Ok. I've been working the following scenario over and over in my head and I need to be sure I have it figured right:
Let's suppose we use a magnetic pick-up (a rod magnet wound with N number of overlapping turns of magnet wire) near an oscillating metallic object to generate an AC voltage. When we measure the source impedance, we measure the DC resistance of the of the magnet wire, correct? For the sake of practicality, let's suppose the DC resistance of N number of turns is 1Kohms.
This can be modeled (within reason) as a voltage source in series with a 1Kohm resistor with a capacitor in parallel (capacitance of magnet wire winding). Simple enough.
Now let's suppose we require as high a signal-to-noise ratio (SNR) from the source as we can get. For this purpose, we desire our AC signal to operate as a balanced source which is then run into a Balanced-to-unbalanced converter of some sort (either passive UnBal transformer or active opamp). Still with me?
Here's where things get tricky:
So the magnetic pick-up is wound in a bifilar fashion: we take two parallel lengths of magnet wire (wires "A" and "B"), each length (naturally) having two ends, or "taps"; an inside winding tap and an outside winding tap.
Because two wires have twice the volume of one, we wind wires A & B in parallel N/2 number of times around the rod magnet. We then connect the inside tap of wire A to the outside tap of wire B (or vice versa).
At this point, here are the facts as I understand them:
1) We are now left with two taps: an inside tap and an outside tap, the instantaneous induced voltage at each tap is 180 degrees out of phase respective to the other. Therefor, these voltages can be said to constitute a (floating) balanced AC signal.
2) Because we used two lengths of wire, each half the length of the example used up top, the DC resistance of each wire is 500 ohms (1/2 of 1Kohm).
Here are my questions:
1a) In my mind, I see two 500 ohm lengths of wire in series. Measured tap-to-tap, the source impedance is still 1Kohm. However, since the winding "switches direction" where wire A is joined to wire B, I imagine a "floating" node forced to reference when the output taps of the pickup are connected to the input of the BalUn stage. Or is the DC resistance of the windings just the real (non-reactive) component of the impedance?
1b) Would grounding the point where wires A & B meet cut the source impedance in half?
2) What happens to the self-capacitance of the windings that are out-of-phase? Is it electrically canceled out? Cut in half?
3) The SNR is increased by a factor of 2, correct?
I think that's about it. Thanks to anyone who can help me out here.
Let's suppose we use a magnetic pick-up (a rod magnet wound with N number of overlapping turns of magnet wire) near an oscillating metallic object to generate an AC voltage. When we measure the source impedance, we measure the DC resistance of the of the magnet wire, correct? For the sake of practicality, let's suppose the DC resistance of N number of turns is 1Kohms.
This can be modeled (within reason) as a voltage source in series with a 1Kohm resistor with a capacitor in parallel (capacitance of magnet wire winding). Simple enough.
Now let's suppose we require as high a signal-to-noise ratio (SNR) from the source as we can get. For this purpose, we desire our AC signal to operate as a balanced source which is then run into a Balanced-to-unbalanced converter of some sort (either passive UnBal transformer or active opamp). Still with me?
Here's where things get tricky:
So the magnetic pick-up is wound in a bifilar fashion: we take two parallel lengths of magnet wire (wires "A" and "B"), each length (naturally) having two ends, or "taps"; an inside winding tap and an outside winding tap.
Because two wires have twice the volume of one, we wind wires A & B in parallel N/2 number of times around the rod magnet. We then connect the inside tap of wire A to the outside tap of wire B (or vice versa).
At this point, here are the facts as I understand them:
1) We are now left with two taps: an inside tap and an outside tap, the instantaneous induced voltage at each tap is 180 degrees out of phase respective to the other. Therefor, these voltages can be said to constitute a (floating) balanced AC signal.
2) Because we used two lengths of wire, each half the length of the example used up top, the DC resistance of each wire is 500 ohms (1/2 of 1Kohm).
Here are my questions:
1a) In my mind, I see two 500 ohm lengths of wire in series. Measured tap-to-tap, the source impedance is still 1Kohm. However, since the winding "switches direction" where wire A is joined to wire B, I imagine a "floating" node forced to reference when the output taps of the pickup are connected to the input of the BalUn stage. Or is the DC resistance of the windings just the real (non-reactive) component of the impedance?
1b) Would grounding the point where wires A & B meet cut the source impedance in half?
2) What happens to the self-capacitance of the windings that are out-of-phase? Is it electrically canceled out? Cut in half?
3) The SNR is increased by a factor of 2, correct?
I think that's about it. Thanks to anyone who can help me out here.
