Kitz Forum
Broadband Related => Broadband Hardware => Topic started by: risk_reversal on September 11, 2009, 06:28:27 PM
-
I was wondering if users with a DG834GT could provide me their feedback as to how hot their units get. I know it is a subjective exercise but please try.
Could you try to describe the heat at the top & bottom also.
Cheers
-
Mine is slightly warm on top, rather warmer on the bottom, but nowhere near to being uncomfortable to hold.
-
Not got mine plugged in atm, so bit difficult to say for sure.
But iirc if the router was flat it was the left side/middleish that got the warmest. - bottom if on its side
It was still ok though, youd touch it and think thats warm.. but not unbearably so.
Where mine was it had plenty of air-flow although was near the PC.
I wouldnt say its any hotter than some other routers that Ive tried.
-
I'd echo the comments already posted. When I was running mine it was warm on the top with the base a bit warmer; this was in bridge mode with the wireless deactivated, and I always stood it up vertical with the stand.
-
Hi
Just warm certainly not HOT. similar to most small appliances like my shaver charger warm to the touch
Regards Jeff
edit I suppose it will vary depending on how hard it is working.
-
I have mine held vertically in its stand. After many months of being switched on it is cool/cold at the bottom and warmish at the middle and top. Certainly not hot - if it was an old cup of coffee I would be throwing it away rather than drinking it at that temperature! :)
-
Many thanks for your feedback. The reason I was asking is that over the last 5 days or so, my DG834GT has been rebooting every day or so. This is an anytime occurrence.
I started a new LLU service about 2 1/2 weeks ago and was trying to compile a list of likely suspects. Router rebooting due to heat was one item that I was trying to get the measure of. I purchased this router to use with my new LLU service, it is an ex Sky flashed router.
jeffbb said:
edit I suppose it will vary depending on how hard it is working.
Good point yes.
Mine has the wireless disabled so just modem/router. The 4 ports are connected up and prior to my posting the unit was lying flat ie on top of a set of bookshelves. The top was getting very warm and the base was very very warm, bordering hot.
It's difficult to stand the unit upright so this is what I have done. I have a couple of pairs off the brackets that are used to stand the DG834GT upright and have used those 4 brackets on the unit and now it sits flat but with about 1 1/2" gap from the flat surface.
The unit is now just very warm on the base so an improvement. Been without reboot for just over 2 days, so lets see.
Your feedback has been been very informative.
Cheers
-
My Netgear router used to get get very warm (not the same model)
However, what I discovered was because of the way I had it in place, either lying flat on the wooden desktop or standing on a cardboard box was not helping to dissipate the heat as the wood or cardboard were storing the heat, what I did then was to lie it flat on top of a old HDD which was also just the right size, this works a treat as the heat is absorbed into the metal and dissipates far more effectively.
(Originally I was looking for a small kitchen trivet, but never found one the size I wanted)
I don't like these types of things standing upright because that way all/any heat has to travel upwards passing all the other heat giving out components.
Whereas lying flat the heat automatically rises to the top of the unit away from other components and is more likely to get expelled quicker out of any vents.
-
Upright position, slightly warm on top.
Cold on bottom.
-
Router site and orientation can be important as some have found.
I discovered a 585V7 flat on a shelf in direct sunlight that was almost too hot to touch, which was clearly inadvisable.
Likewise keeping it "cosy" under papers etc. is to be avoided.
However when manufacturers put side vents rather than top and bottom when flat, and supply feet to stand the unit vertically, they clearly expect them to be used.
In these circumstances if a modem is getting hot when flat on a shelf I suggest it's better to use the feet and obtain a convected air current through the unit rather than trapping the air inside in a pancake.
Kind regards,
Walter
-
Hi
Routers based on the Broadcom chip (equally applies to the other chipsets), regardless of how hot the case feels (which is very subjective) get very hot. The chip itself is rarely heatsinked and I've measured temperatures of 70-80 degrees on the chip itself (using infra-red thermometer).
Silicon can run at quite high temperatures just fine so 70 to 80 degrees is probably okay (the plastic package is likely to fail before the silicone due to excessive heat) but the life of silicon chips is shortened quite a bit for rises in temperature.
