iso124 failure issue

The maximum output current has a minimum value of 5mA, maybe this is the problem (5V / 1kOhm = 5mA). Why not trying to use a higher resistance value?.

The datasheet is rated for a resistance value of 2kOhm, maybe this value is appropiate for you.

Please, post the final solution when you find it.

William,

Ivan is correct, your load is a bit heavy for the assured output rating of the ISO124 but I doubt that would be the problem. This is more likely to just limit the full output voltage swing of the device.

Ivan is working a similar case of ISO124 failure so there is a shared interest here.

Isolation amplifiers often live in difficult environments with isolated supplies, high voltage and noisy signals. So our first guess is that something in the environment is causing the failure. We won't however, rule out the possibility of some type of problem with the part but let's start with the most likely causes:

1. Excessive voltage transients on the power supplies on either the input or output side.

2. Input signal voltage that exceeds the power supply voltages (positive or negative).

3. Excessive voltage across the isolation barrier.

You can do some simple checks for possible causes by making measurements on a failed unit. Measure the input side and output side quiescent currents. Are either of these currents abnormal? Measure with a ohmmeter from the input ground terminal to the output ground terminal. Use a very high resistance range on the ohmmeter. Any sign of conduction across the barrier would indicate a failed isolation barrier.

Regards, Bruce.

Bruce,

 Thanks for your post. 

Further information - The input side of the iso124 is fed by a OP227UA that is powered by the same +/- 12V supply that the input side of the iso124 is powered by.  Its output is divided down to 0-5V.  I would kind of assume that a large transient is not coming through the input side as there would probably be a failure of this op amp as well.

I must correct what load I am driving also, it is not a 1K it is approximately 11.5K.  So I think that it is well within the range of the device.

I have found that if I remove the output +/- 12V supply ot the iso124 the current drops which would lead me to believe that it is the output side that has failed.  I have measured the reisistance between Gnd1 & Gnd 2 and have not found any conduction.    So I would also assume that it is something on the output side causing the issue? 

Is there any "typical" failure mode for these parts?  

Your help is appreciated.

thanks,

Bill

Bill,

You can't necessarily assume that because the OPA227 does not fail that there is no problem with the supplies on the input side. Some careful probing of these supplies may reveal some issue. Also, to diagnose the possible issues with the quiescent current you would need to make actual measurements of the current drawn from the positive and negative supplies on both input and output sides.

You have several ISO124 channels all powered from the same power supplies, correct? Does this mean one supply for all or the same power converter repeated N times? If one particular position is the only offending one, you should carefully investigate the differences between the positions. This could be differences in input signal behavior, power supply, including location of bypass capacitors, or differences in voltage across the isolation barrier.

Regards, Bruce.

Bruce,

 When I said that the OP227 doesn't fail, it was my thought that it was not due to an excessive input signal, not a power supply issue. 

I have 3 iso124 circuits on the board.  The same set of isolated DC-DC converters feed all of them.  One circuit is problematic only.  The other circuits read signals from my HV supply and the problematic one is an output to the HV supply.  It adjusts a reference circuit to vary the voltage output.  So the 3 circuits are different, but power is provided the same way.  Since I also did the PCB layout, the bypassing and layout are fairly consistent.   

This kind of points me back to the environmental issues of transients, etc., I would guess???

thanks,

Bill

William,

Maybe an image with the schematic of the four ISO124 can help to found the problem. Can yo attach one?

Regards,

Iván

Bill,

Ivan is correct, we really need to see a complete schematic to be of any further help.

Bruce

Bruce,

It has been a while.  I have tried to paste a pdf of the schematic but I could not do it. 

We had 2 of the failed parts analysed by a  third party reliability lab.  The results were that the output saw an overvoltage condition.  Question: how sensitive are these parts to an ESD event?  Do you think it is more likely an ESD event or a transient?  What type of protection is recommnded on the output of the device?

