MMZ_TimeLord's System Rev 02 (Worklog - 56k Warning!)
 

Well, it's that time again... after finding so many issues, getting a new video card and deciding on a full RAID 0 + 1 setup instead of just RAID 0, I've begun redesigning the whole system (save the CPU block).

To start with I've been doing LOTS of design drawings to nail down how all this will come together. I'm still working on one or two items so I will not show the full system design drawing ... yet. :D

I've ordered the copper I need for the Video card solution and the HDD cage block solution. While I'm waiting on the copper to arrive, I went ahead and rebuilt the reservior and pump setup per this drawing...

looks good. got any pictures? I'm in the process of building a homebrew acrylic double-bay reservoir right now and woundn't mind seeing yours finished.

edit:
I guess if it were me doing it, I would have made the hole for the pump inlet further down in the res, making the bottoms of the pump and res flush with eachother. This also gives more water above the inlet, lessening the chances of any air getting down to that level somehow. It's already finished though, right?

First order of business was to work on that complex acrylic side that the pump will mount to. This was the most precise piece of the reservior.

First: The milling of the inlet and mounting holes.
http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0038.jpg

Second: Here is the milled off and widened the inlet on the pump.
http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0039.jpg

It used to look like this...
http://pages.sbcglobal.net/mmz_tl_01...%201-EA-42.jpg

Third: Gluing the pieces together (they were pre-cut by TapPlastics).
http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0041.jpg

Fourth: Here it's almost complete, just need to glue the top on. I will probably end up bolting it to the base as I want to be able to remove the pump (just in case :p )
http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0042.jpg

Here's a closeup of the widened inlet on the pump...
http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0043.jpg

More to follow tomorrow... :D

Looking good...
...but, I don't think you should have made the pump inlet so large. As it is now, the rotor will be trying to spin the water in the reservoir, rather than just in the pump volute (casing).
Not good for efficiency.
Might be a good starting place for an experiment though:)

The weight of the water i the res should be sufficient to push the water in the right direction(down). I would have been more worried if it was a small res.

~ Boli

Lolito_fr and Boli,

I already tested the reservior in the water lab (kitchen :D ) and the pump did try to spin the water in the reservior when it was below the top of the inlet. But when the water was above the top of the inlet it was pumping water just fine, so I believe it will be okay in a closed system with a ½ liter or so of water pressure from the top of the system.

It did seem to pump noticeably faster than before the modification. As I have no real measurements before and after, it's just a visual observation from recollection. :p

Whew... long day today... so not as much done.

I received the cap screws today in the mail from microfasteners.com.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0052.jpg

To give you an idea of just how small the "0-80" cap screws are...
(Ati All-in-Wonder 9700 Pro in the back for scale)

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0050.jpg

Close-up... :eek:

http://pages.sbcglobal.net/mmz_tl_01...050-Detail.jpg

And... measurements... :D

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0054.jpg

Also got the Acrylic piece for the top of the reservior and drilled and tapped the holes for the "4-40 x 3/4" cap screws. You can see two of them in place.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0057.jpg

The white "stuff" is white "Silicone II" Bathroom Tub & Tile Sealant. Old, but still good in the tube. I'll let it cure about 24 hours and pull off the intake cap and peel off the "gasket" and trim it up. Then it should be ready to assemble again. I also still have to mill the holes for the two return tubes.

That's it for tonight... more tomorrow. :p

cute lill screws:)
I maintain my doubts about the pump inlet though. It's a good sign if it seemed to be pumping normally (either that or it would be stalled) - however, I wonder if it will still develop as much head...
Seriously, with such a major modification, I think you want to check that it will still pump water through your loop before you finalise that seal.

EDIT: oops, wrong thread :rolleyes:

Well, not much done yesterday or today... busy weekend...

Just a few odds and ends fabricated... Hope to start work on the power supply water cooling solution tomorrow, as well as getting the reservior finished and tested. :dome:

Okay... got some more completed... Specifically the main block of the power supply water-cooling solution. :D

First I cut a piece of 2"x5.5"x0.5" copper from my bar stock.

Then I sprayed on a 3M adhesive and attached the template.

Locked it in my mill and started cutting... First thing I noticed was that a two flute endmill has a tough time routing channels.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0058.jpg

So I put in a 7/16" drill bit and drilled the channel down to the depth I wanted in as many places as I could.

Then I went back with the end mill and cleaned it up.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0061.jpg

After I was done (I decided to leave the dimples in the channel), I took off a few thousandths to remove the template and reveal shiny copper.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0062.jpg

After all of that was said and done, I was going to use my torch to tin that surface to attach to the sheet copper. I decided to try something a little different :dome: ...

I preheated the oven to 450°F and put the copper in on a cookie sheet (non-stick kind).

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0063.jpg

When it was hot enough to melt the solder, I tinned the surface like so...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0064.jpg

After it was fully tinned... I pulled it out and let it cool so the solder would solidify again.

IMPORTANT!!!: If you do this MAKE SURE YOU HAVE A VENT HOLE DRILLED. I almost forgot this which would have ruined my piece.

