Reverse polarity detection for Electrolytic Capacitors?

[BasePointer]

Hi to all,

We are facing some troubles in mass production during the process of assembling through hole electrolytic capacitors.

Do you know a way to catch wrong capacitor polarity other than using visual inspection?

10x,
BP.

 

[karesz]

Reverse polarity detection for Electrolitic Capacitors?

Hi,
If their are THT types, often are the wires different of lengtht _as by LEDs_, but it exists in so many versions, that i can not think on an relative simple & generell possible detection.
I believe, the best one are the mankinds eyes!
Otherwise you can apply intelligent camera system with relative expensive SW; if its a population mashine, but you dont like/use it...
I think, it must be possible to check it with a lower DC voltage as Wvoltage & measuring the currents_by right polarising;_ it must be lower...
I didnt never it, Im thinking only_it must be possible, but needs some time pro component/test! :-(
K.

 

[123jack]

Just a few brainstorms:

Ideally you need to catch the problem before assembly.
That's a tricky issue depending on what you mean by "mass production"
and how the caps are being delivered and inserted. And what your actual problem
is.

As karez said - visual inspection can be expensive or error prone (human eyes miss things) but it can work.

Some things you may want to think about that might or might not make
sense in your situation (you didnt give much info to go on about methods used)

1. Inspect every component manually before feeding - low skilled labour
but must be accurate and not damaging to feed tapes etc

2. Visual inspection - you COULD do this very cheaply depending on how
your components are marked - even if you just have to use a spot of paint manually and re-roll the reel - use a light/dark sensor
... not elegant but it can work at speed if there is enough contrast. You may want to engineer that upwards for reliability !

3. If you do have access to legs as they come off the roll a simple metal
brush to sense the leg length should work (like we used on the old punched tape readers)
That would allow relatively high speed sensing but relies on the manufacturer providing reliable differences in legs

4. If your insertion process is getting it wrong (you didnt say) you'd need to look into that or provide a lot more info.

5. Failing all that a camera solution can be releatively cheap if there is something
that can be seen. Its only an edge detection algorythm you need. Not too hard
and you should be able to get a freelancer to custom write a simple solution pretty cheaply.
I would be suprised if there wasn't already the start of such software on the net somewhere - maybe
sourceforge.The hardware could be a simple web camera/PC.

Rework is expensive so I'd say work real hard catching the problem before insertion even if it takes you longer to solve than a simple electical test afterwards.