Initial state. I bought the sucker new off eBay from the UK for $2500 when it was brand new (pre-"M"). It was a grey market unit (of course), so
Juniper effectively wanted me to buy it again to get support. I managed to wrangle a couple years' worth out of SBCIS's Juniper rep.
Sitting on top is a J2300. I abandoned this at SBC's Lakeside building when we vacated (in 2015), as the CF interface had died.
You can see a couple more Netscreens on the right...
|
...Specifically a 5XT (also with a fan and heat sink mod, as it was a burning weenie roaster) on top of a 5GT.
|
Interior views, as I disassembled the unit...
|
|
|
Yep, just a PC with a custom motherboard.
|
A close-up of the Mini-PCI Nitrox card, which gave the thing its 1Gb/s performance over encrypted tunnels.
|
This is what I removed from the unit: three agressive 80mm Delta fans and companion shroud.
|
Reinforcement plate for the fan and filter, with the crossbar for the filter lid stud.
|
Layout... sort of. Hard to tell what all is upside-down. But it does illustrate position and fit.
|
The reason for the corner divot.
|
Power harness for the fan...
|
...And a view complete with fan and speed control.
|
Detail of the speed control. I adjusted it so that the fan topped out at less than full speed.
|
Layout for cable position and fit.
|
Factory heat sink on the right with Swiftech replacement on the left...
|
...And another view. The factory heat sink was not bad, but not suitable for use with the down-firing main fan.
ScreenOS ran the poor P4 551 at full power (~75W) at all times - no idle reduction. It was a cooker.
|
As installed. I don't recall why I didn't wire up the tach output - perhaps it wasn't used with the Deltas. That would tend to explain how I got away
with eliminating two fans completely without complaints from the system. (A locked rotor signal would have required spoofing. Ask me how I know.)
|
The Rotron Major DC, assembled and ready for installation.
|
Power source for the Major. The SSG550 had SATA power but no SATA devices. I don't even see any SATA headers offhand.
Normally I'd just use an appropriate mating connector, but SATA mates to a PCB edge. There might be some solution out there, but it's never come
up again.
|
Everything worked OK.
|
Interior views of the assembled lid.
|
|
|
Assembled less filter lid. Looks like I'd run out of USA-made AC A697Cs. I'd snagged the last ones a few years earlier for Ares4.
|
This shows the position of the big fan. Not much clearance.
|
Complete. As with all of these projects, it was louder than I'd hoped (but still not as loud as stock). A major culprit was the 40mm fans in the
power supply (I ran it with one installed, as the mounting hardware for the reinforcement plate interfered with the upper supply). I replaced them
with quieter ones that had similar performance but (in particular) much better balance. I had to spoof the locked rotor signal, as I'd gotten tach
fans.
Now that I'm thinking about it, fan balance was a major motivation for this project. Up on its shelf, this thing shook the entire wall. Very
annoying.
The Major eventually pretty much wore out - after a few years it didn't want to start from power-on. I don't know what became of the unit after I
decommissioned Ares4 in 2015. I assume it went to the great scrap heap in the ground.
|