PC Processors

[thansen56]

I built a Ryzen 5900x system using a gigabyte x570s MB and MSI 240 water cooler.  Memory is GSKILL RipJaws  F4-3600C18D-64GVK x 2 (so 128G total).    Stock CPU power and frequency, slight OC on the memory with reduced CAS timing; but no frequency adjustment that could affect the CPU.

System runs fine.   I was surprised at the CPU temperature spikes the minute something CPU intensive happens.   CPU die temp hits 80-90 degrees C almost instantly.   Since the system is water cooled, I have a fan profile that doesn't cause my fans to immediately spin up to max and my temperature probe placed in the radiator fins where the water enters the cooler shows modest temperature increase under load.   When the load stops, the temp immediately drops down into the 60s and then 50s.  Individual core temps do not hit 90 (70s is typical, but have seen 80s) and the cores are low 50s at idle.

Should I just accept this as normal and do nothing or should I start tuning and/or trying to understand why the water cooler isn't doing its job?
I have read this article about lower voltage etc. but I can't see that this should be required for non-OC config with a water cooler.

https://www.reddit.com/r/Amd/comments/kfpele/5800x_adjusting_ppttdcedc_limits_on_pbo_got_me/

So just looking for any advice.  Like I said, the thing runs fine.

[elstaci · Accepted solution]

For a 240 MM AIO Liquid CPU Cooler your processor shouldn't reach 90C. Basically it is overheating.

The Maximum Operating Temperature for your processor is 90C. Which means when the CPU Temperature starts reaching or has reached 90C the CPU will automatically throttle or slow down to keep the temperature at 90C or below.

So either you MSI AIO Cpu Cooler is not installed correctly or not making good contact with the surface of the CPU or simply isn't powerful enough for that processor.

The TDP Rating for your Processor is 105 Watts. Which means you need a CPU Cooler that is rated at least 125 watts or higher.

Do you have good air flow in your computer case?

I would try to have a very aggressive Fan Curve.

Is the pump running at its recommended speed? You can check using the Cooler software or Monitoring software.

Is the part that makes contact on the processor on tight and make even solid contact? It isn't slightly loose or wobbly?

What is the exact MSI AIO CPU Cooler you have installed? is it the MAG CORELIQUID 240R?

EDIT: found this tech review article on the best Ryzen 5900X CPU Coolers. It also explains about the CPU throttling when it reaches it maximum operating temperature of the processor: https://whatifgaming.com/best-cpu-coolers-for-ryzen-9-5900x/

[elstaci · Accepted solution]

For a 240 MM AIO Liquid CPU Cooler your processor shouldn't reach 90C. Basically it is overheating.

The Maximum Operating Temperature for your processor is 90C. Which means when the CPU Temperature starts reaching or has reached 90C the CPU will automatically throttle or slow down to keep the temperature at 90C or below.

So either you MSI AIO Cpu Cooler is not installed correctly or not making good contact with the surface of the CPU or simply isn't powerful enough for that processor.

The TDP Rating for your Processor is 105 Watts. Which means you need a CPU Cooler that is rated at least 125 watts or higher.

Do you have good air flow in your computer case?

I would try to have a very aggressive Fan Curve.

Is the pump running at its recommended speed? You can check using the Cooler software or Monitoring software.

Is the part that makes contact on the processor on tight and make even solid contact? It isn't slightly loose or wobbly?

What is the exact MSI AIO CPU Cooler you have installed? is it the MAG CORELIQUID 240R?

EDIT: found this tech review article on the best Ryzen 5900X CPU Coolers. It also explains about the CPU throttling when it reaches it maximum operating temperature of the processor: https://whatifgaming.com/best-cpu-coolers-for-ryzen-9-5900x/

[thansen56]

Yes it is the MAG LIQUIDCOOL 240R.  It isn't really stated in the specs the max CPU wattage heat dissipation provided.  I am pretty certain the heatsink contact is solid and it was done using artic silver thermal compound following the recommended application pattern over the cores of the CPU (per artic silver recommendations).  I feel like the fact the CPU instantly cools back down to a good temp is an indication that the cooler is working.  I see the post with regard to the gigabyte temperature reporting which is interesting.

[thansen56]

Another note.  I ran PASSMARK on the system.  The CPU scores very high.  So not sure it is throttling.

