Jump to content
Corsair Community

Corsair RMi : Aida64 shows all temperature sensors...but Corsair Link does not?


corsican

Recommended Posts

can Aida64 read CL info better than CL software itself? :bigeyes:

 

Aida64 shows TWO sensors: Power Supply #1 & Power Supply #2

 

(PS. interestingly the #1 seems to be always hotter than #2)

 

But Corsair Link only shows ONE sensor (apparently the one that corresponds to Power Supply #2 in Aida64)

 

I can understand if CL only picks one sensor to show (the one that matters the most) but here it's not the case because Aida's sensor #1 is obviously the important one

 

I know cause I launched Prime95 + Furmark: in Aida64 only sensor #1 heated up, sensor #2 stayed the same (so it's a false sensor???)

 

 

so is that a bug in CLink if so can it be fixed? I rather use CL than Aida64

Link to comment
Share on other sites

can Aida64 read CL info better than CL software itself? :bigeyes:

 

Aida64 shows TWO sensors: Power Supply #1 & Power Supply #2

 

(PS. interestingly the #1 seems to be always hotter than #2)

 

But Corsair Link only shows ONE sensor (apparently the one that corresponds to Power Supply #2 in Aida64)

 

I can understand if CL only picks one sensor to show (the one that matters the most) but here it's not the case because Aida's sensor #1 is obviously the important one

 

I know cause I launched Prime95 + Furmark: in Aida64 only sensor #1 heated up, sensor #2 stayed the same (so it's a false sensor???)

 

 

so is that a bug in CLink if so can it be fixed? I rather use CL than Aida64

 

I think that this is probably true for a lot of Corsair PSU models. With my HX850i, I just pick the AIDA64 sensor that matches the iCUE sensor temp and ignore the second one.

 

Not all sensors are useful. For example, Samsung NVMe drives have two temperature sensors!

Edited by Yemble
Link to comment
Share on other sites

I think that this is probably true for a lot of Corsair PSU models. With my HX850i, I just pick the AIDA64 sensor that matches the iCUE sensor temp and ignore the second one.

 

Not all sensors are useful. For example, Samsung NVMe drives have two temperature sensors!

"iCUE"?

 

you mean the sensor with the highest temp?

Link to comment
Share on other sites

I use Corsair Link which accurately reads my AX860i completely
well here I use CL too with my RMi & it only shows 1 sensor (the false sensor at that) so unfortunately I can't use CL & I've to use 3rd party to show the correct sensor :/

 

(Aida shows both the real & false sensor readings)

 

 

 

 

what's your CL version?

Link to comment
Share on other sites

yeah that's mine too v4.9.9.3 original setup exe DL'ed from Corsair

 

so that version's buggy apparently can't correctly read sensors at least not on RMi series

 

any way to report this sort of bug? so they can fix it

Link to comment
Share on other sites

k never heard of ICUE (sounds like they basically renamed CL) gonna check it out

 

weird that the app with the old name's still available for download on their site though +_-

 

 

edit> WTF that new app is huge over 300mb is there a "trimmer" version? I just wanna monitor the PSU nothing else

Edited by corsican
Link to comment
Share on other sites

  • 2 weeks later...
  • Corsair Employee

Guys... it was a conscious decision to only report one temperature from the PSU.

 

The PSU does measure multiple temperatures from multiple points within the PSU, but that is for the PSU to know! Not the user!

 

We found that with reporting multiple temperatures, many users would freak out on the really high ones, wonder why the fan still wasn't spinning when one reading was higher than the other, etc. So it was determined that the temp probe that reports the "ambient temperature" of the inside of the PSU housing would be the best to report.

Link to comment
Share on other sites

Guys... it was a conscious decision to only report one temperature from the PSU.

 

The PSU does measure multiple temperatures from multiple points within the PSU, but that is for the PSU to know! Not the user!

 

We found that with reporting multiple temperatures, many users would freak out on the really high ones, wonder why the fan still wasn't spinning when one reading was higher than the other, etc. So it was determined that the temp probe that reports the "ambient temperature" of the inside of the PSU housing would be the best to report.

 

I'm calling BS on this. Was going to avoid this but since Mr. "JonnyGuru" decided to, for all intents and purposes, call the user a fool time to roast him.

 

You might want to apologize to the people you just insulted there. For someone who has Guru in the name you seem to have very little respect when you word it like that. And who bloody goes after ambient temp? I can get the inside of something cold as can be and still generate more heat per component than what is reported. Personally I don't care about temps as they aren't a problem for me but fact is you can get an ambient reading while the rest of the components are on the verge of melting.

