When overclocking , you are not only heat stressing the CPU/RAM but the power supply as well, as Lost_Animal has pointed out in a previous post. Consider also adding heatsink/airflow to the power supply chip. I don't have an RT-N16 YET to look at this section, the switching power supply section should either have a controller with integrated switch (the heatsink goes on it), or a controller with external (MOSFET?) switch (the heatsink goes on the external switch).
Most of the SW chips I worked with have some sort of overcurrent AND/OR thermal protection, if stressed and the chip is working near this area, you might get into all kind of problems (random erratic behavior, reboots..), that you might think are software related and thus impossible to troubleshoot from a software end.
I would advise against using the yellow dot. It looks like it is before the protection fuse (white, to the left, labeled F3), if something goes wrong with he fan, it might short the power supply. Assuming the black dot is ground of course.
A cheap effective way to cool is to take an old spare 12v power brick from a Walkman, etc. and splice in a fan to it. I used a 120mm and added a cheap Zalman Fanmate 2 to control the speed. Very quiet. I initially built this assembly to cool a Pioneer receiver, but now it sits on the router. I have heatsinks coming from eBay and may end up drilling bigger holes on the router top cover. With the sinks and fan, RT-N16 will stay nice and cool.
Joined: 23 Dec 2009 Posts: 16 Location: 39°15'58"N, 81°32'32"W
Posted: Tue Feb 02, 2010 23:00 Post subject:
wahran wrote:
When overclocking , you are not only heat stressing the CPU/RAM but the power supply as well, as Lost_Animal has pointed out in a previous post. Consider also adding heatsink/airflow to the power supply chip. I don't have an RT-N16 YET to look at this section, the switching power supply section should either have a controller with integrated switch (the heatsink goes on it), or a controller with external (MOSFET?) switch (the heatsink goes on the external switch).
Most of the SW chips I worked with have some sort of overcurrent AND/OR thermal protection, if stressed and the chip is working near this area, you might get into all kind of problems (random erratic behavior, reboots..), that you might think are software related and thus impossible to troubleshoot from a software end.
My 2 pennies.
All of the voltage regulators, the ram, and support circuitry run cool to touch. Overclocking to 533 results in no noticeable change in heat output. But this is only my observation on my N16. Others results will undoubtedly very with ambient temps and airflow. I have marked the voltage regulator's in white. If your having random reboot or wireless instability check these points.
wahran wrote:
I would advise against using the yellow dot. It looks like it is before the protection fuse (white, to the left, labeled F3), if something goes wrong with he fan, it might short the power supply. Assuming the black dot is ground of course.
Its easier to replace a 12v powersupply (12v ones are very common) then replacing (or even worse dead shorting) a SMD fuse. But I guess that's a matter of opinion. _________________ " I don't care about what anything was DESIGNED to do, I care about what it CAN do. " - Gene Kranz
Posted: Tue Feb 02, 2010 23:27 Post subject: Quick note on my RT-N16
So, I finally (after a month or so) got around to flashing DD-WRT (13491 big) onto my RT-N16. The stock firmware was (how can I say this delicately?) unacceptably poor in every conceivable respect. The performance change was astonishing. I mean, I'd run older, mini versions of DD-WRT on my WRT54G V8 in the ancient past, so I somewhat knew what to expect, but the V24 kernel on the RT-N16 is really Red Hot. Not that I'm overclocking like the guys above - I'm not a power user. But my pings have dropped by 60% and my throughput is improved up to 20X (really) to the WAN compared to the stock firmware.
What is ASUSTek thinking? They ship DD-WRT on other products, don't they? Newegg indicates that they do. Why not on RT-N16?
I'll have more questions as I walk through the massive amount of features on the big binary, but I am really impressed with DD-WRT on the RT-N16. This will be a great WAP / router / file and print server platform for me...
Joined: 23 Dec 2009 Posts: 16 Location: 39°15'58"N, 81°32'32"W
Posted: Tue Feb 02, 2010 23:42 Post subject:
ACE 256 wrote:
wahran wrote:
When overclocking , you are not only heat stressing the CPU/RAM but the power supply as well, as Lost_Animal has pointed out in a previous post. Consider also adding heatsink/airflow to the power supply chip. I don't have an RT-N16 YET to look at this section, the switching power supply section should either have a controller with integrated switch (the heatsink goes on it), or a controller with external (MOSFET?) switch (the heatsink goes on the external switch).
Most of the SW chips I worked with have some sort of overcurrent AND/OR thermal protection, if stressed and the chip is working near this area, you might get into all kind of problems (random erratic behavior, reboots..), that you might think are software related and thus impossible to troubleshoot from a software end.
