Coil connection to PCB

[brumbarchris]

Hello,
I am struggling to find a suitable wire-to-board connection for some coil sensor we are currently working on.

The coils are made from solid 0.25mm diameter copper wire (30 AWG) , isolated by a thin enamel layer. I suppose it qualifies as "magnet wire" although I am not sure... We need to connect these coils to a PCB, and ideally I would do it through a connector rather than through direct soldering.

We currently have a Molex Picoblade connector on-board:
https://www.molex.com/molex/products/part-detail/pcb_headers/0533980467

This means I had to use a correponding crimp and crimp housing on the coil end wires themselves, still from Molex:
https://www.molex.com/molex/products/part-detail/crimp_terminals/0500588100
https://www.molex.com/molex/products/part-detail/crimp_housings/0510210400

Nevertheless, our current experience indicates that when crimping the terminals to the coil wire, the result is a very brittle connection, which can easily break after a very limited number of flexings of the wire. The breaking point is the exact entry point of the wire in the crimp terminal. So clearly, such connectors are unsuited, and I am looking for a different solution. I did find some I did find some dedicated "magnet wire terminal" ranges from reputable manufacturers (https://www.mouser.ch/_/?P=1yzoqgn) but these seems to mostly geared towards connecting a magnet wire further to other wires, and they are very bulky, compared to what we need.

I need to connecting solution to be rather small in size, and with as little metal as possible (the metal adversely influences the readings from our coil inductive sensor nearby) and to allow connection of 4 wires to the PCB. Pre-stripping the enamel isolation from the wire is acceptable.

What solution would you suggest?

Best regards,
Cristian

[Berni]

Yes this is a problem of solid core wire, it just simply breaks once bent around enough. When you put a connector on it you encourage bending at one spot and this makes it break even sooner.

Typically the connector style for solid core is telecom insulation displacement punchdown terminals. They hold the wire a bit more loosely so that it has some wiggle room.

If i was doing this i would give up on any sort of crimp connector, id rather just use a regular troughhole PCB connector, solder wires to it then cover it in heatshrink or some flexible glue to give a gentle support to the wire, letting it bend over more of a radius. Even with regular stranded wires these sort of connectors are the most resilient of them all.

[Ian.M]

As Berni has just commented, crimping a thin solid enameled wire in a terminal designed for stranded PVC insulated wire is never going to be reliable, especially with nothing to provide any strain relief, as its likely the wire will be severely deformed at the point of entry, weakening it unacceptably.

If prototyping, soldering the wires into partially pre-crimped Picoblade contacts, (so as to avoid deforming the wire at the point of entry) + gooping the back of the connector shell with electrical grade neutral cure silicone (or other suitable flexible adhesive) to provide strain relief, may give you sufficient improvement in reliability.

For production it gets trickier and 'gooping' down the wires and soldering them direct to the PCB may actually be the best option.

Note that 30AWG wire is on the edge of being fine enough that weakening due to the copper of the wire dissolving in the solder joint becomes a problem, 'necking' the wire at the surface of the joint.  Use a solder alloy with a significant copper content to avoid this.  Many ROHS alloys are already suitable, but if you are still using Sn/Pb, get a 'Savbit' alloy e.g. Sn50/Pb48.5/Cu1.5.

[mikeselectricstuff]

For 30AWG I think you need to be soldering the wire to something which then plugs in.
Maybe a very small PCB, which could also incorporate some support or strain-relief, with a PCB-PCB connector.
Or perhaps wrap and solder to  pin-header
Crimps and IDCs are out, there may be some sort of spring-contact connector that could work.

Maybe something like this
https://media.digikey.com/pdf/Data%20Sheets/Molex%20PDFs/Lite-Trap_Mini-Lite-Trap_Conn.pdf
https://www.farnell.com/datasheets/2646176.pdf

They spec a minimum o 26AWG solid, but may be useable with thinner.

[T3sl4co1l]

A good idea is soldering the leads to stranded wires, which go on to cables.  This is common enough for coils, switches, thermistors, etc.

Preferably, that join is done up not just with heatshrink and whatnot, but also on a support panel/card/enclosure that's part of the coil itself, to minimize flex.

For example, traditional power transformers terminate the connections inside the windup, stacked with insulating tape and usually stiff cardboard or plastic panels.

Or use very fine flex, like how speakers have that tinsel stuff between coil and basket.

I don't know how many of these are reasonable given existing design and possible limitations, but maybe it's something.

Tim

[jonpaul]

Bonjour brumbarchris

Like most posts re transformes, magnetics inquiries, you do not give the coil details.

Much beetter responses if you can supply   photo, winding sheet or cross section diagram.
Is this a solenoid? Had a core? Voltage/current Used on PCB? Environment eg portable equi, aviation, medical, industrial?
The wire #30 is indeed "magnet" wire.

You can never terminate #30 without soldering.

