Grounding your heatsink. Anyone here ever try it?

[KnightElite]

We've got a thread (http://www.hardforum.com/showthread.php?s=&threadid=486010) over at hardforum about whether grounding the heatsink will allow a higher stable overclock or not. I've tried it, and so far it seems be working, since I would have normally gotten a BSOD about an hour ago at this clock speed.

But I'll leave it on overnight in order to be certain. Any info you guys can add would be excellent.

[QuadDamage]

I've tried it, and so far it seems be working

have you tried it on your axp? how did you do that? have you soldered anything to the HSF itself? would a piece of wire do? any suggestions? i'm willing to check that out, since my air cooling solutions are very much maxed out.

[Theta]

Sounds interesting, to say the least!

I'll test out the theory tomorrow.

[KnightElite]

Well, so far so good. It's run overnight with no troubles, and it used to BSOD after a couple of hours. I just stuck a wire in the heatsink and stuck another wire through the whole in the case. No alligator clips or anything.

[docah]

I tried this befrore when trying to pop ou an extra mhz from an AXP1900+ when i was TEC/water Cooled.

It gained me nothing, and in my opinion an unsoldered unclipped wire sitting around in your case is asking for trouble, and my dead gf3 would agree.

[Smizack]

Originally posted by docah
...an unsoldered unclipped wire sitting around in your case is asking for trouble, and my dead gf3 would agree.

That's kinda what I was thinking.

I haven't tried it, but I don't see where you would get stability from it.

[docah]

I understand the idea of it(i think). Grounding the heatsink allows it to act as a big EMI shield preventing the preventing the "Noise" generated by the chip from affecting the rest of the system. As well as preventing the noise from the rest of the system from possibly affecting the signals on the cpu.

By that logic, a northbridge that is having trouble at high fsb could be helped out by grounding it's heatsink, and the same for the new ddr sticks with heatspreaders.

It simply didn't help in my case, back in the K6 days it did have minor stabalizing effects for some people. Again it didn't help my k6 at all. So it makes sense that for some people it helps on the intel chips, as they too have heatspreaders.

Edit: If done properly with spade connectors and soldering where neccecary, it could prove to be safe. Even if it doesen't help the overclock, if you find your computer interferes with your television, such grounding tequniqes (if effective, which i'm not sure they are) could reduce that.

[KnightElite]

Well, it certainly seems to be working for me, at any rate. Also worked for Squeaky1 on hardforum. Not sure why or how, but it does work ;).

[PMM]

I tried this a long time ago with no improvement
but the theory is sound.

Your heatsink esp. if bonded to your CPU core with Artic Silver
paste can be acting as a very good radiator of RFI generated
from the CPU.

In general electrical waves run on the outer surface of a metal
surface same principles as used in Antenna design and the
Faraday Cage.

Earthing the Heatsink with a good connection will null out
some not all the RFI, the bond to the earthing lead being
on the heatsinks bottom is logically the ideal.

I suspect with the ever increasing MHZ this method could be
showing more rewards esp as we are heading into Microwave
frequencies with our CPU's.

Considering how much power our CPU's take that could be
translated into a lot of RFI if it works is a different matter!!

I like to think it does a bit, but I never tried it since funny enough
when I owned a K6 LOL.

[docah]

the k6 was such a crappy performer i did anything i could to get it up to speed... I even bought 133 ram and pumped the bus to 113 on my epox board at the time, my 450 made it to 560. and the bus speed helped me run UT nearly as smooth as my friends celeron :), i knew better, but they were convinced and that's all i wanted.

[KnightElite]

No one else have any thoughts? The thread over at [H]ard|Forum is already like 15 pages long ;). Seems a few people have had success with it now, and for some it does nothing. One person was able to get 1MHz more FSB on his C3 :D.

[BigDov]

hmmmm...... read through the whole thread and not sure if I want to do this or not. I probably will because I'm game for anything........ I just won't expect it to net me huge gains :) Thanks for the heads-up KE.

[docah]

After reading the thread, which i never initially did. It looks like it seems to help out those who are pushing voltages more. Which fits in pretty well with the assumption that it reduces emi in the system, as more voltage would make the chip "noisier".

