Unsolved
This post is more than 5 years old
19 Posts
1
22690
Alienware R4 Cooling problem
Hello,
Dell just sent me a new CPU cooler, which managed to drop the temperature by 20 degrees C.
However, a component identified as TMPIN0 is strangely very hot. I don't know what it is, and I'm afraid I'm damaging it. Here's a view :
Then I ran ePSA, and I got an error on the CPU fan, even after setting up the whole new stock cooler :
After opening desktop, I got this error (got it only once so far) :
System.TypeInitializationException: Une exception a été levée par l'initialiseur de type pour 'AlienLabs.ThermalControls.Controller.Classes.ThermalControlsTaskbarIcon'. ---> System.Runtime.InteropServices.COMException: L’appel a été annulé par le filtre de messages. (Exception de HRESULT : 0x80010002 (RPC_E_CALL_CANCELED))
à System.Management.ManagementScope.Initialize()
à System.Management.ManagementEventWatcher.Initialize()
à System.Management.ManagementEventWatcher.Start()
à AlienLabs.ThermalControls.DeviceDiscovery.Classes.DeviceWatcherClass.Start()
à AlienLabs.ThermalControls.DeviceDiscovery.ObjectFactory.NewThermalControlsDeviceDiscovery()
à AlienLabs.ThermalControls.Domain.ObjectFactory.get_DiscoveryService()
à AlienLabs.ThermalControls.Domain.Profiles.Classes.ProfileProcessorClass..ctor()
à AlienLabs.ThermalControls.Domain.Profiles.Classes.Factories.ProfileRepositoryFactory.get_ProfileRepositoryforController()
à AlienLabs.ThermalControls.Controller.Classes.ThermalControlsTaskbarIcon..cctor()
--- Fin de la trace de la pile d'exception interne ---
à AlienLabs.ThermalControls.Controller.Main..ctor()
à AlienLabs.ThermalControls.Controller.App.AppStartup(Object sender, StartupEventArgs args)
à System.Windows.Application.<.ctor>b__1(Object unused)
à System.Windows.Threading.ExceptionWrapper.InternalRealCall(Delegate callback, Object args, Int32 numArgs)
à MS.Internal.Threading.ExceptionFilterHelper.TryCatchWhen(Object source, Delegate method, Object args, Int32 numArgs, Delegate catchHandler)
1-Is the heat currently damaging the component, and what component is it ? In any way, what can be done to lower the heat ? I can afford to pay for another small part.
2-How can I fix the ePSA and the desktop error ?
3-Do you find the CPU temperature is too high ? Currently my room temperature is 18 degrees C, but in summer it can easily get to 38 C in the room, so I add 20 degrees and get to 80-100 degrees C, which worries me ...
Thank you for any answer.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 2nd, 2016 11:00
See this old post:
http://en.community.dell.com/owners-club/alienware/f/3746/p/19518281/20419617#20419617
1. If HW-Monitor reading is accurate, I think TMPIN0 might be the temp at cpu-socket on MB.
2. You might want to double-check your work, the Asetek cooler, etc.
3. I only know the temps for my Aurora-R1 system. If you run HW-Monitor again, I like to do it like this:
Value (Current): Gaming
Min: Idle desktop
Max: Prime95 (or OCCT)
dupre_7
19 Posts
0
April 2nd, 2016 15:00
Thank you for your answer.
If it is really the cpu socket, is this temperature a concern ?
The ePSA error happened before I changed anything, I thought the new cooler would solve the problem but it did not. Temperature dropped however.
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 2nd, 2016 16:00
Tesla messaged me, asked if I could run tests here on my baby ALX, an R4, as a favor for you.
