How to test a speaker voice coil(raw part, not assembled to anything) 3

[lixc]

I am dealing with voice coil.
How can I detect a voice coil whether it has any short-circuit or open circuit defect ? My main concern here is intermittent cases, eg. intermittent short/open (sometimes short, sometimes open, some only can be reveal after it is assembled to be a speaker and full load is applied BUT i need to detect it before it is being assembled.).
The copper wire for the coil is very thin , so the tester shouldn't predamage the coil during testing period.
I need a fast test method and can indicate "GO/NO-GO". The price if possible lower then better.
Thank you.

 

[MacGyverS2000]

Open-circuit shoudl be easy enough to test. Put it in a paintcan shaker kind of deal to test for opens during vibration. Not sure of the best method to test for shorts...with an accurate ohmmeter and the known length/guage of wire, you could determine what the resistance should be compared to what it actually is.

 

[BrianG]

Measuring the resistance will certainly find open circuits, but for short circuits it is unlikely to show much of a change compared to the overall coil manufacturing tolerance.

The best way to test for short circuits in the winding is to measure its inductance. A shorted turn will seriously reduce this. You will need to use a reliable L-C-R bridge; measuring L, the inductance, will tell you if it is open or shorted in one test. Modern bridges, with digital readout, can usually be programmed for "go" "no-go" range limits, so should be quite speedy to use.

 

[lixc]

Hello BrianG:
Can you give some example of these bridges that you mentioned?
I agree that a short-turn will decrease the inductance and resistance. BUT an intermittent short-circuit is the main problem. It sometimes show , sometimes doesn't show and, sometimes only show when it is assembled to be a speaker and the speaker is being played for long time under full load.

 

[BrianG]

There are literally hundreds of LCR bridges to choose from - just do an internet search on "LCR bridge".

Your intermittent problem seems to be one of material specifications. If the enamel insulation on the wire is failing in manufacture or on load, then you need to use a better specification wire. You could try to improve the insulation by coating the wire after coil winding, using either a high-temperature varnish or an epoxy resin.

 

[OperaHouse]

Dumb and cheap bridge. Compare it against a known good speaker. Use an audio oscilator to supply the top and bottom of a bridge. Top two resistors of the bridge should be 330 ohm resistors. Use the good speaker on the right bottom side of the bridge and the left "test speaker" on the left. Measure the AC millivolts between the two speakers. Use a frequency appropriate to natural resonance of the speaker. This will show a variety of mechanical and electrical problems. Have used this method in testing RC snubber networks and it is quick and easy. Use test pins in a fixture and you should be able to test as fast as the meter will settle.

 

[sreid]

We build voice coils for fast tool servo aplications. To test for winding shorts, we use a BK Precision 875B LCR Bridge (inexpensive) to test the coil inductance.

Dynamic shorts are difficult to find. Is there any possibility you could use double insulated magnet wire? This greatly improves the possibility of coil shorts during winding.

 

[IRstuff]

The only way to test for a dynamic open or short is, as mentioned above, to shake the coil and hope that aggravates the condition. You might be able to simply place the coil under test on a completed speaker and blast away, although it'll probably be a bit noisy.

A decent 4-terminal Kelvin-connected milliohmmeter can easily measure resistances down into the milliohms.

TTFN

 

[lixc]

A big THANK YOU to all you guys for the fast and nice replies.

Hi OperaHouse:
I have to test the coil before it is assembled to be a speaker,that's why i mentioned "raw part".
If I use the method you said,
How should I decide the good and defect differences using the AC milivolt meter ?
The audio oscillator how should I adjust it? Sinewave ? and the frequency ?

Hi sreid:
The BK Precision 875B good ? How accurate is it ? The voice coil I have to test do not have core(or air-cored).
Can 875B shows even if only one turns is short-circuited ?
Need to do any calibration ? Can be done in-house by ourselves ?
I think the wire cannot be double-coated bcos it will run from the drawing.

Hi BrianG:
The wire is very thin. Any coating too thick then it will run from the drawing and the spec.

Thank you again for all the informations. I will try the methods you all mentioned.Hopefully can settle my problem. Please continue to reply me if got any new ideas. Thank you.

 

[sreid]

The BK875B is $189 at Digikey. Hrere's the link to the spec

http://www.bkprecision.com/download/documentation/100-66.pdf

A shorted turn makes a dramatic difference in the coil inductance so a coil with a shorted turn is easy to find.

 

[OperaHouse]

Guess we all assumed you were testing assembled speakers prior to installation in a cabinet. Just testing the coil would be like testing an inductor. The same bridge method would apply. Use a sine wave and test a number of coils to see where the numbers generally fall. Like I said, this is a very simple test. A better test would be to have a pulse generator put a very large pulse into the coil through a resistor and monitor the decay on a scope. The ringing should be very consistant. The pulse could even be large enough to induce some stress without doing damage for a better test. We did this on electrmagnetic clutch coils where the internal diode connections needed to be tested.

