View Full Version : Alright the first OC results are in...
Using the incredible CuBE (parts as listed in SIG; only thing not listed I believe is the KOYO race car radiator used to cool the system). This is using passive mode, as I currently have no fans...
(Remember, this is a Xeon 3060: 2.40GHz)
Ambient is typically around 20*C in the basement, where the CuBE is kept.
On STOCK:
<22*C while idling in Vista
OC (VCore = 1.5063, DDR2V = 1.95, MCH1.5V = 1.65, ExtClk = 365 ==> Core Speed = 3285.9 MHz)
While Crunching (WCG):
63*C
EDIT. I'm doing this in a separate post because I'd like to summarize my conversation with our buddy serialkiller or w/e the spelling is. Je me pardon.
SUMMARY. "Your OC sux balls. People are owning you on air LOLZ0RZ"
Well, apparantly i'm not OC'g the right way. I think I raised the VCC too quickly. Now I'm noticing that I can do the same OC at a much lower (close to stock VCC)... I think I got owned due to temperature reasons.
Stay tuned =D
EDIT2. 1:18a, GMT-5
I've dropped VCC to stock (1.3 something)... and kept raising fsb with 1:1. I was able to hit ~3.5ghz until it froze. Funnily, this was the same brick wall with a much higher VCC. Tomorrow I will try to get past it using a slightly higher VCC setting.
serialk11r
06-05-2007, 11:56 PM
Now that you've said it already, I don't think you'll mind a little correction LOL no actually summary: you're whoring 1.5KW from your socket and you got owned by someone who used a not so good straight water system :P
Now that you've said it already, I don't think you'll mind a little correction LOL no actually summary: you're whoring 1.5KW from your socket and you got owned by someone who used a not so good straight water system :P
-__-
:doh:
Holst
06-06-2007, 09:22 AM
Im sorry I cant quite understand your post :(
Your running TEC with a PASSIVE auto rad and temps are 63*c
Is this idle or load?
Any idea of water temperature?
Car radiators are poor performers with PC watercooling even with fans, without any fans im supprised you didnt boil the water (If this happens it will be VERY VERY BAD)
Sry; allow me to clarify: the 63*C was TEC with passive auto radiator. However, this is the KOYO radiator... it is huge. Though after much testing last nite, it turns out that temperatures drop drastically with forced conv. on the radiator.
It also turns out that those OC's had stability issues.
Right now with minimal air flow, I'm reading:
CPU Temp: 47*C (Question: how do I tell what the temperature is if it keeps jumping around? It jumps from 39 to 47 to 45 and keeps repeating... I simply took the highest value to be safe).
Using Core Temp 0.95 (Values also jump around significantly, from 40 to 39 to 48 to 51 for Core0, and 37 to 44 to 36 to 46, and back to 37 for Core1):
Core0: 51
Core1: 46
Right now the OC is as follows:
FSB: 370
VCORE: 1.4563
MCH1.5V: 1.65
DDR2V: 1.95
Multiplier seems to jump and oscillate between x6.0 and x9.0 in CPU-Z, resulting in oscillating readings between a core speed of 2219 and 3329... Anyone have any idea what could be the problem with that?
littleowl
06-06-2007, 09:53 AM
what psu are you using for the pc. sounds to me like a power fluctuation
afireinside
06-06-2007, 05:54 PM
You can get a better OC on stock cooler. Run straight water.
It changes between 6x and 9x becasuse you have EIST or whatever on. The speed step thing.
what psu are you using for the pc. sounds to me like a power fluctuation
Silverstone 850W
You can get a better OC on stock cooler. Run straight water.
The temperatures are really bad, huh? At least it does drop below 22*C on idle... -__-
I have a suspicion that I may have a faulty TEC module. I reason thus because at some point my temperatures were far worse -- it would idle at 50 and climb steadily to 60+; I manually shutdown before it had the opportunity to climb any further; I was uninterested in whether the issue was thermal runaway or if there was a discrete endpoint. However, using my fingers to press in various places near where the wires were under the Neoprene insulation -- as if I were giving a massage to my contraption -- caused suddenly the temperature to drop drastically to the case it is now.
