7700k delided CPU delided

8 Pack · 10 May 2017 at 13:46

We have a new QC procedure on all Binned chips now to make sure all are benefiting fully from Delid and none slip through the QC net.

voidshatter · 10 May 2017 at 14:53

SniperDude said:
the one I have is L649f773..... the 9 could be a 0 or a J its hard to make out

Outcasst said:
This is the CPU I returned last week.

I purchased one from the 5.1GHz bin, and received one with exactly this serial number. I don't know whether they re-lidded it and sent it to me, or they simply by chance put the same lid on a new chip and sent it to me. For this sample I received, I consistently get 10C hotter on the third core, which results in unstable overclocking with the recommended settings (BSOD after 15min of non-AVX Prime 95 with 1.36v despite the 1.35v written on the box), and unfortunately had to return it.

Glad to see that 8Pack is now looking into this to improve QC.

8 Pack · 10 May 2017 at 15:12

you fail to mention your cooling is not to recommended standards or your board!!! ^^^^ Ragtag!!!! Also 1.4v is fine on these chips and it says so in the description so 1.36v is not a reason to return...

voidshatter · 10 May 2017 at 15:49

My other CPUs at 1.35v don't get 10C hotter on one core. There was reasons for the others to return this serial number. Maybe the liquid metal inside the lid can move during shipment so to reduce overclockability? I'm really disapponted that I spent a premium to buy something returned at least twice by at least two other customers.

8 Pack · 10 May 2017 at 16:05

But your cooling was not the recommended level nor your board so how can u comment?? Your commenting on others experiences...I can show you many examples of Intel TIM CPU that are 10C delta across cores.

We sold 100's of the 4.9 and 5g CPU and received very few back in the grand scheme of things. We now have much more robust QC to ensure RMA rates are even lower still.

DragonQ · 10 May 2017 at 16:13

8 Pack said:
Also 1.4v is fine on these chips and it says so in the description so 1.36v is not a reason to return...

It is if it was advertised to be stable at 1.35 V.

8 Pack · 10 May 2017 at 16:15

Its advertised to run 1.32v to 1.39v, thanks for your useful input!!!

jrwagh333 · 10 May 2017 at 16:19

Sells premium product - Gets returned - Gets angry at return because it isn't exactly as advertised. Awful

DragonQ · 10 May 2017 at 16:26

8 Pack said:
Its advertised to run 1.32v to 1.39v, thanks for your useful input!!!

OK then? I said if, based on voidshatter's assertion that it said "1.35 V" on the box.

Your hyper-defensive attitude really doesn't help sell these parts.

8 Pack · 10 May 2017 at 16:32

Its not defensive its fact.

voidshatter · 10 May 2017 at 16:45

8Pack may have a point on my cooling and motherboard.

Recommended hardware to get the best from the CPU:
- Motherboard in ASUS Strix, ROG range or Gigabyte Aorus.
- AIO cooler with 240mm Rad or higher, lesser cooling may work fine but is not guaranteed to.
- Quality Gold rated or above PSU providing enough quality power for the entire system.
- Waterloops may see higher stable overclocks

My PSU is a Corsair AX860i - more than enough.

I have an ASUS Maximus VIII Gene motherboard, which is in the ROG range, and was also once used in 8Pack's "Approved - Elite Tier SkyLake MATX Bundle: Intel i7-6700K @ 4.50GHz". Maybe he should have specified that Z170 motherboards are not supported.

My cooling is a Noctua NH-U12S Ultra-Quiet Slim CPU Cooler with NF-F12 Fan, which is well below the recommended AIO cooler with 240mm rad. I agree that this could be a decisive factor affecting the overclocking, but I'm very upset with the temperature discrepancies among cores. I didn't know that 81C could be problematic and introduce BSOD after 15 minutes of non-AVX Prime 95 (the 7-minute screenshot below was taken before the BSOD). Keep in mind that 8Pack claimed to have binned the chips before de-lidding, so I strongly doubt whether 81C should be an excuse here.

I'm fine with lower vcore and lower frequency, and even settle with -4 AVX offset if my motherboard supported it at all, but OCD is killing me for the 10C delta among cores. As a previous user of the i7-980X where temperature was quite even among all 6 cores, I just feel that further quality control can be made for these quad-core CPUs.

Besty · 11 May 2017 at 13:04

just to add another opinion.

You really need a double rad on Kabylake if going for 5GHZ or over, this limitation should not be attributed to 8 Pack.

I got my binned CPU from another source but they also stipulate a Maximus ROG VIII Hero and Corsiar H105 AIO cooler with 240mm rad as minimum specification. If you don't have either of these and are going for 5GHZ or over then don't expect to be able to replicate the same results as in the lab with either clockspeed or voltages. Again unfair on 8 pack if you don't have the recommended board.

For example on my Gigabyte board, they (Gigabyte) undervolt across the range, its just the way they roll so in this example 1.35-1.36v may not be enough at 5.1 even if you have the AIO. With a 5.2-5.3 binned chip it may work but milage is gonna vary.