 

[lixc]

Hi OperaHouse:
The better test that u said is a "surge/impulse test" , is it ?
I found a few tester company like Baker and Wagne Kerr selling this kind of tester. But the thing is these testers are expensive and I don't know how effient they are. The test is it fool proof ?
I heard that the impulse or surge test is only good for the first few turns, is it correct ?
Because it have to be implement into the production line, so the tester is for operator to use. It must be easy and tell the operator whether is GOOD/BAD. Showing number is not suitable for the operator. Human judgement tends to make mistake and influence by feeling.

Thank you.

 

[lixc]

Hi OperaHouse:
Forgot to mention, these Surge tester company claimed that their testers can test for intermittent cases, is it true ??
And they claim that the high voltage surge won't damage/predamage the coil, is it true ?
What is the life-span of these tester ?
Calibration, service or repair is frequently needed ??

Thank you.

 

[MacGyverS2000]

Would two shorted coils next to each other make that large of a difference in inductance? I can certainly imagine two shorted coils above/below each other making a huge difference due to their separation distance along the wire length, but two next to each other?

 

[Skogsgurra]

McGyver,

Yes. It makes quite a difference. The Q will go down the sink if you get a single shorted turn, regardless of where in the coil. The Operahouse method could be applied easily, especially if an elevated frequency is being used. Something like 10 kHz seems to be right.

Use as much power as the coil can take for the few seconds that will be needed for the test. Instead of resistors and a reference coil, I would use three coils and the CUT to form a bridge.

A mV meter sometimes takes too long to stabilise for an automated test like this. A scope with settable limits and GO/NOGO output can do the checking and will help a great deal if you are going to automate the test.

The Bruker equipment mentioned is not for tiny coils, but more for motor windings and tens of kilovolts of pulse amplitudes. Testing with such a device will probably produce 100 percent NOGO (even if you had 100 percent GO to start with). Wayne-Kerr? Yes, can do.

 

[OperaHouse]

A lot of testing is simply to fulfill ISO testing requirements. Rejecting products off the line is not the way to control quality, it is improving the process so defects aren't made. I think a quick test is unlikey ti find anything but major failures.

With simply an air core, I would make the coil part of the oscilator circuit. Probably lowering the gain of the circuit so that a shorted turn might keep it from oscillating. An air core should be very consistant in frquency. This would be more sensitive than a bridge circuit.

A site you might find interesting is hhtp://sound.westhost.com/ A lot of fun projects in audio and general theory. Look up project 59 Self Oscillaing Amplifier. It would be nice to modify this so that more power is dumped into the coil so it is electrically and mechanicaly stressed. You should at least put some voltage on the coil over what it would see in normal life. This could mean switching in a power circuit prior to an oscillator or inductance test in an automated tester.

 

[automatic2]

How about submerging your coil in water when doing your tests. Should indicate any 'latent' shorting conditions. Adds the expense of a dry cylce.

 

[sreid]

1) Do you know what percentage of the coils are failing open or short? I ask because if the percentage is high it is usually better to fix the process that is creating the problem than is is to test the problem out.

2) Is it possible that the coils are not shorting but that the coil has not been completely bonded? What is your bonding method?

3) Could you change to a tougher magnet wire varnish?

4)Another way to find shorts is to put the coil in an AC magnetic field from another coil. A coil with a shorted turn will vibrate due to the induced current in the shorted turn (this is refered to as "Growling).

 

[OperaHouse]

I like to blue sky ideas and a couple have mentioned putting the coil in a magnetic field. Just think if a small cpacitor was actually added in the manufacturing process across the coil. This would have no effect when the speaker was assenbled with the higher inductance of the core. But, it would allow testing without any electrical connections by just placing it in a magnetic field. Process would be similar to the old grid dip meter measuring absorbsion. Tuned slightly off frequency there would be a normal current. Open would be lower current and shorted turns would be higher. This is just a reminder that it sometimes pays to add things if you can make it easier to test. Got our company to add circuit board edge connector pattern so we could do plug in testing. Only had to cut one slot in the board and edge connector used up real estate that couldn't be used because of board cutter.

Is this coil an unusual shape or not have a center core. Coil winding seems like a standard process which is very reliable.

 

[lixc]

I think submerging the coil into water won't detect inner layer's short.

The coil I need to test actually come from supplier. They already did DCR test and open-circuit test. But they couldn't detect intermittent ones.

The failing rate is low, BUT one out of ten thousand is still a problem bcos if one speaker with intermittent short goto the customer then I will be killed.

Now the market is very competetive. All is about cost. *sigh*

The vibrating idea is nice,BUT i don't want the operator to judge by his/her eyes whether the coil is vibrating or not. This will go bac to the issue of human judgement again.

The coil is pure round shape. No core. The core will only be insert during the process of assembling into speaker. (the yoke).

We buy the coil from our current supplier bcos they are cheap. But still sometimes one or two out of thousand or ten thousand will have problem bcos the supplier production line is using man-power, not automated. They are like traditional coil company. No robotic arm or anything. Just a very simple winding machine fully control by human, just the turns of the coil is control by machine.

I hope you understand my situation. *sigh*...