1st, as a metric, I don't know what HWMon thinks TMPIN0 relates to; my TMPIN0 at time of snip was 118F, now it's 109F; this is an idle temp-reading & it isn't hot enuff for me to reach in, feel around, & understand what or where the reading comes from, one spot relative to another. I suppose if someone grounded themselves to the chassis, poked around, a 180F hotspot on the mthrbrd like yours has (or may have) might be sensed by a finger & feel very hot. The fact I have a warm spot on the mthrbrd 30degrees hotter than TMPIN 1/2 (84F vs 118F), I might speculate that if it isn't the CPU-area, it is the southbridge cooler or the top VRM cooler. Maybe verify the top VRM fan is spinning & in fact plugged in; I assume your lower SB fan works, since it's part of ePSA fan testing. Open the side panel, peak in & watch if both of the mini-fans spin up. Your HWMon snip's labeled with the caveat: 'with a video game launched'. One wonders what the idle temps report as, since my 118F max isn't a game play temp, nor will I stress test to heat things up in there. I wouldn't be so qwik to call 80c dangerously hot.
Your cpu temps look ok if that's during gameplay. 90c-100c would be very troubling. Here's a post with basic tips: how to get my Alienware Aurora r3 to run cooler
Ok, radiator/system 'cpu' fan. I'm using a (qwieter) 25mm Everflow. I own two of them, one is new, I used it for testing today also. I own a stock 38mm 1.4amp Nidec too. I did three separate fan tests. The stock 38mil Nidec passed with these numbers:
If you've a 1.5amp Delta fan instead, we might assume similar readouts are reqwired for a pass. To view these numbers, click 'Results' tab before cpu fan test (or after a fail > & retest > watch results happen in real-time) - is it trying to push the fan to some max RPM (3500-4000) & fan can not respond to it? Does it max out @ some particular rpm & can't go above it?
I was forced to do a 38mil fan test because my 25mil Everflows failed ePSA, like yours did:
Hmmm, my brand new fan 'failed' also; It seems ePSA wishes to force the 25mils to 3500RPM but they do not go above 2600rpm by design; it is possible ePSA has a pre-conceived notion that all rad-fans must be capable of 3500RPM (or more). I call bull*** that both my 25mil fans failed, since the spare is the 1st time I've used it. Since I can induce a fail with perfectly good fans, either you're not using a stock Delta/Nidec - or - maybe buying/trying a new fan is a worthy investment (in order to 'possibly' pass test); if cpu temps otherwise seem ok, maybe ignore PSA FAN FAIL. If Dell sent you a new stock kooler, maybe it came with a fan too? - if so - swap fans out, retest?
Lightstrip boards can go bad, sometimes the far right connector isn't fully seated. Inspect & retest?
*note: all radiator fans plug into the top light strip Sys_Fan header > not the Cpu_Fan header
*note: before I hit 'post', my side panel is off & TMPIN0 is 96F:
Q1-Is the heat currently damaging the component, and what component is it ? In any way, what can be done to lower the heat ? I can afford to pay for another small part.
A1: Unknown. Wait for further help & follow cooling post to make it chilly in there
2-How can I fix the ePSA and the desktop error ?
A2: try known-good stock fan or ignore error. Desktop error may go away after a reboot. If not, uninstall CmndCntr & reinstall &/or uninstall all .netframeworks through control panel (uninstall) & reinstall
3-Do you find the CPU temperature is too high ? Currently my room temperature is 18 degrees C, but in summer it can easily get to 38 C in the room, so I add 20 degrees and get to 80-100 degrees C, which worries me ...
A3: not too high during gameplay, 90-100c would be alarming. After a good bit of testing, Falcon Northwest swears by the fact a CPU can (& should) run its best at 85c all day long. A very hot cpu can be a good thing, so they tell us. Not letting it get above 85c is the #1 goal.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 2nd, 2016 19:00
Thanks Cass-Ole for running those test on a like-machine and posting all that.
Good to know that a stock Aurora-R4 will pass the tests.
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 2nd, 2016 19:00
When you replaced the Asetek cooler, did you connect the (4-pin) fan to the proper header on the "Top Lighting Board"?
According to my notes it should ne plugged into "Sys_Fan":
"Top Lighting Board" Connections:
CPU_Pump (Asetek Cooler) pump
Sys_Fan (Asetek Cooler) radiator or main system fan
CPU_Fan - No connection
Mem_Fan (small case fan in top for memory)
Arrow - "Rear of case" external white lighting with switch (to help view external connections)
... this is for Aurora-R1
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 2nd, 2016 21:00
Bing/google, 'what is TMPIN0'; general consensus is it reports the cpu socket temp, what we would know as a sensor on the mthrbrd, not a sensor inside the cpu (which reads core temps). I use Intel XTU for all core temps; I delight in using an Intel program to monitor my Intel Ivy-E, & trust it implicitly.