Concordantly (or rather sadly), it may very well be that the TEC module's wire is not well-connected to the TEC module's lead.
It changes between 6x and 9x becasuse you have EIST or whatever on. The speed step thing.
Yea many things fluctuate. Strangely, sometimes the voltage options are wrong for Vcore. If you go to the BIOS, and select the voltage for Vcore, sometimes there will be a "shifted" list... a list that starts at a lower voltage than normal, and ends with a lower voltage as well. However, in such a case, the position of the list item determines the actual value, so it appears.
Tomorrow I will see if there is any way to turn EIST off... Is EIST a good thing to have, or should it be turned off?
Holst
06-06-2007, 10:15 PM
our situation with the temp skyrocketing sounds like a bad connection to me.
This sort of high current DC wiring can get hot and catch fire easily so please check your wiring carefully. The fault is more likely to be in your own wiring extensions than on the TEC itself.
How exactly is the TEC wired up to the PSU ?
If you get the hot side of the TEC too hot then the solder will melt
I dont know enough about speedstep to help, but if you can turn it off then please do, on AMD this sort of thing killed OC stability for me. This probably explains the temp jumping as well. It shouldnt step at load unless your overheating, which you might be.
Things I would like to know to help you.
TEC voltage and PSU used.
Water temperature after 1hr at load (any OC)
Although you probababy think your big car rad is good, but its probably rubbish. I tried one (VW golf rad) and its performance was poor. The reason for this is that car rads are designed to keep the water quite a bit above ambient at high wattage and airflow. Not close to ambient at low wattage and low airflow.
To make things better put some big fans on the rad (big office fan or soemthng would be ideal)
littleowl
06-07-2007, 03:58 AM
as holst said we need some more info and a pic or 2 would be nice :D
I would say check your mount onto the cpu also check to make sure you have the tec crushed between the cold plate and water block.
The main big question is what holst asked. What is water temp after 1hr of running wcg?
Alright I'll work on getting what you guys want. Previously, I have no yet run the computer for anywhere close to 1 hour. Anytime I must leave the computer for an extended amount of time, which I choose to define as 15 minutes, I turn off that computer. Currently due to the difficulty of finding a specific molex-like connector, I am unable to allow the entire system to electronically turn on and off. This is of course, okay as long as every time I am diligent... Nobody else uses my computer; everyone else knows (they have their own computers anyway). Though this problem will eventually need to be fixed, or else there will be hell to pay at some point -- or it might be the case that I will never be able to run the computer for extended periods, which is quite annoying.
In any case, I will work on getting the information you guys need.
For a hint, I would say that without a fan, the tubes leading from the system to the radiator get warm enough to feel rather limber. The tubes, when cool, are rather stiff. They are the cheap ones from Home Depot. Only the internal tubes are the nice tubes.
Also, without a fan, the radiator gets quite warm to the touch after just a couple of minutes.
littleowl
06-07-2007, 09:53 AM
yeah I would say your getting way to hot and back lashing the tec. You need them fans soon.
aaronjb
06-07-2007, 03:02 PM
Now that you've said it already, I don't think you'll mind a little correction LOL no actually summary: you're whoring 1.5KW from your socket and you got owned by someone who used a not so good straight water system :P
I lurk here a lot and post infrequently. If this public shaming earns me a vacation, so be it.
I see you in every thread, without anything constructive to say. Do what a mature young person does: sit back, listen, and take it all in. You haven't become an expert in a matter of months, regardless of your post count.
afireinside
06-07-2007, 03:22 PM
Also, without a fan, the radiator gets quite warm to the touch after just a couple of minutes.
You're trying to run a TEC passive cooled? You'd have a hard enough time running a straight watercooled CPU passively...
yeah I would say your getting way to hot and back lashing the tec. You need them fans soon.
