Its a bit late but I hadn't the time until now to read in detail, but I suspect Fluke designs its DMM to far exceed IEC impulse test voltages to allow margin for error in production, be it component quality, to assembly. We should consider that when producing a batch of devices, a range of performance can be expected due to parts and construction tolerances, but if the Fluke design criteria far exceeds required IEC criteria, then we can expect nearly if not all production to pass IEC criteria, with individual devices being more or less robust than a sibling.
In the JQS video, the 87V malfunctions in ohms at 13kV, but that impulse is still above its 8kV rating and other functions appear normal. Alas, there was no function test at under 13kV. The Fluke educational video posted earlier in this thread shows an 87V dying at 17kV in volts mode, again far above the IEC requirement, but they did not show how well it functioned between each kV impulse test.
I think the clearest take home message is the lowly entry level Fluke 101 DMM, which appears to have only CE Fluke safety certification, no 3rd party safety certification, is likely to take surge's well, and likely meets CAT III 600V criteria. Chances are the 101 was designed like other Fluke DMMs, to survive far over the required IEC surge voltage.
That prompted me to search if indeed the 101 and other non-sold in the US Fluke meters have had 3rd party testing, as I noticed some 101 have a CSA mark, others do not. I found the CSA listing for the 101 and several other Fluke meters not marked CSA, which means these Flukes were tested to a point independently. The devices made it far enough in the certification processes [safety issues] to receive a listing. Why Fluke doesn't mark them after they've been granted a listing can be due to many speculative reasons, e.g. not paying recurring fees to keep production costs low, not completing less critical items in the CSA procedures, etc.,
I expect low. But we can't know for sure unless we try a surge test and demonstrate it, then its not just an educated guess. Many things can change over time and cause problems in new DMMs versus prior runs of the same model. A test of just one meter can be criticized, but its better than nothing.
This is one reason in the past, say in the US military, samples of a procurement were tested per batch by independent military labs to insure they live up to their specification, but I don't know if they still do this.
The 87V is a defacto standard in that class DMM, so these tests will attract attention from a lot of professionals, particularly if it fails
What are the chances of failure?
This is an excellent point. I was very happy when another member took it upon themselves to run similar tests on the 101 (Well, that is until I stepped things for that last round ). I would like to see a second 87V tested as well just so we have two data points. Even then, that's too small of a sample size. I'm sure Fluke already has the answer as that video made it sound like they test every design to failure.
From what I understand from all of the posts I have read about the amount of money Fluke has invested in making their designs robust, and the 87V being a very popular meter and how long they have had to improve their designs, and again we are talking about it just doing as well as the lowest cost meter Fluke offers. It doesn't even have to exceed it! I assume the chances of a failure are very low.