, reactance is a group effect.
I hate using an IPad to post this stuff..
Jn
I'm not violating conservation of energy or Kirchoff.
By losses I didn't think you were referring to Re. If so then they are linear losses, and not very useful to know about. If you are talking about eddy current or hysteretic losses then you are not talking about Re, which is constant and fixed as is it's voltage drop according to Icoil*Re=Vre.
These losses are on the other side of the flux loop. They appear as in-phase impedance in series to Re, but they don't change Re. The voltage across these losses is expressed equally in the EMF from both coils, thus they cancel out. Leaving you with nothing.
By losses I didn't think you were referring to Re. If so then they are linear losses, and not very useful to know about. If you are talking about eddy current or hysteretic losses then you are not talking about Re, which is constant and fixed as is it's voltage drop according to Icoil*Re=Vre.
These losses are on the other side of the flux loop. They appear as in-phase impedance in series to Re, but they don't change Re. The voltage across these losses is expressed equally in the EMF from both coils, thus they cancel out. Leaving you with nothing.
I'm sure that your motor work 50 years ago was truly "groundbreaking", that rotary stuff is just so challenging.
Nowadays, I control torque, angle, accelleration and velocity to achieve nanometer resolutions on multi-ton devices. But hey, tape decks are so much more "advanced".
Yah, brushless three phase servo's, micro-stepped two phase steppers, brush servo's...small potatoes compared to your experience..
You know, this discussion can use people with good analog knowledge. I thought you might be able to contribute, but I guess I'm wrong?
Well, I had hoped..perhaps you could prove me wrong, and that you can engage in a technical dialog? After all, this is not something you are competing for in the market..so no excuse.
Jn
Yes, but the feedback is high impedance so we cannot measure the temp directly.
All losses which consume energy from the flux are communicated to the pickup coil. Anything that affects the flux is communicated to the second coil. They are global effects.
And, any effect which alters the forces without communication by flux are missed. Eddy dissipation does not alter the flux directly. The dragging will affect the end and therefore the current drawn.
I know this is difficult to understand, so please be patient with me. I rarely have to explain this outside of my circle of peers. My explanations...sigh, I'm trying.
I know that sounds condescending, but it is not meant in that way.
At work, as here, I rely on others with far more capability than I for other aspects of the problems. Circuit modeling, noise analysis, test knowledge, machining.. I am used to collaboration there, and hope fo same here.
Jn
Last edited:
I don't think you're condescending, I just think you're wrong and you don't realize it. But if not, I would be learning something quite fascinating.
Both coils couple to the same eddies. So why would the eddies show preference to one coil or the other?
If you have a 1:1 transformer, it doesn't matter whether there is a 1k shunt resistor on the primary or secondary, the impedance measured at the primary will still be 1k for both cases. Same as with 3 coils.
This system has 3 coils. The voicecoils and an eddy coil. The voltage across the driven coil will drop (due to the voltage source's impedance) due to the current drawn by the eddy current in the 3rd coil. This voltage drop will appear across both coils, as they are all in the same flux loop.
Both coils couple to the same eddies. So why would the eddies show preference to one coil or the other?
If you have a 1:1 transformer, it doesn't matter whether there is a 1k shunt resistor on the primary or secondary, the impedance measured at the primary will still be 1k for both cases. Same as with 3 coils.
This system has 3 coils. The voicecoils and an eddy coil. The voltage across the driven coil will drop (due to the voltage source's impedance) due to the current drawn by the eddy current in the 3rd coil. This voltage drop will appear across both coils, as they are all in the same flux loop.
Last edited:
Communicating is dead serious difficult, even for those trained for years to educate others. Explaining things within your specialty always seems deceptively simple but everybody beyond your peers will reward you with blank stare without exception.
Ah, Perhaps that is the problem of understanding. Eddy currents dissipate energy as heat, and they fight dB/dt by the creation of a counter magnetic field which works to try and exclude flux change within the conductive surface. The difference is, when the stimulus ends, the eddy collapse does not return magnetically stored energy to the system like an inductor. So, it does not change the flux in a way the pickup coil can see.
Your 1:1 example. The 1k on the primary will be read by the source as an additional current in phase, summed with the inductive currents of the coil creating flux in the transformer.
If the resistor is on the secondary, then there are additional losses in the primary because the primary has to support the current required for the power transfer, as well as any magnetic non linearities which occur as a result of the additional flux required for that transfer.
