A closer look at the AF108 bi-polar drive circuit

[GeoMax]

Hi Carl,

thanks for the information.

[Carl-NC]

180uH per side (also posted in your other thread)

[GeoMax]

Hi,
the last post to this thread is quite a while ago. But maybe someone can remember.

Question:
1) How is the headset connected to the audio circuit? (see schematic page 1, post #1) There are two external sockets labeld with "Audio Socket" in the diagram.
Audio Socket "A" and Audio Socket "B" & "C".
2) What is the coil induction of one site of the bifilar coil?

dbanner
What is the last progress on the AF-108 project? Have you acomplished it?

[6666]

dbanner thanks for your post, it reminded me of something I had forgotten to read, cheers.

[dbanner]

I look back at my old posts and wonder what the hell was I thinking. Ha.

The relationship between number of turns and inductance is non linear. So halving the number of turns does not halve the inductance. In fact you end up with much less than half. The formulae proves this( and also the coil calculator).

However for a bifilar wound coil, I suspect that the overall inductance might be ever so slightly higher for the same number of turns due to the sequential side by side laying of the turns next to each other. Don't know if it would make any difference, but in the case of a centre tapped coil, a bifilar wound coil would make both halves extremely similar and achieve a better symmetrical balance.

[6666]

If I remember correctly, each half had a calculated inductance of around 1mH( this was the coil I wound for the aquasound) . However when I connected both ends to a tgsl collpits oscillator (leaving the centre tap floating) I calculated the end to end inductance as 3.7 mH. So the total was more than the sum of the halves.
This I attributed to mutual inductance and high coupling coefficient.

I have just wound a 6 inch bifilar coil each half is 82uH, but connected as centre tap total is 320uH, interesting phenomenon.

[dbanner]

If I remember correctly, each half had a calculated inductance of around 1mH( this was the coil I wound for the aquasound) . However when I connected both ends to a tgsl collpits oscillator (leaving the centre tap floating) I calculated the end to end inductance as 3.7 mH. So the total was more than the sum of the halves.
This I attributed to mutual inductance and high coupling coefficient.

[6666]

What value of inductance does each side of the bifilar coil have ?, thanks

[dbanner]

K would be approaching unity. However there is never 100% mutual inductance due to flux leakage.

If you know the formulas, then it is possible to calculate precisely the coupling by testing the coil.

Important thing here is not so much the coupling coefficient, but that the coil is dynamically balanced so they give a symmetrical response about the centre tap. The tighter the coupling between both halves, the better.
Winding in bifilar arrangement simply makes for a better "dynamically" balanced coil.
A randomly wound coil would not only have less coupling but would be less balanced because of random vector dispersions of magnetic flux.

[eclipse]

Any idea on the coupling (k) on the bifilar wound coil vs randomly used and joined coil ?
For the bifilar wire maybe for test purposes a speaker cable (0.35mm2, 0.5mm2) red and black can be used?

[dbanner]

Yes.
The bifilar coil should encourage a slightly higher flyback voltage too, as the flux density is increased. Would be interesting to wind two coils, the first is wound in traditional manner and the c.t taken at the middle. The second one is wound using bifilar windings. Both coils having same number of turns.
Then the flyback voltages are measured for each coil for comparison.
You would expect critical damping would be slightly different too.

[6666]

From WM6

[dbanner]

Just recently I wound a bifilar centre tapped coil. I can tell you that the mutual inductance is certainly enhanced. I wound the coil using two wires simultaneously. This ensures that the coupling is best because the wires end up next to each other throughout the coil. The start ends end up connected to the drive circuit, the finish ends are joined for the c.t.
You get more magnetic flux density for the same DC series resistance and number of turns.(I think)

[Carl-NC]

Is the center tapped coil wound with bonded bifilar wire type or you wind TX1, cut the wire at the end, and start winding TX2 ?

You can do it either way, but bifilar wire gives the best coil matching. No idea how the '108 coil is actually wound.

[eclipse]

You need to add a center-tap ground on the coil... the '108 uses bipolar pulsing by alternately pulsing each side of a center-tapped coil. The SK2 side therefore generates both a negative flyback and a mutually-coupled positive flyback that are used as the RX signal. Same thing happens on the SK1 side, except those drive comparators used for (it appears) fault detection.

Is the center tapped coil wound with bonded bifilar wire type or you wind TX1, cut the wire at the end, and start winding TX2 ?