Why does the ICs(AD620AN) gets damaged?

[samaaa]

Hi:
I bought six instrument amplifiers(AD620AN),
i tested these ICs,all ICs was work properly
but later the ICs gets damaged(5 five of 6 ICs gets damaged)

i was very carefully(by reading datasheet) about power supply and input voltage

i used this circuit(to make AD620AN work as just amplifier):

so i wondering:
Why does these ICs gets damaged(not working properly)!??

 

[jim hardy]

That's a nice part and i'd hate to be failing them, too.

It's hard to troubleshoot without observations on the actual setup.

Two things that happen a lot:

1. Grounding through oscilloscope's reference lead - be aware it connects through 'scope chassis and power cord to building ground and eventually to earth.
If either your function generator or your power supply tries to send any current to earth, it'll flow through your circuit and 'scope probe reference lead on its way there.
If you ever see a small spark when connecting 'scope ground lead you have this problem. Look for it with a meter.

2. Power supplies - generally they have an active regulator inside that might go through a transient state on powerup or powerdown.
Connect a meter, preferably an analog one , to your supply and switch power off and on a few times.
If the meter reverses or shows a brief overvoltage then that's likely the culprit. It may not happen every time, and may depend on which supply comes on first..

This is rare but real on dual supply systems-
If one supply forces any current backward through the other one at powerup, usually through capacitance,, well,, I've seen that cause "regulator latchup" which causes overvoltage.
That's a sneaky one.

good luck,

old jim

 

[berkeman]

@samaaa: One other possibility might be latchup due to ESD. It isn't usually a big problem with analog circuits, but it still is a possibility.

Has the humidity been low in your area while this has been going on? If the humidity is low, and there is a source of ESD buildup (like when you slide off your chair to stand up at your lab bench), then if you touch the unprotected leads/wires in your circuit, you can introduce an ESD "hit" into the circuit, and this can lead to latchup and chip damage.

You should make a habit of grounding yourself to Earth ground before touching any exposed, powered-up circuitry. Even better is to wear an ESD grounding strap that you can buy inexpensively at electronics stores.

Do the failed chips get hot when you power them up? Do your power supplies have current meters on them? If so, watch for high current events that correspond to when the circuits fail. That would be an indication that latchup may be the problem. The source of the latchup would be ESD or the power supply sequencying issues brought up by Jim.

 

[berkeman]

An update -- I just checked the datasheet, and the very high input impedance and low operating current imply extensive use of CMOS circuitry inside this part. That would make it more vulnerable to ESD hits causing latchup.

 

[samaaa]

That's a nice part and i'd hate to be failing them, too.

2. Power supplies - generally they have an active regulator inside that might go through a transient state on powerup or powerdown.
Connect a meter, preferably an analog one , to your supply and switch power off and on a few times.
If the meter reverses or shows a brief overvoltage then that's likely the culprit. It may not happen every time, and may depend on which supply comes on first..

i like you, i think the problem is from power supply(overvoltage)!may be!
i will check the power supply as soon as.

 

[samaaa]

@samaaa: One other possibility might be latchup due to ESD. It isn't usually a big problem with analog circuits, but it still is a possibility.

Has the humidity been low in your area while this has been going on? If the humidity is low, and there is a source of ESD buildup (like when you slide off your chair to stand up at your lab bench), then if you touch the unprotected leads/wires in your circuit, you can introduce an ESD "hit" into the circuit, and this can lead to latchup and chip damage.

Maybe this is why ICs gets damaged!

i touched the input lead"input pin" many times

and many times the input lead"input pin" was floating,
so:if the input pin was floating ,does the IC get dameged?

Do the failed chips get hot when you power them up?

no , the IC has normal temperature
[/QUOTE]

Do your power supplies have current meters on them? If so, watch for high current events that correspond to when the circuits fail. That would be an indication that latchup may be the problem. The source of the latchup would be ESD or the power supply sequencying issues brought up by Jim.

ok, i will check it soon
thanks...

 

[jim hardy]

Berkeman has a good thought there on ESD.

I grew up in South Florida and never experienced static electricity until i moved up north. Humidity in near-tropics is so high it's just not a problem there.
So i just don't think about it enough.

Do you live in a dry climate?

There exists special conductive mats for workbench surface. They are somewhat expensive.
http://www.isconline.com/product/8213/3M-8213-Gray-2'X4'-Table-Mat.html?cid=14

Here's a DIY approach for static conrol, you might place a carpet remnant that's been treated this way on your workbench and Earth ground it..
http://www.888400floor.com/flooring-resources/carpet-static-electricity-control.html


EDIT just a thought...

with your gain of 50, a DC offset of just 0.04 volts would drive the amplifier to its limit and make it appear "latched".
Be very sure your function generator has no DC offset.

old jim

 

[AlephZero]

i touched the input lead"input pin" many times

and many times the input lead"input pin" was floating,
so:if the input pin was floating ,does the IC get dameged?

Those are two good ways to kill static-seisitive chips.

While you are experimenting with the circuit, a 1M resistor connected between
IN- and IN+ probably won't affect how the circuit works but it will prevent static problems.

If your circuit is designed to have an input that can be "unplugged", you need to include something like that in the design, so the input can never be left floating.

As well as the static sensitivity, RF electrical noise or EM radiation might induce enough voltage on a high impedance "floating" input to overload the amplifier and make the output "latch up" to one of the supply rails.

 

[dlgoff]

Berkeman has a good thought there on ESD.

I grew up in South Florida and never experienced static electricity until i moved up north. Humidity in near-tropics is so high it's just not a problem there.
So i just don't think about it enough.

Here in Kansas it's even dryer; especially now during winter. My rule of thumb for handling static sensitive ICs is to leave them in their conductive pad until you are ready for them. Once I've discharged myself by touching a know Earth ground, I make sure never to move the feet. (You can get charged with Kvolts of energy just by lifting your feet up.) Then I retrieve the ICs, place them on the board and solder, all without shuffling my feet. I don't think I've ever blown an input or had any failures doing it this way; albeit it's been years ago.

The energy stored as static electricity on an object varies depending on the size of the object and its capacitance, the voltage to which it is charged, and the dielectric constant of the surrounding medium. For modelling the effect of static discharge on sensitive electronic devices, a human being is represented as a capacitor of 100 picofarads, charged to a voltage of 4000 to 35000 volts.

http://en.wikipedia.org/wiki/Static_electricity

 

[berkeman]

Maybe this is why ICs gets damaged!

i touched the input lead"input pin" many times

and many times the input lead"input pin" was floating,
so:if the input pin was floating ,does the IC get dameged?


no , the IC has normal temperature


ok, i will check it soon
thanks...

SCR Latchup can occur when an input to an IC is driven beyond the power supply rails. That's why either a problem in power supply power-up sequencing, or an ESD hit can cause SCR latchup with the associated high current thermal damage.

 

[NascentOxygen]

While testing these (open loop bandwidth?), a 1:10 potential divider (say, 100k and 12k) at the input might give adequate protection.

 

[samaaa]

jim hardy,yes i live in dry climate

thanks for all, i will take your advice .