Your 3 coil model.. As I said, eddy currents do not return energy to the system, so cannot be considered a "coil" per se.
Eddies are weird like that, as they are small loops of current that overlap. If you could see current in the conductor, you would see zero current everywhere despite the fact that eddies are there.
A shorting coil, like that used to try to linearize Le(x), that has loop currents from the vc, and it has eddies due to motion of the coil.
Jn
Edit: Malcolm Hawksford fell into that misunderstanding back in '85 in his Essex echo article. He believed that eddy currents within the wire would return their energy to the system when the current collapsed. He was incorrect of course.
Last edited:
But amongst the blank stares here are really great and well informed questions.Communicating is dead serious difficult, even for those trained for years to educate others. Explaining things within your specialty always seems deceptively simple but everybody beyond your peers will reward you with blank stare without exception.
A very good thing
Jn
Try as hard as I can, bubble tree eludes me.
I assume it has to do with the fact that you are still homeless and your bubble tree (?) is still packed?
Me, all I wish for is a year (even a few days) without personal challenges. Hitting the basement to make some metal or wood chips, compromise some electronic components through inept soldering, or even dropping a hammer through a 15 inch woofer...all preferable scenarios at this time..
That, and I'm gaining weight, stress weight. I hate that..
I hope you get what you want for the holidays. I would refrain from wishing you a white Christmas, but in this climate, who knows...18 inches in North Carolina????
Jn
Last edited:
50 years ago, nobody really cared about line currents from motors, sure they distributed correction cap banks for power factor...but motor design back then was pretty rudimentary compared to current tech.. As always happens.I'm sure that your motor work 50 years ago was truly "groundbreaking", that rotary stuff is just so challenging.
Nowadays, I control torque, angle, accelleration and velocity to achieve nanometer resolutions on multi-ton devices. But hey, tape decks are so much more "advanced".
Yah, brushless three phase servo's, micro-stepped two phase steppers, brush servo's...small potatoes compared to your experience..
You know, this discussion can use people with good analog knowledge. I thought you might be able to contribute, but I guess I'm wrong?
Well, I had hoped..perhaps you could prove me wrong, and that you can engage in a technical dialog? After all, this is not something you are competing for in the market..so no excuse.
Jn
Now, ya put a half megawatt motor on the line, somebody always complains..
Darn it, bolluxed it...
Meant to say, positive value is energy heading to the load..
Duh.
Jn
What a doofus I am...
Meant to say, positive value is energy heading to the load..
Duh.
Jn
The magnet plate is a 1 turn coil. When current goes through the voicecoil, current flows in a circle in the plate (I was calling this eddy current, my mistake). The eddy currents themselves determine the current distribution in the plate, something like skin effect. Correct? Since the eddies themselves are only coupled to other places inside the magnet plate in a distributed fashion, they are really only responsible for the current finding the path of least impedance. This impedance is what is seen as the load on the 1-turn coil which is the magnet plate.
But, that one turn coil is coupled equally to both voicecoils. The EMF that appears across all coils is the same (after scaling) because they are part of the same flux loop. The shunting effect of the magnet plate will drop the drive voltage, which will appear on both coils and cancel out.
Eddy currents inside the wire in the driven coil might not cancel, but I don't think you're referring to that.
The circular currents of the plate, since it is a shorted one turn secondary will indeed store energy, that will return as well during collapse and both coils see it.
Eddies in the plate, no. No return. They add, but do not alter the single turn currents nor stored energy.
The eddies dissipate energy and fight change of flux, causing a lowering of the dynamic inductance.
The bottom line, I'm hoping the pickup coil used to cancel magnetic non linearities in the feedback loop are worthwhile.
Jn
Eddies in the plate, no. No return. They add, but do not alter the single turn currents nor stored energy.
The eddies dissipate energy and fight change of flux, causing a lowering of the dynamic inductance.
The bottom line, I'm hoping the pickup coil used to cancel magnetic non linearities in the feedback loop are worthwhile.
Jn
Last edited:
I still don't see why the EMF of both coils would not be the same. They are in the same flux loop, anything else would violate KVL (assuming leakage is negligible, which you ensure with a bifilar coil). Given the EMFs cancel out, what's left is just the voltage across Re which is linear with coil current (excluding any nonlinearity caused by non-uniform current distribution in the driven coil wire, which IIRC is a unicorn which probably doesn't exist).
