Project "Groove Scribe"- A DYI stereo 45/45 head.

[Snug Music]

I was also at the university, but if you watch here, what are you doing here with the magnetic fields, because I will really dizzy .... smile
Greetings Scotty

[opcode66]

Split coil is what I've alluded to the entire time.

[DjMikiRia]

For better high frequency response, try slowing the source track from 1 (normal 45 equivalent) to 0.67 (33.3 equivalent) and playback at 45rpm. That is one thing I learned from another lather on here. I suppose if possible to get even better highs for 33.3rpm one could do the same to the source track, but my turntable only had 2 speeds so I have never tried that slow.

sincerely,
Mikiness

[Fela Borbone]

Hi! congratulations for the great work and promising results ,
The split feedback coils are usefull for cancelling hum and some drive flux leak, they work like this hum killing guitar pickups. I made some experiments with the usual configuration une above and one bellow, but it was very frequency selective, so developed a " double D" situated in in the same plane. There is more info in a thread called " DIY feedback" somewhere . But I found an affordable rekokut and this project is waitin... maybe wake up one of this days!

Ciuens, is your shield really a ring or "C" shaped? (I looks like it in the picture) a true ring will bring the shorted turn effect. If not.,only small eddie currents will form.

[Ciuens]

Ciuens, is your shield really a ring or "C" shaped? (I looks like it in the picture) a true ring will bring the shorted turn effect. If not.,only small eddie currents will form.

Its a ring.

[markrob]

Hi Bryan,

Using the two magnets in your second drawing won't quite work. The problem is that if you orient them N to S from left to right, both sides of the coil will cut through the same field direction. If you look at the speaker diagram, you can see the center pole is S and the outer pole is N. So, as you travel from left to right through each side of the coil you go N-S and S-N.

Perhaps using the fringing at the ends is the best bet here. See what signal you get and determine if its above the background and coupled pickup from the driver (using the shielding you have in place). If you get enough signal, then don't sweat it. Also, don't forget to make sure the input impedance of the amplifier you connect the coil to is high enough such that the L/R time constant does not cause high frequency rolloff (and the phase shift this brings). As you wind more turns, you increase inductance which gives you more signal, but also means higher input impedance to avoid the rolloff. You could also run into self resonance if you go too far with the number of turns due to parasitic capacitance . Inductance goes up as the square of the number of turns, while induced voltage is only proportional to the number of turns. You can see the implications.

Mark

[EpicenterBryan]

...so developed a " double D" situated in in the same plane. There is more info in a thread called " DIY feedback" somewhere.

I totally forgot about that thread Fela!
http://lathetrolls.com/viewtopic.php?f=15&t=4231

Looking back over that thread, why did you have different number of turns on one coil vs the other (first version - not double D version)?
By the way I looked at Femm 4.2 that you mentioned (Finite Element Method Magnetics) from this link:
http://www.femm.info/

I love that it's free! Is that what you used when you worked on your magnet replacements for your Presto head?
It looks pretty complicated to set things up. Are you interested in doing a simulation for this project?

Bryan

[EpicenterBryan]

The problem is that if you orient them N to S from left to right, both sides of the coil will cut through the same field direction.

Yes, but I was going to try to orient them so that the same polarities are on each side of the coil (two opposing N faces). That would bend the forces but I see that is really what the ring is doing inside if I'm visualizing that correctly.

Perhaps using the fringing at the ends is the best bet here.

I'm going to print some spacers to go on top of the magnet (open in the center). Then I can put the film over the magnet at known distances and see how many mm the white zone is usable on the end. That way I know how long the coil should be. Next up, I'll try to wind two coils on the same core - one CW and the other CCW and try out Fela's Humbucking like dual coil idea. This is pretty cool, not only do we get signal from both ends of the magnet, but in theory if we pick up some noise from the surroundings it should help cancel it out.

Ignore the text on this image - it's the only one I found that showed dual coils wound opposite and connected in series. Imagine the magnet placed between them. This is just to show the winding directions.

humbucking.jpg

This is where I'm going to add a little twist.
Since the white area (the fields I need to cut through) gets smaller in diameter as I move slightly further way, there is no point in having the same number of wires (layers) throughout the length of the coil. If out at the far end, there are really only 2 layers cutting through the usable field, the other 8 (if it's a 10 layer coil as an example) are just dead weight (mass). So why not taper the coils in a cone shape?

As you wind more turns, you increase inductance which gives you more signal, but also means higher input impedance to avoid the rolloff.

It brings up another interesting idea that I may not mess with, but the bobbin core could actually be tapered and we may only need a few layers over the form, and the rest of the shape is filler material (only if it's lighter than the copper it would replace). That not only reduces non-needed copper mass, but also minimizes the number of turns Mark mentions. If it's a mass wash Mark, do you think it's still worth doing to keep the inductance as low as possible?

revised dual coil.jpg

Chime in!

Bryan

[EpicenterBryan]

Split coil is what I've alluded to the entire time.

Todd, your inputs are always welcome and very much appreciated. If you alluded to split coils on this thread please point it out because I would like to re-read it. Some times you speak in puzzles and I don't always catch them.

Bryan

[opcode66]

Page 11

I'll let you know of any of you identify the other two requirements correctly.

Page 13

There are a couple other tricks. But, I'm keep them to myself. This is already giving you all a lot for nothing.

Page 14

If the interaction between the coil and the magnet isn't optimal you can change the magnet or change the coil. Not saying anymore than that.