The main problem is the circuit board and surrounding components are used as a crude heatsink so they all warm up. Capacitors which are not far removed from being just cans of spiralled wetted blotting paper age very quickly in continuous heat, and the soldered legs conduct heat from the copper tracks into them like dry central heating.
Most problems (rebooting, inability to hold a connection, freezing, not turning on) with routers after 12 months or more of continuous operation are most likely due to capacitors having dried out and aged so they longer do their job correctly.
Very occasionally a router with a Broadcom chip will ship with a heatsink (I've seen them D-Link 2640Bs) however they are only tiny and without very good venting or active cooling the heat stays in the box anyway. I've fitted a slab of a heatsink from an old Nvidia graphics card to my Broadcom chip and even with this huge chunk of metal and the top lid completely removed it gets hot and reaches 40 to 50 degrees. Turning the router on it's end may help but mostly it just stops a hot spot on the plastic case, inside the Broadcom chip will still be running very hot.
Really for 24/7 operation active cooling is required or a very good heatsink and venting design to remove the heat from the box, but adding this would increase the cost considerably and an active design would not be silent and so wouldn't be liked by everyone. Tiny cheap stick-on heatsinks little bigger than the chip itself don't really help and can become detached during shipping (unless a more complicated fixing is used raising cost again) causing increased returns.
These routers are very cost sensitive, essentially you are buying a small Linux computer with 4 network cards, a Wi-Fi dongle, some RAM and flash memory, powersupply and an ADSL modem, plus software for around £40.00 to £100. The manufacturer isn't going to want to erode profit or be uncompetitive by implementing a decent cooling solution if it can get away without doing.
Having an inbuilt obsolescence (they don't last long!) with a product that hasn't changed or advanced really for many years (the same Broadcom chips are used in the very latest models that were used in models 4 or 5 years ago) is no bad thing for the manufacturer as it creates ongoing sales where there would be few, as once bought few will replace them unless they break. I personally think SpeedTouch deliberately case their routers with very poor ventilation in order to cause more early failures. As they supply ISPs who supply them to their customers, I bet their sales model is to sell each unit extremely cheaply, to the ISP, when they fail the ISP then needs to buy a replacement, providing ongoing income for SpeedTouch or their distributors. This is the only explanation I can think of for ISPs (such as Be/O2 and others) for supplying these boxes that most would rate as very unreliable.
Regards
Phil
-
You make some good points there phil.
Price is a major issue and they do appear to cut corners.
Whether its true or not I dont know, but I heard on the grapevine that the TG585v7 has an inferior board over the ST585v6, due to cost constraints of a major ISP who they supply to.
The large ISPs can drive a hard bargain and the manufacturers will go for it & cut pennies where they can because they obviously want to get the order.
-
Hi All
I have a DGN2000 which gets quite warm, not hot though. I thought it might benefit from being stood on two bits of wood about 10mm high, this helped and the underside doesn't get so hot. But then I thought id try and make it cooler, so, even though it's new i took it apart, it's only 4 torx screws under the rubber feet. From new the Broadcom chip had a little heat-sink on it as did the chip behind it. I fitted a bigger, flatter chipset heat-sink and it didn't have an improvement, so i returned it to original.
Next mod attempt is to get a flat fan in there to move the air outwards. I'll let you know if it works 8)
Hammy
-
I thought it might benefit from being stood on to bits of wood
The problem with wood is that does not dissipate the heat but stores it.
Try using metal as it will spread and dissipate the heat faster than wood, or in my case a old HDD lying flat.
If memory serves me right, heat is drawn into cold objects, so a flat plate of metal will help to draw the heat away evenly and avoid hot spots.
-
:lol: http://forum1.netgear.com/showthread.php?p=128605 (http://forum1.netgear.com/showthread.php?p=128605)
-
Hi
quote : If memory serves me right, heat is drawn into cold objects, so a flat plate of metal will help to draw the heat away evenly and avoid hot spots
True ,Would probably be even better if supported on 4 small posts to allow air circulation underneath .
