Loving this thread and the continuing push for performance in Ant-Man. @CubanLegend is really squeezing every ounce of performance out of it. How is the fan noise on the 1080 Mini while gaming? Can you compare it to say a 1070 or 1060 ?
-
Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!
Log NFC S4 Mini Classic S4MC #193 "Ant-Man" (7700k/1080mini) *7/11/2020 BRICKLESS 4.2L*
- Thread starter CubanLegend
- Start date
Thanks! I'm pushing for the best performing PC in the under 4.3L category, and I think I've got it down, so far.
Regarding the 1080mini's fan noise, the noise it fans make are pretty quiet, Josh noted that its a little louder than the 1070mini's because these run at 30% minimum without a 0% silent option... (my only gripe with the card to be honest). the MOSt noise though, comes from the front bezel GPU vent, as the fan is VERY close to it and there's a slight woosh noise caused by air turbulence since the fan is in close proximity to the inner chassis. IF i had the new 7MM bezel I might not get that much noise since there'd be more of an air gap, causing less turbulence, but I've got the beta 5mm Acrylic Bezel.
--------------------------------------------------------------------------------------------------------------------------
[Main Build UPDATE]
STAGE 3: CPU Temp Testing [Post-Delid] (Termolabs/Cooltech LP53 vs Delidded Temps)
NHL9i vs NHL9i* vs LP53* vs LP53 w/i7-7700k Delidded* - RESULTS! (+bonus alternate "mini" VRM heatsinks installed!)
PICS of the Delid process adventure, skip these and look at the results below if you like
(theres a bonus video in there showing the ACTUAL delid moment, too!)... Well, wonder NO MORE!Getting ready to delid, got all my tools out.
Removing the 7700k revealed my LP53's mounting to the A9x14 PWM fan was on point:
Also my NT-H1 application was on point:
Video of the Delid itself (the IHS pop was not super loud and was not bad at all!):
Here's my verification pic for the overclockers.net forums delidded club:
The bare CPU PCB/Die and the back of the IHS, note the GAP in the sealant Intel uses, i marked it with a red pen on the IHS to not lose track of it and remember to replicate it when i relid the CPU:
Cleaned the integrated TIM from the CPU's die and the black sealant glue with my fingernail (perfectly safe on the CPU and I used the flat end of my old Discover card on the IHS's glue. Here's the bare CPU PCB/Die and the IHS:
Made sure to tape off the CPU die for applying my CLU (CoolLaboratory Liquid Ultra) I also tapped off the IHS:
Applying the CLU, was interesting since its got this strong surface tension, I placed a drop of it on the surface and had to spread it around with the little brushes provided with the CLU TIM until it covered the whole surface:
Both surfaces nicely covered, i used the brush to lap up any extra to make the surface as uniformly flat as possible, removing any excess CLU TIM:
the tape surprised me and it worked prefectly once I peeled it off, any CLU that overapplied onto it, came off with the tape and it was a perfect application. And according to other's CLU application results, a good application like this, will last the life of the CPU.
Reinstalling the CPU WITHOUR SEALING THE IHS...
NT-H1 applied (NOTICE ANYTHING???) I was able to reapply the Heatspreader to my RAM stick closest to the CPU socket as my LP53 wasn't pushing it aside that bad.
LP53 back in.
Everything back in place!
So, how did the Delid go, you may wonder? ....
PRE/POST DELID TEMP RESULTS DISCLAIMERS:
1) AMBIENT TEMP was 74F/23C at all times
2) * means my Z270i's VRM heatsinks were removed
3) Results were taken with the S4 Mini Side Panels On, in a vertical position (power button near the bottom)
4) Synthetic Benchmark temps taken after 10 minutes, or until i hit 99C (with timestamps included)
5) Gaming all done at 1080p in most CPU intensive area I could find, or using the game's built-in benchmark
Here's how I gathered my temps: My temp numbers were taken from the hottest core (core0) noting the highest temp that core had hit during the 10 minutes of continuous load from said benchmark/game. If the benchmark/game was showing erratic spiking of temps up/down within a range, I would take the average hottest temp of that range (ex: if spiking from 75-80c I'd average that to 77c).