Found word of advice: "I use compressed air for finding motherboard sensors, just blow it on the suspect area and see if the temp drops". To me, this could be as simple as borrowing a mattress air pump. Regardless, HW Monitor appears to be infamous for reporting TMPIN0 at 115c-120c, preposterous. I wouldn't lose sleep over this. If the socket's marginally higher than the package temp, so what. The cooling advice given prior should suffice.
dupre_7
19 Posts
0
April 3rd, 2016 03:00
Thank you very much.
TMPIN0 might be the VRM cooler then. Still not sure. I'm sure it's somewhere on the motherboard, not in the CPU. 80 degrees C worried me because the max CPU temperature for optimal use is only 67 degrees C.
I don't know which fans I have. All I can say is they are all stock coolers. Some are from "Nidec". I already replaced the cpu liquid cooling system and as before, there is an ePSA error. Always fail. Perhaps it is a bug from the application ?
I will try removing and re connecting the power connector. Not sure if I should plug it into the top light strip Sys_Fan header and not the Cpu_Fan header, or not in Sys_Fan but instead following this :
"Top Lighting Board" Connections:
CPU_Pump (Asetek Cooler) pump
Sys_Fan (Asetek Cooler) radiator or main system fan
CPU_Fan - No connection
Mem_Fan (small case fan in top for memory)
Arrow - "Rear of case" external white lighting with switch (to help view external connections)
Maybe a fan connected in CPU_Fan instead of Sys_Fan triggers the ePSA error and creates heating problem.
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 3rd, 2016 05:00
a two-part photo of an unboxing, showing us the factory as-shipped connexions on delivery
master i/o board relays Sys_Fan signal to/from mthrbrd as 'cpu fan' signal, over far right connexion. How to Replace Aurora Cooling System > *note Remove/Replace video is wrong: use Sys_Fan.
100c is the shutdown value (fail-safe) > 80c plenty of thermal headroom > your package temp is 67c
Confirm for yourself that the new cooler came with pre-applied thermal grease / paste on the cooler coldplate, present there before you cinched the retention bolts down (if not, Gelid, ICD, Arctic 5, use a decent thermal paste).
Tesla1856
8 Wizard
8 Wizard
•
17K Posts
0
April 3rd, 2016 13:00
1. No, Cass-Ole proved that ePSA diags app works fine (with stock working Asetek cooler and MIO-Board hardware).
2. Yes, that's very likely ... because it would be connected incorrectly.
The Asetek-Cooler's radiator fan should be connected the "SYS_FAN" header on Top-Lighting-Board. I really can't type it any clearer than that. Also, see the pic above that Cass-Ole was nice enough to post.
dupre_7
19 Posts
0
April 4th, 2016 11:00
The thermal paste was there, the CPU pump was plugged in the CPU PUMP connexion, the cpu cooler fan was connected to Sys_Fan as expected, so this is not the cause of the overheating component and ePSA error.
Should be noted that Cpu_Fan (expected) but also Mem_fan have nothing plugged in.
That's strange that the video shows something wrong...could it be in fact right and the reason why there is an error and an overheat ?
I really don't know what to do more now. The customer service can't even make sure where this temperature TMPIN0 is coming from :P
dupre_7
19 Posts
0
April 12th, 2016 12:00
TMPIN0 is reaching 90 C / 190 F all the time, any idea what can I do ?
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 12th, 2016 23:00
all the time? u mean when gaming/load on CPU, or at idle.
I posted a link to ways to improve case cooling, do you actually think that whimpy 'memory' fan up top is enough? At a minimum, simply dangling a small fan at that area using bread-ties to see what happens. You can entertain the idea of buying the R4 01YGW premium cooler or upgrade to Corsair H80i or similar 38mm radiator with improved copper base. I think we're all aware that the stock 25mm cooler-rad with aluminum base was Alienware's way of dishing out minimum cooling for Sandy/Ivy-E to keep their cost low & profit high.
