winding band pass coils

[Erik Mandaville]

These are for a new network I've been working on and refining. I needed coil and cap values that aren't common, so wound my own. The green magnet wire is from a c. Klipsch 1960s Heresies. Brittle enamel was really hard on finger tips, but it was good wire I could use for this project. Just a couple of pictures to show the process. BTW: It's possible to *arrive* at a needed odd value, but you have to buy larger and unwind and measure until you get the needed amount of 'L' (Inductance). This technique gave me exactly what I needed essentially for free (since I wasn't doing anything with the wire).

Erik

Note: Meter indicates a long way to go, still. This is winding more than 150ft of wire onto a 1" former.

[Erik Mandaville]

Taken off form, wire-tied, and drying after dipping in a poly varnish. The inductors bubbled in the varnish, so I know it went deeper than just an outer coating. Keeps the windings from moving with respect to one another.

DCR on these were the same as those I found available retail online for the same AWG -- in this case 16.

[Erik Mandaville]

Since I went from a 4500Hz. crossover point to 4000Hz, I needed to change both inductor and cap values for the bandpass -- which was then altered again because of the interaction between the two components. That requires a certain percentage subtraction from the calculated capacitance, but an added percentage of the calculated value for inductance. The only way I could do this was to remove the former motor run cap, and then parallel capacitors to get what I need. In this case, I hit it right on the head, luckily. The rectangularly shaped cap is of course an early Klipsch, which measured exactly 8uf on my meter. Solen are fine for me.

I'm rebuilding the new networks tomorrow morning after the varnish on the coils has dried. If you are familiar with Bob's type 'A' with the modification for the 4,500Hz crossover, than you understand this new network I made. It's really very similar. I'm just using a slightly lower crossover point, and have replaced the autoformer with a fixed L-pad. The inductor for the woofer is vintage Klipsch, and measured a little under 2.5mH, which is what I wanted. The band pass changes result in a sharpening, or more discriminating circuit (aka: 'narrow band pass).

What makes this easier is that we have known entities to work with -- the drivers. As the 'A' does not use notch filters or impedance compensation, nor does mine. The only anomaly I'm aware of has to do with that little peak or notch around 9kHz for the K-55V with spring terminals. That's been repaired with a series notch filter, although I never really found I needed one. They are simple and easy to install if you some basic related skills; I just chose to leave it out since I don't have a problem with it.

Here are the piggy-back caps:

Erik

[Guest " "]

do you do custom work.

need two 14mh inductors with 6 leads 2.8hm apart

they only need to handle 75 watts

[Erik Mandaville]

The schematic is nothing -- so very simple. I would be glad to post a hand-drawing of it, however until I've tested it, I prefer to leave the values out. I have so much other work I need to get done to get ready for school to start again, and I didn't even plan on making a new network. It came about accidentally by way of an experiment. I just happened to like what I heard so much that I decided to take the time to refine it. See that doggy in my avatar? That's Sarah. She's undergone three surgeries for cancer, and just completed her last round of chemo-therapy today. We don't have pet insurance, and it can be expensive. My point is that I could make the changes in the network if I made what I needed instead of spending more money for parts and shipping. So I hand-wound the exact value of inductance required, and dug through bins for the capacitors. I made and found what I need to complete these networks for the time being.

Erik

[Erik Mandaville]

do you do custom work.

need two 14mh inductors with 6 leads 2.8hm apart

they only need to handle 75 watts

I'm not winding coils for anybody but me -- sorry 'bout that. It's time consuming and hard on the hands, and if you want a multi-tapped choke at 14mH, which is what it sounds like you're looking for, that would be big work for slow hand winding.

Erik

[Guest " "]

sounds like we need a new invention.

[Erik Mandaville]

"sounds like we need a new invention."

That invention already exists. For small inductors such as coils in loudspeaker crossovers, hand winding works. For me, it's just more in the spirit of true DIYing, and tolerances are obviously important, but not so critical as when working with components designed for RF frequencies. As it is, these coils are exactly what I need, and I didn't have to buy them at retail prices and then end up wasting money and wire by unwinding a larger value in order to get the required inductance.