You're forgetting that by implication that any coil of wire is an inductor so your source impedance consists of series R and L with shunt C. Generally in a MM cartridge the R will be a modest few hundred ohms or so.
For balanced operation you don't need a bifilar set of windings at all, you would simply float the single winding. Having a center tapped winding is not a good way to get good CMRR in a transducer as balance can't be assured and it's not necessary, all you need is a winding that is not grounded at either end and has reasonably similar levels of shunt capacitance relative to ground at both ends. Keeping mass down and output voltage up is usually an important requirement particularly in moving coil type transducers. (cartridges, accelerometers, etc.)
For balanced operation you don't need a bifilar set of windings at all, you would simply float the single winding. Having a center tapped winding is not a good way to get good CMRR in a transducer as balance can't be assured and it's not necessary, all you need is a winding that is not grounded at either end and has reasonably similar levels of shunt capacitance relative to ground at both ends. Keeping mass down and output voltage up is usually an important requirement particularly in moving coil type transducers. (cartridges, accelerometers, etc.)
If you did wind it bifilar (not necessary, as kevinkr says) then the winding does not change direction at the join.
The SNR of the cartridge itself is not changed. You have the same thermal noise from winding resistance, and the same total voltage. Looking at each half separately, you would have half the ohmic resistance (-3dB on noise) and half the signal voltage (-6dB on noise), so 3dB worse off. But then you take signal from both halves: two lots of noise give you +3dB on noise, and two lots of signal give you +6dB on signal - you end up back where you started.
The reason why a balanced connection is good for a magnetic cartridge is to balance out noise induced in the cable. The pseudo-balanced connection usually used achieves this with an unbalanced input, provided that the cartridge has negligible capacitance to ground.
The SNR of the cartridge itself is not changed. You have the same thermal noise from winding resistance, and the same total voltage. Looking at each half separately, you would have half the ohmic resistance (-3dB on noise) and half the signal voltage (-6dB on noise), so 3dB worse off. But then you take signal from both halves: two lots of noise give you +3dB on noise, and two lots of signal give you +6dB on signal - you end up back where you started.
The reason why a balanced connection is good for a magnetic cartridge is to balance out noise induced in the cable. The pseudo-balanced connection usually used achieves this with an unbalanced input, provided that the cartridge has negligible capacitance to ground.
Thanks for the help guys, but I never intended this to have anything to do with MM or MC phono cartridges.
I understand the underlying principles behind magnetic circuits are the same regardless of the application, but my curiosity is in regard to an electro-magnetic pickup such as seen on an electric guitar. Such things as permanent magnet cores and reversed-polarity electromagnets (humbuckers) are conventional and that is the reason for their inclusion in my line of questions.
As such, there is no cantilever and mass considerations to be concerned about.
A primer for anyone who still cares
A single coil pickup conducts all induced voltages, including EM interference.
A humbucking pickup is two single coil pickups next to each other. The windings are in reverse with respect to the other, as is the polarity of the magnetic core(s). When the signals are summed, the result is common mode noise reduction through destructive interference (electrically 180 degrees out of phase) and a signal boost through constructive interference (electrically 180 degrees out of phase + 180 degree magnetic pole reversal = 360 [in phase]). This would not be possible without the magnetic core, as it must be reversed to re-align the phase of the summed signals.
The trade-off of the noise reduction of the humbucker is double the parts count and double the theoretical parasitics of the coils. Depending on whether the coils are wired in series or parallel (either increasing the inductance or capacitance, respectively) the resonant frequency of the circuit is lowered and the extra resistance of a series wiring configuration decreases the Q of the circuit. There is also the real world issues of the two pickups not sharing the same exact physical space as each other, as well as unlike winding irregularities. The effect is that no matter how you wire them, humbuckers will have a different tone than single-coil pickups.
Musicians have heated arguments about the superiority of single-coil vs humbuckers, much like audiophiles might argue the superiority of MC vs MM. I'm curious about a "middle path."
See pages 6-8 of this link on common mode chokes.
A humbucking pickup would be equivalent to the "sectional winding" common mode choke, the difference being the two "sections" are on separate, reversed-polarity cores.
I understand the underlying principles behind magnetic circuits are the same regardless of the application, but my curiosity is in regard to an electro-magnetic pickup such as seen on an electric guitar. Such things as permanent magnet cores and reversed-polarity electromagnets (humbuckers) are conventional and that is the reason for their inclusion in my line of questions.
As such, there is no cantilever and mass considerations to be concerned about.
A primer for anyone who still cares
A single coil pickup conducts all induced voltages, including EM interference.