I would include this from Page 14 as well since I'm trying to point out the underlying reasoning behind splitting the coil.

What you are not understanding is how the coil is moving in relation to the dividing line of N/S in the magnet.

I've hinted other places as well. There are still two pieces of the puzzle you don't have detailed here. Just so you know. I know two more critical design features that have not been documented anywhere here or in other threads.

Other than that I can show you historical changes to my push rod which very early on in feedback incorporated a split coil design.

[markrob]

Hi Bryan,

If you place a second magnet with the pole flipped so that there is a N pole on each side of the coil, I think you'll be in worse shape. The will be no flow from N to S and in fact, at the center of the coil, no field. You need the S pole in the center to allow a return path for the flux.

The headphone transducer might be a good thing to try.

Mark

[Fela Borbone]

Its a ring.

Ooops sorry I was mislead by the black stripe

Bryan, no problem in simulate this, but it has some limits... no 3d,no forces,and no coil and magnet simulation at the same time... I dont have exact measuresbut the following screenshots can give a clue about magnetic fields around. This is an axissimetric view,( a slice if the magnet were a cake ) I merged some shots in two images. first comes the coil and ring magnet at the same level force lines dont head across coil . then magnet is lowered and some lines pass more perpendicular but still poor.

[Fela Borbone]

...and with an iron washer on top of the magnet it drives flux better oriented.
last, a magnetic circuit (Just a a qwick and dirty one, can be optimized) adds lot of wheigt in the push rod, but increases efficiency. note that one air gap is wasted bellow. if other coil were there it can use it.

[EpicenterBryan]

Like usual Todd had valid points. He also kept a few critical points to himself. I just wish Todd would say more than "change the magnet or change the coil" or "closing the gap is critical". How about a diagram man, I'm a visual learner!

So let me share what I've learned and what I'm going to do:

First off, the magnets most of us are now playing with are polarized axially and are hollow in the center. Good start if we don't have the luxury of having a center pole like in ported headphone drivers ect... The field we need to cut through (the white zone in the green shots) is near the inside edge with this magnet except near the end where the white zone curves and gets smaller with small distance from the end. Closing the gap inside the tube is hard for many of us home made guys to do. It makes aligning stuff more tricky because of the tight tolerances and small gaps. But we all know now that any coils in that zone produce voltage when moved, any that are in the black zone do not and are useless.

So here is my revised idea for one end of a coil, using my existing magnets. This is not to scale.
This magnet is 5mm thick. Based on spacer and green Magna-vu observations, there is really only about 1.75mm of usable fields inside the magnet along the axial polarization prior to field reversal, and about 1.75mm outside fields concentrating to around 3mm. For something more usable for me, it's really more like 1.25mm at about 6mm diameter outside of the magnet.

The coil I was using before was on a 5.4mm core, with a center that was 3.2mm. It was 5mm long. It had 4 layers of 50 winds on it for a total of 200 windings.
This tapered coil will have a core that varies from 4mm on one end to 6.8mm on the other. When covered with 4 layers of 0.1mm wire it will have much higher efficiency with fewer windings and less mass.

revised end coil.jpg

Here is what the CAD file looks like on the right, but what the form would look like when actually printed with a 0.4mm nozzle on a 3d printer. Since the wire is only 0.1mm thick, I have no clue what this will really look like but I'll give it a shot. The idea here is fewer windings, and closer to the white zone both inside and outside of the magnet with non-necessary windings eliminated.

former.jpg

Yes, I have wider magnets on order (10mm) so in theory I can double output with existing coils if moved...but I'm still going to change things when they get here... So again, Todd can say he was right when he said to "change the magnets".

Double it and wind the other direction for the other end and increase output but also take advantage of the Humbucking effect...


Who knows - Wish me luck.

Bryan

[Bahndahn]

Hey Bryan just wanted to say that I have been really appreciating your updates and am learning literally way more from your experimentation than I am in university right now ! You are a noble person, man.

Keep it up, and good luck! I'm riveted to the updates.

[Fela Borbone]

Another interesting point for design is that when the coil moves in uneven magnetic fields, the signal generated may be not a true representation of the motion ( or the opposite if we are talking about drives).In my experience is not so terrible if "just fine" results are expected.,but think you are going to mix it with your carefully recorded, mixed and masterized program...
Overhang configuration(when the coil is long compared with the gap,or magnetic flux fringe, )is npt problematic as when a coil turn moves in, another moves out, whatever uneven field is in the middle.
In underhang, when the coil is shorter, a gap controlled with a proper magnetic circuit brings an even magnetic flux and puts all copper at work.
As FB coils are not too heavy, an extra length is a safety net to avoid this kind of distortion.

[Ciuens]

Guys, I confess that I am very confused, hahaha. Too much information, my English is bad, and long to understand all the theories involved, especially when Mark says. Is valuable information to be digested. But I'm reading and re-reading to understand all the theory involved. Thank you all for sharing.

Ciuens

[juba bc]

its is possible for a cutter head construction??

[Snug Music]

?? 10mm ....Am a bit confused, how does it work juba??

Scotty

[Snug Music]

Guys, I confess that I am very confused, hahaha. Too much information, my English is bad, and long to understand all the theories involved, especially when Mark says. Is valuable information to be digested. But I'm reading and re-reading to understand all the theory involved. Thank you all for sharing.

Ciuens

Hey cius,..
Have u made a new test with your actualy driver uodate? How much is the stiffnes on your driver?

best regards..
Scotty