When the solder was solid again... I put a sheet of 4"x10"x0.025" copper on the cookie sheet and put the tinned piece upside-down on it and lined it up with the corner (NOTE: USE AN OVEN MITT!!!) and put it back in the oven for about 10 to 15 minutes.

The bigger the copper piece, the longer it takes to remelt the solder.

Here is the piece after it came out...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0066.jpg

I will have to leak test it and see if I got a good seal. If not... I can take it and put it back in the oven and flow some more solder around the edges.

That's about it for today... Will try and get the rest of the power supply solution done tomorrow.

Great idea dude with your oven!!
It's like in micorelectronics factory he he he.

Btw, One suggestion/idea... have you ever seen 'solder ring' plumbing fittings? They have a kind of groove in which solder sits and if done correctly it creates very clean and nice solder joints. I was wondering if you could do the same thing. Mill a grove in your block around the edge to melt some solder into it, it's sth like O-ring rubber seal grove type idea as well.
Well done m8!

Jabo,

I have not seen those type of fittings, but it makes sense. I don't know if I will try that with other blocks or not. We will see... :D

I've seen these fittings. They look like a normal copper elboy or tee or whatever, but they have a ring of expanded area halfway across the mating surface. This groove is filled flush with solder, so that all you have to do is flux the pipe, stick this thing on, and heat it up.

http://www.screwfix.com/sfd/i/cat/99/99202_l.jpg

I still think that the oven method should give better results when two flat surfaces are being mated instead of using some kind of groove.

Any word on how that seal came out? I like how easy and foolproof the oven technique looks. I'll have to try that once I get my hands on some copper. :)

Cool thread timelord and not a bad job. The two flute just tries to cut too much and bites into the copper. Try a 3 flute or 4 flute with a little more rpm. I'd use the 2 flute for rough work and finish up with a 3 or 4 flute.

Nice project you got going, the first problem I see with your copper cutting is that your using HSS! :eek: carbide is almost a must for much copper cutting, secondly, you should get ahold of some good coolant, and thirdly, whats your spindle speed at? for a 3/8" carbide it could be up 5000+ the faster you spin it the easier it is and the more heat gets transfered to the chips.

The only hss end mills I have are ones given to me, I learned after my first set of hss end mills.

Power supply cooling, I have wanted to do that for the longest time! Maybe this summer Ill hop on it;)

Jon

SysCrusher,

I've found that if I try and do channels with the four flute, it is even worse. Heat just builds up quicker because of the additional friction. The process of predrilling the area to be milled cut down on the chatter and heat CONSIDERABLY!


JFetting,

HSS is okay for me because of two things... no coolant system at present, and my spindle speed doesn't go much higher. This mill tops out at UNDER 1000 RPM. I MAY try a different motor down the line, but not for a while...

I just have to be patient with this and it seems to keep the cutting edges pretty sharp.

Here's today's update...

Cut out the block from the sheet copper... and made a pass on the edges.

Then I pilot drilled and then milled a 0.50" hole for each of the copper tubes.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0067.jpg

Here the tubes are being soldered in place...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0069.jpg

you must be careful with assemblies like this as you can have the existing tube pieces fall off if you over heat the whole thing... The way around it is to put everything except what you are soldering under water. I will be doing that tomorrow as the "T" joint has a small leak.

The block had NO leaks from the oven proceedure... YAY! :D

More tomorrow... :cool:

Even with low rpm, carbide will help a ton,

Check out jtsmachine.com they have some cheap carbide end mills.

that powersupply block is looking pretty good. Im looking forward to seing results.

Jon

Well, another slow day on this front...

I did get the return holes milled and the acrylic return tubes in place on the reservior. I also ran it with the top on and about an 1/8" of air in the top. The water was churning towards the pump side, but didn't draw in air at all.

I also received the copper for the video card and hard drive cage solutions... the smallest pieces are 4.5" x 4.5".

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0071.jpg

I will have more pictures tomorrow on the power supply conversion.

mmm copper drool*

Welp, got a lot more done today...

First, I already had been working on the power supply water block. Got it finished as far as the unit being sealed. The oven trick worked, but soldering those tubes on was a pain. Every time I tried to solder them to the main block they would "pop" off. Even with them submerged in water I had other problems. But, I perserveered and it is water tight.

I then began to dig into the power supply and pull the old heat sinks out.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0073.jpg

As you can see there is quite a bit of space in there to work.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0074.jpg

It's almost as if this power supply was laid out for water cooling. :D

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0075.jpg

Here, I have the board pulled and you can get a better look at those heat sinks.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0076.jpg

There were two screws in each of the heat sinks that mounted them to the circuit board. I've removed two of them and the third is loosened, Red circles. The pins for the various regulators are outlined in Blue.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0077.jpg

Here's the first row desoldered and removed with it's heat sink. I've written the part numbers down and mapped their location so I won't forget where they go.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0079.jpg

Here's the second one...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0080.jpg

I've put the regulators back in their spots just so I don't loose them. All but one have insulating heat transfer pads (rubber like substance) and insulating washers for the screws. When I put this back together I will have to put a meter on the block and make sure it's not providing a small voltage.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0081.jpg

Here are the two naked heat sinks on which I based my "fingers" for the water block.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0082.jpg

The first "finger" is done...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0083.jpg

Both "fingers" along side the original heat sinks... pretty close :cool:

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0084.jpg

This is what I hope the solution will look like when installed. Just mocking it up here to get an idea of whether I've gotten everything to fit or not. Looks like I got it pretty close.