- averaged temps marked with a "~" RealBench, AIDA64,and Watch_Dogs 2 fluctuated within a Temp Range, and the temp noted is an average.
- the orange temp delta shows the difference between the LP53 and the LP53 with the delid
- the GREEN temp delta is used to show the difference between my "stock NHL9i" temps and my new "LP53 w/Delid" temps
Synthetic Benchmarks:
Prime95_v26.6____IDLE__________LOAD
NHL9i Stock_______39c_________99c@5min (FAIL) on core0, cores1-3 averaged 93c
NHL9i *___________39c_________97c on core0, cores1-3 averaged 91c
LP53 *____________36c_________92c on core0, cores1-3 averaged 86cc
LP53 & Delid*______35c_________79 on core0 , cores 1-3 were ALL at 77c (13c from pre-delid LP53) (-20c from "NHL9i Stock")
Realbench Stress___LOAD
NHL9i Stock_______~99c@2min (FAIL)
NHL9i *___________~94c@
LP53 *____________~90c@
LP53 & Delid*______~76c@ (-14c from pre-delid LP53) (-23c from "NHL9i Stock")
x264 Custom Load___LOAD
NHL9i Stock______99c@2min (FAIL)
NHL9i *__________97c@10min
LP53 *___________ 95c@10min
LP53 & Delid*_____ 78c@10min (-15c from pre-delid LP53) (-21c from "NHL9i Stock")
Intel Burn Test (Standard) Hottest Core after 10 runs
LP53 *___________99c@1min (FAIL)
LP53 & Delid *____75c (-24c from pre-delid LP53)
AIDA64 System Stress Test (15 minutes)
LP53 *___________~92c
LP53 & Delid *____~78c (-14c from pre-delid LP53)
Thats an average 14c temp drop (on the hottest core) from from pre-delid LP53 to Delidding, ALONE.
But let's look at where I started with the NHL9i w/VRM heatsinks on... then this becomes a total average of 20C temp drop (on the hottest core) on these synthetic benchmarks, going from an NHL9i, to an LP53 without VRM heatsinks, and then a CPU delid!
Gaming Benchmarks
Watch Dogs 2:
LP53 *___________~83c
LP53 & Delid *____~69c
Mirror's Edge Catalyst:
LP53 *___________77c
LP53 & Delid *____63c
Batman Arkham Knight:
LP53 *___________70c
LP53 & Delid *____57c
Gears of War 4 Benchmark:
LP53 *___________61c
LP53 & Delid *____55c
Rise of the Tomb Raider Benchmark:
LP53 *___________64c
LP53 & Delid *____59c
That's an average 13c temp drop (on the hottest core) in these Real-world PC-game tests!
BONUS Pics of adjustments to the Motherboard VRMs (which were bare after removing the stock VRM heatsink towers). You can buy them here
Here's the little heatsinks I saw another user from the DanA4 Thread on [H]ardOCP use on this Z270i STRIX. I bought them and my order of 10 came in. They have a paper you peel with some black thermal glue on the back:
Bare VRMs before, after, and what I used to apply the heatsinks:
I used tiny pliers to apply them, then once in place i pressed them down lightly with my pinky until i noticed they were set.
The heatsinks have not fallen off after 24H of use sitting in my S4 with the vertical position.
BONUS pics of adjusting my rear M.2 mounting with the proper spacer:
When I initially installed my boot M.2 drive on the back of my Z270i I didn't use a spacer. This had now been fixed.
So... the delid was sucessful. I'm going to attempt to go for 5GHz now and see if I'm stable at a low enough vcore and temps. If not I'm fine with thie 7700k at stock turboboost clocks (4.5GHz @ 1.2vcore)
Thanks to all of you who helped me with this build in this SFF community of ours. This brings the main phase of this build's progress to an end... FOR NOW!
Things that might make me update this thread in the future:
- A new 1080ti Mini from ZOTAC that fits in the S4 mini ...? (praying for one at least, if not for me but so Josh can get more customers! and more can join the SFF or S4 Mini community!)
- The release of PC 4K/UHD Bluray PC Players, so I can drag all my 4K blurays that havent been ripped yet along with me when i take it with me for movie nights.