You're trying to run a TEC passive cooled? You'd have a hard enough time running a straight watercooled CPU passively...
Both very good points. I did notice that the temperature drops drastically when I point a "stand-up fan" at it, even at the lowest setting. At the medium setting, the temperature drops a bit more. Diminishing returns are in effect, and the highest setting doesn't seem to bring much more to the table except added noise pollution.
I lurk here a lot and post infrequently. If this public shaming earns me a vacation, so be it.
I see you in every thread, without anything constructive to say. Do what a mature young person does: sit back, listen, and take it all in. You haven't become an expert in a matter of months, regardless of your post count.
:owned:
littleowl
06-07-2007, 04:46 PM
I lurk here a lot and post infrequently. If this public shaming earns me a vacation, so be it.
I see you in every thread, without anything constructive to say. Do what a mature young person does: sit back, listen, and take it all in. You haven't become an expert in a matter of months, regardless of your post count.
That is a very good point that I keep trying to say in a nice way. :D Thank you very much aaronjb.
I am not pointing fingers at anyone but IMO you don't have a right to post how to do something unless you have done it 2 or 3 times yourself. I posted some stupid questions to nol when I was building my first tec. I just was asking them to double check myself. I also refused to help people out until I had built my second tec setup. (pic of it in gallery)
MaxxxRacer
06-07-2007, 06:47 PM
I see you in every thread, without anything constructive to say. Do what a mature young person does: sit back, listen, and take it all in. You haven't become an expert in a matter of months, regardless of your post count.
You got a point there. Although, in serials defense he does help alot of ppl.
Serial, there was no personal attack.
aaronjb
06-07-2007, 06:55 PM
Serial does seem to offer a lot of advice, but sometimes a little tact goes a long way. ;)
Sorry to take it off track, L33T. And, yeah, get some fans on that radiator. Really. :)
You got a point there. Although, in serials defense he does help alot of ppl.
Serial, there was no personal attack.
Whoaa, it's MaxxxRacer! ^__^
Serial does seem to offer a lot of advice, but sometimes a little tact goes a long way. ;)
Sorry to take it off track, L33T. And, yeah, get some fans on that radiator. Really. :)
NP =D
Yes I'll definitely look for some fans; I do hope also that the TEC unit is not busted.
serialk11r
06-07-2007, 09:49 PM
Serial does seem to offer a lot of advice, but sometimes a little tact goes a long way. ;)
Sorry to take it off track, L33T. And, yeah, get some fans on that radiator. Really. :)
sheesh that was a joke...we were just chatting on AIM and he decided to post that here, meaning he is okay with it :P
Apology accepted (the implied apology that is). I would also like to make it clear for the LAST TIME that I don't give a f***ing crap about my post count if that's what you think.
And it seems not very many people around here like me, I think I'm really going to disappear for a while (by disappear I mean only make occasional posts when I feel like I really need to). You've succeeded in convincing me to make this decision. As ironic as it is, thank you, I think my grades are going to go up next year :) (not that they're low but...)
littleowl
06-07-2007, 10:50 PM
I am with you L33T ...
Holst
06-07-2007, 11:15 PM
Alright I'll work on getting what you guys want. Previously, I have no yet run the computer for anywhere close to 1 hour. Anytime I must leave the computer for an extended amount of time, which I choose to define as 15 minutes, I turn off that computer. Currently due to the difficulty of finding a specific molex-like connector, I am unable to allow the entire system to electronically turn on and off. This is of course, okay as long as every time I am diligent... Nobody else uses my computer; everyone else knows (they have their own computers anyway). Though this problem will eventually need to be fixed, or else there will be hell to pay at some point -- or it might be the case that I will never be able to run the computer for extended periods, which is quite annoying.
In any case, I will work on getting the information you guys need.
For a hint, I would say that without a fan, the tubes leading from the system to the radiator get warm enough to feel rather limber. The tubes, when cool, are rather stiff. They are the cheap ones from Home Depot. Only the internal tubes are the nice tubes.
