Project BSG Scorpion Fleet Shipyards (Lian Li PC-A77B)

[luke997]

Rev. 1.2 done.

Good news:
- T-Virus with screw-in caps is awesome!
- Flow is now 200l/h - no idea why, only change is the res and orientation of flow meter

Bad news
- New battery - old problem - still BIOS reset every time.
- Had only 1 more spare BP link so no parallel this time.

Not sure now what to do (anyone?).
Probably will try rebooting some more tomorrow and if can't get it to work - RMA or a new board...

Pics after dusk

[Waterlogged]

Rev. 1.2 done.

Good news:
- T-Virus with screw-in caps is awesome!
- Flow is now 200l/h - no idea why, only change is the res and orientation of flow meter

Bad news
- New battery - old problem - still BIOS reset every time.
- Had only 1 more spare BP link so no parallel this time.

Not sure now what to do (anyone?).
Probably will try rebooting some more tomorrow and if can't get it to work - RMA or a new board...

Pics after dusk

So wait...your saying the flow meter was backwards? That might be part of the reason for the drop. Another might have been a bubble stuck somewhere, they can easily kill flow rate.

[luke997]

So wait...your saying the flow meter was backwards? That might be part of the reason for the drop. Another might have been a bubble stuck somewhere, they can easily kill flow rate.

Nooooo Beside - it can be used in any orientation.
Just previously it was mount horizontally, between rad & res, now it is vertical between rad & rad - and actually new res required 2 more 90 fittings.
Probably like you're saying, bubble - can't see other explanation.

[kvadzio]

fabulous work down there luke

but if i'd to change a thing - cpu area is crowded a little :P

[luke997]

fabulous work down there luke

but if i'd to change a thing - cpu area is crowded a little :P

Thanks

Still have BIOS problems.
Tried 0813 - no joy, back to 0703 - no joy.

Looks like MemOK is kicking off every time - so tried 1 RAM in first slot only - same on each so looks like RMA

I'm a bit tempted to get Rampage 3 - if only EK waterblock was all black...

[ScottALot]

Just noticed those were 480s and not 5870s... I'd like to see some benchmark scores, as I've heard the tri-SLI from 480s is great!

[luke997]

Just noticed those were 480s and not 5870s... I'd like to see some benchmark scores, as I've heard the tri-SLI from 480s is great!

46159 Vantage P PhysX OFF

[luke997]

Last pic before mobo goes out for RMA.

Apologies for the poor quality - but had to show the T-Virus.
In reality though it looks sooo much better - just couldn't capture it right.



Thanks to FrozenQ for smoothly solved RMA - I can recommend the new screw-in res - brilliant product!

[ScottALot]

So much work and you have to send the mobo in... one of the bigger ever.

[luke997]

So much work and you have to send the mobo in... one of the bigger ever.

Yeah, that sucks but I can't complain, was expecting more RMAs with that many components.

Should be back in business next weekend - P6X58D-E on the way

[luke997]

I've finalised spec for the Rev 1.3, all the gear is coming on Saturday, if time permits I should have all work done by Sunday night.

1) 2xDDC 3.25 (18W) + EK Dual Top
(yes, decided to listen to the wise ones

2) New mobo - P6X58D-E

I'm looking forward to see how flow is going to improve - as I'm not changing any blocks/tubing, it should be a good Dual D5 vs Dual DDC comparison.


I'm was also considering swap from MCR220 to RX360 in the back, but unless someone can convince me it'll be much better (>50% ?) - I'll skip on this one.

[luke997]

What the heck on a second thought, getting RX360 for the back

It's XS after all

[luke997]

Quick update to the Rev 1.3 spec

3) GPUs in parallel

As couple of you guys advised me to try it (thanks again!), I'm going to include this in this revision too.

Originally wasn't in plan only because I had nothing else to order from highflow but since I can get DDC 3.25 (PMP-400) only from there, I've got more sli links too.

Really looking forward to see the flow with these improvements. Just hoping DDCs won't be too loud.

The only downside is this will push completion date for next week as I won't have all parts until Monday - but hey, good things come to those who wait

[luke997]

Quick questions - I'm doing parallel flow for GPUs - or at least how I understand it

Could you please confirm that this is right and I will not end up having GPU2 temps higher than GPU1 or GPU3 for example?

Basically I plan to place SLI links into the blue (inlet line) and red (outlet line) areas:



Sorry to ask it again but I want to do it right this time and with the photo it should be clear what I'm doing


On a side note - back MCR220 has been swapped for RX360 and mobo transplant has been completed too - but the patient is still in the coma
Wake up time planned on Tuesday when my PMP-400 come from NL.

[Church]

Yup, you are right on parallel connection. Example.

[luke997]

Yup, you are right on parallel connection. Example.

Thanks man!

[Johnny87au]

Yup, you are right on parallel connection. Example.

So having the 2 middle ones is parallel? i thought just running 1 d-plug left then bottom right then bottom left is fine, Does it help with temps running them together in parallel?

[Church]

Johnny87au: imagine that those middle connections are like two F type connectors that are embedded within blocks, that have one inlet and few side branches within full cover. Just take care to not make mistake and don't put both inlet and outlet on one side, this way is almost straight pass-through, with all liquid going throught that instead of going through more resistive waterblock inside path
(Yeah, i recall one guy mistaken that way and how he wondered why so high temps for gpus )
As for what you thought - from three gpus one d-plug on the left and among other two one to the right - it's what serial connection is (where water goes in 1st block, through it, then through 2nd, then subsequently through 3rd. - in series, with full summed restriction of all three full covers affecting flow). In parallel flow divides three ways with 1/3rd of flow going simultaneously through each block, and then all three uniting back again. This way total flow restriction of all gpu part of the loop is even less then that of single block because of much larger summed up cross-section. Yeah, you get part of flow through each block, but gpu water blocks are not that much flow sensitive, and it partially is compensated by better flow overall in loop, and with probably a bit better cooling of CPU if it's in same loop.

[Johnny87au]

Ah thanks churchy didnt realise parallell SLI would actually run better, I was thinking of just using D plugs left right left right etc as i see a tonne of people doing it that way, I might try both setups and see if their really is a difference in temps, as for inlet and oulet their really isnt a inlet or outlet on gpu blocks.. thanks mate

[luke997]

Rev, 1.3 done - still bleeding the loop, but so far I'm very disappointed - to say the least.

MCR220 --> RX360 (less restriction)
GPU serial --> GPU parallel (less restriction)
2xD5 --> 2xDDC (more head)

And what's the result? 205 l/h (previously I've had 200l/h with fully bled loop).

I'm fairly sure it won't get better as it actually started at ~290/h, the more bleeding -> less flow.

So unless I'll see jump to 300l/h tomorrow when I'll finish bleeding (or I''m doing something terribly wrong) - I say the DDCs are much overrated with all the head pressure.

The only good thing about them is they are not as noisy as I thought, just a little, little bit more noise (if you put the ear near them).

[luke997]

In all fairness to DDC I should have added that the flow is better with 1 pump on - 160l/h vs 130l/h on D5.

[Biffa]

Heh you are having a hell of a time mate. Sounds like we are in the same boat, very similar heat loads, little bit similar cases but of course your look and artistic audacity far outstrip mine

Look forward to your modifications, I'm just about to strip mine out for a rethink.

BTW I know flow is the goal for you but what are your temps like? What I mean is do you have such bad temps that you are worried about flow? My temps are not great on the CPU so I'm worried about everything But I'd be less worried about flow rate if my temps were good.

[luke997]

Heh you are having a hell of a time mate. Sounds like we are in the same boat, very similar heat loads, little bit similar cases but of course your look and artistic audacity far outstrip mine

Look forward to your modifications, I'm just about to strip mine out for a rethink.

BTW I know flow is the goal for you but what are your temps like? What I mean is do you have such bad temps that you are worried about flow? My temps are not great on the CPU so I'm worried about everything But I'd be less worried about flow rate if my temps were good.

Thanks

I can't complain about my temps, are quite good (posted earlier) but I got the WC bug now hence the better flow pursuit

Looks though I've did all I could so it'll stay the way it is

Will post new temps once I've done some benches with the new rad at the back.

[Johnny87au]

IM sure you would notice a difference in overall temps if you try and hold off on the 45 and 90 degree fittings, try and use as least as possible.. I really only use them when the timings right, for rad or reservoir, hardly ever for chipset, mosfets etc..

[josk1]

IM sure you would notice a difference in overall temps if you try and hold off on the 45 and 90 degree fittings, try and use as least as possible.. I really only use them when the timings right, for rad or reservoir, hardly ever for chipset, mosfets etc..

couldn't agree more... but hell your rig's sexy