Frequent failure of LM358 !!!

Jitendra Suthar

Intellectual 295 points

Other Parts Discussed in Thread: LM358, TL074

I use this circuit (see figure/attachment ) to generate 4-20mA for 100ohm load.

But I am suffering from frequent damage of LM358 on site as well as in my lab.

over 9 years ago

0 Ron Michallick over 9 years ago

TI__Guru**** 166436 points

Jitendra,

Can you describe what is damage, it is any particular pin?
I'd like to see a diode across Q1 base to emitter to protect the transistor if pin 7 goes negative.

The LM358 doe not have high ESD tolerance. Could this be the reason for damage?

0 Jitendra Suthar over 9 years ago in reply to Ron Michallick

Intellectual 295 points

Thanks Ron Michallick for your kind reply.

second stage opamp gets damage.

In healthy condition, voltage at 5th pin and 6th pin are identical and voltage at 7th pin is 0.7volt more than voltage @6th pin.

but in damage condition, voltage @6th pin become +10.27V and @7th pin become -14.32V when voltage @5th pin is 0.97V.

but Transistor doesn't get damage in this condition.

0 Ron Michallick over 9 years ago in reply to Jitendra Suthar

TI__Guru**** 166436 points

Jitendra,

The voltage measurements are helpful. However, I also need to know the voltage at NPN's base and collector.
At this time, I don't know if TL074 is bad or if it found an alternate operating point.

With an input voltage differential of -9.3V, the output should be low and it is low.
I don't know from where the 10.27V/ 250 ohm = 41mA comes. Voltage on R5.

0 Jitendra Suthar over 9 years ago in reply to Ron Michallick

Intellectual 295 points

I have mentioned voltage at different nodes. Hope this would be helpful.

0 Ron Michallick over 9 years ago in reply to Jitendra Suthar

TI__Guru**** 166436 points

Jitendra,

There does appear to be a lot of current flowing out of pin 6.

My only other idea is that the -15V supply doesn't power up correctly.
Does the -15V supply ever goes positive during power up or power down?

The CL100 is very impressive, most NPN transistors only allow up to 8V or less reverse VBE.

0 Jitendra Suthar over 9 years ago in reply to Ron Michallick

Intellectual 295 points

Ron Michallick,

I use linear power supply for this circuit (Using LM7815 and 7915 regulator ICs). so I don't think -15V supply goes positive or it doesn't power up correctly.

Does absence of LOAD (100ohm at collector) may be the reason of failure ?

0 Mike Tripoli over 9 years ago in reply to Jitendra Suthar

Genius 3295 points

A couple of questions:

Is the load resistive as shown? No inductive component?

As Ron mentioned, a diode from the base of Q1 to ground is recommended.

Is the circuit oscillating? I'd put a small capacitor across R2 and from pins 6-7.

Adding a resistor (100) between the emitter and pin 6 limits the current into the device.

Lastly, did you see this: https://e2e.ti.com/support/amplifiers/etc_amplifiers__other_linear/f/18/p/431250/1546348

Mike T.

0 Jitendra Suthar over 9 years ago in reply to Mike Tripoli

Intellectual 295 points

Hello Mike,

Thanks for your reply.

The load is inductive in nature (this circuit is used to drive control valve ).

Can you describe which kind of diode I should use like, general purpose diode / zener or TVS diode or any other ? and according to Ron, diode should be connected from base of transistor to emitter and you said base to gnd.

The circuit is not oscillating.

Regards,

Jitendra

Design Engineer

0 Mike Tripoli over 9 years ago in reply to Jitendra Suthar

Genius 3295 points

Ahhhh... I believe this will turn out to be the root of the issue. I'm assuming, maybe not correctly, that there is a diode across the coil of the valve? If there is not, you are seeing very large transient spikes, potentially of hundreds of volts, being driven right into your '358. Yes, there is a transistor there; I suspect if you were to do a test with a fresh transistor and one that has been run in the circuit for some time, you will see a difference in the beta of the device. Each time the collector gets "whacked" it destroys a bit of the silicon, which usually shows up as a change in gain (until outright failure). If you're using the transistor in a "saturated" mode this will be harder to detect.

Bottom line; make sure there is a diode across the coil, a fast schottky diode should do it. Not to disagree with Ron, but I would put a diode on the base to ground, otherwise you are going through the 250Ω resistor. It doesn't sound like your signal "goes to the rails" (correct me if I'm wrong); a pair of clamp diodes on the output of the opamp to each rail will clamp the transients. Again, use some fast schottky's. Your cabling (if any) between the output transistor and the valve is important as well. Not sure we have the time and space to cover all the "gotcha's" there.

Mike T.

0 Jitendra Suthar over 9 years ago in reply to Mike Tripoli

Intellectual 295 points

Mike,

sorry for wrong feedback, this output (4-20mA) is feed to smart positioner which is not inductive in nature.

0 Mike Tripoli over 9 years ago in reply to Jitendra Suthar

Genius 3295 points

Out of curiosity, what technology is the smart positioner based on? I'm not aware of any "motion" system that doesn't use a motor of some sort at its core.

Mike T.

0 Jitendra Suthar over 9 years ago in reply to Mike Tripoli

Intellectual 295 points

smart positioner contains microcontroller circuitary which reads 4-20mA through some signal conditioning circuit and generates analog output accordingly to drive control valve. smart positioner has its own control valve drive circuit.

so basically, this 4-20mA signal is not going through any kind of inductive load.

Jitendra

0 Ron Michallick over 9 years ago in reply to Jitendra Suthar

TI__Guru**** 166436 points

Jitendra,

Adding a resistor in series with pin 6 is a precaution that can be implemented.
2.2k should be plenty.

Amplifiers

Amplifiers forum

Frequent failure of LM358 !!!