... and now that the latest version of CEMU is relased, I'm off to play Zelda Breath of the Wild in Glorius 4K! and then take the S4 to a friends house for some Mario Kart 8 4-player couch-multiplayer in Glorious 4K on his projector!
Last edited:
Sorry but I don't quite follow you... 19c?
I will clarify in that post that the "(-c)" measurements in green are comparing the TOP "NHL9i Stock" temp result with the final "LP53 & Delid* " result. I just did quick math of 99 - 77 = 22, i believe that is correct, right? Woops I put 21 instead, my bad. Fixed to -22c
I got the 19c difference from the 97c@10min minus 78c@10min, but the way you got it works too.
oh,crap, youre right i was measuring the wrong timestamped temp for that. Thanks for catching that!
- I've fixed that and added clarification in my Delid temps post. I've been looking at a scribbled peice of paper with all my temps on it for hours so I've been typing them out and probably got confused.
- I also edited my post to reflect the 2 different temperature deltas im showcasing in the synthetic benchmarks. the orange temp delta shows the difference between the LP53 and the LP53 with the delid, the GREEN temp delta is used to show the difference between my "stock NHL9i" temps and my new "LP53 w/Delid" temps I also noted where I used averaged temperatures with a "~" symbol
- and I removed the redundant temp timestamps of temps 2min into a benchmark when the benchmark completed for the full 10min without hitting 99c
- OH and I also added a link to purchase those mini VRM heatsinks in my VRM heatsink BONUS adventure, for my Z270i bros!
Last edited:
So it looks like you have done some extensive testing there! Nothing beats the lp53 with the noctua fan? I am assuming the lh9i doesn't but I didn't see much about the id-cooling is-vc45. I am guessing that wasn't a contender either ?
oh,crap, youre right i was measuring the wrong timestamped temp for that. Thanks for catching that!
- I've fixed that and added clarification in my Delid temps post. I've been looking at a scribbled peice of paper with all my temps on it for hours so I've been typing them out and probably got confused.
- I also edited my post to reflect the 2 different temperature deltas im showcasing in the synthetic benchmarks. the orange temp delta shows the difference between the LP53 and the LP53 with the delid, the GREEN temp delta is used to show the difference between my "stock NHL9i" temps and my new "LP53 w/Delid" temps
- and I removed the redundant temp timestamps of temps 2min into a benchmark when the benchmark completed for the full 10min without hitting 99c
- OH and I also added a link to purchase those mini VRM heatsinks in my VRM heatsink BONUS adventure, for my Z270i bros!
@CubanLegend : You know if anyone has tested the LP53 w./ a Noctua 92 25mm fan? My LP53 in enroute, I can test it sometime when it has arrived.
It's not that nothing beats it, its more that nothing is more efficient at under 50mm tall, (because it has to be that small for it to fit inside an S4 mini). There's other coolers that are 60 mm that fit inside the DanA4 case that are better than the LP53, but they obviously wont fit inside the S4 because of height, and also some of those coolers have RAM clearance issues.
And about the ID-Cooling IS-VC45... a vapor chamber is a nice concept, but @Josh | NFC already tested the ID Cooling VC45 here in his video, and it DID NOT outperform the NHL9i, so I never bothered with it since I already had an NHL9i.
No I have not tested the LP53 with my 25mm Noctua NF-B9 Redux PWM fan, but in post #12 I tested it with the NHL9i and it BARELY fit inside the S4 mini, the fan was FLUSH against the side panel and cause serious air turbulence, and as a result it did NOT perform better than the stock A9x14. And since it barely fit with the NHL9i, I can tell you FOR A FACT that ia 25mm fan WILL NOT FIT on top of the LP53, inside the S4 Mini & with the side panel closed, Sorry!
Thanks for all the info you've put up here on your build!
I'm in line for Josh's 3rd batch of S4's and already have my 1080 on order and looking to send my CPU off to Siliconlottery for a de-lid as I do not trust myself to do it ha. Did you ever try the de-lid with the Noctua perchance? Unsure if I want to go the LP53 route right now, worst case I thought about maybe slowing my 6700k down to 3.6-3.8 and see if it can stay nice and cool off the stock NHL9i.