Regards Jeff
-
Sorry it wasn't very clear about the wood. This is it :-
(https://forum.kitz.co.uk/proxy.php?request=http%3A%2F%2Fi37.tinypic.com%2F21aefxs.jpg&hash=99b585170f26b7aaa0ad9f7ec6106a5cad48ace7)
I tried it on an old HDD and all that happens was the HDD got hot as well.
Ref the post above, the DGN2000 already has two little heat-sinks on but is still gets rather warm.
If the internal fan idea works i'll post a pic of that as well.
Hammy
-
I tried it on an old HDD and all that happens was the HDD got hot as well.
That would just go to show how effective the metal HDD was at drawing away the heat, rather than still waiting for it to rise and go away.
-
I have my DG834GT mounted vertically, sort of "hanging" on the wall (using the clip-on feet over some small hooks) about a foot above my master socket in the hallway (well ventilated).
It's just warm not really hot. You could leave your hand on it indefinitely.
-
:lol: http://forum1.netgear.com/showthread.php?p=128605 (http://forum1.netgear.com/showthread.php?p=128605)
OMG.... router modding :lol: :lol:
/me laughs - good idea though if you have the time and patience I suppose.
beware that you will lose your warranty for yur router if you do this:
erm... yes I think you would :D
... nor would you be able to stand it on itside any more either :lol:
-
@ Kitz,
Very good advice !
"Das maschinen ist nicht for gefingerpoken und mittengrabben. Sitz back
und vatch das spitzensparken."
or more fully at :-
http://onlinedictionary.datasegment.com/word/blinkenlights
Kind regards,
Walter
-
I love the mangled German :lol:
-
Hi
Ref the post above, the DGN2000 already has two little heat-sinks on but is still gets rather warm.
Yes those little fin heatsinks don't really do very much as the heat produced by the Broadcom chip is more than can be dissipated by the heatsink. Once the heatsink is hot it needs to pass heat into the air and out the box, but without forced convection or much better ventilation the heat just remains in the box and heats up just as much. I'm surprised they bothered really :)
Whether its true or not I dont know, but I heard on the grapevine that the TG585v7 has an inferior board over the ST585v6, due to cost constraints of a major ISP who they supply to.
I can believe it, most new models that arrive now are just cost cutting measures by the manufacturer, for example the DGN2000 model is basically a DG834N but uses the cheaper 6348 chip rather than the faster 6358 that was in the DG834N. Netgear can even produce the same model but with a completely different chipset just because they can get it OEM'd cheaper. So the model number appears the same but you can be buying something completely different. They do add a Version number to the model number to help tell them apart, but most people would assume the newer model is a better revision of the previous one, and not something altogether different internally.
D-Link have just done the same thing, the 2740R version that will replace the 2740B is a completely different piece of equipment using a Trendchip chipset and not Broadcom.
Regards
Phil
-
Hi,
I've always felt my DG834GT runs too hot, and after acquiring another unit I have found the difference in temperature between the two has been very distinct. I've located the source of the excessive heat to be coming from the BCM5325 network chip near the network ports. On my old router, the chip gets too hot to touch, where the new one is much cooler.
I set about replacing the capacitors on the old router with better quality versions, and after fitting I tested continuity across capacitors to check for short circuits. All were fine, apart from 'C107' at the bottom of the board. This cap showed a short circuit / resistance reading of about 20ohms, but the cap was brand new so I'm thinking the problem lies with the mainboard or some other component.
Testing the other DG834GT, this cap showed no such short (as you'd expect) and the BCM chip runs cool. Hmmm I thought, perhaps the old routers' original cap had gone faulty, causing a short circuit and damaging the BCM chip enough for it to run excessively hot, but not enough to stop it working completely. While testing, I accidentally bridged the cap connections rendering my router useless while the BCM chip temperature soared, AND now we're left with a short circuit across the cap!
It looks like any issues with that capacitor that could result in a short circuit across its connections causes permanent internal damage to the BCM5325, resulting in extremely high running temperatures and sometimes, but not always, router death.
It would be interesting to see measurements across the capacitor of those who consider their routers to run hot, perhaps it's possible to prevent overheating of the BCM5325 by replacing the capacitor BEFORE it goes faulty. Anyone here handy with a multimeter and feeling extremely bored, feel free to give it a go :p
Thoughts / comments? :)