Having read the horror stories about the previous slide feeder (SF-200), it was with some trepidation that I purchased this for the LS5000ED. I have thousands of slides to scan, and feeding them one at a time doesn't seem like an option.
I was pleased to find that the first batch of Kodak cardboard mounts went through without a hitch. These were some Kodachromes from the early 70s, normally exposed. The scanner took the stack of 40 without a problem and completed the scans. However, my good luck didn't really continue. About every third stack manages to jam, normally just after I watch the first 3 slides go through without a hitch and I leave for my real job.
This slide feeder is just plain poorly designed. The feed path is designed such that each slide must pass under the stack, giving the possibility of catching the edges of the window in the mount above. Additionally, for some reason, the slide gate adjustment, which adjusts for the thickness of the mount, wants to move a bit on it's own. Failure to re-adjust after stack has caused a couple of jams, where the gate moved off of vertical and closed the opening too far. These problems all occurred with rather uniform stacks of slide in good to excellent condition.
I've had no problems with Gepe plastic mounts for slides I've mounted myself, but that is normally only a small fraction of slides that people need to scan.
One solution to the jamming seems to be to use a shim to put the feed follower pressure on the outside of the slide stack. This fans the stack a bit on the inside, which has allowed a balky stack to feed okay. Be careful how you engineer this, however, as if the shim gets caught in the feed mechanism you'll have another set of problems getting the feeder back to normal.
An additional failure I have had once was for the autofocus to go totally wrong. It managed to scan an entire stack of slides badly mis-focused. I put the single slide adapter in and the slides focused and scanned fine. I put the feeder back, only to find the problem persisted. I was about to pack it off to Nikon, when I power cycled the whole thing and found that correct operation had been restored.
The error handling on the batch scanning is broken, with the software thinking that scans have been completed which haven't. This is just a nuisance, resulting in the wrong file number part of the saved files if you don't catch it when you restart the scan. However, my software does detect that the slide didn't feed, and just shuts down, requiring exiting and restarting the scanner software to resume.
And then there is the general issue of software stability. On Windows 2000 with the SP4+ stuff, I cannot use USB 2.0 (scanner software looses communications with the scanner, and restarting the software leads to a blue screen of death). I also have to restart the software after each roll of film (approximately) or the application crashes. I'm going to try XP one of these days and see if it's any better. Hard to say if it's Nikon or MS that is screwing this up. Probably a joint effort.
Update: now running on a fast machine with XP Pro SP2, USB2.0 - after solving the XP ROC-GEM problem, I can report that the software is reasonably stable, typically going several rolls or sets of slides without crashing. There are no system crashes under XP, just application faults. Slide scans take around a minute to 90 seconds each in the stack feeded.
One drawback of the slide feeder over the film strip batch scanner is that there is no way to do different scan settings for the slides in the stack. With film strips you can tweak the settings for each frame, but there is no preview capability in the batch scan from the slide feeder. This is an oversight that they could correct in the software. However, the feeder reverses the slide order, so you would have to restack the slides between the preview pass and the full scan pass, or the software would have to be smarter than Nikon.
So I just use a default of ICE on and DEE of 30 and rescan manually any very difficult shots. With the 16-bit channel depth (actually only 14 unless you do multisampling - correction, the 5000 does 16-bit scanning with or without multisampling, the V is 14-bit) most exposure and shadow/highlight problems can be compensated in Photoshop afterwards.
All in all, I give Nikon a C for this effort - they could do some software improvements to raise the grade to a B- with a preview option, and possibly get a real solid B if they modified the pressure plate (needs to be adjustable where the pressure is placed). But the horizontal stack design will always have jamming problems. Why they couldn't use a tray feed with an open acceptance path like working slide projectors is beyond me.