lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 ^ yes also use Gen 1 or 2 PCIE and 127ish bclk helps to go colder. I think you need a bencher to help you! OK. I'll be benching again later this week (thursday or friday). Let's chat before. Quote Link to comment Share on other sites More sharing options...
xxbassplayerxx Posted February 1, 2016 Share Posted February 1, 2016 Definitely need to tune the voltages manually as well as use better paste. You need to be using either Thermal Grizzly Kryonaut or Gelid Extreme for paste. This. As I mentioned above, some of the IC's around the socket sometimes have cold bugs of their own. It may benefit you to at least put some sort of gasket around the head to prevent the cold from spreading as much. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 This. As I mentioned above, some of the IC's around the socket sometimes have cold bugs of their own. It may benefit you to at least put some sort of gasket around the head to prevent the cold from spreading as much. Thanks. The design of the head has changed, it is pretty much fully insulated now. I'll be posting a video from set-up to benching in a bit. I'm pretty sure voltage and TIM is the issue, and I was duped by NewEgg into buying a non-extreme Asrock Extreme4, thinking that it was designed for cold. That's 3 strikes. Any suggestions on an x99 MOBO I can ebay for +-$300? I found the batch number on the box. It's J513D142 Quote Link to comment Share on other sites More sharing options...
nickolp1974 Posted February 1, 2016 Share Posted February 1, 2016 (edited) Gigabyte X99 SOC champion, not in stock at newegg but maybe OCUK ship to the states https://www.overclockers.co.uk/gigabyte-x99-soc-champion-intel-x99-socket-2011-ddr4-eatx-motherboard-mb-520-gi.html Edited February 1, 2016 by nickolp1974 Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Gigabyte X99 SOC champion, not in stock at newegg but maybe OCUK ship to the states https://www.overclockers.co.uk/gigabyte-x99-soc-champion-intel-x99-socket-2011-ddr4-eatx-motherboard-mb-520-gi.html Found a few on eBay - local. Think its a bad idea to buy used? Quote Link to comment Share on other sites More sharing options...
xxbassplayerxx Posted February 1, 2016 Share Posted February 1, 2016 Buying used should be fine. That and the Rampage V Extreme are pretty much the two best X99 boards. It should treat you nicely. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Buying used should be fine. That and the Rampage V Extreme are pretty much the two best X99 boards. It should treat you nicely. Awesome! Thanks. I have a lead on an RVE in my price range. Will be picking it up! Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 1, 2016 Share Posted February 1, 2016 (edited) I highly doubt that the temp probe is in the block based on temps on videos and your own statement from another site stating that the temp probe was at the valve. This is 98% of your fail, the controller is responding to the valve temp not the CPU IHS. I bet your CPU overheating based on controller failing to respond to load and you are banking non existent capacity of the block unloaded to hold a load. Your AMD results on LN2 is also huge fail, thus proving that the location of the temp probe is at the valve. Move the temp probe and try again. You will find out that you will need to remove the controller completely to maintain the perfect temp. I already know that you will overshoot and hit cold bug with the probe at the block in your current design unless you have a bug free chip or move the valve to the back of the block. You might be able to adjust your in and out points to get better functionality but as of right now, no one is impressed. Been here, done this 10 years ago. SVID on that board works great, all you needed to set was the multiple. No professional overclocker needed. Post a drawing of your block design. If you used my block that is ok, it is free to use. I can do 5Ghz XTU on chilled AOI with positive temps. (AOI is the cheapo "all in one" water cooler) to give you an example of how fail your cooler design is, my pressure fed LN2 loop hit 6.2Ghz on 5960X Edited February 1, 2016 by FUGGER Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 1, 2016 Share Posted February 1, 2016 To better clarify the overshoot. The distance from the valve to the block is excessive and that distance will have too much latency for the stupid controller to back off soon enough (your in and out points set on controller). I do believe you have a block heater based on the power leads going towards your block, amuse us... How many watts are these heaters located at the suction end of your block? Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 To better clarify the overshoot. The distance from the valve to the block is excessive and that distance will have too much latency for the stupid controller to back off soon enough (your in and out points set on controller). I do believe you have a block heater based on the power leads going towards your block, amuse us... How many watts are these heaters located at the suction end of your block? Wow. I thought we'd been through this. The temp probe is in the head. The head is not overheating and the OCC works so well that it holds the temp at whatever setting within 1 degree of setpoint and within 2.5 degrees under load. I really don't consider any of this a fail at all as I have just started doing benchmarking with these benchmarks this weekend and I am not a seasoned overclocker by any means. I am going to post up a video of the process from set up to benching here in a few. I respect your standing as a top level benchmarking competitor. Maybe you are mistaking me for someone else as it would be really out of character for me to say that my temp probes are in places other than they actually are. The system wouldn't work at all if they were. BTW this is not a closed loop system. I remember you actually complementing this system on your site xtremesystems.org after a second look and then telling me that I'm welcome on the site. Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 1, 2016 Share Posted February 1, 2016 Looking at your XTU results you are about to hit positive temps at 5040Mhz -2c. http://hwbot.org/submission/3111747_ Try some Gelid Extreme or Thermal Grizzly and you might be able to squeeze a few more Mhz from that chip. RIVE is the best board, good luck and 2v + input voltage might be needed on that chip Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 1, 2016 Share Posted February 1, 2016 (edited) You are claiming -128c and your XTU shows -2c and your board tops out at -5c based on the many XTU results. You are getting a > 120c swing in temps. I am assuming based on temp probe location and poor choice of thermal paste. Paste will get you a few more Mhz but as well all know the TIM under the IHS is solder so that is not an issue at these low speeds. You are only at 5Ghz... Let me quote you again: "the temperature readout on the display of the OCC is from a thermocouple placed at the base of the OCC" You are very welcome at XS, and we would love to help you work out the issues but your claims are not accurate at all. Post a pic showing your temp leads going to the block and not to the OCC please and I will apologies. Until then, I will still be cordial and offer my 2 cents. For the LN2 guys here, please take a minute to visit: http://www.lnlcooling.com/ You have the ability to build nice stuff no doubt but your claims are failing short, based on many samples of 5960X your cold bug is somewhere around -120c unloaded, you need to keep the cpu close to that temp to compete with pour method and a 120C swing is not even close. Edited February 1, 2016 by FUGGER Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Looking at your XTU results you are about to hit positive temps at 5040Mhz -2c. http://hwbot.org/submission/3111747_ Try some Gelid Extreme or Thermal Grizzly and you might be able to squeeze a few more Mhz from that chip. RIVE is the best board, good luck and 2v + input voltage might be needed on that chip Getting one and some grizzly this week. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 You are claiming -128c and your XTU shows -2c and your board tops out at -5c based on the many XTU results. You are getting a > 120c swing in temps. I am assuming based on temp probe location and poor choice of thermal paste. Paste will get you a few more Mhz but as well all know the TIM under the IHS is solder so that is not an issue at these low speeds. You are only at 5Ghz... Let me quote you again: "the temperature readout on the display of the OCC is from a thermocouple placed at the base of the OCC" You are very welcome at XS, and we would love to help you work out the issues but your claims are not accurate at all. Post a pic showing your temp leads going to the block and not to the OCC please and I will apologies. Until then, I will still be cordial and offer my 2 cents. For the LN2 guys here, please take a minute to visit: http://www.lnlcooling.com/ There are thermocouples at the base of the OCC (the surface of the head) as close as possible to the CPU surface. Not posting up photos of the inside of the cooling head. This is not a closed loop system. There are no MOBO connected temp probes and from what I've heard the CPU internal probe does not read accurate sub-zero temps. Time will tell whether "claims" made are legit. As I recall, you have seen the videos of the OCC running at -180C on the FX8350 - so yes it will maintain any sub-zero temp LN2 is capable of achieving. Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 1, 2016 Share Posted February 1, 2016 Yes, the on die cpu temp is 100% correct at -2c in XTU, beyond -5c your board fails. Post a pic where the temp probe enters the line set please. Also, what TIM did you use for this so we can better guesstimate. And you never answered, how many watts are your heaters?? Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Yes, the on die cpu temp is 100% correct at -2c in XTU, beyond -5c your board fails. Post a pic where the temp probe enters the line set please. Also, what TIM did you use for this so we can better guesstimate. The "temp probes" are internal thermalcouples. They are in the cooling head. Wouldn't be able to photo without disassembling. With this system, you don't insert a probe as with a pot. The probes are fixed internally. But you have me convinced on the TIM now. It is Arctic Alumina by arctic silver. Inexpensive in bulk, apparently ineffective at sub-zero. You'll see it in the info video I'm about to post on youtube. You'll see the head mounted and through to the benches. Thank you very much for the info on XTU having accurate temps. That really tells what is going on. I do have a little arctic silver on hand. May try that tonight. I'm ordering thermal grizzly and a RVE friday. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 (edited) Can someone clarify? At -5C boards fail? Why do we see - all the time pots at <-140 and certainly cooler than -120? I'm not trying to be a bugger - just want to understand. I've had Mobo's way down -0 temps and haven't had that experience. Here is a pic of the case with the OCC mounted and the rig running smooth like butta, while the case is a popcicle. Edited February 1, 2016 by lnlcooling Added photo Quote Link to comment Share on other sites More sharing options...
xxbassplayerxx Posted February 1, 2016 Share Posted February 1, 2016 (edited) Keep in mind that XTU will not read lower than -5°C. However, if it reads higher than that, you will know that your temperatures truly have reached those levels. EDIT: No, boards don't necessarily fail at -5°C. However, the fact that the chip likely started the bench somewhere around -100°C and peaked at -2°C means that the temperature swing was absolutely massive. This rapid change in temperature can definitely lead to instability. Edited February 1, 2016 by xxbassplayerxx Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Keep in mind that XTU will not read lower than -5°C. However, if it reads higher than that, you will know that your temperatures truly have reached those levels. AAAAHHHH!!!! Master teacher! So the thermal paste completely failed. Quote Link to comment Share on other sites More sharing options...
