I did the same thing, sucked out the solder in the IDE header with a Radio Shack desoldering iron (the $9.99 one with a red bulb). To make things easier, I *added* a little blob of low-melting 63/37 solder to each hole with a regular soldering iron, then sucked it all out with the desoldering iron. Then I soldered in a snappable Berg 40-pin header.
I could not find 4x22 resistor packs in that size, so I bridged them with solder (very tedious). Added R99 and R100, but used a slightly larger sized chip resistor and had difficulties positioning it exactly. Added a 2-pin header of the +5v for a notebook hard drive.
Connected a notebook hard drive using a 2.5->3.5 adapter to the on-board IDE header. I could autodetect the hard drive from BIOS, but I could not boot. Tried everything, including removing the DiskOnChip. I suspect that the 22 ohm anti-reflection resistor packs are crucial to reliable reads from the IDE drive. Either that, or keep the cable length down to a few inches.
Anyway, I gave up and used an ISA IDE card, which worked like a charm (see other posting).
BTW, what is the value of the 9 slew-rate limiting capacitors around the onboard COM1 header (CN18)? I've fixed the header and I'll be adding the capacitors later.