 

So how about this, instead of wording things in such a way that you insult people and their intelligence, and I say this based off several things you, Mr. "Guru", have posted here. But I do thank you at the same time because it shows just what some people at Corsair think of the people PAYING THEIR SALARY.

 

The best thing you can do is walk into HR and tell them you screwed up and need to have your wrist slapped. Then not be so pretentious and not treat people like idiots.

Link to comment
Share on other sites

  • Corsair Employee

I enjoy a good roasting from time to time. Thanks.

 

I've been around a long time and am mature enough to come clean with what I know, and what I don't know. Because, if you pretend to know something, you fail to learn. If you know what you don't know, you see many things as learning opportunities.

 

My favorite pie chart:

 

http://www.raymonds.com/blogs/jr/files/2017/05/Pie-Chart-of-Knowledge.png

 

Not sure what makes you an expert on this. But clearly you haven't researched this through. I would like to make this a learning experience for YOU.

 

First off, a good PSU design should have multiple thermistors because the primary side typically runs hot in most units, especially half-bridge, but there are times where the secondary side may run close to spec under high loads, so you want to keep an eye on that too. You'll also want to keep an eye on the DC to DC temps (the circuit that converts +12V to +5V and +3V we just call "DC to DC" for short even though, in reality, every PSU has DC to DC after AC is converted to DC).

 

The "i" PSUs from CWT use different thermistors for OTP (Bit 3 and Bit 5) and the fan controller (Also Bit 3). In addition to this, you have the thermistor reporting to the user is located just around the area of the fan so the user gets an ambient temperature of the inside of their PSU (Bit 4). The CWT solution is an analog product outside of the MCU that takes the analog information and converts it to the I2C data. The Flextronics made units, since they are a fully digital solution using all MCUs, don't have an "ambient temperature probe" like the CWT ones because they can take the temperatures from all of the points within the PSU where the data is taken and calculate an ambient temperature to the user that, in similar situations to the CWT units, align fairly accurately.

 

Why do you want that alignment? It's simply called "consistency in end user experience". Theoretically, two Platinum PSUs at the same load should produce the same amount of heat. After all, your efficiency is based on the heat, or lack of heat, generated by the PC from losses from converting AC to DC. So if you have one analog 850W PSU made by CWT running at 50% load in a 35°C environment, you certainly wouldn't want your user to see a substantially different number if they have an 850W made by Flextronics that is also running at 50% load in a 35°C environment.

 

Since you're using the same software to report data from two or three different platforms from two different vendors, you have to come up with a way to deliver consistency in your reporting. You have to decide on a single temperature report and a clear cut definition to the end user of what that temperature is. For us it was the PSU's "ambient temperature" as if it were measured between the fan and top of the highest heatsink (which it is in the CWT made products, and is "simulated" in the Flex made products).

 

It also makes engineering a little easier if you define your OTP at an "ambient temperature" and then make sure you spec your other components from there. For example, if you define OTP to latch the PSU at an "ambient temperature" (which again, is the same thing that's reported in LINK and iCUE) of 80°C, you then have to spec your other components and make sure that while the ambient is increased to that threshold, that those components are not pushed to their thermal limits. If they are, you need to either upgrade that part to a more robust part, or reconsider your design.

 

We'll usually test a unit with 12 different thermal probes. To give you an example of how alarming some of those temperatures can get, at a 50° ambient temp, we can measure T1 (the main transformer) getting as hog as 109.8°C! You can drop that puppy in a pot and boil water with that sucker!

 

 

 

Not too long ago, we had a user here called Redray. Snarky... not as snarky as you.. but he did serve the user base with a good tool called "SIV" and I tended to get along with him because he didn't **** on Corsair's PSUs... just Corsair's software. :D You can find his software here: http://rh-software.com/

 

Any way... Redray's software actually got the data from the Flextronic's MCUs and reported ALL of them. And.. as predicted... one of the side effects of that was people freaking out that they had temperatures as high as 80°C+ and yet their fan wasn't spinning. This lead to support calls, people assuming their fans were dead, people assuming their PSU must be bad, people assuming that Corsair was intentionally showing an artificial temperature instead of an ambient temperature, etc. etc.

 

 

 

So... I hope this helps you understand some of the work that goes on behind the engineer's curtain. But if we're going to continue the roasting, I would like to request Amy Schumer, Steve Martin and Steve Colbert to participate.

Edited by jonnyguru
Link to comment
Share on other sites

I prefer Gary Busey, Martha Stewart and William Shatner personally. And for a foursome lets throw Mike Tyson in there as well. That said, well played sir. Definitely got some brownie points with me on that reply.

