sag wires

It's too long a story to tell here, but there's a party that builds a gadget that we designed, under license from us. They use contract manufacturers who make near-zero profit. Two years ago, an NPN RF transistor went EOL, and they knew it but did nothing. No CM is going to buy and stock a couple reels of EOL parts. So now they can't build the gadget or ship their product.

So, after over 100 emails involving an average of about 15 people, they found some from a broker. But then they sent them to their QC people, who x-rayed them and found a couple with "wire bond sag wires", so rejected the lot.

formatting link

Does anybody know about wire bonding? Is this a real quality issue? Seems like it wouldn't matter, and that you might see straight or saggies at various x-ray angles. I could also imagine that the encapsulating process could push the bond wires around a little.

One irony is that this transistor buffers a scope monitor output that is probably never used.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

the extra inductance due to sagging bond wire may impact the match or performance at a very high frequency.

These parameters may not be critical in your particular application.

mark

Reply to
makolber

You designed the circuit, is there no equivalent transistor available?

Reply to
<698839253X6D445TD

Sure, that's easy, 5 minutes on Digikey. But the customer would require up to a year to qualify a new part.

This is actually sort of fun. The more people that get involved, the less likely that anyone dares to make a decision.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Don't all wire bonds make a loop of some sort? I think the straight ones are basically optical illusions.

formatting link

formatting link

formatting link

This looks like it might be a SOT-23 transistor:

formatting link

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
.

How very true!

--
This email has been checked for viruses by Avast antivirus software. 
https://www.avast.com/antivirus
Reply to
TTman

That's how the broker ended up with them, the parts were rejects no one wou ld accept or the manufacturer rejected, then some sleaze stole them out of a dumpster somewhere and sold them to the "broker." I wouldn't touch them o r anything else the "broker" sells.

How unique can an RF transistor possibly be???

Reply to
bloggs.fredbloggs.fred

I think the parts are legit (it's a very high-end EOL parts broker) and the wire bonds just sometimes look that way.

This customer is deep into the Intel-style "copy exact" mindset and would take "months to a year" to approve any replacement part. Even the same part with a different beta binning (top mark) was rejected. A BFS17 would work fine, 7.5 cents each by the reel.

5M+1E stuff.

There aren't many NPN rf transistors around any more. BFS17 is still a good one, fast but not too fast. The PNPs are basically all gone.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Does anyone else have a "copy exact" requirement for their gear? Intel invented this for their fabs, and lots of semiconductor equipment manufacturers require it. Some interpretations say that any piece of gear must be form/fit/function/appearance identical with the original. The extreme version is that you can't change anything at all, not a resistor part number, not the supplier of a lockwasher, absolutely nothing without requalification and approval. An "improvement" of any sort is not allowed. This extends to the manufacturing process; if a table is moved in the production area, that must also be approved and requalified. Seriously.

Of course, all employees must have Copy Exact training, and that has to be documented.

One of our customers assigns a 12-digit stock number to everything, and places orders accordingly. The slightest change spins that number, and requires requalification, and must then be purchased under the new number.

formatting link

formatting link

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Having participated in several technology transfers, remote debugging is the second-hardest part. (The first-hardest is getting the rock-throwers to shut up.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

I remember Philips uses a long number like that. Their service center one day needed some small parts for a repair.

They entered the number. Few days later a big truck arrived full of big transformers or something.

What happened? In their computer system / terminal you HAD to add the leading zeros. So

000123 was something else than 123

Maybe for quantities too... left justified? did not ask. Eighties.

Reply to
<698839253X6D445TD
[...]

Can't really see how this could be an issue.

--
This message may be freely reproduced without limit or charge only via  
the Usenet protocol. Reproduction in whole or part through other  
 Click to see the full signature
Reply to
Cursitor Doom

Not to John. The customer's QC is the real problem.

Reply to
John S

Some of my smaller clients will not change one iota on a design unless I bless it.

Sounds "Dilbertesque".

As for the bonding I don't believe there is an issue with what they call "sag" in the xray.

That is actually normal and has to be that way in companies that produce mission-critical gear. Med-tech, aerospace, oil/gas, and such. I have personally seen the results when this wasn't done. For example, a CM took it upon themselves to change a capacitor on one of my designs. No formal ECO process was followed which was a big mistake. The sales rep told them that the new cap is actually a higher quality part. He failed to notice, however, that the leakage spec was much worse. They had to scrap almost 4000 built boards and eat the cost.

[...]
--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

I've only seen wire bonds for ~1 of (or maybe 100 of) custom type parts. ('hand made' wire bonds.) And there was always a loop or sag. I have no idea what things look like when making millions.

George H.

Reply to
George Herold

You make me happy to be in a small company, where I get to decide on replacement parts. I can switch from film to ceramic caps and no one cares, (but me.)

George H.

Reply to
George Herold

like

IME you always leave a loop, because otherwise the bond is liable to fail u nder thermal cycling. Things might be different in the plastic package worl d though--the wire isn't free to move.

OTOH tension caused by sideways pressure is maximum when the wire is straig ht.

Cheers

Phil Hobbs

Reply to
pcdhobbs

I didn't think that a bond wire could be pulled flat like a guitar string. It just looks that way if it's x-rayed from above.

formatting link

formatting link

If you bring in enough quality experts, you'll never get anything done.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Right. We would have bought a reel when it went EOL, or just bought a similar/good-enough part now. 10 minutes maybe.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

k like

under thermal cycling. Things might be different in the plastic package wo rld though--the wire isn't free to move.

ight.

You don't mean a loop, like a circle, you mean you leave some slack. Yeah, that makes sense. As long as the coefficient of expansion of the wire is greater than the plastic and/or leadframe, there won't be any stress on a s traight wire... unless it gets cold. Ok, you need slack.

Rick C.

Reply to
gnuarm.deletethisbit

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.