The waiting game is killing me >.<
I'm in line for Josh's 3rd batch of S4's and already have my 1080 on order and looking to send my CPU off to Siliconlottery for a de-lid as I do not trust myself to do it ha. Did you ever try the de-lid with the Noctua perchance? Unsure if I want to go the LP53 route right now, worst case I thought about maybe slowing my 6700k down to 3.6-3.8 and see if it can stay nice and cool off the stock NHL9i.
The waiting game is killing me >.<
You're welcome, I always wanted to share AS MUCH info as possible to help others looking into buying S4 Mini to see what's possible.
I'm glad any info I shared helped you out, and I can't wait to welcome more bros into the S4 Mini/1080 Mini club! Onto your questions: I did not try the delidded CPU with my Noctua NHL9i... Sorry! Here's why: only because I was looking to find out if the LP53 or the NHL9i was better alone (pre-delid), then after determining that... I would use the winner from that point forward for my delid and put away the losing CPU HSF. The LP53 won by 3 degrees at idle and 2-4 degrees in benchmarks. So yeah.
What motherboard will you have? That's the key to determining if the LP53 will fit. And you shouldn't need to slow down your 6700k at all, as long as you dont OC it before your delid you'll be fine. And once you DO delid your temps will be perfect, and you might even gain some overclocking headroom inside the S4 Mini with that K chip.
By that way is this a new build or are you transferring from another PC? I recommend the 7700k if it's a new build. It had slightly better temps than the 6700k, has a higher turboboost clock and more overclocking headroom, plus futureproofing for Intel Optane and 4K UHD streaming, h264 and h265 encoding is better than the Skylake series, and PC 4K UHD Bluray player support in the future.And yes, the wait will kill you, take the wait time to plan out your build, get all your cable mods done and parts ordered. Use the time wisely to plan it out.
What motherboard will you have? That's the key to determining if the LP53 will fit. And you shouldn't need to slow down your 6700k at all, as long as you dont OC it before your delid you'll be fine. And once you DO delid your temps will be perfect, and you might even gain some overclocking headroom inside the S4 Mini with that K chip.
The board is a Gigabyte H170N Wifi.
It's a transplant from my Current itx build. Wallet can't survive another upgrade for a while so this'll have to do sadly.
Ah i see that board, from the look of it, it doesnt have any VRM heatsinks to interfere, you should be good, just try to mount the LP53 the same way I have mine mounted. What fan will you be using? I recommend the NHL9i's tock fan, the NF-A9x14 PWM, whisper quiet at idle.The board is a Gigabyte H170N Wifi.
It's a transplant from my Current itx build. Wallet can't survive another upgrade for a while so this'll have to do sadly.
And if you ever plan to delid it with a Rokkit tool (~$30), you can extend the life of your 6700k as it'll run at lower temps at load and overall, so you can match 7700k performance and not have to spend more money ($340).
Ah i see that board, from the look of it, it doesnt have any VRM heatsinks to interfere, you should be good, just try to mount the LP53 the same way I have mine mounted. What fan will you be using? I recommend the NHL9i's tock fan, the NF-A9x14 PWM, whisper quiet at idle.
And if you ever plan to delid it with a Rokkit tool (~$30), you can extend the life of your 6700k as it'll run at lower temps at load and overall, so you can match 7700k performance and not have to spend more money ($340).
I may give the LP53 a go then if I do not need to modify or move anything (I thought I'd have to pull the heat sinks off my ram)
I didn't really trust myself to mess with the de-lid so I was thinking of having siliconlottery.com do it, they charge $50 but also replace the chip if they break it. Definitely going to get it done tho, I want this running as cool as feasibly possible under the circumstances
I might have overlooked a picture that shows how the LP53's stock fan is attached. Is the Noctua fan a different size? Or, why did you need to use the zip-ties?
Did you end up testing different fans with the LP53? Writing this out I feel like you did, but I didn't see it after a quick look. I'll go double check again...
Did you end up testing different fans with the LP53? Writing this out I feel like you did, but I didn't see it after a quick look. I'll go double check again...
