@jeremy:
Funny little machine this is. I did notice however that when it runs, telnet becomes slow so power must be limited. Also its wlan signal is a lot worse than that from my SpeedTouch modem. Level as reported by Windows is fine but a connection is only possible rather close to this box. A more powerful antenna probably would improve this.
The engineer that came up with the idea of holding the boards (more or less) in place using this bar/spring construction should be sent back to boarding school. Gee; what a kludge. The bar on my unit was too long so that the box didn't close properly.
A shame it can't mount cifs/smb volumes... would turn it into a little file server.
The firmware 1.9.6.4 with (as the only modification) telnetd enabled for WL-530g V2 is at this moment available here:
http://www.euronet.nl/users/ernstoud/wl530g/WL530g_1.9.6.0_EN-V2-TELNETD.binEither copy it to your download site and point there, or point to my link.
When I take your firmware apart (in zImage and cramfs) and use asustrx to create a bin file with the proper system version and checksum, it will load into the WL-530g V2, the device starts (power and wlan led turn on) but it doesn't work. No SSID found, no telnet possible, no device on my network.
Luckily enough recovery works

I ftp'd (to /tmp) some binaries that you compiled (for instance 'cpu' or 'busybox') into the box but also these don't run (ERROR 11 occurs, whatever that means - I found that as a kernel error it means 'try again', someone must have had some serious fun when this was defined

).
UPDATE: this was due to the ftp client on the wl-530g defaulting to ascii transfers. Giving a 'binary' command before the download solves the problem. I can now download working executables to the device. 
I noticed that the GPL on the CD that came with it is 1.9.5.2 where for your experiments 1.9.4.2 was used. On GPL 1.9.5.2 there are loads of changes from beginning of 2008 where the latest firmware (1.9.4.6) for the W-530g is dated june 2007. So asus changed a lot for V2.
I had a more thorough look and the two PCB's are different than those in older pictures. (Someone should tell the FCC that asus changed the board

). Also the used SOC has an indication BAN1 instead of BAN. The rest of the chips are more or less the same but I reckon that a WL-530g and WL-530g V2 are different enough that binaries need to be recompiled. Perhaps some other ARM version is used...
I just untarred the GPL (> 60.000 files, more than 1 Gb - blimey!) and need to devote some serious attention to this. I doubt whether I have the time...
Will post on progress... although with the loss of your archive I guess a lot of attention for this box was lost also. On
http://wl500g.info/ this box also gets very limited attention.
UPDATE: In het latest GPL 1.9.5.2 there are a lot of different versions of the Marvell wlan object files (no source available, so the nature of these changes is unclear). I reckon the wl-530g v2 has a different 802.11bg radio chip or setup. Recent pictures show a totally different layout of the pcb under the cover that covers the radio part compared to the FCC images. So that might explain why db90h's firmware loads, runs but no wlan is possible. I plan to replace the mvlan.o file from the db90h's firmware with a more recent version within the GPL 1.9.5.2 CD-ROM.
@tsveten:
See link mentioned above.
Ernst
P.S. db90h = define byte NOP