Also a suggestion to use an air pump to pinpoint where TMPIN0 is (air shoud drop temp). If it represents the top VRM heatsink you might remove/repaste it with Arctic Silver or better. If your CPU cooler is in fact 'correctly installed' & this temp represents the heat in the CPU socket assembly area, I'm not sure what can be done other than to inprove overall case cooling as listed above. We also have to make room for the possibility it is an abberant temp reading, therefore you might look into different software that has access to this same onbard temp sensor, & compare. HWInfo64/SpeedFan, I wouldn't know what software has access, I do not use these sorts of things personally.
If you aren't willing to improve case cooling then just ride it out. If the board dies someday from it, FPV4P & 7JNH0 are the part #'s, try eBay. Your CPU core temp deserves the spotlight, improve that & your related TMP0 reading 'should' drop also.
dupre_7
19 Posts
0
April 13th, 2016 04:00
It is on load but my computer is always on load.
I don't know if the small memory fan is doing anything. The CPU temp is fine, 71 degrees C all the time. With a Noctua thermal paste (to replace the default thermal paste), and a new CPU cooler Alienware sent me, I don't think the CPU is the problem. There is no way to get the temperature lower. I thought 65-70 degrees C with the CPU at 100% with no sound that you can hear is the work of a good cooling system.
I really don't know what this TMPIN0 is. If I knew I could find a way to cool it.
Does the top VRM use thermal paste ? I didn't know that. If that's the case, the thermal paste that remains is certainly not of any uses to help cooling.
The motherboard is expensive and I would like to avoid replacing it at all cost. If I have to replace it, I would be willing to replace it with a better motherboard that will guarantee future component compatibility going forward.
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 13th, 2016 05:00
true, an optimum price for R4 motherboard is $100-$200 eBay, whereas the best X79 board to try is the Rampage IV Gene which is $300+, so the notion of expensive is relative I guess. The VRM has a heatsink so yes it has paste under it. If it isn't TMP0 then repasting would add an element of risk for no known current benefit (if the VRM's happy as pasted now then it's happy)
This is your photo & I presume they replaced kooler with identical unit. If so, that is the base kooler
aluminum koldplate, the choice of aluminum is a kost kut
kopper on the other hand is a superior thermal material; absorbs heat faster & sheds it faster
The size difference between radiators is also a factor, The 25mm base kooler holds less fluid & has less metal surface area (to wick away heat from the 'water'). The fatter 38mm koolers hold more fluid (50mm kooler may hold double the fluid of a 25mm kooler) & both have far more surface area, where we expect the larger volume of fluid & larger surface area of the radiator themselves to become superior sinks of heat when compared to the base model. We'd expect the cpu temp to drop, by how many degrees comes on a one-by-one basis, but the idea of an upgraded kooler's obvious: to drop the cpu temp; the kooler a cpu runs the slower & qwieter the fan should run, so if kool & qwiet is the name of the game, keep your fan & trade up to a better unit; the base Aurora kooler is as I said it was; I know it for what it is, a basic entry-level non-high-performance part. I doubt you would hear 'pump' noise coming from these other types of non-base stock koolers, therefore you can probably buy a better one & not suffer for it, so long as you re-use your fan. Perhaps invest in a qwieter fan, since the stock 38mm Nidec/Delta's are loud, aren't engineered for qwiet in mind, just 1.5amps of brute force.
Most gaming pc owners have to toss $ at kombating heat, so, there it is. I wish I could help more. Good luck.
Cass-Ole
1 Rookie
1 Rookie
•
1.8K Posts
0
April 13th, 2016 05:00
Try Arena again, there may be some R4 owners there using HW software or willing to. I can't stress my stuff in order to compare notes but someone there might be able to help figure out what TMP0 & 'normal' readings under stress are in their rig.
edit: try to simply tighten down the VRM h-sink screws 1st in case they were loose by nth degree, re-check temp, then go from there