Project: TYPHOON (Stacker+ GTX480) **2010 REBUILD**

WOW, I feel like I need to go take a technical engineering course now. :D

Very impressive work!
 
Here's a quick update. I probably won't have much of a chance to do anything this week, but I had to at least post some pics of what arrived in my mail box today... :D

1600.jpg


The front panel PCB was ordered on 5/7, shipped on 5/18, and arrived today (5/19). Not bad! And the quality is excellent! They even printed two of them when only one was ordered. Very nice of them! I think I need the second one for soldering practice anyway. ;)

I've collected most of the components, and now it's just a matter of soldering them on. Here's a few of them in place. So far, they all fit perfectly!...
1610.jpg


1620.jpg



The next update will be about my soldering experience, and then powering it up for the first time. Can't wait for that! (man, do I hope it works ;))

-- stay tuned...
 
I made some more progress with this front panel. Those surface mount components are actually not as difficult to mount as I first thought. One or two of them could've gone down a little straighter, but for my first go around at it, I'll take it! :p Of course, all celebrations are on hold until the circuit actually works ;) ...

1650.jpg


The rest of the pieces (except for the switches) have to be mounted at a specific height, so I figure might as well get the panel done first. craigbru had the front panel for his LOSIAS project done by FPE which turned out awesome, so I thought I'd see what they're all about. Here it is in their free software...
1640.jpg


It's also going to need sides, so I might just have all three pieces done at a local machine shop. Either way, here it is assembled...
1630.jpg


That's it for now, enjoy the weekend! :)
 
You're a sick, sick human being, and I thank you for it!

Excellent work! You are putting my worklog to shame!
 
Mr. G, thanks for the compliment. I think I'm in very good company ;):p.
 
I find it extreamly hard to believe, and disturbing that you can do all this extremaly technical stuff and pay so much attention to detail, Cad Designs, Getting perfect aluminum bends etc, but not know how to polich that aluminum stock.

Your kidding right? Oh yea how are you getting those perfect Bends with that thick aluminum, you can't be getting those perfect bends fromt hat bunch vice alone?

Very good attention to detail here you have one of the most complicated stackers I have sceen and you have managed to cram a TON of stuff into it!
 
Justintoxicated said:
I find it extreamly hard to believe, and disturbing that you can do all this extremaly technical stuff and pay so much attention to detail, Cad Designs, Getting perfect aluminum bends etc, but not know how to polich that aluminum stock.
Click to expand...

LOL It's true. I've never polished aluminum. Well, I tried once on an old motorcycle clutch cover, but it didn't work so well.:rolleyes: I think I even have the mother's polish that warmace recommended, but I've yet to get try it on any of these brackets.

Justintoxicated said:
Your kidding right? Oh yea how are you getting those perfect Bends with that thick aluminum, you can't be getting those perfect bends fromt hat bunch vice alone?
Click to expand...

Yep, believe it or not, it's just the vice and a hammer. But there's a trick to getting the same bends everytime:
  1. Determine the before and after measurements of the bend (or the "developed length" of the bend) with a piece of scrap that has a clean, square edge.
  2. Measure in 1/2" or so, mark it square, and take note of the exact distance from the mark to the edge.
  3. Mount the piece in the vice taking care that it's perfectly lined up with the mark and perpendicular to the jaws (I use a combo square for this).
  4. To bend it, keep a steady force on the free end, and with a hammer, tap the piece right near the bend until it reaches 90 degrees ( or whatever angle you need). During the bend, throttle the force on the free end to help the piece through the bend, but not so much that the free end also bends. This way all the bending happens right where you want it - at the jaw line. Also, depending on how wide the stock is, you may need to tap with the hammer evenly across the piece to get it to bend squarely.
  5. Measure the distances from the mark to the inside and outside surfaces of the bend. For a 90 degree bend, the inside distance should just be the inside bend radius, and the outside distance the inside bend radius + the thickness of the material. The difference between the two = the developed length.
  6. Now, let's say you need a "U" bracket to fit exactly into a 3" opening and have 2" legs on each side. To achieve that, place a mark at 2" minus the inside bend radius and bend it. Then measure from the outside face of the first bend to 3" minus the developed length and bend it. Finally, cut the other leg to 2". Voila! :D
It usually takes me a bend or two to start getting consistent bends. Also, if you have more than one piece to bend, it helps to draw out a template and mark the pieces off it instead of off a ruler every time.

Justintoxicated said:
Very good attention to detail here you have one of the most complicated stackers I have sceen and you have managed to cram a TON of stuff into it!
Click to expand...