CPU was 41097 (99% percentile)

The AIO pump is measured at max RPM (4000 rpm).  Increasing FANS to max have no impact on the CPU temperature and as I stated before the actual radiator temperature isn't really climbing significantly.  

I never see core temperatures hit 90.  Perhaps internal temperature throttling of the CPU is based on those values vs. the CPU case temperature being reported by the MB as CPU temp.  The core temp is max in the 70-80 range during cpu benchmarks.

[thansen56]

Here is a good view of what I get while running a benchmark.  Core temp vs. the case temp.

[elstaci]

@ryzen_type_r 

Seems like the Ryzen 5000 series Processor run much hotter than their TDP ratings.

I noticed that under stress your 5800x reached its Maximum Operating temperature of 90C which means that it was throttling under those temperatures.

This could be similar to what happened when, I believe, the 3000 series CPU first came out. The way the processor were built or made, at that time, most CPU Cooler were ineffective since it wasn't covering the area where the processor got hot. Later on this was rectified by the manufacturers.

But it is fine as long as the processor Maximum operating temperature doesn't run at 90C or above for extended periods. It is safer to keep the temps below 85C at all times no matter how heavy of a load.

[DRDOS]

@thansen56
What is the tool are you using for monitoring a per-core temperatures?
Those i used provide only one equal value for all - Package/CPU temperature and all of the per-core temperatures (from 1 to 8 ) - all are equal. Or maybe 5700G doesn't have a dedicated per-core temperature sensors unlike X-versions processors...

[elstaci]

Yes it could be a faulty Motherboard CPU Thermal sensor or it can also be a faulty CPU Thermal sensor.

But when you mentioned it reached 90C almost instantly under load made me believe it was your AIO not doing it job effectively or it isn't powerful enough to maintain the processor from overheating under heavy loads or installed incorrectly.

But I would think if it was a bad thermal sensor it would read bad all the time.

The maximum Operating Temperature is read from, according to AMD, is the TJMAX Junction on the processor.

I guess one way to find out if it is a bad thermal sensor either on the Motherboard or CPU is by installing a different CPU Cooler and see if you get the same results.

Or installing your Processor on another compatible motherboard and see if you still get the same temperature spikes.

But then again I realize that isn't very realistic since most Users don't have more than one CPU cooler nor another PC to install test their hardware. Unless you take it to a computer repair shop and have them test it for you.

I would open a Support Ticket with MSI Support about your AIO and see what they say could be the problem.

Basically you need to start eliminating hardware. The best place to start is your CPU Cooler than both the Motherboard and Processor.

You might want to open a AMD Support Ticket under Warranty category and see if they believe you need to RMA your processor to be checked from here: https://www.amd.com/en/support/contact-email-form.

That way you won't have to go through the hassle of opening a AMD Warranty Request first.  IF AMD Support believes you processor is defective then they will tell you to open a AMD Warranty Request to have it RMAed.

[thansen56 · Accepted solution]

I switched the CPU cooler to the EVO 212 V2 with push fan rear facing.    As part of the switch, I checked existing contact patch on the CPU and it looks good (thermal compound was evenly distributed).   With the EVO, I now have idle temperatures in the 30's and while running benchmark I only reach 72 degrees C.  No more 90 degree spikes.   My plan is to RMA the MSI AIO and see what happens.  Honestly not sure I will switch back to an AIO, but doubtful I will ever switch back to the MSI unless there is some description of a clear defect.  If not, I will add top mounted case fans blowing OUT which were present with my previous MB in this case.  I will also verify that my side mounted case fans are BOTH blowing IN as recommended.  I know the one behind the MB is; not sure about the other one.

Here is my contact picture:

Here is my idle view from ryzen master:

So thanks everyone for the input and I believe we have identified the issue: the MSI AIO.

