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]

Looks like you found your problem wihtout us...

"We buy the coil from our current supplier bcos they are cheap."

If you cannot afford even 1 in 10,000 coils to fail, maybe going the cheapest route isn't the best path to take. I'm being serious. You'll spend quite a number of pennies setting up a test jig, paying somone to stand by it and use it, etc. Seems the more logical solution would be pay the extra pennies to have a quality coil sent to you in the first place and be done with it.

 

[sreid]

You have a tough but not intractable problem. First you need a PLAN to discover where the problem is. Recording coil resistance and inductance at incomming inspection would be the first place to start. Log the data so you'll have a record. Retain coils that don't pass so they can subsequently be analysed. The point of this is that a problem can't be solved unless it is first understood.

Since what you really want to do is to improve the quality of speakers shipped to the customer, perhaps a better final test of the speakers could be implemented. What comes to mind is an automated high power frequency sweep of the speakers with an automatic FFT analysis of the sound comming out of the speaker. This might identify any number of pathologies that might occur in the speaker assembly.

You can never attain 100% perfection but you can improve it. And there is always some point where the cost does not make economic sense.

Pragmatically, here's your position;

1) Management want's the failure rate improved.
2) You can't fix a problem until you know what it is.
3) Here's your plan including time and cost.
4) Management can either accept the plan or reject it.

 

[lixc]

Hi MacGyverS2000
I totally agree with you.
But I am just a tiny person in the company. They gave me this problem to settle, I cannot just tell them to find a better supplier and make sure the supplier supply 100% good products.
My job is to design or find something to can detect the intermittent short/open circuit of voice coil. That's my task, and I am facing a big headache.*sigh*
The tester will not only be used in production line, but also for customer rejected items. We can use the tester to reveal what is the problem that occur,eg. why speaker no sound come out after a few times of playing by customer.
Thank you.

Hi sreid
Sorry, but can you explain to me about the high power frequency sweep ? And what is FFT analysis ? Maybe I can recommened to my supervisor for the final testing of speakers.

We do have test using Polarity Checker for the speakers, the speaker is connected to the probe then a sound from speaker go into the mic of the tester for analysis. It is for polarity only.

We have random check using nominal load check for the speaker, eg. 1000-hours of nominal load test.

I think we have DCR checking for the incoming coil, but no inductance check.
I went to the supplier side to see they process. They got DRC test for each of the coil, megaohm test for dual-winding coils(test for short between the two winding), polarity checking(using a not fool-proof method by human judgement).

Still, my main task is on the coil itself.*sigh*...

 

[lixc]

The main problem now is how to check the voice coil for [/b]intermittent short-circuit[/b] and intermittent open-circuit.
I think both is related.

 

[IRstuff]

Perhaps you can also quantify how much such a procedure is going to cost the company. If you had a full time tech running this exercise, you'd be talking in the neighborhood of $60,000/yr wages/bennies/G&A. That would seem to be a lot of speaker coils.

You didn't seem to have mentioned how many you need to test in what amount of time. That would be a major input into the design equation.

A small shaker, ala:

http://www.vibtest.com/vr520.htm

might be adapatable for single or multiple part testing.



TTFN

 

[lixc]

Hi IRstuff:
U mean u suggest I use a shaker to shake the coil during test ?
I don't quite understand the "$60,000/yr wages/bennies/G&A". Sorry, can you pls explain ?
I don't have the exact number of voice coils to be test, but I think it is around 500k per month.

Thank you.

 

[stevenal]

Put enough current through it to raise the temperature above ambient while watching its infrared image. Poor connections and shorts should show up as hot spots. Windings bypassed by a short will be cold. Since they are cheap, you can burn a few to establish some testing limits.

 

[IRstuff]

So, in a single shift operation, that's roughly 3200/hr, or about 54/min, assuming 155 hrs/month.

You're going to need something fully automatic and darn fast to keep up with the product flow. That means a full-time operator. Assuming wages around $15/hr, that's $30,000/yr. Assuming some benefits and overhead of 100% of wage, That's $60K/yr for labor and a rather huge investment in machinery plus maintenance.

Assuming your 1/10,000 bad, that's 50 per month. So unless they're worth more than $100 each at the stage where they're otherwise caught, you won't break even financially.