[DarkLord666]

Anyone wanna post some pix ... ?
Coz I got myself some Aircooling, but also maxed out. So I'm game ... anyone also wanna tell me what can go wrong ?

Running my XP1700+ @ 1700Mhz
Not too bad .. it's at 170x10
And since it's am AGOIA Y core, it should do better ... right :)

At acceptable temperatures & noize ... for me anyway ;)

[morphling1]

This definetly works for me, I grounded my waterblock and I gain 3mhz in fsb speed, that before grounding hard lock as soon as I started seti.
Tomorrow I'll be installing new cpu block, so I'll solder pin to both cpu and nb block for grounding and from now on I'll have M3 threaded hole in all the future blocks. This thing didn't get me to much (4mhz) cpu frequency overclock but it sure did bump my fsb from 184 to 187 (most agressive timings) on my 8kha+ and Corsair pc 2700 ddr

[KnightElite]

That's about the same gain as everyone else is reporting as well. Awesome that more people have gotten results. As to pics, there are several in the hardforum thread.

[DarkLord666]

Ah ... cool.
Let me go read through that then. Gotta work today anyway ;)

[seversphere]

:bounce:

you guys still doing this trick?

Does it help out with vapo cooling?

[Tweaked!]

Put it this way, could it hurt? Seriously though, With the amount of electricity flowing so close together, seems to me like grounding would definitely help. When electricty flows seperately so close together like it does within chips and on mobos, it has to generate some kind of energy field, which grounding would help in that scenario. Not sure if voltages this low actually do create any kind of energy field, but higher voltages definitely do, so I would think that this would too, to a certain degree.

[muzz]

I believe AMD has a write up on their site about this, like trying to get rid of an RF loop( ground), which I've had to do many X b4. I would somehow attach it solidly ( thats just me ) as not doing so could lead to an intermittent situation, which could actually be WORSE than nothing at all.

muzz

[muzz]

It's not so much the voltage, but the freq. Oscillation is tough to track down.

[KnightElite]

Actually, it's good of you to bring this back to life. I'm still doing it, and I posted the same thing over at ProCooling:
http://forums.procooling.com/vbb/showthread.php?s=&threadid=4417
However, someone over there used an oscilloscope to measure pulses, you can check out his data in the thread. Quite interesting.

[Penguinsfan]

gonna try this tonight :)

should be interesting if it works since i live in an apartment with crappy ass wiring

[SupaMan]

ugh, 5billion pages, and the 1st post has no how-to data.

So what do u do? just make a metal connection from HSF to case?

[wwilliam54]

Originally posted by SupaMan
ugh, 5billion pages, and the 1st post has no how-to data.

So what do u do? just make a metal connection from HSF to case?
yup

[felix88]

heh, i remember that thread over at the hardforums. i was going to try it, but forgot. maybe i'll remember this time. ;)

[Penguinsfan]

i'm going to solder a wire to the bottom of my heatsink on the copper slab then attach the other end to a black lead somewhere in the rat's nest of wiring going on in my case lol

[wwilliam54]

I would try it, but one of my procs doesnt have a hsf on it
(p2 and I dont care that it runs at 72c)

[KnightElite]

Originally posted by wwilliam54
I would try it, but one of my procs doesnt have a hsf on it
(p2 and I dont care that it runs at 72c)

I'm sorry to say that there must be at the very least a heatsink on it. If it was just the silicon chip exposed to air, it would have burned up long ago.

[Haut^Karl]

All the heatsinks on Macintosh computers have been grounded since at least 1994 and is done to reduce static buildup in the aluminum heatsink. Imagine a very small but high voltage arc from the heatsink to the cpu die. Not a good thing. I doubt grounding would help overclocking.

[wwilliam54]

Originally posted by KnightElite
I'm sorry to say that there must be at the very least a heatsink on it. If it was just the silicon chip exposed to air, it would have burned up long ago.


I know It "Should" have but it hasnt (its a crap 266)

really wierd cuz my p2 350 I had would lock up if the HSF fan was off

still a POS ill be glad to replace soon

[Penguinsfan]

Soldering a wire to my heatsink didn't go as planned (surface area to contact ratio = way too large lol need a larger end for my iron)

I managed to thread a wire into a screw hole on the heatsink and force the screw through the hole, then connected the other end to the negative (black) terminal of a fan connector....however it did not help...pushed the cpu to 1693MHz and it still could not run 3dmark.