[Sunni]

Hey, 

So I have a similar situation and I just wanted to check in that thats still the case?  You didn't swap back to an AIO with any luck?  Would you mind posting a HW info that was just idle-light stuff over the course of 30 mins to see what your averages are.  So my situation:  I have a 5900x and 3080 in a Lian Li 011 Mini.  Case has 6 intake fans and 2 exhaust, and cpu has an EK 240 AIO.  These are my temps with some moderate usage over an hour or so:

(by moderate usage I mean having a handful of tabs open, few streams up, second desktop with some light school work up, discord, hwinfo - no games)

So, I've been worrying for months that my temps seem high, but I've seen way worse on  numerous forums and everyone just gives ya the old "its not thermal throttling..its under well under max".  Ambient temp is about 20C.  I repasted the thing 2 times out of paranoia.  So my question is are your temps holding true for 30C TCTL/TDIE and staying that amount lower under light loads/idle etc...tldr is it worth swapping to an air cooler.  I do know a good air cooler is only slightly less efficient as say a good 240 aio but 20C, kinda seems like something else is going on.  Just to clarify, pbo on, slight undervolted via scalar...basically a -10  all core as I was having some BSODS with increased drops on the higher cores.  Can you/would you mind posting your hwinfo for replicating light work for around the same time.  Thanks in  advance either way.  Also,  just to add, I'm never thermal throttling or going above high 70s under load,  maybe getting to low 80s at worst but yea just seems a little hot if your saying your getting your consistent idle down to 30/sub 30 tctl/tdie.

[DRDOS]

I'll just remind what already written before (here or in another related threads).

When i compared REAL temperature taken from CPU lid by an external temperature measurement devices, it was 20-30 Celsius degrees lower than what was reported in BIOS and in software tools. This difference depended on the temperature: higher is the temperature - higher is the difference.

I don't think such a high thermal resistance between the CPU die and the lid takes place, especially that another thermal sensors (e.g. iGPU sensor in G-versions CPU's) returns very correct temperature value that completely suits to what external measurement devices showed on their display (when i did that comparison).

Such a huge difference would be normal if you measure a temperature not on the CPU lid surface, but on the motherboard surface under the CPU Socket.
Exactly such situation we had over than 20 years ago with a Socket-A motherboards which used a thermal resistor mounted on the motherboard itself under the CPU socket, instead of integrated on-die thermal diode in the CPU (those CPU's had it already, by the way). IIRC, a difference between the temperature from that on-board and on-die sensors was somewhat in a range of about 20-30 Celsius degrees (depending on the on-board sensor implementation and else factors...).


Well, here we have seen at least 2 another reasons of too high CPU temperatures:
1. Defective/broken cooling system;
2. CPU hardware failure.
So, incorrect values of reported temperature is not necessarily the case. Just keep this in mind that such a case is also possible.

[elstaci]

@Sunni 

Your processor's maximum operating temperature is 90c. So if the temperature (tctl/tdie) doesn't go above 80c at any loads you have no problems. Your CPU cooler is doing a very good job of keeping the processor cool.

It might spike above 80c or even 90c in certain situations but if it is just a very short term spike that isn't a problem. You processor was engineered to work at a maximum 90c temperature. Once it reaches 90c or above the processor will automatically start to throttle or slow down to keep the temperature at 90c or below.

I would use Ryzen Master to verify any temperatures reading you have with other monitoring software.

[elstaci]

add the following statement to my edited reply:

unable to get past the "Highlighted" error just for inputting one sentence in my edited reply.

[Sunni]

@elstaci 

I understand everything that you said.  I even said in my post that "I've seen everywhere on tons of threads of people asking similar type questions to gauge such as  "what their cpu should be running at with similar cooler, ambient, etc" and that most people will give nearly the exact answer you did, lol.

Thanks for trying, but me saying that in my original was to avoid that.  I'm someone who's last PC was an OC'd 6700k/gtx1080 and was used to much lower temps, and was trying to get ballpark what others are seeing with 240/280/360 AIOs and setups.  I know the chip isn't going to deal damage to itself, I realize they have countermeasures that if it gets that hot it will stop prior or ramp down prior, but I was looking for "What temps should I be getting given my info"  not if it was an issue or not.  I know its a non issue even what my PC is running at now.  I mean **bleep**, during most games its not breaking 80C, but I truly thought it should be lower is my point.

That said, I said what pc I was coming off of, I know Ryzen runs much hotter, I know the cpu is a higher watt cpu off the bat but yea, I was wondering what others are seeing bc the OP also thought this range was a bit high.

[DRDOS]

Again temperature issue and again with Gigabyte motherboard... Very interesting...
I bet if an AIO system would be manually turned on to its maximum performance, a temperature issue will remain anyway. Sure a temperature would be less then, but still higher than it normally should.