TTFN

 

[sreid]

The basic idea would be to put enough power into the speaker (if only briefly) to stress the speaker in an effort to find bad speakers (no matter what the problem might be).

An FFT (Fast Fourier Transform) is a mathematical technique to convert a signal in time into a frquency spectrum. Bad speakers will probably have a spectrum different from a good one. Stepping the frequency would show up problems that are frequency sensitive.

We had a problem with voice coils making noise in the system. We locked the bad coils in a magnet housing and did a power frequency sweep. The bad coils made distinctive horrible noise at certain frequencies. The problem turned out to be that the coil was not bonding to the aluminum bobbin we wind the coils on.

 

[OperaHouse]

Another calculation. At 500,000 units testing at 1 minute each comes to 8,000 hours a month or 2,000 hours a week. If an employee works 40 hours a week, that is 50 extra people to hire. Drop the time to 30 seconds is still 25 people. Managers always think in terms of people. Any manual operation is always goint to take about 30 seconds.

While you have speaker failures, You don't indicate they are a majority of one type. This means you have to do multiple tests and probably a significant number are not even related to the voice coil. In this high volume application the problem should be controled at the outside vendor.

I remember some voice coild being wound on an aluminum sleeve, possibly for dampening. This would make some testing difficult. Fastest testing would be if you didn't have to attach anything manually. I keep thinking of that tuned coil idea. Just think of the coil mounted to the speaker cone. A metal foil on each side of the speaker cone forming a capacitor. Advertizing guys could call it the "ring of fire." These could pass down a conveyor at a pretty good speed and be inductivly tested with the bad ones kicked out. Induced frequency could sweep eliminating need for close tolerances. Think you have quite a challenge ahead of you getting testing down to 10 seconds and eliminating just 50% of the failures.

 

[IRstuff]

It "simply" means that he needs to be able to test about 30 at a time, which is why I caveat "huge investment". since that requires a fairly high-speed switching matrix and a very expensive parts handler.

One thing that could bring down the cost would be to have your supplier deliver the coils on some sort of carrier that would clamp the pigtails on known positions for the test setup to connect to.

TTFN

 

[lixc]

hi all:
I need some time to "digest" all your ideas...haha...I am still a newbie. I have to read a few times to understand a little bit. Appologize for my stupidity.
What my boss have in mind is a test that can be conducted on the voice coil for about 2 to 3 seconds.
We got around 9 production lines.

Sorry that I still cannot understand the FFT method and the "ring of fire" method, can you please explain clearer(including what tools/equipment needed) ? These methods is for speaker test, is it ?

I am now looking at a tester, Wagne Kerr 6815 Impulse Winding Tester (or the Microtest TF-6815, they are the same). It costs about US$7000. From taiwan. The salesman claim it can detect intermittent cases. The theory is that it inputs a HV but low ampere pulse into a master coil and record how the waveform decays. Then a reject percentage is set. The master coil can be remove and all coils for test will be test at the same method but the waveforms are compared to the master's one.It claims that the test can be conduct for each coil for less than one second.
Anyone ever use this kind of tester ? any comment ?
The claims can be trusted ?
IS this test method suitable for me ? Will it damage the coil ?
Someone said that the impulse can only test the coil for the first few turns, is it correct ?
It claim the voltage is 200V to 5000v and the ampere is very low only(some claim nano amp).
Other company like ECG-Kokusai or Baker sell for more expensive, almost double the price.

The website is

http://www.waynekerrtest.com/html/products/lcr/6815.htm

and

http://www.microtest.com.tw/english/products/tf-6815.htm

.
Hope you guys are free and have a look there and give me advice.

Thank you.

 

[OperaHouse]

OK, last post. The problem exists with the coil winder because you have beat every penny out of him. Before you spend these thousands of dollars for testing, go out and buy a couple spools of high quality wire from a known manufacturer. Then make a test run and see if your problems go away.

I remember a plastics manufacturer that asked us to inspect plastic pellets for specs of carbon. We told him that this seemed like a process problem and controlling the ramp up of the extruder heaters would eliminate the black specs. He said he knew that, but he had budget money for inspection and not new capitol investment.