Normal practice in magnetics/transformers is use of a bobbin, with PCB pins.

The coil or transformer is wound on the bobbin, with solderable wire eg Soldereaze, Nylease, etc.

The wire  termination is wrapped around each bobbin pin, and the pins are soldered with solder pot or iron.

Then the finished part is mounted to the PCB.

Both SMD and TH pinned bobbins are available.

https://mhw-intl.com/products/bobbins/

Bon chance,

Jon

[mikeselectricstuff]

Or just wind and solder the 30AWG round a pice of solid-core tinned-copper wire, 24AWG or more, and poke this into one of the connectors I linked above, or perhaps a turned-pin contact like an IC socket.

[brumbarchris]

Thank you all for your valuable (and prompt) input.

@mikeselectricstuff: the Molex lite-trap series seems on its path to obsolescence, only the vertical variant and the bare terminal contact are still available; but admittedly, it would have looked interestingly. The WAGO series you suggested is something that I am familiar with, however they are only suitable for thicker wires. Of course, I could use an intermediate piece of wire, like you and Tim suggest, but I'd rather go without this additional intermediate piece of material.

@jonpaul: the coil is an aircore coil which sits directly on the PCB (fixed to it through other mechanical means), something like this:  http://www.jantzen-audio.com/air-cored-coil/
The PCB together with the coil environment will be completely sealed, so there is no need for this connection to withstand any specific humidity or dust hardships, but it has to be mechanically solid, as the portable equipment in which it is employed can be dropped. Voltage is +/-5V and current is some 200mA.

At the moment, I am seriously contemplating to just design in some soldering pads on the PCB (perhaps also with a through hole) and directly solder the ends of the coil to the PCB. Sort of ugly... but then anything else apart from a dedicated connector looks ugly, in this case. And there seems to be no dedicated wire-to-board connector systems for attaching such thin solid wires directly to the PCB. Amazing!

[langwadt]

RJ11 ?

[SeanB]

How about making a flexible PCB with the one side having solder pads for the wire, and the other being there for a common pin header that then solders to the board. Then the coil is wound and soldered to the flex, and the joint is sealed with some form of varnish to protect the wire. You cou7ld ven make the flex PCB to fit in a coil former, and if few enough turns a multilayer PCB could replace the wire entirely, just using the flex PCB as the coil itself. Multilayer flex boards are relatively cheap, though you could also do the same with a multilayer PCB as well with rigid board, and simply have the pins there as well.

[brumbarchris]

Hi, I am not sure I see the advantages of this suggestion, over just soldering the coil wire ends directly onto the original PCB.
Maybe I have not made it clear (enough) above: the coil sits on the same PCB to which it needs to be electrically connected. This means there is no relative movement between the two elements (coil vs PCB), so the additional in-between board only adds a layer of material with no specific benefit.

[jonpaul]

This is a speaker crossover coil, the Jensen uses a self bonding wire, to form the coil.

If you Wind them with regular wire it will need impregnation, encapsulation or a bobbin to hold it togather.

Many pins are available for PC insertion, that can be staked into the PCB hole and soldered to.

Concord and Keystone are good USA suppliers.

The "dedicated connector"  you seek  a coil bobbin.

I would wind it on a bobbin with PCB pins, like a potcore but no core.

Alternate is to make one as the Jensen, then use epoxy or hot melt to fix to the PCB, very common in the production consumer hifi soeaker passive crossovers.

Jon

[T3sl4co1l]

Yes, that is perfectly acceptable, wires tacked to solder pads.  For another example, RFID coils are normally terminated this way.  A bit of goop over the connections, and preferably up the wires to the coil (or tape over them, etc.), will keep them from shaking due to vibration.  Or potting, since it's sealed.

Tim

[Terry Bites]

Crimp connectors are ok as long as you use the right tool and not a crappy pair of pliers.

[brumbarchris]

Crimp connectors are ok as long as you use the right tool and not a crappy pair of pliers.

Nope, unfortunately not, already tried that. We crimped with the original Molex dedicated crimping tool. Crimping quality looked good, but then if you bend the wire a couple of times a the exit point from the crimp it easily breaks. So the wire breaks, not the crimped connection (which holds firm)

[T3sl4co1l]

Part of the trouble is the terminals are made to grasp plastic insulation, as strain relief.  With enamel instead, there's nothing to grab, and stress is applied directly to the crimp section, which is already stressed from having been crimped.

Even with [grasping from the end prongs], it's solid wire without strain relief, nothing to increase its radius of curvature when bending force is applied.  So it fatigues easily.

You may've also nicked it while stripping off the enamel, depending on exactly how you did it and all.  But since it broke at the crimp, that's probably fine, and it's simply as above.

Whereas, as intended, the double whammy of plastic insulation acting to pad the bend, plus being divided into strands, the bending radius per strand is much larger in relative terms, so the flex life can be massively higher.

Tim