Thanks! I thought at first there'd be plenty of room, but I'm slowly learning how wrong I was! :p
 
Thanks Craig! Thanks for the reference to FPE. I just place the order with them for the final version... :D

1660.jpg
 
I'm always happy to spread the word about FPE. They're a great company, with a great service. I've referred enough people that I should be in their marketing department!

I can't wait to see the panel once you get it.
 
JonCl said:
LOL It's true. I've never polished aluminum. Well, I tried once on an old motorcycle clutch cover, but it didn't work so well.:rolleyes: I think I even have the mother's polish that warmace recommended, but I've yet to get try it on any of these brackets.



Yep, believe it or not, it's just the vice and a hammer. But there's a trick to getting the same bends everytime:
  1. Determine the before and after measurements of the bend (or the "developed length" of the bend) with a piece of scrap that has a clean, square edge.
  2. Measure in 1/2" or so, mark it square, and take note of the exact distance from the mark to the edge.
  3. Mount the piece in the vice taking care that it's perfectly lined up with the mark and perpendicular to the jaws (I use a combo square for this).
  4. To bend it, keep a steady force on the free end, and with a hammer, tap the piece right near the bend until it reaches 90 degrees ( or whatever angle you need). During the bend, throttle the force on the free end to help the piece through the bend, but not so much that the free end also bends. This way all the bending happens right where you want it - at the jaw line. Also, depending on how wide the stock is, you may need to tap with the hammer evenly across the piece to get it to bend squarely.
  5. Measure the distances from the mark to the inside and outside surfaces of the bend. For a 90 degree bend, the inside distance should just be the inside bend radius, and the outside distance the inside bend radius + the thickness of the material. The difference between the two = the developed length.
  6. Now, let's say you need a "U" bracket to fit exactly into a 3" opening and have 2" legs on each side. To achieve that, place a mark at 2" minus the inside bend radius and bend it. Then measure from the outside face of the first bend to 3" minus the developed length and bend it. Finally, cut the other leg to 2". Voila! :D
It usually takes me a bend or two to start getting consistent bends. Also, if you have more than one piece to bend, it helps to draw out a template and mark the pieces off it instead of off a ruler every time.



Thanks! I thought at first there'd be plenty of room, but I'm slowly learning how wrong I was! :p
Click to expand...
Polishing:
  1. I just sand aluminum with 400 then 600 (or just 600).
  2. Next, using a buffing wheel from lowes or HD, and some white rouge, turn on the buffing wheel, apply the rouge.
  3. Hold object to be polished against the outer corner, and viola mirror reflective aluminum.

Mothers is best for polishing things that are already polished :p It takes forever by hand otherwise. You can use a drill press for polishing as well. I don't have a machine in my room, so no more polishing for me...

I read your post above 3 times and I'm still totaly lost. Not sure what a you mean by a bend radius, or marking the scrap and putting it in the vice perpendicular to the mark?

Actualy you lost me on step 2...I'm pictuing drawing a straight line on a piece of aluminum (say a piece of aluminum strip/bar) from one side to the other. Sticking it in the vice so the mark is parallel to the jaws (you said perpendicular but I don't understand how to do that and still get a good grip). Do you mean like put the sides of the aluminum strip in the vice? If so I don't see how you could hammer it into shape?

Ok so I mark something and bend 90 deg, clamp it in the vice and hammer it flat down to the jaws (for 90deg) then take measurements? haha yea I'm lost..

It's ok though since I no longer have access to a bench vice, all my moding has to happen in the driveway or lawn area now with no vices or anything...Someday I will own a house, I just need 200k a year here in cali to even consider it.

One more thing; does the quad rad have more performance than the Thermochill PA 120.3?
 
Justintoxicated said:
Mothers is best for polishing things that are already polished :p It takes forever by hand otherwise. You can use a drill press for polishing as well. I don't have a machine in my room, so no more polishing for me...
Click to expand...

Oh, now I know why it didn't work! :rolleyes: lol

Justintoxicated said:
Ok so I mark something and bend 90 deg, clamp it in the vice and hammer it flat down to the jaws (for 90deg) then take measurements? haha yea I'm lost..
Click to expand...

Basically, what you're after is the developed length of the bend. With a piece of scrap, you can determine the developed length experimentally by measuring first, bending it, then measuring again. The difference is the developed length. When that bar bends, it stretches some amount, which is mostly dependent on the type and thickness of the material. You can use the developed length to determine exactly where to mark other pieces so that the finished dimensions are predictable.