Also, without a fan, the radiator gets quite warm to the touch after just a couple of minutes.
I think this is enough info for the moment.
If the rad is getting warm to the touch and the tubes are going soft then its safe to say that your water temps are MUCH MUCH MUCH too high to make your system run propperly.
Water temp in a TEC system is almost directly proportional to CPU temp.
If your water temp is 50*c+ then there is a potential 25* to be gained by making the water cooler.
25* water temp is acheivable with the right rads and fans, this will drop your load temps by allot and get you a much better OC.
Firstly you should try more fans or bigger fand on the rad, but I think to make it work propperly you will need to take a different approach entirely.
Either replace the big rad with a propper watercooling rad with some propper fans (or run both rads, more is allways better)
You might also want to consider making a bong cooler, this is the most efficent way to keep water cool in a TEC system, you should be able to get water temps close to ambient (possibly even slightly below!) and you will get a big performance gain.
I dont think you will be able to make the TEC system "quiet"
Also what do you mean by "molex connextor"
I have repeatedly said that molex connectors are not suitable for TEC connections and are a FIRE RISK!!
Please solder the TEC direcly or use an alternative connector.
And it seems not very many people around here like me
Not true at all.
Sometimes you post things that are a bit wrong but I think you help allot more than you hinder.
Please keep posting, but try and post the right info and search if your not sure :)
Please try and keep this thread on topic guys, I hate having to wade through a page of spam to get to something interesting.
littleowl
06-08-2007, 11:48 AM
I think this is enough info for the moment.
If the rad is getting warm to the touch and the tubes are going soft then its safe to say that your water temps are MUCH MUCH MUCH too high to make your system run propperly.
Water temp in a TEC system is almost directly proportional to CPU temp.
If your water temp is 50*c+ then there is a potential 25* to be gained by making the water cooler.
25* water temp is acheivable with the right rads and fans, this will drop your load temps by allot and get you a much better OC.
Firstly you should try more fans or bigger fand on the rad, but I think to make it work propperly you will need to take a different approach entirely.
Either replace the big rad with a propper watercooling rad with some propper fans (or run both rads, more is allways better)
You might also want to consider making a bong cooler, this is the most efficent way to keep water cool in a TEC system, you should be able to get water temps close to ambient (possibly even slightly below!) and you will get a big performance gain.
I dont think you will be able to make the TEC system "quiet"
Also what do you mean by "molex connextor"
I have repeatedly said that molex connectors are not suitable for TEC connections and are a FIRE RISK!!
Please solder the TEC direcly or use an alternative connector.
Not true at all.
Sometimes you post things that are a bit wrong but I think you help allot more than you hinder.
Please keep posting, but try and post the right info and search if your not sure :)
Please try and keep this thread on topic guys, I hate having to wade through a page of spam to get to something interesting.
I agree with that!!!!!!!
I think this is enough info for the moment.
If the rad is getting warm to the touch and the tubes are going soft then its safe to say that your water temps are MUCH MUCH MUCH too high to make your system run propperly.
Water temp in a TEC system is almost directly proportional to CPU temp.
If your water temp is 50*c+ then there is a potential 25* to be gained by making the water cooler.
25* water temp is acheivable with the right rads and fans, this will drop your load temps by allot and get you a much better OC.
Firstly you should try more fans or bigger fand on the rad, but I think to make it work propperly you will need to take a different approach entirely.
Either replace the big rad with a propper watercooling rad with some propper fans (or run both rads, more is allways better)
You might also want to consider making a bong cooler, this is the most efficent way to keep water cool in a TEC system, you should be able to get water temps close to ambient (possibly even slightly below!) and you will get a big performance gain.
I dont think you will be able to make the TEC system "quiet"
Also what do you mean by "molex connextor"
I have repeatedly said that molex connectors are not suitable for TEC connections and are a FIRE RISK!!
Please solder the TEC direcly or use an alternative connector.