xxbassplayerxx Posted February 1, 2016 Share Posted February 1, 2016 Sounds like it, yes. Hopefully it's only the TIM that failed and not a lack of mass in the head of the unit. When using a phase change unit that gets overloaded, the temperatures quickly shoot to positive numbers, similar to what you're seeing here. That is usually a result of the unit not being charged for the specific load or not having enough mass in the head. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 Interesting. The temps are being read right at the surface of the cooling head so if the cooling head is not making contact with the CPU then it definitely would read a temperature different than what the actual CPU temp is if the thermal paste failed. The entire inside of the case is being cooled by the gaseous nitrogen, likely somewhere just below zero overall, so the cooling on the CPU being at -2 with no load but no thermal thransfer connection with the headvwould make sense. The head is extremely accurate at surface temperature sensing. But if there is no surface temperature connection of course it would just read temperature of the head with no load. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 1, 2016 Author Share Posted February 1, 2016 We've been developing this machine for over two years, and we know how well it holds temps especially after extensive testing. Holding the -100 or 120 temps during the benches just seemed normal :-) Quote Link to comment Share on other sites More sharing options...
Strong Island Posted February 2, 2016 Share Posted February 2, 2016 We've been developing this machine for over two years, and we know how well it holds temps especially after extensive testing. Holding the -100 or 120 temps during the benches just seemed normal :-) not trying to be a jerk but I am a little confused, how do you know how well it holds a load if your not even sure if the head was making proper contact. During the testing stages did someone use correct paste or is it possible it was never used? Instead of testing it yourself when your unsure of what paste to even use, you should have given it a well known person over here for some real testing. That would be the best marketing tool you could have. I'm sure you have already invested a ton of money so to loan out a unit for testing and marketing would probably be worth the cost. Because I think right now everyone is a little confused with the testing that has gone on so far. Quote Link to comment Share on other sites More sharing options...
lnlcooling Posted February 2, 2016 Author Share Posted February 2, 2016 not trying to be a jerk but I am a little confused, how do you know how well it holds a load if your not even sure if the head was making proper contact. During the testing stages did someone use correct paste or is it possible it was never used? Instead of testing it yourself when your unsure of what paste to even use, you should have given it a well known person over here for some real testing. That would be the best marketing tool you could have. I'm sure you have already invested a ton of money so to loan out a unit for testing and marketing would probably be worth the cost. Because I think right now everyone is a little confused with the testing that has gone on so far. We already have a unit in the hands of a pro whose planned benching was delayed. We'll have a couple other well-recognized pros benching too within a couple weeks. Arctic Silver was recommended in another forum. I used it. No one was confused until today when Fugger confused the OCC system for a cascade or something with temp probes and "suction lines." I have already had the same exchange with him and, in his own words after railing on my system then reviewing what it was he said: "My apologies if I come of harsh, I am the owner and I also a builder and often skeptical of wild claims. We had a guy come in, know it all and tried to fake us with a vortex engine. I do research as you have seen, just seemed fishy that you had different names everywhere. [i have one other name "phasechanger" from a couple of years ago when R&D began... still on overclock.net] We know your knowledge of OC is limited, we know not to knock you on that and support you on any questions you might have. My dissection of your unit proves it is very real and you built a environmental case from a pc case..." BTW I have seen lots of posts by pros complaining of having lost the TIM contact and having subsequent issues. 1 XTU out of 12 in two sessions is not bad. That's during about twelve hours of chilling. Lots of folks have been following the past few months of development since the Patent Pending and revealing of the project. I have known Splave for a couple of years and he actually was very helpful even a couple of years ago in the beginning with answering some CPU dimension questions and wattage questions. Kal El introduced us back in '13. Of course there has been skepticism and there should be. But when really reviewed and tested (especially by pros which is right around the corner) it is plain to see that there are no mirrors and no smoke, save some LN2 fog. I liken it to when horses and wagons were replace by motor vehicles (although I know this is nothing near it) but all of the horse trade folks said it was impossible to replace the effeciency of the well known method of equine transportation. Here is a video of the OCC from this weekend with setup to benching with three of the benches I submitted to HWBOT this weekend: Quote Link to comment Share on other sites More sharing options...
FUGGER Posted February 2, 2016 Share Posted February 2, 2016 It appears you are using my block, I bet you are feeding LN2 in the suction port. http://under-the-ice.com/product_info.php?cPath=25&products_id=126 I built pressure fed LN2 setups a decade ago, I used them in exhibitions with lots of media coverage. http://www.washingtonpost.com/wp-dyn/content/article/2007/06/15/AR2007061501296.html My Kryotek has an Omega controller, also precise to 2c unloaded. Even with AS, you should not have ~120C temp delta. No complaints in your mounting, that should be fine since it is based on my retention design. I know Kal El, had dinner with him and his wife, met his family and been to his house in Hawaii, been to his parents house, hung out with his dad, checked out his awesome collections, and that doesn't make one bit of difference. Ask Kal El about me. I don't really know Splave, I am not sure he even exists. "1 XTU out of 12 in two sessions is not bad. That's during about twelve hours of chilling." From your screens you have really good chip based on vcore @ 5Ghz so are battling a lack of capacity as you breach 5Ghz Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.