Also measured and marked the holes for the regulator mounting screws. I will drill and tap those tomorrow as well as drill and tap the mounting holes for the "fingers" to the main block. More then... :p

Fantastic job dude!
Looks great!
What about it's weight?
Is it not going to be too heave for this little board to hold?

why dont you take those 2 peices you made and drill holes strait through almost all the way, then one so the holes are facing eachother so basicly the water goes in one big side, across to the other and back out?

Jon

looks kickassish!

Jabo,

The power supply is installed with the block "hanging" from the circuit board. I'm trying to keep the weight off the regulator pins and on the screws.

If I see that the board is bowing, I will put some "feet" on the main block to the casing to support it.

JFetting,

I'm not running passages into the "fingers". The main block is a single loop passage and is already milled and sealed.

It will be very difficult at this point to solder anything else to the main block and expect it to be water tight. As it is, I had originally planned to solder them to the block with two 4-40 cap screws holding them in place. This way I can keep the rest of the block from de-soldering by submerging the tube side in water.

Well, I got a little more done...

I had already measured and scored lines to center the holes for the regulator anchor screws in the "fingers". So I drilled the holes and began to tap them...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0085.jpg

One technique for keeping the tap from "wobbling" is to use a pin in the chuck and let it sit inside the hole in the back of the tap handle... this aligns the tap to the hole and makes it easier (at least for me :p ).

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0086.jpg

I then drilled and tapped the holes in the bottom of the "fingers" for the anchor screws for the circuit board.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0087.jpg

Did some light sanding... here is the results... almost done... :D

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0088.jpg

Tomorrow I will mark, drill and tap the holes in the "fingers" for the anchor cap screws that will hold the "fingers" to the main block.

Very nice work!

I wish I had access to your kind of tools.... buuuu :(

Well... one step closer to a finished system... The power supply cooling block is complete. Now I just need to trim up the power supply casing and install it back into my tower. Last thing I'll do to the power supply is custom lengthen the power connectors just for my drives, etc.

On to today's workload... :D

First off is a word of warning... if you're in a hurry, STOP... take a break and think it through or something like this may be the result!!!


Let me explain... First: I was in a hurry, second: I forgot the rule for smaller drills and taps in copper "USE LUBRICATION!", third: I got in a hurry.

Result: Broken drill bit in the first hole... offset and drilled a second with success...

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0089.jpg

...only to break the tip of the tap off in the hole.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0092.jpg

The tap was still usable with a little grinding down, but I could not get it out as the hole does not go all the way through and there is no way to punch back out the way it went in. :mad: :cry: :shrug:

The good news is I believe that one cap screw will be sufficient on this "finger" as it has three small regulators and not that big one and two small ones.

Here are two tips for those planning to tap very small holes...

1. Use a screw of the same type you are going to use for a tap by filing a cutting edge in the end. This will make it MUCH easier if you break it off. These are soft enough you can drill back through them.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0090.jpg

2. Use the "T" handle with out the cross bar, this forces you to use your finger tips and NOT apply too much torque.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0091.jpg

Here is the other "finger" with the regulators re-installed and ready to fasten to the circuit board.

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0093.jpg

Here is that "finger" being fastened to the circuit board... notice the holes actually line up for the most part. :p

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0094.jpg

Here is the block with the fingers installed and secured to the circuit board and all the regulators in place and ready for final leak testing. (Yes, I also messed up on which hole to move over for that one I tried to correct. DOH! :eek:

http://pages.sbcglobal.net/mmz_tl_01...s/Dscn0095.jpg

The final leak test revealed one last pinhole leak near one of the cap screw pass through holes. So, out came the torch and reheated that one area with the tube end of the block in water. All fixed and tested with the completed reservior, NO LEAKS!!! YAY! :D

The pump does get pretty warm, but even after an hour with only the power supply block and reservior in the system, the pump could be held without burning you. (No radiator of any kind, not too bad. ;) )

That's it for today, I'm going to re-solder the regulators and re-fasten the circuit board into the power supply casing tomorrow and begin the custom cut-out on the power supply casing. Hopefully, I can get a finished power supply out of this mess. :p

Lookin pretty good so far, is that one inlet/outlet have 2? it looks kinda like a T

One thing you might want to consider investing in is some cold forming taps or 'fluteless' taps. They push the material rather than cut it. You start with a little larger hole than normal and a little lube and you can turn it right into the peice of copper with no problems at all! These things rock! I have tapped proabably over 100 holes with my 4-40 and 50 or so with my 6-32. you can get them at any larger tool place like mcmaster, jtsmachine, and mscdirect

Jon