Thermal Grizzly Mycro Direct-Die CPU Water Block - Sockel AM5

[hiroofheroes]

8 Pack "shower head" watercooling

I recommend everyone watch KitGuru's video of 8 Pack showing his new "shower head" water cooling tech. A genuine new approach and very interesting. Hopefully we will get some more insight from him on this tech when he gets back from Computex on this forum.

forums.overclockers.co.uk

I saw him talk about it either last year or early this year when he popped up in a YouTube video so it doesn't seem like the project is completely dead, but I wouldn't hold off any purchases waiting for one at this point.

 

[Jamin280672]

8 Pack "shower head" watercooling

I recommend everyone watch KitGuru's video of 8 Pack showing his new "shower head" water cooling tech. A genuine new approach and very interesting. Hopefully we will get some more insight from him on this tech when he gets back from Computex on this forum.

forums.overclockers.co.uk

I saw him talk about it either last year or early this year when he popped up in a YouTube video so it doesn't seem like the project is completely dead, but I wouldn't hold off any purchases waiting for one at this point.

Thanks, looks good, 2019 unfortunately, wouldnt hold my breath

 

[Jamin280672]

So I took the plunge and brought one, not from OCuk unfortunately as they have just been on pre-order forever, got the de-lid tool too, im a bit nervous about the delidding, anyone done it yet and hows it going ?, im going to use liquid metal to remove the excess solder and for the mount, do I need to cover the SMDs with anything ?

 

[Tobywon]

So I took the plunge and brought one, not from OCuk unfortunately as they have just been on pre-order forever, got the de-lid tool too, im a bit nervous about the delidding, anyone done it yet and hows it going ?, im going to use liquid metal to remove the excess solder and for the mount, do I need to cover the SMDs with anything ?

hes done a good youtube video, yeah you need to use some nail varnish. the thermal grizzly stuff is just that relabled.

They sold out in Germany on pre order alone, so I wasnt able to get one, im waiting for the next batch

 

[Jamin280672]

hes done a good youtube video, yeah you need to use some nail varnish. the thermal grizzly stuff is just that relabled.

They sold out in Germany on pre order alone, so I wasnt able to get one, im waiting for the next batch

Caseking.de have them (OCuk sister company, im sure they wont mind me naming them based on that), but they wont ship to the UK, I was going to ask my sister.....who lives in Germany to buy me one and send it over, but managed to source some from a UK store elsewhere.

EDIT: Nope, the store in Germany are out of stock now too.

 

[Tobywon]

Caseking.de have them (OCuk sister company, im sure they wont mind me naming them based on that), but they wont ship to the UK, I was going to ask my sister.....who lives in Germany to buy me one and send it over, but managed to source some from a UK store elsewhere.

EDIT: Nope, the store in Germany are out of stock now too.

Yup, I live in Germany, that’s where I was going to buy from. And all the other German stores are out as well.

 

[RSR]

Mine arrived today

 

[Jamin280672]

Mine arrived today

Nice

My delidding seemed to go quite well, I was missing one set of bearings in the delidder on one side, however as I fly RC helicopters, they were exactly the same size as the thrust bearings for the feathering shaft on my Trex 500XT, so stole the bearings from my spares, I didnt think it was ever going to give way though, after about 20 winds and unwinds, i decided to do what debauer did and move to the drill, another 10 or 15 goes and it finally gave out, I could move the delidder by hand, opened up the delidder and it fell apart, all looked in tact when finished.

I used liquid metal to remove the remainder of the solder, just left it on for 15 mins and then rubbed it off, I had to do this twice, all the solder seemed to be gone, then mounted it in the socket to protect the pins and removed the SAM, I was a bit disappointed when I mounted the block, I triple checked myself, all images on the website show the inlet at the top and outlet on the bottom, but when it's mounted, because of the layout of the SMDs, you end up with the inlet at the bottom and outlet at the top, not sure if there is someway to change this, ie take it apart and turn it around, but if you look at the inside of the block, you'll see the longer raised part basically has to go at the top of the CPU (Gold arrow) where there's no SMDs at the edge, this puts the inlet at the bottom.

You can see all the gaps around the raised parts of the under side of the block to allow for the SMDs on the CPU, basically every website is advertising the CPU block upside down.

I'm leak testing right now for 24 hours so have yet to find out if it works lol, fingers crossed.

 

[Jamin280672]

So a bit of an update, SHE LIVES !!!, ive limited the max CPU temp in the bios to 75oC, so its no longer hitting 90oC and throttling, and its pretty much sitting / boosting to 5.5ghz constantly with the slightest move of the mouse, PBO at work, core temps idle are 28oC-30oC average and 65-70oC running Intel burn in test, but package temp is 45oC average, and max temp is 75oC, winner, winner, chicken dinner.