 

My issue with what you said was not that I don't know how stuff is made. Not going to pretend I know what you do about PSUs, and I prefer to stay consistent and harp on crap hardware when it is crap. That said I do not believe that the PSUs are, just a lot of the other hardware (and you can chalk most of it up to the software). It was with how you said it.

 

To come out and say:

 

"Guys... it was a conscious decision to only report one temperature from the PSU.

 

The PSU does measure multiple temperatures from multiple points within the PSU, but that is for the PSU to know! Not the user!"

 

The issue I have with that is that you have no reason to end it that way unless you are trying to point something out. And I come from a different school where if you explain things in plain English for someone, they tend to understand and realize you might be right. Not all but most. Again my issue was with how it was ended as it basically says "You don't need to know because you don't understand." That is all.

 

"We found that with reporting multiple temperatures, many users would freak out on the really high ones, wonder why the fan still wasn't spinning when one reading was higher than the other, etc. So it was determined that the temp probe that reports the "ambient temperature" of the inside of the PSU housing would be the best to report."

 

If I might make a suggestion here, some people want to see that crap. Personally I do but only on certain things. Its how I found out one of my 2Tb WD storage drives was going bad since it was idling at 59°C or so (granted the fact it was dropping offline constantly was a clue). So after spending a day of transferring some files off (yeah I know better when it comes to backups) a little at a time I pulled both of them and retired them. Sister drives since they came off the line back to back. 9.8 years (or was it 8.9? Either way a long damn time) of operational time on them. My suggestion is this:

 

BURY the setting but if someone wants to view it then they can enable, it but by default, it is off and buried. That way you get the best of both worlds if you want.

 

But I did mean it when I said I could get ambient reading what it should while the components melt. No LN2, Dry ice or A/C needed to do it either. Heat and me have a painful relationship, you try having half the bloody Kia/Hyundai logo burned into your forearm by a cat. Part of why I quit mechanics. Now I'm all about being nice and only calling people out when needed......

 

Anyway, just look at the optics of how things are worded and how they might be read. It's my primary issue with text communications since you can't HEAR how something is done.

 

And yeah, I am snarky (which also tells me you are probably read in on my issues as of late...and have seen my posts) but if yall get your **** together on the software front and catch those people who keep having similar issues/failures, or are asking for help from support and getting none (referring to the dropped tickets here) then we will be fine.

 

At the risk of this being blocked, I'll put it this way;

 

If I didn't give a **** about Corsair I wouldn't have stuck with you guys for longer than my account had been around and I wouldn't be wasting my time trying to get better interaction from you guys for people on here. I mean really? No open kegger at the HQ for the 25th anniversary? Hell there isn't any announcement at all. Pretty poor showing you ask me :)

Link to comment
Share on other sites

  • Corsair Employee

Well, you saw how the thread started, right?

 

My response was in the context of those posts that assumed the Corsair reporting was wrong or unintentionally incomplete.

 

I simply wanted to be straight to the point.

 

Yes: There are multiple sensors.

 

No: LINK/iCUE doesn't report them all.

 

Yes: If you do read the raw data from those sensors, they might not match what you're reading.

 

Sorry my delivery does not meet your approval (I'm not really sorry).

 

Do I want to get all technical in a user support forum? No. I reserve that for more technical forums where I'm then sent to HR the next day for a slap on the wrist for exposing too much IP.

 

Just yesterday, we just had a discussion about the whole AX860i failure thing. I wanted to get all technical on the how and the why. But I'm keeping out of it because at the end of the day, the customer wants a working PSU and telling them what's wrong with their current PSU isn't going to help that.

 

My response here was akin to that. You see something that doesn't jive. Assume something is wrong. I'm telling you it's not and to move on.

 

It's what I call my "middle ground". The alternative, for me, is to not participate at all and there's no fun in that. You'll note that my handle in these forums isn't "Corsair Jon". I'm not here because I'm required to do so. I'm here because I genuinely enjoy what I do and sometimes get a little offended when people **** on the fruits of my labor.

Edited by jonnyguru
Link to comment
Share on other sites

Guys... it was a conscious decision to only report one temperature from the PSU.

 

The PSU does measure multiple temperatures from multiple points within the PSU, but that is for the PSU to know! Not the user!

 

We found that with reporting multiple temperatures, many users would freak out on the really high ones, wonder why the fan still wasn't spinning when one reading was higher than the other, etc. So it was determined that the temp probe that reports the "ambient temperature" of the inside of the PSU housing would be the best to report.

 

I found using a custom fan on CL was able to stabilize my AX860i from faults. Unfortunately the iCue package cannot do that with my PSU. Thankfully the CL software works with my PSU properly.