Probably, we have to address this issue to a Gigabyte motherboards developers.
Especially after number of similar cases with 5700G processors - which (unlike X-versions) is a single die (mono-crystal), and a CPU temperature values on a Gigabyte motherboards appear to be up to 15-20*C higher than integrated GPU temperature located on the same die.

I already don't pay attention to a CPU temperature value in my 5700G at all, using iGPU temperature sensor instead which is the only one that provides a reliable/correct temperature value in this CPU (in my PC with Gigabyte motherboard). But it's not the case for X-versions processors of course, so need to find another way to obtain a correct CPU temperature value.

Another question raises here - is this invalid temperature value on a Gigabyte motherboards used only as information for user (displayed in BIOS menu and provided for applications), or maybe it is also used, e.g., for overheating protection mechanisms, thermal throttling, and so on, which potentially, in some cases, may cause a performance degradation by no actual reason?

So, if you have double-checked your AIO system and sure it's working properly and its performance is sufficient for your CPU, then keep in mind there are already many people experiencing similar issues with similar CPU+Motherboard combinations.

In my case i personally have checked everything a lot of times to eliminate every single reason that potentially may cause such an issue. Also, measured a temperature in a different points on the sides of the CPU lid by an external thermal sensor to make sure a CPU cooler is installed properly and a thermal contact between the cooler and the CPU lid is good/reliable. Now, at this point, i'm completely sure in my case it's a Gigabyte motherboard issue that just provides invalid temperature value for the CPU.

[ryzen_type_r]

Your temps shouldn't instantly spike up to 80-90 unless you apply a serious computational load.  I never see temps behave like this on my 5800X unless I'm running a Prime95 (or similar) torture test.

Cores idling in the 50's also isn't normal, unless your ambient (room) temps are very high.  You shouldn't be seeing more than 10-15C over ambient just sitting at the desktop doing nothing, unless:

- There is something running in the background putting a load on your CPU.  Try a clean Windows install with no other software installed except for the necessary drivers.  Don't install any motherboard utility software.  If you still see idling in the 50's, then maybe:

- There's something off with your cooling system.  Whether it's the AiO itself, or maybe cold plate mounting pressure, or thermal paste application.  Personally I'm not a fan of Arctic Silver's instructions of a single vertical line for Ryzens, because if you look at the chiplet layout underneath the IHS, there's no silicon in the middle.  The SoC is offset to the left and the CCX's are on the right.  The way I did it on my build was a small vertical line over the SoC and a dot over each CCX.  Had to look at delidded CPU's to find out where the points would be on the IHS, but it worked out fine.  Maybe it's not a big deal and a single vertical line works fine, I didn't do any real comparisons, just checked to make sure temps weren't ridiculous for my cooling setup.

And yeah, try an aggressive fan profile just to rule that out.

[thansen56]

Question:  So if the cooler wasn't working would you think the temperatures would immediately return to normal (well not normal based on your assessment that 50 is too high) when the load stops?   What would cool the CPU?

[elstaci]

@thansen56 

The CPU Cooler might be fine under normal loads but once the processor is under heavy loads and become very hot, the CPU Cooler may not be powerful enough to remove the heat before the processor starts to overheat.

But like the others mentioned it could be a faulty Thermal Sensor either on the processor or motherboard. That is why you first need to eliminate the CPU Cooler as the culprit from your troubleshooting list.

As an example only, I had a stock CPU Cooler on my FX8350 processor. It was working fine under any loads. Then one day I noticed my PC shut down by itself. But it would only shut down when I was under very heavy loads. I noticed the temperature would start climbing past the Maximum Operating temperature (68c) until the PC shut down to protect itself.

Turned out the my stock CPU Cooler had  become defective. Seem like the fins were damaged after several years of use. Replaced it with a CoolerMaster Hyper 212 EVO CPU Cooler and the temperature never goes above 68C under stress or very heavy loads.

@DRDOS  made a good point about the monitoring software you are using. I would try 2 or 3 different types of monitoring software to see if you get the same results.

Also I would use Ryzen Master to check you temperatures since that was created for the Ryzen Processors.

[thansen56]

I think just installing an air cooler vs. the AIO as a test is a good idea.  I will purchase the EVO as you suggest and then if the temp issue is the same, I have eliminated the AIO as a problem.   It seems to be relatively inexpensive.  If this doesn't resolve the issue, I can take it up with AMD and/or GIGABYTE.   I do use Ryzen master as well, FYI   It just doesn't give as much data.

