One by one...
Well just a longer test phase would be great.. As I read it you did 10 test's overall right? I recall Jamie doing hundreds of hours of testing in order to finally reveal there were issues because just a few tests didn't show them.
I've discussed with Jamie - he felt 10 x 3 hour tests was adequate. I can certainly rack up a few more hours but perhaps it would be more valuable if someone with a different firmware version ran a few tests.
Also testing on a fragmented drive (as you had mentioned) might be warrented.
One thing I want to stay clear of is
trying to make it fail for no good reason. You can do something like this with a laptop too (fragment the hell out of the drive or load a crappy config and tons of background programs until it fails) - but why bother? There are simple, good practices take place even on the high end professional equipment and DAWs. Keep free space on your drive (though I did run it down to a few gigs left) and periodically defrag i.e. run the built in disk cleanup routine every once in a while.
You make no mention of how your powering your JB3 which could be worthy of testing/mentioning. Was it plugged into AC power during all your tests? Have you tried removing it from AC during recording allowing the battery to kick in, and vice versa? How about doing it with a battery only half charged? Etc.. etc..
It was AC powered for the tests. From some previous power experiments, it does switch back and forth without a glitch. I can verify the switching back and forth again for bit accuracy and let you know.
As far as a half charged battery goes, the highest current draw occurs on the bootup (when you see the EAX screen and the drive spins up). Over 1 amp as I recall. It may be possible to continue a recording in progress on a battery that doesn't have enough juice to start the JB3 if you pull a switch after startup (kinda desparate though!). The good news is when an internal battery dies, the player shuts down nicely and saves your file. The bad news is that same safety isn't built into the external power source - so if you pull the AC without the internal battery, or run external power (and it dies) in it, you just lose the whole file (whether you have the internal battery or not).
I've got a whole pile of stuff on power / external power that I can post as well (had a lot of fun playing with this machine ya know!)
I'm just thinking that really giving it more of a "stress test" overall would help solidify your theory. Trying to imagine what all a taper might do.. Do Firewire transferred recordings differ from one transferred via USB? What software are you using to transfer from the JB3? Creative PlayCenter? If so what version? Have you tried different versions? Have you tried Notmad Explorer (looks really kickass, don't have it myself)? What version of the drivers are are you using and on what operating systems as I can imagine different OS, different drivers, different version of PlayCenter might result in different issues/problems perhaps?
All of the above has to do with the transfer of a data file - it works or doesn't work. There is no opportunity to create a file with missing samples as it won't pass the basic error correction required to move the file. We're no longer in the audio realm on this one. FWIW, I use the Creative Play Center 3.02.52, Win2k, driver 1.21.01a (which lists as 1.20.4.51 in device manager)- but all that stuff does is allow me to transfer a data file from one device to another.
Just some thoughts popping into my head. I know I have had problems with transfers which were either linked to USB (now using firewire) or (believe it or not) having the unit plugged in while transferring instead of using the battery (which seems to work better for some people)...
I can also do JB3 > USB > PC but I almost never get a larger file transferred before getting an error "Player Not Connected - " at a random interval during transfer.
I've also had occasional transfer issues that always seem to be fixed by disconnecting the firewire cable, going into record mode, recording a few seconds of anything (like silence thru the line in or mic in), stopping the player and trying again. Likely some file system issue in the JB3.
I'd gladly lend a hand with testing.
My chain is like 'dis:
UA-5 (optical) > JB3 (1.20.07) > Firewire > PC
Looks like the main difference is the firmware - and that would definitely be worth a test. As you may know, I can't rollback to an earlier version with 1.32.02 so go for it.
The catch is you need a
reliable USB transfer for comparison. That's what shut me down the first time around - I pretty much knew the JB3 was behaving but my control sample (the USB copy) was totally flaky. Then I got a laptop and did dual boot with a totally stripped config for recording and all is good.
So try a few runs and see if you get the same results or not - each time I had anomalies, I could zoom in on the USB copy and see that it was the bad one, but it's still better to have absolute proof. You can look at the link to my original tests on that page to see the kinds of glitches and stuff I was getting.
Good luck!
david