Stock fan uses rubber mounts with a fixed undersized spacing:
The Noctua fan does not line up with these more closely spaced mounts, so cable ties (or similar) need to be used:
What RAM will you be using by the way? And the LP53 shipping period is 3-6 weeks (got mine in 4) so you'll have plenty of time while you wait for your S4 mini.I may give the LP53 a go then if I do not need to modify or move anything (I thought I'd have to pull the heat sinks off my ram)
I didn't really trust myself to mess with the de-lid so I was thinking of having siliconlottery.com do it, they charge $50 but also replace the chip if they break it. Definitely going to get it done tho, I want this running as cool as feasibly possible under the circumstances
Good idea with the Delid, if you dont feel comfy doing it yourself, no harm in paying for it to be done properly. The 6700k is a great CPU do you should be all set once its delidded. And by the way GamersNexus did a nice revisit of the old 2600k pitting it against today's top CPUs like the 7700k stock with HT on and OFF, and also with a 5.1GHz OC. And the difference between the stock 7700k and the OCed one... was only like 4 fps out of liek 115 FPs (at 1080p), which is nothing... So the stock 7700k performance is really decent that if you delid it or a 6700k you might not even need to overclock it and can just reap the rewards of a more cool running CPU! Yeah what @EdZ says below is correct, the LP53's stock fan is WAY too big for the S4 Mini, and the A9x14 is perfect for it once you mount it with zip ties, I've also seen others mount it with rubber bands, but Zip ties are much more secure IMO since you can tighten them down securely click by click.
And regarding your question about if i used other fans with the LP53, I did not. The A9x14 was the only one that I had that would fit inside the S4 mini so i only tested it. I didnt test the 120mm slim scythe or the Noctua Redux I had previously tested with the NHL9i, sorry!
Stock fan uses rubber mounts with a fixed undersized spacing:
The Noctua fan does not line up with these more closely spaced mounts, so cable ties (or similar) need to be used:
I mean, you'd only end up pushing the A9x14 up towards the S4 mini's side panel if you added any more mounting hardware... it WOULD cause more turbulence and probably lower performance and raise sound due to more wind "wooooshing" close to the case side panel. You could always try beige colored zip-ties like in that Dan A4 picture EdZ posted, it's from the LP53 thread on [H]arcOCP btw, where I learned about the Z270i fitment possibilities.
I used black zip-ties because those were the ones I had available, plus they looked nice and stealth like and dont stand out.
Last edited:
Curious, has anyone thought about the Thermolab ITX30 and slapping a Noctua 80mm Redux on it?
If my math is right, the Noctua L9i is 23mm without the fan, the ITX30 is 20mm, if the L9i with a redux barely fits, then the extra 3mm of space should clear with the ITX30 (total of 45mm)? It is also rated up to 100 watt procs. CubanLegend did you look at that at all before the LP53?
Partly want to try and get as much air flow in as possible but staying relatively quiet, and also avoiding the brown
If my math is right, the Noctua L9i is 23mm without the fan, the ITX30 is 20mm, if the L9i with a redux barely fits, then the extra 3mm of space should clear with the ITX30 (total of 45mm)? It is also rated up to 100 watt procs. CubanLegend did you look at that at all before the LP53?
Partly want to try and get as much air flow in as possible but staying relatively quiet, and also avoiding the brown
Thats a very creative idea. I will admit it did cross my mind, but I'd only really seen tests of the NHL9i against the LP53, none against the ITX-30 and the NHL9i... you could order that ITX-30 and zip tie that 25mm Noctua Redux PWM fan on there, I'd be interested to see the results, I'm sure others would too.Curious, has anyone thought about the Thermolab ITX30 and slapping a Noctua 80mm Redux on it?
If my math is right, the Noctua L9i is 23mm without the fan, the ITX30 is 20mm, if the L9i with a redux barely fits, then the extra 3mm of space should clear with the ITX30 (total of 45mm)? It is also rated up to 100 watt procs. CubanLegend did you look at that at all before the LP53?
Partly want to try and get as much air flow in as possible but staying relatively quiet, and also avoiding the brown
If the ITX-30 is a full copper like the LP53, it'll probably be a great combo with that fan, and it'll avoid the beige A9x14 fan.
Last edited:
