SG-3100 - how to reduce it's temp [quietly] ;-)
-
@msf2000 - on my very first attempt to cool the SG-3100 I tried 2x 80mm fans and a very mvp (min viable product) makeshift air-guide to blow air into the side of the unit, but the impact was so unnoticeable that I moved on to cooling from below. I did try also with the fans the other way around to draw off any warm air from the side vents, but that was similarly uneventful.
To be honest, what I'd love to do, is take the lid off the SG-3100, and pop it onto my CNC, where a little bit of Fusion 360 and a 3mm or 4mm drill bit later would result in about 80-100 perfectly aligned and spaced vent holes in the top of the case. And that's where I'd place the 120mm fan with a grill, and dust filter etc.
I wonder if anyone on this forum has a dead or surplus SG-3100 with a top cover no longer needed ;-) Though I've seen 3D models of the mechanics of the SG-3100....
One thing I didn't mention was why I went with a 15v PSU and the variable voltage regulator. Well, my plan is/was to run the fan much lower than it it's full speed, but I couldn't be sure if that would have enough of an impact, so I knew that I'd start testing with the fan running on 12v, but the regulator has a voltage drop of 1.5v, so a 12v psu would have only given me 10.5v due to the voltage drop of the regulator.
Anyway, here’s a picture of the current as requested, it's still pretty mvp.
-
Hello!
I just setup an sg-3100 and had a similar question about the temp.
I have had good luck cooling other equipment with these:
https://www.acinfinity.com/quiet-usb-fans/
Maybe you could stick some rubber feet to the top lid, run the sg-3100 upside down, and put the fans on top?John
-
@serbus - agreed, a USB powered fan is such a sweet spot of a solution, no ext psu required, no additional trailing leads, and with a couple of power/speed settings the fan you shared looks like a great fit. Plus with heat naturally rising, drawing it off from the top could actually produce better results. I might try that, thanks ;-)
-
i use three wired usb fans connected to the bottom of a laptop prop and blow air into the bottom of the device. my temp went from 62-65 to 56-58
-
This "double" USB fan kit should do the trick, one below, one above:
https://www.amazon.com/AC-Infinity-MULTIFAN-Receiver-Playstation/dp/B00JLV4BWC
Jeff
-
@akuma1x - thanks for taking the time to post. The whole cooling to noise ratio situation has changed a little bit because my ISP has changed. Moving away from Virgin Media’s DOCSIS 3.0 350Mbps service to BT’s 910Mbps FTTH meant that I was able to put the SG-3100 into my 27u cabinet, but most importantly out from behind the TV and out from the living room. In the cab with the fan running the SG-3100 ticks over at a respectable 50°C, which is fine for me.
Way back I did try the Hi/Med/Low/Off USB fans but they were just too loud for the living room. But I loved the fact that they didn’t need an external power supply like my current setup does. -
@pmk_mck said in SG-3100 - how to reduce it's temp [quietly] ;-):
to BT’s 910Mbps FTTH
Grr. g.fast only just reached me.
-
probably a fire hazard
-
@beachbum2021 - so what temp does the Netgate run at during a hot day or worst case for you please?
-
@stephenw10 - if you’re close to the street cabinet then the speeds should be pretty respectable. What does your CPE train to in upstream and downstream please? And do you have to supply reverse power to the DSLAM in the cabinet?
-
@pmk_mck it runs about 53 on a lite load day, 61 tops on a hot day.
-
@pmk_mck
I am awkwardly positioned relatively close to the exchange but quite far from the g.fast 'POD' cabinet. I actually have two DSL lines here and the regular FTTC is pretty good though contention is quite high these days. I used to get close to the theoretical maximum in that line. The sync speed varies but today I'm seeing:Max: Upstream rate = 26378 Kbps, Downstream rate = 75080 Kbps Bearer: 0, Upstream rate = 19999 Kbps, Downstream rate = 66999 KbpsThe other line with g.fast enabled is faster, ~140Mbps, but that's towards the low end for g.fast as I understand it.
I have no idea what's syncing at since the modem is locked up tight.
Kinda straying off topic there.....
Steve
-
So yesterday I placed the SG-3100 upside down, so that the vents were facing upwards, and the white lid down. Same fan as before sucking the air away from the vents on the SG-3100. And left it for a day to settle down.
Results: There's no difference in the temps reported by the SG-3100 when upside down with the same fan in the same place, sucking air away from the unit. It's still around 49C any time I look at it. And then when I run a speedtest over ethernet (~650Mbps) the temp peaks at 57C and appears to normalise around 54C.
-
@pmk_mck yours is still running less hot than mine. I moved the 3 fans. one blows air in the bottom vents and two on each side suck the air out. The issue I believe is that while we're blowing air in, the chip sets are on the opposite side of the case. Mine has never gotten to the low levels yours are.
-
When my SG-3100 gets busy, it warms up for sure too, at 41% CPU the temp is 55C.
Maybe my next step is to 3D print or CNC a new top cover/lid for the SG-3100 and make it with a load of holes in....or a fan cage?
Or, actually I could probably take the top lid entirely off the SG-3100, and leave the fan sucking air as-is today?
-
@serbus - So yesterday I placed the SG-3100 upside down, so that the vents were facing upwards, and the white lid down, with 3mm tall rubber feet to create an ambient. Same fan as before sucking the air away from the vents on the SG-3100. And left it for a day to settle down.
Results: There's no difference in the temps reported by the SG-3100 when upside down with the same fan in the same place, sucking air away from the unit. It's still around 49C any time I look at it. And then when I run a speedtest over ethernet (~650Mbps) the temp peaks at 57C and appears to normalise around 54C.
-
@pmk_mck take the case off and let it run with a fan blowing on it, i mean..if you're blowing air in the dust is there anyways. Just dont want to electrocute or short anything during the test. I can do the same if you want to know the result.
-
S serbus referenced this topic on