 

[RSR]

So a bit of an update, SHE LIVES !!!, ive limited the max CPU temp in the bios to 75oC, so its no longer hitting 90oC and throttling, and its pretty much sitting / boosting to 5.5ghz constantly with the slightest move of the mouse, PBO at work, core temps idle are 28oC-30oC average and 65-70oC running Intel burn in test, but package temp is 45oC average, and max temp is 75oC, winner, winner, chicken dinner.

What's the core to core delta like that's usually a good indication of an excellent even contact?

 

[Jamin280672]

What's the core to core delta like that's usually a good indication of an excellent even contact?

Not too bad, im probably going to do a remount at the weekend and plan the tubing out a bit differently when im not rushing it and got a bit more time, ie go straight from the pump res to the inlet at the bottom of the block and then the outlet at the top up to the rads etc, the coolant pretty much equalizes anyway so it doesnt really matter the order of things.

 

[RSR]

11.2c core-to-core delta isn't great, as I would expect around 4 to 6c for excellent contact. I assume all TIM has been removed and enough LM has been applied?

I'll have to test this at the weekend and see what it achieves.

 

[Jamin280672]

11.2c core-to-core delta isn't great, as I would expect around 4 to 6c for excellent contact. I assume all TIM has been removed and enough LM has been applied?

I'll have to test this at the weekend and see what it achieves.

Yes, I actually thought I put too much LM on, it was pooling a little.

Idle core temps were all roughly about the same, it may of been the benchmark I was running playing on different cores, it could also be difference in curve optimizer, as im pretty sure 2 or 3, and 7 or 8 were my best cores (I cant fully remember as it was a long time ago now lol), therefore have the lowest CO settings, something stupid like negative 3 or 5 whereas other cores have like negative 20, reduction in voltage = reduction in temps, however, it is strange that its 2 cores on each CCD.

When Derbauer was showing the results of the water block on his Yotube video, he was running an All Core overclock at 1.3v

 

[Jamin280672]

So I messaged Thermal Grizzly and asked them if there was anyway to disassemble the block and change it so the input was at the top and out at the bottom, heres, the messages PMSL:

 

[SirConfused]

Classic German efficiency in that customer service response though!! I'm not surprised, given the flow path they've had to engineer into it.

 

[RSR]

Mycro Direct-Die Install & AM5 Short B/P​

Loaded Temps - R23 - LM Conductonaut

CCD0 C2C Delta - 3.7c
CCD1 C2C Delta - 6.3c

All seems good and it seems to have an excellent contact, which is really good as this is the first mounting as well. When I get a bit more time, I will disassemble it and replace the LM with LM Conductonaut Extreme which from my testing leads to a 2 to 3c drop in overall temps.

 

[Jamin280672]

Mycro Direct-Die Install & AM5 Short B/P​

Loaded Temps - R23 - LM Conductonaut

CCD0 C2C Delta - 3.7c
CCD1 C2C Delta - 6.3c

All seems good and it seems to have an excellent contact, which is really good as this is the first mounting as well. When I get a bit more time, I will disassemble it and replace the LM with LM Conductonaut Extreme which from my testing leads to a 2 to 3c drop in overall temps.

Whats you load temps like ? you idles are pretty much the same as mine, all even'ish.

Ive re-laid out my cooling loop so tubes arnt crossing each other, i'll remove the block over the weekend to check the mount and remount.

CCD0 Delta - 1.9c idle
CCD1 Delta - 1.6c idle

 

[RSR]

Whats you load temps like ? you idles are pretty much the same as mine, all even'ish.

Ive re-laid out my cooling loop so tubes arnt crossing each other, i'll remove the block over the weekend to check the mount and remount.

That has load temps on the max column, which the C2C deltas are from.

 

[Joxeon]

The biggest thing for me is that now CPU's have gotten so expensive the voiding of the warranty is now a major problem. It was different when you could get a top i5 for £125 and my first delid on a 4670k was done in a vice with a hammer. Even the lowest AM5 cpu is almost £220 with the majority well over £300 up to £670. That's not worth voiding the warranty in my opinion as not everybody has that sort of money as pocket change to replace a failed/broken cpu.

It’s not even like overlocking yields much performance these days compared to times gone by, sure you can lower temps and increase the clocks a couple of hundred mhz but realistically how much actual performance will that translate into outside of benchmarking.

 

[SirConfused]

It’s not even like overlocking yields much performance these days compared to times gone by, sure you can lower temps and increase the clocks a couple of hundred mhz but realistically how much actual performance will that translate into outside of benchmarking.

You're not wrong, but it's not always about the real world performance gains for some folks. It's still a hobby, and some people just like tinkering to make number go big .