Tin whiskers and other packaging problems

[Pat Pending]

I'm not familiar with the 2N706, but the early silicon transistors did have a failure mode not related to dissipation, SOAR or the like. I killed a load of BCZ11 transistor's in a switching circuit that I'd built for a school intercom project because I had allowed the b.e junction to draw a pulse of current through a small speed up capacitor. "Newer" devices seem to zener it away successfully. It left the devices totally o/c. In the original post on this topic I mentioned the "tin creep" that shorted the TO18 packages across the base. I've just been informed this phenomena is called dendrites (?) And have been given a link to a NASA who are constructing a web page devoted to explaining the difference between this and tin whiskers. http://nepp.nasa.gov/whisker/dendrite/ . Its amazing how something so small can cause thousands of pounds worth of down time for industry time and again just because an accountant didn't see a bit of varnish as being necessary! Andy.

[ppppenguin]

I don't have the reference to hand but there was an interesting failure mode in line output transistors such as the BU105. You had to control the switch off very carefully to avoid SOAR problems. If you switched them off too quickly there was a problem that I don't remember very clearly. I think it was excess carriers trapped in the very large collector region. If you switched them too slowly then you got the usual SOAR problems.

[Pat Pending]

Yes Jeffrey. Can't remember what that one is called, but I was talking about some similar failures in HV bipolar transistors to a guy from TI in Dallas, (on a discussion forum). Who called it punch through. A similar kind of thing commonly happens to power MOSFETs (but much more easily, static builds up on the gate region capacitance), US term is gate rupture and is an out of circuit handling problem, and not found very often in devices that are in circuit. Had a problem once with a switching DC-DC converter I was designing, for a high power off grid fluorescent lighting system. It ended up with a circuit suggested in some TI. application notes I found that used BUX xx bipolar transistors in the inverters. I bet the 320volt dc output surprised a few careless electricians. Easy money! The royalties have stopped now, f*****g LED's! Back to topic, is the failure mode that you are referring, the one where the transistors BE junction ends up open and the C to E ends up short? Regards. Andy.

[ppppenguin]

Honestly don't remember the nature of the failure. This was when I was at uni in 1978. Got trhough a fair number of BU105 before resolving the problem.

Just found the TI book online. https://archive.org/details/Semiconducto...gnVolumeIv I've attached an extract from p29. There's no way I would have found the problem without this book.

[Pat Pending]

Jeffrey. Tried to open from your thumbnail but with no success. So I'm downloading a copy. If it's Ok there is also a book on there about failure modes in semiconductors, looks like proper bedtime reading, a cross between doctoral dissertation and b*lls**t, where there are twenty references per page, I haven't had much sleep recently, I think I'll get it and stop taking the pills! Thanks. Regards Andy.

[Mark Hennessy]

Hi Andy,

What happens when you try to view the attachment?

Try before replying, because when you get to 15 posts, your permissions are automatically upgraded. But you should currently have permission to view attachments prior to 15 posts. Having said that, it's possible something has gone wrong with the permissions - it's quite complicated.

All the best,

Mark

[Pat Pending]

Mark. How did my pic of the transistor insides look to you? Wasn't ever David Bailey but every thing on the board just opens, for me, in its own window at thumbnail size? If you want a better res one for your website let me know.
BTW. I will be going to see George about the middle of january, with my MP3 recorder, so if you or anyone has any  questions (non personal) I'll pass them on. Bear in mind he worked for the Philips group, in their semiconductor divisions in Europe and the USA between 1958 and 1977. Start a topic for it maybe? Regards. Andy.

[Mark Hennessy]

Your picture looked just fine to me.

Images should open in a new tab, and they should show at the size they originally were at - we don't resize them or anything. Exact behaviour will depend on your browser and operating system. From what you say, it could be that the server is sending you the thumbnail versions only, which is weird.

Try it again in about 15 minutes - preferably without making any new posts in that time. Once a "job" has run on the server, I'm expecting your user account to change from "New User" to "Technical Assistant". Here's hoping

Cheers,

Mark

[Mark Hennessy]

OK Andy, you're now a "Technical Assistant". Has anything changed?

The only change I expect is that you can now access attachments in The Archive. But, let me know if you notice anything different...

PS: I'll delete these OT posts later so that they aren't cluttering this thread...

[Crackle]

I recently bought a Decca TP85 Radio, ah! I thought when I opened it up at the boot sale, AF116 transistors, that's not going to work, I thought to myself, but for £2 where can you go wrong.
Very wrong, it worked perfectly after giving the capacitors a chance to reform.
All 5 AF11x transistors are working perfectly with no sign of a crackle if tapped. Stamped on the transistors was "Made in Holland"

I have "treated" loads of failed AF11x transistors that were made in "Great Britain" to get them working again.
But this is the second radio I have come across with Dutch transistors which all work fine.
I wonder if there was a different manufacturing process between the 2 countries..


Mike