My gut feeling, unproven in any way but discussed extensively, is that the quick format routine in the DR70D is buggy.
That's a very interesting theory, but if it were true I would expect it would negatively impact all cards formatted in the unit.
You'd think. But part of the discussion has been about the possibility of different production runs, potential for poor quality control because it's a low cost unit, etc. I don't think those debates need to be rehashed and I'm not claiming that the gut feeling I expressed above is the source and solution. At the same time, I want to try and replicate the issues I experienced with an unapproved card that was problematic and the first place I'm going to focus is the quick format. If I can find some kind of repeatable action that reproduces the glitches onto that specific card, I can get to the bottom of what's causing the problem...and perhaps determine if the problem is sourced from my unit or from the card. Don't know exactly how I'm going to do that right now though...maybe I'll 'stress' the card to simulate that I'd used it over and over by doing 20 or 30 quick formats while recording a few minutes between formats. Don't really know and I have no idea if or when I'll even do this.
The idea of possible QC issues in this price range certainly is possible. It happened to me with my first Sony M10. The built-in mics were almost 3dB mismatched, and Sony service determined it was manufacturing defect and sent me a replacement.
Your "stress" test sounds good. I imagine you'd to this at 4 channels, 24/96 to tax the media to its fullest.
If you're testing the reliability of the media itself and not necessarily the format routine in isolation, I would propose a similar test again, but run 4 channels at 24/96 and fill the card completely. When storage media goes bad, it often starts in specific places which could be anywhere across the entire writable area, and not just the "beginning". By "going bad" I am talking about bad sectors, but your test could throw up some filesystem corruption (caused by a variety of factors) which is the problem I experienced. If the 70D is somehow causing otherwise good cards to have problems, I think it would be filesystem corruption; not bad sectors which are a much more severe issue.
Your test sounds good for expediency and makes sense for testing the reliability of the format routines, but keep in mind that you'll be putting much more wear onto the "beginning" of the writable area that way. That would of course make a borderline bad sector on that portion of the card more likely to be hit, but one later on less likely to be "seen" by the recorder. Since quick formats don't look for bad sectors, writing to a previously-unknown (and thus unmarked) bad sector will definitely cause an error in the recording. The natural thing to do after this happens is to run a full format, and your formatting program should tell you (but it doesn't always) if it found bad sectors. Otherwise