SiDeards73 · 11 May 2017 at 13:17

You should all be grateful that OCUK even offer this service tbh, according to Intel its not even supported to OC a CPU, regardless of them selling locked and unlocked models and advertising them for OC'ing potential etc...

Personally i doubt i will be buying any more Intel CPU's for the time being, but if i bought a pre binned OCUK CPU i would make damn sure i met the recommendations of the seller with regards to other hardware to achieve that clock, without that in place you have no leg to stand on in regards to arguing.

8 Pack · 11 May 2017 at 17:28

4.9ghz - 5.2ghz CPU now reduced prices...

voidshatter · 11 May 2017 at 18:57

Besty said:
just to add another opinion.

You really need a double rad on Kabylake if going for 5GHZ or over, this limitation should not be attributed to 8 Pack.

I got my binned CPU from another source but they also stipulate a Maximus ROG VIII Hero and Corsiar H105 AIO cooler with 240mm rad as minimum specification. If you don't have either of these and are going for 5GHZ or over then don't expect to be able to replicate the same results as in the lab with either clockspeed or voltages. Again unfair on 8 pack if you don't have the recommended board.

For example on my Gigabyte board, they (Gigabyte) undervolt across the range, its just the way they roll so in this example 1.35-1.36v may not be enough at 5.1 even if you have the AIO. With a 5.2-5.3 binned chip it may work but milage is gonna vary.

The only requirement I miss is the 240mm rad. However, if 8 Pack did the binning before delidding, how come the CPU is unstable at 81C? Do you mean all cores of the CPU can stay below 81C without delidding if it's with 240mm rad? As someone who has tested a couple of 6700K's, I strongly doubt that.

Besty · 11 May 2017 at 20:30

Here is my experience so far. I will get back to you on final results tomorrow.

I got a Corsair dual-rad AIO with the smaller fans inside a very small MITX portable, lan-party build.

I just dropped in a delidded 5.2 bin 7700k from a non-delidded or binned 6700k (4.8GHZ @ 1.36v).

Looking at the hwbot sales pages I thought my 6700k was just above average - not great by any means, the 6700k imho is a very hot chip at 4.8GHZ+

Have not had time to play today as GF is here but the 7700k benchloops 5.1 @ 1.34v with 3600 C16 (stock) average temp is <70C on quiet pump mode with a fan curve. I need to stress test it so volts may go up.

This is way beyond the reaches of my 6700k which would temp spike at 4.8GHZ >1.36v.

I went for 5.3GHZ @ 1.44v but as the chip goes past approx 78C on Intel bench my USB starts to reboot itself so on face value, I will need to look to see whether its the temp causing that or system voltages.

On the 5.1 bin you have I would imagine that the chip will be slightly hotter than mine but not massively so if its been de-lidded . Running on a dual-rad AIO with a good fan and pump profile I cant see why <80C is not possible - I will get back to you.

Build log for my rig:

https://forums.overclockers.co.uk/threads/bumblebee-build-miniitx.18726318/

voidshatter · 11 May 2017 at 21:06

You completely missed my point. If 8 Pack did the binning before delidding, then the CPU should have passed 48 hours of stress test at above 81C (based on my experience with Intel's TIM, that's the sort of temperature you get at above 1.36v regardless of whatever 240mm rad you use).

Now, the CPU has been delidded, (despite my inferior cooling) it does not stabilise at 81C. 81C is supposed to be a very realistic temperature before delidding. This is very suspicious! How did it pass the quality control?

Journey · 11 May 2017 at 22:10

voidshatter said:
You completely missed my point. If 8 Pack did the binning before delidding, then the CPU should have passed 48 hours of stress test at above 81C (based on my experience with Intel's TIM, that's the sort of temperature you get at above 1.36v regardless of whatever 240mm rad you use).

Now, the CPU has been delidded, (despite my inferior cooling) it does not stabilise at 81C. 81C is supposed to be a very realistic temperature before delidding. This is very suspicious! How did it pass the quality control?

I'd imagine that the pre delid quality checks were/are very fast, and not that labour intensive. Whack it in a pre-built test rig/bench and see if it works at 4.9GHz or whatever, then down clock until you find out where it is stable for 15 minutes, then pop it in the correct tray.

It may have become damaged during the handling, it has happened before, all the caps looked OK etc? No one is perfect, and human error can creep in to anything, especially when you are repeating the same task for hours on end. I'd chalk it up to a bad experience, but it helps highlight the QC Process and where improvements can be made.

Besty · 11 May 2017 at 23:30

well only one way to find prove it.

void needs to strap a proper cooler to his CPU and get a better MB

voidshatter · 11 May 2017 at 23:49

Besty said:
well only one way to find prove it.

void needs to strap a proper cooler to his CPU and get a better MB

Not an AIO water cooler again cuz I hate the clicking noise issue that may randomly appear from the water pump after some usage.

As for motherboard, yes, please update the product page to clarify that my £199.99 ROG motherboard is too cheap for the CPU. After all, it's £10 cheaper than your vendor's, but still £40 more expensive than yours.