A humbucking pickup is two single coil pickups next to each other. The windings are in reverse with respect to the other, as is the polarity of the magnetic core(s). When the signals are summed, the result is common mode noise reduction through destructive interference (electrically 180 degrees out of phase) and a signal boost through constructive interference (electrically 180 degrees out of phase + 180 degree magnetic pole reversal = 360 [in phase]). This would not be possible without the magnetic core, as it must be reversed to re-align the phase of the summed signals.
The trade-off of the noise reduction of the humbucker is double the parts count and double the theoretical parasitics of the coils. Depending on whether the coils are wired in series or parallel (either increasing the inductance or capacitance, respectively) the resonant frequency of the circuit is lowered and the extra resistance of a series wiring configuration decreases the Q of the circuit. There is also the real world issues of the two pickups not sharing the same exact physical space as each other, as well as unlike winding irregularities. The effect is that no matter how you wire them, humbuckers will have a different tone than single-coil pickups.
Musicians have heated arguments about the superiority of single-coil vs humbuckers, much like audiophiles might argue the superiority of MC vs MM. I'm curious about a "middle path."
See pages 6-8 of this link on common mode chokes.
A humbucking pickup would be equivalent to the "sectional winding" common mode choke, the difference being the two "sections" are on separate, reversed-polarity cores.
Yeah, I should probably just prototype the damn thing and stop confusing myself.
for links to some pickup theory, modeling, measurements:
UIUC Physics 498POM Guitar Pickup Measurements
UIUC Physics 498POM Guitar Pickup Measurements
Well, like I said, your help is still appreciated as my original questions still stand.
In fact, my reply took several hours to compose because of the excellent points you guys raised.
Seeing as how far I actually am from prototyping the pick-up (just moved, no test bench setup), I would like to have as strong an understanding as I can get of the circuit from a theoretical standpoint.
As it goes, only today I wound a bifilar pickup (non bucking) - then found this thread! ach end of the coil feeds into a differential type amp....the centre tap provides the bias at 1/2 VCC to the +ve pins of two opamps..
That said, it was a bit of a faff (wire is very thin to work with!) , so I'm curious as to kevinkr's initial reply....
"For balanced operation you don't need a bifilar set of windings at all, you would simply float the single winding"
I'm assuming here, you mean rather than have one end tied to ground (as per a normal pickup), you simply have each end of the coil at a common DC voltage (supplied via a highish resistor)....are we saying that this would yield a signal with opposite polarities at each end of the coil?
If not, could someone please describe it more?
btw re a bifilar pickups - there is no middle path here, if you want to buck hum, you need two coils that each have their own reversed magnets!
That said, it was a bit of a faff (wire is very thin to work with!) , so I'm curious as to kevinkr's initial reply....
"For balanced operation you don't need a bifilar set of windings at all, you would simply float the single winding"
I'm assuming here, you mean rather than have one end tied to ground (as per a normal pickup), you simply have each end of the coil at a common DC voltage (supplied via a highish resistor)....are we saying that this would yield a signal with opposite polarities at each end of the coil?
If not, could someone please describe it more?
btw re a bifilar pickups - there is no middle path here, if you want to buck hum, you need two coils that each have their own reversed magnets!
Last edited:
In this context "floating" a winding means connecting one end to signal and the other end to signal return, neither being grounded at the winding end of the cable. No DC or resistors mentioned. At the other end of the cable it either goes into a balanced input, or one side is grounded and the other goes into a singe-ended input. In either case magnetic induction into the cable is balanced out. If the winding is connected to anything else (i.e. not floated) then this cancellation no longer works.
My idea for implementing a balanced input was an mic input transformer wired as an BalUn with a 1:1 or 1:2 turns ratio. Or some ratio inbetween.
It sounds like I may be making an assumption about bifilar pickups which is untrue. Perhaps we can contrast them with humbuckers which, in my mind, seem to be identical to a bifilar pickup once it is connected to a BalUn.
I just discovered Gibson applied for and subsequently was awarded a patent on this a few months before I started this thread.
https://patents.google.com/patent/US8802959B2/en
https://patents.google.com/patent/US8802959B2/en
Read the small details of the 8802959B2 patent. It's not at all what you are proposing. "Bifilar" is not novel and Gibson does not try to Claim it. In fact they have many words saying it may be an UNequal number of turns (without actually saying why). The two windings are NOT connected (again, no clue what benefit). I think they don't know what it is good for, but got the patent to protect against others using the specific ideas in their Claims. (Which suggests your idea is either Prior Art or Obvious To One...)