My 2 pennies.
All of the voltage regulators, the ram, and support circuitry run cool to touch. Overclocking to 533 results in no noticeable change in heat output. But this is only my observation on my N16. Others results will undoubtedly very with ambient temps and airflow. I have marked the voltage regulator's in white. If your having random reboot or wireless instability check these points.
I tested my n16 with no air flow to see how hot the voltage support circuitry gets. They do get hot enough to be a concern to those with no active cooling or air flow. Seeing as cooler is always better I used my tried and true "guetto sinks". Simply super gluing fan screws to the Vregs. If you sink your Vregs don't forget the memory Vreg :wink:
_________________ " I don't care about what anything was DESIGNED to do, I care about what it CAN do. " - Gene Kranz
That is not a USB Header, is just a serial port header. You won't be able to power the fan from there, as between VCC and GND is only 3.3v and very low amps, so it's not enough power... That was my original idea, but afterwards I had to connect the fan to the original 12v power connector of the router...
Interesting. Has anyone hooked up a terminal to the serial port? What can we do with that?
Always good to have the serial port, it allows you to access the router if the network isn't working (if you changed something on the settings which you shouldn't have or something not working right and you locked out the network ports) and gives you full shell access. Once you back in you can change the wrong setting back via nvram command and you won't loose all settings (good if no nvram backup exists). In addition you will be able to access the router on the CFE level if you interrupt the boot sequence with control C (erase nvram etc.).
Nice on the RT-N16 is the fact that the header is already there and you don't have to solder it on first.
If you are interested in building a cable here is some info: http://www.everythingplug.com/ttl-to-rs232-serial-port
I actually found an old fan (3.3v) from an old server at work. It blows sideways, not down, and it is working perfectly off the serial port line (via old CD-ROM audio cable).
I tested this for days on with extensive file transfers in both dd-wrt, stock Asus fw, and Tomato fw, and all chips kept cool with nice flow.
Joined: 24 Aug 2009 Posts: 2039 Location: South Florida
Posted: Wed Feb 03, 2010 4:41 Post subject:
My current setup (80mm Fans powered via USB, one sucks, one blows )
Running at 532,266
DSC00407-1.png
Description:
Filesize:
714.61 KB
Viewed:
17336 Time(s)
_________________ Optware, the Right Way
Asus WL520gU
Asus WL500gPv2 x2
Asus WL500W
Asus RT-N10
Asus RT-N12
Asus RT-N16 x3
D-Link DIR-825B
Linksys WRT320Nv1
Linksys WRT600Nv1.1
Linksys WRT610Nv1
Linksys E2000
Netgear WNDR3300
TP-Link WR1043ND
I might have to join the 'fan club', pun intended :p
My router rebooted while I was watching the Lost through my slingbox while at work. 20 minutes before it rebooted, my video kept pausing, and was getting very low throughput. So maybe it was running hot or something.
A little update on he cooling side... Since on the first try, the 40mm fan mounted on the cpu heatsink was a little bit noisy at night (the router is placed in a bedroom) because the fan was very close to the top cover of the case, I decided to cut a hole in it, changed the fan to a bigger one (50mm) and also reduced the rpm by using two 10 Ohm 3W resistors. In addition, I've covered all the top and side air holes, so now there is a good air flow on both top and bottom sides of the board.
_________________ ASUS RT-N16 with TomatoND USB Mod
Joined: 30 Nov 2009 Posts: 90 Location: Kristiansund, Norway
Posted: Thu Feb 04, 2010 6:47 Post subject:
Bg8780 wrote:
Anyone else having issues with QoS?
Yeah, I tried QoS the other day. Internet became very erratic and speed dived to a few percent. Had to reset the router to get it back to normal. _________________ RT-N66U, WZR-HP-G300NH
DD-WRT v24-sp2 (01/02/10) big
(SVN revision 13575M NEWD-2 K2.6 Eko)
I've been running this firmware for a couple days now. Haven't really tested a whole lot but plan to once I get a little more time.
I have USB working with optware packages installed. Nexst up I'm goign to attach a 1TB drive and get Samba running.
/proc/swaps is missing but I can live with that for now. This router *does* have 128 MB of RAM after all.
Current uptime is 2 days 5 minutes.
Thanks Eko and everyone else working on DD-WRT! _________________ =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Asus RT-N16
DD-WRT v24-sp2 (03/24/10) mega
(SVN revision 14144)
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
If you sink your Vregs don't forget the memory Vreg :wink:
)