Some history - the circuit on this PCB is used to adjust  a power supply output by adjusting a reference circuit on the power supply.  The output of the ISO124 is connected via a cable directly to this power supply.  The power supply is used to charge a hybrid battery(275V).  The ISO124 is powered by an Isolated DC to DC converter that supplies +/-12V.  The input side of the ISO124 is powered by a different Isolated DC-DC converter at +/-12V.  The input signal to the ISO124 is 0-5VDC.  There are 2 other ISO124 circuits on the same board powered the same way with similar bypassing, layout, etc.  The difference is that they are reading back information from the power supply such as current and voltage readbacks.  The problematic circuit is an output to the power supply that adjusts the reference.  The maximum current supplied to the adjustment circuit is approx 7-8mA.  I don't think that the isolation voltage is being exceeded, the current drive is not exceeded, the supply voltage is not exceeded(the other 2 would fail).  My only prblem is with the one device in the one circuit.  THe other 2 have had no failures.

Any ideas?

thanks,

Bill 

Hi Bill,

Do you have a photo of the die from failure analysis? Sometimes, the appearance and location of the failure can provide a clue as to whether it was an ESD or EOS event. My guess would be Electrical Over-Stress. You say that the failure was in the output... is that the output terminal or somewhere in the output side IC? Details of the failure analysis may help.

Are you saying that the output current of the ISO124 could be 7-8mA? While this magnitude of current should not, on its own, cause damage, it does suggest that you have a pretty low impedance connected to the output. If voltage transients occur, the output circuitry could provide a path for much higher current. No protection is normally necessary on the output of this device. It is the conditions of a particular application that might demand special precautions. Often, a resistor in series with the output would provide some protection against voltage pulses that might exceed the rails. If you are pulling 7-8mA from the output, you may not be able to tolerate much series resistance. Diode clamps to the supply rails may help.

You may  be able to probe the system to look for voltage spikes during power up/down or other transient conditions such as load variations, switching events, etc. You may be about to narrow down the source of transients and eliminate or mitigate the behavior in some way.

Again, schematics, photos or what ever you have could help. You can zip multiple files into one zip file and attach it using the paperclip button.

Regards, Bruce.

7065.T100101421_P9.pdfBruce,

Thanks for the reply.  I will try to attach thte failure report( finally got an electronic copy), and the schematic.  U42 is the only one of 3 ISO124U parts that fails. 

I stand corrected on the output drive, it takes about 3-4.5mA to drive the power supply reference ckt on the output , not 7-8mA. 

We have looked for switching transients upon turn on, turn off etc without much luck.

Best regards,

Bill Ernst

TDI Power

Bill Ernst

5481.FR-061201.pdf

Bill,

This looks like some pretty serious overstress of some type. I cannot see where U42's output is connected. Is it on this schematic somewhere?

Regards, Bruce.

Bruce,

No it is not.  It goes off board via P1-5, P2-5 and P3-5.  I do not think that I am at liberty to send that schematic out.  What I can tell you is that it goes to a series 1K resistor which branches to a 10.5K resisitor that goes to the poitive input of a LM358 amplifier.  Also going to that positive junction is 4.096V reference through a 10K resistor.  Also from the 10.5K resistor, there is a .1uF capacitor to return. 

As I have said in previous posts, there are 3 rectifiers that have this circuitry on them.  they are connected to this PCB via 3 separate cable assemblies.

Best regards,

Bill 

Bill,

Based on the FA report, I'm suspicious of what could be happening on this output terminal. I suggest that we take further discussions off this forum. You can contact me directly at trump_bruce@ti.com.

Regards, Bruce.

Hi Bruce and Bill,

When problem is solved, please post the conclusion here (if it is possible).

At job we use the ISO124 and have had problems in the past. We would like to know were are the potential problems and how to avoid/solve it.

Thanks,

Best regards

Ivan,

Bill's application has a pretty low impedance network on the output of the ISO124 with a possibility of pulses that might pull the output beyond its power supply under unusual circumstances. The solution is likely to involve clamp diodes to the power supplies and possibly raising the impedance of the output network.

Isolation amplifiers generally live in difficult environments. Consider any possible transient conditions that could cause excessive power supply voltages or pull inputs or outputs beyond the supplies. Current limiting resistors on inputs and outputs and clamp diodes may lead to a more robust design.

Regards, Bruce

Thanks, Bruce.

Regards