Either it's not making good contact which i doubt or it's the fact that my apartment isnt properly grounded *shrug* heh

Kept me busy for an hour and a half at least

[SupaMan]

solder? why not just use alligator clips?

[STEvil]

That would be me that got 1 more mhz on the c3... 8-)

Did some more testing today (have put in 62 hours of combining in the last 4 days) and these are some more results.. 8-)

XP1600+ AGKGA Y & Dr. Thermal Ti77n w/ 60mm delta 7k.
No ground - 1585mhz
Ground - 1626

XP1800+ AGKGA 4 & Dr. Thermal Ti77n w/ 60mm delta 7k
No ground - 1725
Ground - 1782

I wonder if this would get OPP over 3g? 8-)

[Penguinsfan]

i dont have any alligator clips and am too lazy to buy any heh

[KnightElite]

I may have an explanation as to why this works. I mentionned it to a 4th year EE student today, that I was able to clock higher with a grounded heatsink, and he was not at all surprised. I didn't entirely understand everything he told me, but he said it has to do with the fact that at the clock speeds that modern computers run at, all the traces have to act as waveguides, so one electrical signal creates a companion signal in the other traces.... (I probably got that completely wrong :D). But in any case, that design feature causes an electric field in the heatsink, and that grounding it gives a better waveguide. He says that in theory, connecting the heatsink to the positive voltage line should work just as well as a ground, but it's not quite as safe :D.

[Ranger Loki]

OK so just to be sure...

You attach a wire from your HEAT SINK, and connect teh other end to your case? I have an Aluminum Case and a Swifty, so what I am thining is that I simply have to TIE the wire to the HSF since it uses RODs and not Fins, and the same for the case.

Am I correct? CUase I want to hit 3GHz!

[SupaMan]

Originally posted by Ranger Loki
OK so just to be sure...

You attach a wire from your HEAT SINK, and connect teh other end to your case? I have an Aluminum Case and a Swifty, so what I am thining is that I simply have to TIE the wire to the HSF since it uses RODs and not Fins, and the same for the case.

Am I correct? CUase I want to hit 3GHz!

yeah, u got it.

[GreenBeret]

Might try this weird method later :)

Edit: grounded the base of my MCX462 but no luck.

[antipop]

What about doing it with a wc sys, will this work? Has anyone tried?

[SupaMan]

I just grounded my swifty MC462A, didnt help worth a damn....

[JBELL]

Originally posted by SupaMan
I just grounded my swifty MC462A, didnt help worth a damn....

how are you grounding? from base or the pins?

[SupaMan]

copper base to steel case backplane, via 1/8inch wide copper insulated wire

[Dissolved]

mines now grounded via alagator clips on the fins of my ax-7 to my case.

[STEvil]

Make sure the place you are attaching to is not painted, annodized, or some other thing that may reduce the connectors condiviness...

Also, if the "earth" ground in your house is no good, this will likely do nothing at all.

I dont know if grounding with watercooling is going to have a lot of affect. If the loop is already grounded then the water should be acting as a conductor.. unless your using perfectly distilled water.. 8-)

[BigDov]

I have a real grounding strap that I foraged out of the garage and tried that for a while...... with no noticeable results. I'll be trying again though!

[felix88]

damn, i may go to napa tomorrow and buy some stuff so i can try this. maybe i'll be able to squeeze an extra 50 mhz out of this chip.

[antipop]

Originally posted by STEvil
I dont know if grounding with watercooling is going to have a lot of affect. If the loop is already grounded then the water should be acting as a conductor.. unless your using perfectly distilled water.. 8-) I'll try it and see how it performs

[antipop]

I've grounded my wc setup and that didn't help me at all. So i removed it

[DisposableHero]

Originally posted by antipop
I've grounded my wc setup and that didn't help me at all. So i removed it

how much voltage are you using on your chip?