Obviously hand-winding isn't going to result in an exactly 'perfect lay' wind, but it's possible to get quite close. What a signal sees is the inductance and resistance, not whether a winding drifts slightly to one side or another on the former, and both of those elements are equal to commercial options.

Creating a multi-tapped autoformer at 14mH creates some complexity. You have to be able to work on the coil at different levels, while mainting wire tension at the same time, then rewinding, stopping again to measure (probably several times), adding the next tap. And so on.

Erik

[Erik Mandaville]

The last two I need for a network that's taking me in a direction I haven't gone since using more powerful SS amplifiers and dynamic mid-bass drivers and dome tweeters (and higher order crossover networks). There are some advantages to a little more power, I think. In any event, I'll end up with two complete networks, but setup for slightly different crossover points and orders of slope.

I'm putting together a second pair on a piece of black board.

Winding the last, slightly small inductors: I have to wind slightly above the desired value, since the inductance changes when the coil is removed from the dowel former/bobbin.

[Erik Mandaville]

I've got some ideas for setting up a winding machine using a drill, as well as something to maintain tension on the magnet wire as it goes on the former. My problem is that I don't really like inductors like these, store-bought, or what have you. They're sort of a necessary, not-so-good thing. Kind of like output transformers in certain respects.

Here's the chunk of wire after it's been taken off the spool. I'm using vintage magnet wire from my old Heresy low-pass inductors, and the wire bends a little as it's unwound. Kinda hard to keep it perfectly 'true' as it's wound 'round-and-'round.

Both this coil and the next one I make for the other channel will be soaked in varnish and then allowed to dry.

The autoformerless networks will be done in a day or two.

[boom3]

There are coil-winding machines, which use mechanical counters to keep track of the number of turns. The ham radio community has produced a number of these, some with electronic counters, tachs, micro-processor control-the whole magella. If you surf around the ham sites, you can find these, or, go to a hamfest swap meet. Lots of interesting gear there.

An electrical shop which rewinds motors might be able to do this for you. Winding coils is pretty simple compared to winding motors, so they could probably make a multi-tap coil pretty easily.

A 14 mH coil (if an aircore) is going to be very large if wound with good-size wire.

I've bought commercial (Erse) inductors that were just over what I needed and then slowly unwound them, taking a reading with a good LCR meter every so often. Key word here is "slow". The inductance changes faster as you approach the core.

[Erik Mandaville]

Good point! My dad builds his own radios, and has mentioned the same, although the tolerances are a bit tighter at RF than audio, which is really, really low by comparison. I guess I sort of like doing the work entirely by hand, but if I keep going at this rate, it might be worth it. It's hard on the hands and fingers after awhile.

That's great that you unwound those Erse coils for what you need, and you're right, the inductance can change more quickly than expected. "Slow" IS the way to go.

Erik

[whamo]

That's very cool Erik. Thanx for sharing that. Mark

[Erik Mandaville]

Thank you, Mark. I just want to try something new, and an earlier experimental crossover without the autoformer surprised me (in a good -- IMO -- sort of way).

Erik

[Guest " "]

"Creating a multi-tapped autoformer at 14mH creates some complexity.  You have to be able to work on the coil at different levels, while mainting wire tension at the same time, then rewinding, stopping again to measure (probably several times), adding the next tap. And so on. "

so maybe the best inital approach would be for me to put 5 lower value inductors in series, and use their connection points as the taps.

[Erik Mandaville]

Sure, you could could connect them in series. You'll just have some winding in front of you, but there are people who have hand-wound giant crossover inductors, Tesla coils, etc. Making some speaker inductors isn't a big deal, really, but it's helpful to be able to make exactly what you need.

PS: two in series isn't so problematic, but you may run into some problems with five, particularly since you're trying to create a multi-tapped choke.

[boom3]

"Creating a multi-tapped autoformer at 14mH creates some complexity. You have to be able to work on the coil at different levels, while mainting wire tension at the same time, then rewinding, stopping again to measure (probably several times), adding the next tap. And so on. "

so maybe the best inital approach would be for me to put 5 lower value inductors in series, and use their connection points as the taps.

Yeah, that's an approach. However, I'm not sure if the efficiency of 5 individual inductors will be as good as a tapped coil using one core. OTOH, even with that, you may have less insertion loss than with an L-pad. Try it and see, and please let us know.