BreakPoint

cancel
Showing results for 
Search instead for 
Did you mean: 

question about Insulation Resistance measurement

I have a Keithley 6487 Picoammeter/Voltage Source that failed its calibration check; the current measurements are slightly low.  We don't use the current measurement, only the Insulation Resistance (IR) measurement.  Is it reasonable to say that the IR value is a function of the meter's ability to measure current and that the erroneously lower current will result in a proportionally higher IR?

Jim
You're entirely bonkers. But I'll tell you a secret. All the best people are. ~ Alice
For he does not know what will happen; So who can tell him when it will occur? Eccl. 8:7

0 Kudos
Message 1 of 8
(8,212 Views)

That would be my thinking - it's just Ohm's law with a good voltage source and the picoammeter measurement. Is it a linearity issue or an offset one?

---
CLA
Message 2 of 8
(8,206 Views)

@thoult wrote:

[...]  Is it a linearity issue or an offset one?


Linearity.

Jim
You're entirely bonkers. But I'll tell you a secret. All the best people are. ~ Alice
For he does not know what will happen; So who can tell him when it will occur? Eccl. 8:7

0 Kudos
Message 3 of 8
(8,194 Views)

Will bear that in mind, as we're building one into a test process at the moment - thanks for being the sacrificial lamb 🙂

---
CLA
0 Kudos
Message 4 of 8
(8,164 Views)

@thoult wrote:

Will bear that in mind, as we're building one into a test process at the moment - thanks for being the sacrificial lamb 🙂


It's fairly robust gear.  However, like anything, it can partially fail.  Calibration cycles note when readings are in greater than expected error.  

 

Then you have to go back and look at the products shipped that passed tests Dependant on that gear and the out of tolerance function.  It SUCKS. But, has to be done.  Artificially high insulation resistance readings do not mean that the product is bad but,....well,the experiment designed with the intent to prove the product was flawed was compromised.  Warranty RMAs or recall rates may rise.

 

Its not fun when you get those OOT reports and the OOT measurements are used.

 

 

 

I would not object to this gear in a test system based on one failed cal.


"Should be" isn't "Is" -Jay
0 Kudos
Message 5 of 8
(8,144 Views)

It's nice knowing what the failure modes seen in third party kit are, and putting in checks to spot them in preventative maintenance schedules.

 

There's always the difficulty too of having nice expensive kit that has to go for external calibration - it means having a second unit to swap in to cover any downtime, which isn't always appreciated by the bean counters due to cost and warranty concerns.

---
CLA
0 Kudos
Message 6 of 8
(8,122 Views)

A cow-orker suggested that I implement a periodic sanity-check by testing a known resistance value.  I won't do that.  I'll, rather, shorten the QA calibration cycle now that we've demonstrated that the current cycle may be too long.  Also, that "known" resistor will need to be added to the QA system.  Fortunately, this device is more accurate than the device the test was designed to use, so, even out of calibration, we're in good shape as far as recalling product goes.

 

I've never had trouble getting spare equipment to use when stuff gets sent out for calibration.  It doesn't apply to this case, where the measurement was used in the acceptance test, but I see situations where a small part of a device's functionality is used but the entire functionality is scrutinized.  This is easiest for us to implement, but it's often more appropriate to write a specific calibration procedure to validate only the functionality used in a particular system.  This also make the device un-stealable by someone looking to "borrow" it for some other application. 

 

This is part of my new role as a Manufacturing Engineer.  I'm discovering so many opportunities as I "peel the onion" in my new domain 😄

Jim
You're entirely bonkers. But I'll tell you a secret. All the best people are. ~ Alice
For he does not know what will happen; So who can tell him when it will occur? Eccl. 8:7

Message 7 of 8
(8,109 Views)

I have a laser encoder in my test platform which requires calibration every two years. Fortunately, as it's a product that another arm of my company manufacturers, it's not going a long way for cal. The downside is that the lead time for both cal and for delivery of a new unit are too long for comfort, meaning I will basically have to stock a spare that sits on the shelf for two years doing nothing, then goes into use, whilst the one that is removed for calibration will come back and sit on a shelf for two years.

 

Have you peeled back the layer that deals with process change management yet? That's my favourite one - I saw an approval request the other day which stated as its validation "this code has been running in production for the last month and there have been no problems", which roughly translates as "dear higher powers, please approve this change that I already unilaterally made to my process about a month ago but haven't told anyone about yet".

 

---
CLA
Message 8 of 8
(8,102 Views)