wongck wrote:I just transfered 3 x 2MB ZIP files over from my PC to my Falc for the test.
All came over fine with zero data corruption on STZIP test.
May be it's the software you use to transfer them.
I am using FTP to transfer them.
I first noticed it when I tried to grab a Quincy3 demo off the web using CAB. So...I grabbed it again with Opera on my PC. Fired up gaftp on the falcon and tried to get it from there via ftp. Yes, double checked that it was in binary mode. Tried both passive and port modes. Also tried aftp. Tried storing it to a couple of different partitions on the Falcon (maybe I better double check with a SCSI drive..for this session I had only the Dynaport on the chain). STzip coughs every time.
Next I tried a different file. Thing127e.zip demo version. Same deal. This archive I know was fine because I had used a floppy earlier and unzipped it on the Falcon to get the English bits out for my fresh N.AES setup. Quincy and Thing both unzip fine on the PC. Hmmm...I got everything I need using floppies or flash media, but I'm still a bit miffed at this. Further tests...if the zip file is small...say 200k or less it comes over uncorrupted. The one's in the 512k range and up came over botched. We're still talking rather small files in the grand scheme of things. The largest zip I tried to move this way was well under 1,400k.
This is a CT2b Falcon, but at this phase of getting back into Atari I've left the CT2 switched off. I did try various bus timings in STing...ranging from 10 to 50.
I use HDDriver 7.93, and presently run a very simple Geneva/NeoDesk environment in cooperative mode. Tried the same tests both with and without the CAF_FIX patch.
It does seem that my SCSI bus resets itself alot when using the dynaport...maybe every 6 to 10 seconds during a transfer; however, that minor skip doesn't seem to slow things down a bit...or cause any other system issues. I did try some lower recieve windows and such in STing...not sure what the ideal settings are for this device...but I started with the recommended settings for eNEC. SCSILNK.STX docs do not go into much detail about optimal buffers and such (unless I overlooked it all somehow).
Note...when I put this falcon back on the desk the first thing I did to test the SCSI was load up CuBase Audio and max it out playing 6 while recording two 44.1k audio tracks for a good 30 minute run. That seemed to go well...tho' I didn't stress test it with a lot of punch starts and stops...just a steady run.
Next thing I'll try when I get a chance is moving some files around in the internal structure of the Falcon and STing itself. I.E. Run ftpserver and move some files from localhost to localhost. I suppose if that works fine then I'll know it's probably not a local STing or hard drive issue. Guess I can pull the ole eNEC off the STacy and give that a try as well.