No the TEC is not connected via molex. Molexes do not refer just to the +5/GND/GND/+12 plugs commonly found on computer PSU's, but also on other shapes and configurations as well. This particular PSU has two places to make connections. They are both 12-pin connections. They allow, amongst other things, the ability for the PSU to be electronically turned on and off. To turn the PSU on when the computer is on, two pins must be shorted to activate ON => ON logic (by default: OFF => ON logic). Then there are two more pins: only when +5V is sent through pin J1-6, the PSU will turn on.
Concordantly, the method is to use an existing molex from the computer PSU and extend wires from +5 and GND. Connect those to J1-6 and the return pin, respectively.
Now the TEC PSU will only turn on when the +5V line is active. Since the +5V line is only active when the motherboard is turned on, this is a good way to turn on and off the TEC PSU without implementing a separate relay system.
Though given Zenjirou's thread back a long time ago, such a system in general should work. At least his system worked. On the other hand, times have changed: he wasn't exactly running TEC's; nor was he using an 8800GTX.
Holst
06-08-2007, 11:03 PM
As long as your not using molex on the TEC power lines its ok :)
What TEC PSU are you using, as I dont quite understand your above post.
I just read through the thread on the car rad by Zenjirou, if you copy what he was doing with fans then you will get much better performance.
Do you have any idea of the folwrate your getting with this rad?
To test flowrate do the following.
Fill a big bucket with water.
Put the pump inlet tube into the bucket, same with outlet (use longer tubes to keep the water from PC.
Turn on pump and bleed air.
Then with the system running put the outlet tube into a 2L coke bottle and time how long it takes to fill up.
Do this a few times for an average.
You can then work out your litres a second flowrate.
If you want to find out how much difference the huge rad makes then try the same test with the rad out of the system.
I dont think flowrate will be affected much by the rad personally, but it might be interesting to find out.
Yep. The PSU is an ASTEC VS1-C5-04, which is a version built specifically for a certain buyer (there are actually many extras of this particular build/configuration). It provides a slightly adjustable 24V output with 37.5A draw on the DC-DC, 1500W theoretical on the AC-DC. Since there isn't space to attach another DC-DC, we won't worry about the extra few hundred watts of power available. Basically, this is 24V 37.5A.
Very powerful.
Very loud.
Make that extremely loud.
Turbojet loud.
Delta Fans loud.
Alright over the coarse of the next few days I'm going to be taking various SS-OP temperature readings while varying the parameters:
o IDLE vs. LOAD
o Various voltages on CPU/Mem/&c
o Various FSB settings
I'll post data here
THE JEW (RaVeN)
06-13-2007, 09:34 PM
I'd be interested in your card rad results as I have an old Ford Aspire radiator I keep glancing at for a computer simply because it's sitting there doing nothing.
Fr3ak
06-25-2007, 04:51 AM
I would like to see some more numbers after testing :)
But temps posted at the first page are not that great considering you are using a TEC. I have a E6600 ES at 3600Mhz 1.45V, a NB @ 1.55V and 2 Nvidia 7300GTs at 840/850Mhz in SLi @ 1.45V with a triple rad using 3 Yate Loons at 7V with a roomtemp of 24C and I have 57C max using coretemp under orthos dual load. WCG is not stressing the cores as much, so temps are even better running WCG.
Well my IHS is lapped and I am using liquid metal TIM. Unfortunately I have no idea what my water temp is. I have a temp probe, but nothing to read it with :/
Your temps would be somewhat interesting nonetheless, cos I am thinkin about getting a TEC too, but I have my doubts that it will be strong enough to cool a C2D, apart from the fact it is not that energy efficient ;)
goranm
07-03-2007, 12:04 AM
Question: how do I tell what the temperature is if it keeps jumping around? It jumps from 39 to 47 to 45 and keeps repeating... I simply took the highest value to be safe.
I think your cold plate is possible cause of temperature jumps. How big it is?
With 2 pelts cold plate must be ~10mm tick (so, at the same time massive) to be able to accumulate enough heat.