Indigo Xtreme reflow xtremely bad

rsquared

Limp Gawd
2FA
Joined
Jul 5, 2007
Messages
497
So I took the plunge and bought some Indigo Xtreme. I hadn't been overly impressed with the performance of my EK Supreme HF waterblock, and hoped the better TIM would improve matters. Followed all the directions, including turning the screws exactly five turns, and using SpeedFan and everything. It went exactly as the instructions said it would. My temperatures didn't improve, though.

A week or so passed. My motherboard had been flaky, but it finally crossed the line; time for a new one. I took everything apart in preparation for the rebuild and found this:

indigo.jpg

It had just barely flowed onto the CPU.

What did I miss? Was the recommended "five turns" not enough? Did I not let it run long enough with the pump off? Does the waterblock need to be dry for this? Is it worthwhile to try again (on the new motherboard) with the other Indigo Xtreme application?
 
The loop was filled, and the motherboard was level.

I was suspicious about the loop being filled, that the water would absorb too much heat, but the Official Instructions say only to turn off the pump, nothing about having the waterblock be dry.

So evidently I will need to do like this:
  • build the system with a stock heatsink and the on-chip video (since my video card is watercooled, too)
  • install Windows and SpeedFan
  • take everything back apart to install the CPU waterblock and its backplate—but no tubing or water
  • do the IX reflow
  • install the rest of the water-cooling system, including the video card
  • then fill and bleed the water loop
Is IX really better enough to justify this? And how will I know if it's reflowed correctly?
 
The loop was filled, and the motherboard was level.

I was suspicious about the loop being filled, that the water would absorb too much heat, but the Official Instructions say only to turn off the pump, nothing about having the waterblock be dry.

So evidently I will need to do like this:
  • build the system with a stock heatsink and the on-chip video (since my video card is watercooled, too)
  • install Windows and SpeedFan
  • take everything back apart to install the CPU waterblock and its backplate—but no tubing or water
  • do the IX reflow
  • install the rest of the water-cooling system, including the video card
  • then fill and bleed the water loop
Is IX really better enough to justify this? And how will I know if it's reflowed correctly?

The above is why I just use MX4.
 
I could not get mine to flow with a filled loop, I had to disconnect the CPU block to get it to work. It apparently had enough volume of water to keep it from flowing properly.
 
Rather than put it together and take it apart and put it together, I did the steps up through the IX reflow caseless on my kitchen floor, with spare-parts monitor and PSU.

34rt9ms.jpg

It seems to have worked admirably—my stock-speeds i5 2500K is currently near-idling at 21°C. Running four instances of CPU Burn-in, as I did for the IX reflowing, peaks out at 38°C. This is a vast improvement over before, using PK-1.

So it works as well as claimed, but I don't know if I'd buy it again. What bugs me is that PK-1 is well-regarded. Why did this same CPU / waterblock combo perform so meh with PK-1? (Idling around 35°C. Not really bad, but not as good as it should've been.) It looked fine when I took it apart. I guess I used too much? Not that I expect anyone to have answers to that...
 
IX isn't a paste, it's an indium-bismuth alloy that works very much like a low temperature solder. It has a low melting point to allow it flow between the IHS and heatsink and create a superior thermally conductive interface. If you appreciate this nature and the basic physics of its application, I think it's easier to understand your problem. For reference, I use IX on my rig and it probably only improved temp 2-3C over MX4, since it was already a great mount. However, compare to your using the stock heatsink and you got a much better improvement. The bottomline is that IX is much more conductive than paste-based TIM's, and so the poorer the mount, the more it will improve your temps.

If you're going to tackle it again on your waterblock, make sure you follow the instructions to good effect:
image.jpg


Your graph needs to look like that. Note I did this with my loop filled and the pump off. My i5 2500K was at stock speeds/voltages throughout and it took only a minute or so. The important thing to note is the the initialization of prime 95 (temp spike) followed by a big drop ~1 minute later as the IX changed phases and reflowed. You'll notice that after I cut prime 95, my idle temps took a nose-dive compared to before. Did you get the same difference when you first tried it? First indication that it flowed well.
 
IX isn't a paste, it's an indium-bismuth alloy that works very much like a low temperature solder.
Um... right. Thanks for the "Intro to Indigo Xtreme" lesson. I'm sure it will be very useful for those who started reading this thread wondering What the crap is Indigo Xtreme?

Perhaps I have been unclear. I followed the instructions, including stopping the pump with a "wet" waterblock, also including a temperature graph in SpeedFan that matched what the instructions called for, and got crap reflow. Yesterday I tried again, with a "dry" waterblock, and it seems to have worked well.

The consensus in this thread shows that dry waterblocks work better for IX reflow. What blocks are y'all using?
 
Back
Top