[elstaci]

@thansen56 

The EVO is around 30-40 US Dollars but there are newer versions of the Hyper 212 at CoolerMaster. The one I have is an old model.

But you probably will need to replace your Motherboard's CPU mounting plate in the rear of the motherboard to be able to install that CoolerMaster CPU Cooler.

In my case, I have a CoolerMaster Computer case where I was able to replace the rear metal mounting plate without having to remove my motherboard. The Case had an opening so the User can replace the mounting plate on without removing the Motherboard from the computer case.

Maybe your Computer case has a similar feature to it.

That goes for most high quality powerful Air Coolers. They usually have their own mounting plates that you will need to install.

By the way, the Hyper212 EVO is rated at 150 watts TDP. I would purchase the newest Hyper212 model since it was probably made for the Ryzen processors. They are probably more expensive.

Here is all the compatible CPU Coolers for the Ryzen 5900X from PCPARTPICKER: https://pcpartpicker.com/products/cpu-cooler/?compatible_with=KwLwrH

There you can find a inexpensive Air Cooler that for your processor.

[thansen56]

I ordered the hyper 212.   I have access to the rear of the motherboard from my case.   My plan is NOT to keep this cooler; it is for a test only.  I want to water cool this system.  If the test shows better numbers, I will file a warranty request with MSI unless I see a reason why my installation was bad (not much can be done wrong, other than not correctly installing the heatsink.  My radiator is mounted at the TOP of the case as recommended by many.

[elstaci]

Yes I believe a AIO would be a better CPU Cooler for your high TDP Processor especially since it is a 12 core processor. You really need a heavy duty CPU Cooler.

Good luck.

NOTE: Make sure the Hyper212 will fit in your computer case. Also I have 2 fans on my Hyper212 EVO cooler for my new Ryzen 3700X. It keeps it at or below 72C under stress testing. The Maximum Operating temperature of my 3700X is 95C with a TDP of 65 Watts. Ryzen definitely run much hotter than the previous Non-Ryzen processors.

So if you have a second CPU Cooler or high speed computer case fan that fits on the Hyper212 I would definitely install it for a Push-Pull configuration to get maximum Heat removal.

[thansen56]

Here is my IDLE situation.    Doesn't seem like much is going on.  Temp in the low 50's.   Seems like throttle down is all working as expected with low current draw and low clock speed.

[elstaci]

Your Ryzen Master looks really good.

But 52c at idle with a AIO installed seems kinda warm to me. Especially since all the cores are sleeping except one core.

I would think your Processor's temperature with only one core running at 444Mhz would be in the low to mid forties.

What is the Ambient Room temperature where you PC is located at? is it fairly warm? or around or below 80F?

Remove one side of the computer case and see if the temperature is slightly cooler. If it is then that could indicate poor air circulation inside the computer case.

NOTE: Just for troubleshooting purposes try disabling PBO if it is enabled. Normally it is by default. PBO is AMD safe way of overclocking the Ryzen processor automatically. 

If the processor run much cooler with PBO disabled then the mean you will need a stronger CPU Cooler.

[DRDOS]

I've inspected that tool (HWiNFO) a little bit, with help of it i've figured out one part of the problem...
A CPU temperature value being shown in BIOS, in Ryzen Master and in another tools/utilities is that "Tctl/Tdie" value in HWiNFO. Inspecting its behavior in different conditions and comparing with values provided by another temperature sensors monitored by HWiNFO i've come to a conclusion that "Tctl/Tdie" has nothing to do with real temperature value of the die.
According to description of the "Tctl/Tdie" in HWiNFO - it is a "Actual highest temperature among CPU sensors in the die". Behavior of this value is exactly as it's described. In fact it's a peak indicator that just takes the highest value from any sensor immediately and keeps it within specific amount of time. This way it may show temperature values that are much higher than real average temperature of the die at the moment.

So, in idle condition i see all per-core sensors are at about 35-37*C, but "Tctl/Tdie" is always at about 45*C, because in a very short intervals of time a temperature of one or another core spikes up to about 45*C and then back to 35-37*C. Whenever this occurs a "Tctl/Tdie" takes and keeps these maximum values. While real average temperature of the die is 37*C or lower, having wrong understanding of the "Tctl/Tdie" user is misled by its value, thinking that current CPU temperature is higher than it should.

