LM319-N: LMx19 High Speed Dual Comparator

Thank you so much, I found the information.
I was repairing class D amps and found leaky lm319s by comparing with a new module with my huntron tracker.
I set up huntron to see if it becomes as new condition. I used soldering paste and 400F iron for soldering. The 4 out of 6 became leaky while instantaneous soldering with my iron.
Usually even higher temperature works fine for other opamp chips so I thought there may be something additional caution requires to this comparator. The soldering time was like a second.

Hello Masayuki,

Are you using the tracker *before* you solder? Are you doing ICT?

Be careful with the tracker. Be sure that the current limit is 1mA or less, or that the open circuit voltage is less than ±800mV. Avoid the lower ohm ranges (<200 ohm) as they use the highest test currents.

These older devices do not have ESD protection diodes on the pins, so any reverse voltages applied to the pins go through the random die substrate junctions and this can cause damage. Particularly the input pins. The V- pin should be at the most negative potential. Avoid sweeping below ground.

What do you mean by "leaky"? Can you provide a "good" and "bad" waveform?

Are you cleaning off the flux afterwards? Burnt flux can cause a restive leakage path.

Since it is in a class D amplifier, there is the possibility for inductive kick-back or transients. Have you looked at the pins during operation to watch for spikes or other out-of-spec conditions? Again, can you provide a schematic so we can check for any possible electrical problems?

I have not heard of the LM319 being damaged by soldering - usually any damage is due to electrical over-stress due to taking the pins above V+ or below V-.

You are very helpful, thanl you so much.
I can provide pictures and the schematic via personal email but not on this forum.
Basically similar to this circuit www.ermicro.com/.../comp_09.jpg.
I`m doing this in circuit and ground to negative input. Sure I know in circuit has parallel components but good one I see good diode shape but bad one I see shape like "8".
First I replaced 6 bad ones in one shot. Some of the amps still drawed too much current so I stopped and checked with tracker. "8" shapes were still there. Good working one I saw good diode shape. Wonder if it happened when I soldered so I wired huntron while soldering then I saw shape got changed from diode shape to "8". Powering up and of cause it`s faulty.
"Leaky" means it works if the volume is low but if I send higher signal, it`s eventually goind to protection mode.

Hello Misayuki,

You can email me directly at:

 

Please send your schematic, showing input and output circuitry and loads, and, if possible, a photo of the setup.

Remember that "diode" shape is most liekely not a real diode but a PN junction that is probably not designed to handle current like a "real" diode. Again, the LM319 does NOT have ESD protection diodes.

An "8" shape would indicate a lag, so there may be some capacitence somewhere after the "diode" blows.

Have you looked at the power-up and power-down sequence to look for transients and overshoots? It sounds like you are getting some sort of input tranisent.

Soldering alone should not "kill" a device - the die undergoes much higer processing temperatures during manufacture. We sell millions of these devices, and we have not seen any "death by soldering" isues with this device. Overheating by soldeing usually damages the package and/or bondwires.

Careful soldering while powered-up. Your iron is probably grounded, so you may be shorting something when soldering. It may also be possible that the grounds of the iron and your circuit chassis are differrent potentials and there could be tens of volts of difference between the chassis ground and the iron tip. Measure the votlage between the iron tip and the chassis using both a AC and DC handheld voltmeter (DMM). Even a votlage of as little as 1VAC could do damage.

Your last sentence sounds like you are blowing the diode with the tracker. Remember that below -500mV, the "diode" is high impedance and should look high impedance. Above 600mV, the diode is conducting with increasing current. If the junction is not sized to handle the current, that junction can blow. Limit the current to 100uA or less.

Taken off-line. Failing part was from another manufacturer. TI replacement device is working fine. Thread closed.