Hi all,
Unfortunately we have encountered a problem with some MULTICOMP brand resistors that seem to change value once assembled and deployed out into the field. The component in question is:
Supplier Part No. . : 1170974
Manu./Brand . . . . : MULTICOMP
Manu. Part No.. . . : MC 0.063W 0603 1% 73K2
The fault:
After one to two months of operation, we are finding that the circuit that this is in no longer functions correctly. After diagnosis on three returned boards, it was found on all three that the component in question (R6) had drifted from its specified value. The value should have been 73K2 but in reality we measured 28K - 58K over the sample of three measured using an extech CMM-17CMM-17 calibrator and pogo probes.
The maximum stresses that this resistor will be exposed to:
1/ Power: the resistor has a power rating of 1/16th (0.0625) of a Watt at 70 degrees Celsius. Assuming you manage to connect 24V (this is the absolute worst case being connected to it, though it should be 15V ), the power dissipated through this resistor if directly connected would be 0.00822 W – It’s quite clear this is not the cause. Many units have failed from 10.8V solar panel supplies.
2/ Temperature: The resistor is designed to operate up to 155 degrees Celsius – even with ambient of 80 degrees Celsius inside the enclosure it will not reach this temperature. At above this temperature, I would simply expect it to fall out of tolerance but return at lower temperatures again. Details of thermal dissipation can be found in another thread I posted a few months ago:
3/ Humidity: These are in a IP67 rated enclosure that is installed usually under shelter. Each enclosure has a breather gland to prevent the build up of excessive humidity.
Our investigation so far:
1/ After examining the various methods that the resistor could be damaged and finding no obvious reasons, we decided to replace the resistor on existing boards due for dispatch with an equivilant from vishay. Out of many units (more than where originally shipped with MULTICOMP part) that have now gone out into the field with the replaced resistor for much longer periods, none have failed.
2/ Many more units with the original MULTICOMP part have now failed out in the field and have had to be returned to us for rework, so these where not isolated cases.
3/ We have been in touch with farnell, who forwarded this case to the manufacturer of the resistor (ROYAL resistor). The three boards that had the faulty resistors where forwarded for examination. The manufacturer has confirmed that the values are out of range and not to specification. They have taken resistors from the same batch and performed a variety of tests, each with no unexpected results. I have attached their results to this post. Farnell have not been pro-active in dealing with this and have not even offered a replacement of these parts yet.
Obviously this has left us in a very awkward situation where we have had to retrieve these faulty units from the field, rework and replace them.
What do you think?
Given this very strange set of circumstances, I'm left unsure how to proceed. Do I chalk it up to experience and avoid MULTICOMP? Take it further and have these independantly analysed? Part of me wants to understand the root cause of the failure.
I have no shortage of faulty units from the field with this component drifting from it's initial value.
Your thoughts would be greatly appreciated.
Thanks
Oliver
Attachments:
| Short time overload test.pdf | |
| Pulse characteristic test.pdf | |
| 0603 1% 73k2-llh-3.pdf | |
| 0603 1% 73k2-ll-3.pdf | |
| 2010.10.19 RF102-01-11-01 0603 1sK2 value drop.pdf |