Alright, though it is now clear what a "Tctl/Tdie" is, there's still another question with no answer...
(Just to remind here - i investigate the same issue but with another processor - 5700G, which in this regard has significant differences towards 5900X, such as the count of dies and their physical locations under the lid - one has a single die, and another has 3 standalone ones - IOD and 2xCCD).
In idle CPU condition i see in HWiNFO all of the temperature sensors except "Tctl/Tdie" stay at almost the same value (+/- 1*C) most of the time, which is completely normal. But, once CPU is loaded by a stress test, a per-core temperature sensors values raise up to 15-20*C more than another sensors. E.g., a per-core temperatures are 75-77*C, and a GPU and SoC temperatures are about 55-56*C, and an external sensor that i use to monitor the temperature on the CPU lid outside shows 53-54*C.
From this point i see 2 possible explanations:
1. For some reason, a per-core temperature values are correct for low temperatures, but incorrect for higher temperatures.
2. A high temperature difference in a different parts of the monolithic die may take place.
Here a question raises for a case #2: is this actually possible? Especially considering that a monolithic die is covered by a lid and a massive copper cooler above.
Somebody with a non-Gigabyte motherboards and a similar CPU's could probably help to figure out this question by checking out their per-core temperatures...

[thansen56]

MB temp sensors are in the mid 30's.  I have ambient temp sensors in the case; one attached between RAM chips and one on the AIO radiator; RAM mounted sensor shows 42C and radiator mounted sensor shows 38-40C.  Room is 20 degrees C (as it is winter here and we heat to around 68F).  I have front 280mm case fan blowing IN (highest speed fan I have at 2000+).  I have rear mounted 24mm case fan blowing OUT.   I have a case side 240mm fan blowing OUT.   I even have a bottom mounted (below MB) 240mm fan blowing IN to cool under the CPU (probably silly, but the space was there to add it and the RGB color it adds is cool).   Then I have the AIO top mounted with its two 240mm fans blowing OUT.  As I said in a prior post, I report back results with the Hyper 212 EVO V2 cooler.   Note that there are LOTS of similar posts on MSI forums on this AIO with rapid CPU temp spikes.

[qbtheslayer]

Side fan should be intake...  I have never seen a side exhaust be effective.

QB

[2loki4u]

Hey, i know this is an almost 2yo thread - but i just recently upgraded from a 7x3700x to a 9x5900x and the impact to my AIO was demonstrable. Where previously, I would sit with the 3700x (running avg 4100mhz) was in the mid 40s and even with cinebench running for 30mins the all pkg number would rarely get above 55C or 60C at max. 

with the 9x5900x at near idle it sits around 54-59C and spikes regularly up to 65-70C under sporatic loads (normal usage) though hovers at around 70-75 (again pkg - individual cores are ~10C lower) under 30m cinebench test...

My AIO is like 2yo now - it's a 240mm ek AIO. can't find the thermal ratings (or know if they had them rated that way then) but given i'm using essentially the same artic silver thermal compounds i used back then - i'm beginning to think this is just a factor of the higher TDP of 105, up from 65 of the 7x3700x

Anyone have any insight now that it's 2yr later? quantifying typical AIO 240mm temps on the 9x5900x specifically? trying to decide if i need to replace my AIO or maybe just do another application of the thermal compound.

Issue for me is the increased fan noise from the fans kicking up everytime something pulls a little hard from the CPU... listening to them ramp up is distracting and i'm not adjusting to it well (i'm a pricey b!*ch) everytime i open an app or go to do something that requires the least amount of processing power for 5-10s is annoying - but i don't want to cook my CPU by turning it down. Considering swapping 2 or 3 of my front intake fans to 140mm ones to increase fresh air intake at lower RPMs but I don't have the room in the top of the case for a 280mm fat boy AIO - i'm setup in a lian li lancool case - sadly my old config with the radiator front mounted isn't an option anymore as I added ram and the water block had to be reversed so there isn't enough loop hose to allow me to put the I/O of the radiator at the bottom anymore. Figured having it at the top vertically wouldn't be very good for the pump or the water / air gap in the system, so now it's top mounted as noted earlier with the fans exhausting out the top. 

I've attached some current photos for reference...

Thanks in advance!

 

Any fresh thoughts?