[WannaGoFast]

Tried it, gained nothing. But left it attached because it isnt doing any harm, and I'm lazy:D

[Hausmeister]

Hmm

I have something crazy to tell ya... Some time ago I played with my multimeter and measured evrything i could find (yeah, i was bored) and I also measured the Heatsink of my 2nd Rig on Ground and it was grounded:confused:
(Taisol HS and Abit KT7aR)
The strange thing is there is no other connection from the Heatsink to the Rest except the CPU ( i unplugged the Fan to be sure) but it still measures Resistance against the Case....

Don't know what to think about this except that the DIE is somehow connected to ground ??

[antipop]

I'm at 1.85V (the highest i can go without a vmod)

[felix88]

Originally posted by Hausmeister
Don't know what to think about this except that the DIE is somehow connected to ground ??

the die is non conductive. if it was conductive there would be some major problems.

any electrical signals your heatsink could be picking up would be too small for most multimeters to detect, but in theory they could cause interference with your cpu.

[Hausmeister]

true, but how is it connected to ground then? The soket is plasic, the Fan has no connection either...
btw its the same wether it's running or not...

Anyone else experienced something like this? It just makes me wonder because i can't imagine where it should be connected to the case.. only thing left is the cpu

[felix88]

it isn't connected to ground. ground generally is 0v, and your hsf should be at 0v. even if it isn't grounded.

what kind of heat sink is it?

[KnightElite]

Originally posted by Hausmeister
true, but how is it connected to ground then? The soket is plasic, the Fan has no connection either...
btw its the same wether it's running or not...

Anyone else experienced something like this? It just makes me wonder because i can't imagine where it should be connected to the case.. only thing left is the cpu

Just for the hell of it, I checked the voltage and the resistance between my heatsink and the case. I get -0.15Vdc and infinite resistance.... rather interesting.... Obviously, when it's grounded, the resistance is close to zero, and the voltage between the two is zero.

[antipop]

The core cover is made out of ceramic so it's not conductive, even the p4 who have a metal cover there is an isolant underneath it

[Liquid3D]

In hopes there are those who subscribe to the theory; "There are no stupid questions..." as watercooling also creates the same grounding environment, perhaps this is another (hidden) reason WC works so well?
Has anyone noticed if this grounding has any effects on 12V, 5V or 3,3V readings? Perhaps the "Antec issue" (low readings on the 12V and 5V rails) could be ameliorated?

[antipop]

Originally posted by Liquid3D
In hopes there are those who subscribe to the theory; "There are no stupid questions..." as watercooling also creates the same grounding environment, perhaps this is another (hidden) reason WC works so well?
I don't think it's a reason when you switch from aircooled to wc, it's mostly because of the drop in temp

[LightningCrash]

This is a little unrelated, but Matrox did ground the heatsinks on their G450 series cards.

From A-Vault:
"Dear Client,
In testing of our G450 boards we noticed that under certain conditions, noise was induced on the analog signals -- a common challenge with integrated circuitry. The wire added to the heat sink is used to ground the heat sink. The screw is needed to allow for conduction through the non-conductive coating of the heat sink. Together, they eliminate the noise."

Did anyone think that maybe the grounding eleminates some level of noise within the voltage signals fed to the CPU?

[STEvil]

Several times, this thread and the one on the H...

Mac's are grounded, too.. as are Xeons and some high end servers, if im not mistaken..?

[LightningCrash]

Originally posted by STEvil
Several times, this thread and the one on the H...

Mac's are grounded, too.. as are Xeons and some high end servers, if im not mistaken..?

uh, I read the thread, nobody ever mentioned noise affecting the power feeds themselves

[STEvil]

Well, I kind of thought the posts reguarding line noise and whatnot would have meant that... Maybe im just thinking too hard, though.. 8-)

It wouldnt be the power feeds, though. Those are strong currents whereas signals to the ram/agp/pci/northbridge/southbridge/etc are all relatively small signals and easily disrupted...

[Ranger Loki]

OK I built a system for a friend using spare parts of mine, I did this trick on it to overclock the system for him, and the CPU which was a 1.6A did 133MHz FSB at stock Voltages, before it could only do 127MHz. It might do 134MHz, but it won't do 135MHz, it BSOD's on startup, but 133MHz is perfectly stable, didn't crash EVER. So I say the trick works.

I did the trick to my personal system also, I hate restarting so I haven't tested yet, but I have nothing to do right now anyways.