Link to comment
Share on other sites

Guys... it was a conscious decision to only report one temperature from the PSU.

 

The PSU does measure multiple temperatures from multiple points within the PSU, but that is for the PSU to know! Not the user!

 

We found that with reporting multiple temperatures, many users would freak out on the really high ones, wonder why the fan still wasn't spinning when one reading was higher than the other, etc. So it was determined that the temp probe that reports the "ambient temperature" of the inside of the PSU housing would be the best to report.

alrite so when PSU load increases (eg. run prime95 & furmark at same time) how come PSU sensor #2 (the one reported by CLink) barely goes up? it stays around 35°C maybe an extra °C or so

 

on the other hand sensor #1 (the one not shown by CL) goes up by about 7°C

Link to comment
Share on other sites

alrite so when PSU load increases (eg. run prime95 & furmark at same time) how come PSU sensor #2 (the one reported by CLink) barely goes up? it stays around 35°C maybe an extra °C or so

 

on the other hand sensor #1 (the one not shown by CL) goes up by about 7°C

 

As described above, the sensor reported in Link/iCUE is measuring the PSU internal ambient air temperature or effectively exhaust air temp. Sensor 1 measures some other specific component within the PSU. When your PSU is under load every single component in the PSU likely has a different temperature, measured or not. The ambient air or exhaust temp is a measurement of the total heat being produced from all of the components. It is a reliable indicator of the current operational state and a logical variable for fan control. There may be some additional triggers for fan action related to the other sensor, but that is the proprietary part referenced and it is likely specific to both Corsair and OEM. Regardless, it is not going to be an alterable behavior. At least with "i" PSUs you can see this stuff at all. On most others these are not reported values and you wind up doing what I usually do -- tape a 10K thermistor to the exhaust side of the PSU. If the exhaust temperature is not increasing, that means the fan (or the PSU) is doing its job. A steady exhaust air temp means you are removing the heat, but I would not expect sensor 1+2 to match very often, except when the PSU is in a low power state. Anything with current running through it is going to warmer than something else that is not electrified -- like the air around it.

Link to comment
Share on other sites

As described above, the sensor reported in Link/iCUE is measuring the PSU internal ambient air temperature or effectively exhaust air temp. Sensor 1 measures some other specific component within the PSU. When your PSU is under load every single component in the PSU likely has a different temperature, measured or not. The ambient air or exhaust temp is a measurement of the total heat being produced from all of the components. It is a reliable indicator of the current operational state and a logical variable for fan control. There may be some additional triggers for fan action related to the other sensor, but that is the proprietary part referenced and it is likely specific to both Corsair and OEM. Regardless, it is not going to be an alterable behavior. At least with "i" PSUs you can see this stuff at all. On most others these are not reported values and you wind up doing what I usually do -- tape a 10K thermistor to the exhaust side of the PSU. If the exhaust temperature is not increasing, that means the fan (or the PSU) is doing its job. A steady exhaust air temp means you are removing the heat, but I would not expect sensor 1+2 to match very often, except when the PSU is in a low power state. Anything with current running through it is going to warmer than something else that is not electrified -- like the air around it.
alrite then can someone @Corsair confirm if the fan's triggered only by ambient temp or by any of the other sensors too? (for those who enable 0 fan mode)
Link to comment
Share on other sites

  • Corsair Employee
alrite then can someone @Corsair confirm if the fan's triggered only by ambient temp or by any of the other sensors too? (for those who enable 0 fan mode)

 

There's a fan controller MCU that uses a combination of inputs to determine if/when the fan should turn on and for how long.

 

Load, temperature and duration. If the temperatures are high, the fan turns on. If the load is high, the fan turns on. If the fan turns on due to temperature; depending on how long the fan has to run to bring temperatures down, the fan may run a few seconds after temperatures come down to a desirable temperature to avoid the fan immediately ramping back up again too soon.

 

It's a more expensive solution as you're adding an MCU and you need op-amps to measure the output current, but it yields the most desirable effect.

 

Most fan controllers only use a thermistor and only change the fan speed based on temperature, so when they advertise that the fan increases at a particular load, they're using an "example" of how the fan would behave at a particular ambient temperature (typically 25°C).

Link to comment
Share on other sites

  • Corsair Employee
I found using a custom fan on CL was able to stabilize my AX860i from faults. Unfortunately the iCue package cannot do that with my PSU. Thankfully the CL software works with my PSU properly.

 

Sounds like you figured out part of the AX860i failure issue and why it's so intermittent. ;)

 

Hopefully Kevin will respond to that thread soon.

Link to comment
Share on other sites

×
×
  • Create New...