LMC662: Input offset current appears to vastly different between lots

Tom,

In addition to the symbolization on the package, I also need the full part number.

Both parts are labeled as LMC662AIM.

Tom,

How much offset current are you observing?

Through simulation, I believe it to be about 0.2pA versus the advertised typical of 2fA where our prototypes were running. It also various sightly in time.

Tom,

I got a response back; 'HKAN' is not a valid part marking.
If the issue is a 0.2pA offset, could that be due to the circuit board or contaminates on it?
The data sheet has a section for "PRINTED-CIRCUIT-BOARD LAYOUT FOR HIGH-IMPEDANCE WORK"; were these practices followed faithfully?
Also the lower the current the longer the time to measure it, therefore the test program likely only tests long enough to a verify a resolution of 100’s of femto-amps. There would be no way to know the difference between 2fA and 20fA

• See pictures...
• Contaminants in the board layers from multiple panels is unlikely, but it is part of our investigation.
• The datasheet guidelines for board layout were followed faithfully and as this is our 4th run of these boards, has been proven through test to work. Three previous builds of this assembly with three different LDC of this op-amp (8CX7, M6AB, HLAB) have all performed perfectly
• If the TI LMC662 spice model is accurate, this circuit is that sensitive as I've been able to simulate the observables in spice.
• I'm looking for some of the older LDC parts now to replace one of the HKAN parts to validate my theory and will post back once those results are known.

One follow-up question, please confirm that the HKAN marking is or is not a TI marking. If it's not, I need to trace this part back to see where it came from. Not necessarily related to this issue directly, but I want to make sure I know where these parts are coming from.

Thanks.

Tom,

TI uses the first symbol (0-9) to reflect year's last digit and second symbol as month (1-C). Prior to National Semiconductor acquisition, the symbol format was likely  different. Therefore the 'HKAN' device predates November 2012 (oldest record in database).