Go Down

Topic: How did I fry my transistor? (Read 7299 times) previous topic - next topic

raschemmel

I wonder if the OP was walking on a carpet before poking his finger at the transistor. Static discharge varys with the weather conditions and the floor material but is much less likely in the winter when the humidity is high since it is at it worse when it is very dry.
Arduino UNOs, Pro-Minis, ATMega328, ATtiny85, LCDs, MCP4162, keypads,<br />DS18B20s,74c922,nRF24L01, RS232, SD card, RC fixed wing, quadcopter

oric_dan

I don't know I buy 'winter and more humid'. I've always found static electricity to be 'much' worse in winter.

MarkT

#32
Dec 29, 2014, 12:15 am Last Edit: Dec 29, 2014, 12:18 am by MarkT
In winter the outside air is colder and can hold less water vapour.  The air is brought
into buildings and heated up to 20C or so, reducing the _relative_ humidity because air
at 20C can hold much more water vapour.  This makes for much more static.

Relative humidity depends on the balance between water molecules wanting
to evaporate (greater at higher temperatures) and wanting to stay liquid (or
adsorbed onto surfaces (due to hydrogen bonding).  100% relative humidity
and water molecules condense into drops as evaporation loses the battle.  This
also means surfaces have more adsorbed water molecules on them, which are
mobile and conduct charge (killing static).

Sunny winter weather outside and being in a heated building is the worst for static.
The relative humity is lowest, and surfaces have much fewer water molecules on them,
which eventually can no longer sustain a current as they evaporate before being
able to transfer charge laterally.

I think the reason water is so good at carring charge on surfaces is because they
can bind a proton to make H3O+ ions
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

oric_dan

Yes on winter static.

Quote
I think the reason water is so good at carring charge on surfaces is because they
can bind a proton to make H3O+ ions
Maybe it's due to the dipole moment geometry of H2O molecules.
http://www.bing.com/images/search?q=water+dipole+moment

runaway_pancake

Open base tie.  Clearly the emitter tie is functional, as quite likely is the die itself.
So tedious.
"Who is like unto the beast? who is able to make war with him?"
When all else fails, check your wiring!

oric_dan

#35
Dec 29, 2014, 05:19 am Last Edit: Dec 29, 2014, 05:22 am by oric_dan
❝Open base tie.  Clearly the emitter tie is functional, as quite likely is the die itself.
So tedious.
Quote
BE Vf 1 (Can't get a reading)
If BE is open, it's 1000:1 more likely he accidentally tied Vcc voltage to the Base, and it wasn't ESD at all. Especially if it were me poking my finger into the breadboard. I've probably fried 100 components that way over the years, and none by ESD. I pick up boards and chips everyday, and they always work afterwards. Never killed one just by grabbing it.

KenF

Never killed one just by grabbing it.
Some folks are just born lucky.

oric_dan

Some of us are just born low-static.

KenF

Some of us are just born low-static.
You should get a job with an orchestra.  You're obviously a good conductor :)

oric_dan

I could go on, but going to end it now.

MorganS

Open base-emitter junction.
Yes, but what causes that type of failure? Over voltage? Over current?
"The problem is in the code you didn't post."

jotapeh

A lot of interesting conjecture here but the two main reasons provided seem equally unlikely (or likely, depending on your perspective.)

For the case of ESD damage: low humidity and carpeted floors/movement were cited as likely factors. My house is electronically controlled to 45% humidity. I was working on a solid wood table sitting on a solid wood bench with my feet on a solid wood floor. I didn't actually move very much between setting up the circuit on the breadboard, having it function for several button presses, and then it dying. Certainly no carpet charging involved because - no carpets. So static isn't super likely.

So what about the case of tying the base to Vcc, perhaps even momentarily? As you can see from the attachment in post #5, that would be a difficult task. The leads which are even somewhat likely to short (eg., the legs of the capacitor and the resistor to Gnd) are electrically common anyhow. The 4.7kΩ resistor through a pushbutton switch is the only thing providing +5V to the base, and that ought to be plenty of current limit.

As you can tell - I'm not super convinced by either of these theories. I appreciate all the thought process folks have put in, but none of it really adds up for me. With that said there are orders of magnitude of experience beyond mine in the responses to my original post and I respect that it is extremely likely one of the two scenarios is correct. I just don't know for sure what happened, so I'll leave it alone :)

oric_dan

#42
Dec 29, 2014, 07:25 pm Last Edit: Dec 29, 2014, 07:27 pm by oric_dan
Yes, but what causes that type of failure? Over voltage? Over current?
As mentioned in reply #35, by far the most likely scenario is accidentally putting Vcc on the BE junction. If you're playing with a powered-up ckt, this is easy to do and without even realizing it, when plugging things in and out.

The BE junction is a diode, nominal Vf = 0.7V, and even momentarily applying 5V will drive a large current and produce an open-circuit [the OP's situation] when there is no current-limiting R in the emitter [the OP's situation]. See the graph, current grows "exponentially" with applied voltage, above the threshold voltage - and yes, a transistor BE jcn has the same characteristic.

http://www.powerguru.org/rectifier-diode-basics/

PN2222 transistors are old-style large-geometry [600 mA] devices, and have been around for 30 or more years, not new-style small geometry like modern CMOS processor chips. Static damage is a theoretical possibility, but much less likely, given that the breadboard scenario is extremely prone to operator-error of many kinds, and that the probability of this goes up in direct proportion to how long you play with the ckt.

At this point, you can only play with probabilities. Breadboards and operator-error go hand in hand, even for people who have been doing this stuff for decades. It's only a 5-cent transistor here, but the next time a breadboard is used, you have the same gotchas all over again.

Go Up