Last edited:
Back to basics, folks! '-)
Normal motor drive servos use current drive, rather than voltage drive, to simplify the design equations, and improve performance of the motor drive.
Of course, this is different from voltage drive, usually recommended for direct radiator loudspeakers that was first proposed back in 1925, and is still normally used today. Before negative feedback was invented, it might have been somewhat difficult to create a true voltage drive, in fact. Valve output impedances, or Rp, transformer coupled with optimum (for best power transfer) might give less than optimum drive impedance for direct radiator speakers. Power triodes might be an exception to this problem.
However, back to servos. One of my basic jobs while at Ampex, in both the audio and research departments, was phase locked or velocity servos for various tape transports, both audio and video. 50 years ago, almost to this day, I had just finished my phase locked (xtal clock) phase locked motor drive using a DC motor and installed in future AG-440 tape transports. Over that year, this was the only audio product that survived to another year. The real challenge in making this servo, (besides my learning how to do it, mainly from help from Research people, and a crash course given at Ampex by Dr. Dorf), was to keep the cost down. I had to use some common sense on how to make the circuit work cheaply (very cheaply) and yet work OK. For example, I took circuit equations for a 5 pole low pass filter, and made it with just 2 (hi beta) transistors, and assorted caps and resistors. (It worked!) In order to remove the op amps (IC's even then) as was used in the more expensive video products. I insisted on an RTL phase discriminator, over my boss's objections, even though it was at the time a little more expensive to implement than a simple discrete one, but keeping the cost down overall was a real challenge, because we were just replacing an AC capstan motor and had to produce the whole system at a comparable price.
Now, what about really fancy servo systems? Nope, I didn't design them, but JN, I doubt that you have either. You simply did not recognize the basic circuit for a voltage controlled, current output amplifier that YOU drew up! It is literally the same basic circuit as I used 50 years ago to drive the automotive window opener motor that we used for the capstan phase locked servo. It worked OK, but much better was possible.
Normal motor drive servos use current drive, rather than voltage drive, to simplify the design equations, and improve performance of the motor drive.
Of course, this is different from voltage drive, usually recommended for direct radiator loudspeakers that was first proposed back in 1925, and is still normally used today. Before negative feedback was invented, it might have been somewhat difficult to create a true voltage drive, in fact. Valve output impedances, or Rp, transformer coupled with optimum (for best power transfer) might give less than optimum drive impedance for direct radiator speakers. Power triodes might be an exception to this problem.
However, back to servos. One of my basic jobs while at Ampex, in both the audio and research departments, was phase locked or velocity servos for various tape transports, both audio and video. 50 years ago, almost to this day, I had just finished my phase locked (xtal clock) phase locked motor drive using a DC motor and installed in future AG-440 tape transports. Over that year, this was the only audio product that survived to another year. The real challenge in making this servo, (besides my learning how to do it, mainly from help from Research people, and a crash course given at Ampex by Dr. Dorf), was to keep the cost down. I had to use some common sense on how to make the circuit work cheaply (very cheaply) and yet work OK. For example, I took circuit equations for a 5 pole low pass filter, and made it with just 2 (hi beta) transistors, and assorted caps and resistors. (It worked!) In order to remove the op amps (IC's even then) as was used in the more expensive video products. I insisted on an RTL phase discriminator, over my boss's objections, even though it was at the time a little more expensive to implement than a simple discrete one, but keeping the cost down overall was a real challenge, because we were just replacing an AC capstan motor and had to produce the whole system at a comparable price.
Now, what about really fancy servo systems? Nope, I didn't design them, but JN, I doubt that you have either. You simply did not recognize the basic circuit for a voltage controlled, current output amplifier that YOU drew up! It is literally the same basic circuit as I used 50 years ago to drive the automotive window opener motor that we used for the capstan phase locked servo. It worked OK, but much better was possible.
Last edited:
My scheme is based on the emf's being identical as well as the Re's being identical.
Unfortunately neither is exactly identical. Close, but not exact, we are chasing the minutia.
My caveat against perfection is the non linear asymmetric stuff.
The local effects won't cancel, but global(flux) will.
What was most interesting to me was the 4th plot from demian, where the h2 and h3 very closely tracked the acoustic output. If my pickup scheme works as expected, we will have different minutia to discuss...
Jn
- Status
- Not open for further replies.