When the bar bends, there's always some "round" left at the inside corner. That small little round is the inside bend radius. If the piece was bent exactly from the mark, the inside bend radius should be the distance from the mark to the inside surface of the bend (for a 90 degree bend).

For step two, see if this helps:
2. Measure in 1/2" or so, mark it square (perpendicular across the bar), and take note of the exact distance from the mark to one end.​
Those dark marks near the bends are from the hammer ;)... oh, look there's the hammer and square right there on the bench...
1060.jpg


Those are the lines I keep referirng to. These were actually bent one at a time, then mounted all together to "tweak" them into alignment.
1050.jpg


Here's the template I used to mark the bends on each piece...
1070.jpg



Justintoxicated said:
One more thing; does the quad rad have more performance than the Thermochill PA 120.3?
Click to expand...

That I don't know. I've read the 120.3 has less pressure drop. I'll have some temperatures posted eventually... :rolleyes:
 
I've read at Procooling that the PA120.3 has less pressure drop, and out-performs the GTX480 with low rpm fans. I believe it was said that with higher rpm fans, the GTX480 was slightly superior. (Basically, the PA120.3 has less air-flow and water-flow resistance, so it takes less fan and pump power to make it perform.)

In a case like this where the fans are not low-rpm, and are set up in push-pull, the GTX480 should out-perform the PA120.3 by a handy margin since air pressure is not an issue. If you don't have to worry about air resistance, all that's left is surface area and water pressure drop; the GTX480 wins big in surface area and only gives up a little with water pressure drop.

If we were going for silence rather than balls-out push-pull performance, the PA120.3 would be the better choice.
 
Extremely impressive. I've spent the last half hour going through the whole thread and I'm amazed at your attention to detail. Keep up the good work :)

btw - if you don't mind, how much did those custom PCB's cost? If they're not too bad, I might use that company for printing custom amplifier boards.
 
Nicepants42 said:
I've read at Procooling that the PA120.3 has less pressure drop, and out-performs the GTX480 with low rpm fans. I believe it was said that with higher rpm fans, the GTX480 was slightly superior. (Basically, the PA120.3 has less air-flow and water-flow resistance, so it takes less fan and pump power to make it perform.)
Click to expand...

Yup, that's an excellent summary of the comparison. I wanted as much surface area as I could get to hopefully keep it as quite as possible, but still get excellent cooling. But the GTX480 doesn't flow air quite as easy as the 120.3, so how do you solve that? 7 120mm low-speed fans! :eek:

Nicepants42 said:
In a case like this where the fans are not low-rpm, and are set up in push-pull, the GTX480 should out-perform the PA120.3 by a handy margin since air pressure is not an issue. If you don't have to worry about air resistance, all that's left is surface area and water pressure drop; the GTX480 wins big in surface area and only gives up a little with water pressure drop.
Click to expand...

Actually, I'm pretty sure those Yate Loon fans are the low-speed variety. Even with a full 12V, one can hardly be heard - to the point where the fan controller is almost not needed. When I bench tested the rad with all 7 fans, I was quite pleasantly suprised at how quiet they were even while push/pulling a healthy amount of air through the rad. The temperatures will tell the rest of the story once I get it running.

Nicepants42 said:
If we were going for silence rather than balls-out push-pull performance, the PA120.3 would be the better choice.
Click to expand...

I don't know - even with 7 fans, it's just a whisper. I don't know yet how it cools though.

Citizen _insane said:
how much did those custom PCB's cost?
Click to expand...

Thanks! The PCB was I think around $34 shipped. I ordered one but they printed and included two! I also looked at pcbexpress, but they wanted I think $50 for a minimum quanity of 3, and they didn't come with the solder mask or silkscreen which BatchPCB includes on BOTH sides. I have to say, it's a really impressive service with an awesome price.
 
Man If I had your skillz I would have my own side buisness right now. I actualy have had something in the works for 3 years that requires bending sheet metal, and an SMT Circuit.

I had someone bend the metal for my prototype and used Pin Through components and made a cheap circuit (Actual ynot much cheaper than doing it the right way)...

Oh well that stuff is too complex for me to understand and model etc...I already wasted a good chunk of my free time trying to learn it on my own.

I still don't follow your bending tutorial but I bookmarked it...

What program are you using to model the circuits?
Bend Radius is the thickness of the metal at the bend then?

Well keep up the good work, that custom Panel is really nice, about 100,000 times more complex than the simple regulator I was trying to design to control Current and filter spikes...I'm jealous of your knowledge both in making perfect bends, and in circuit design.

For me I just bend stuff with pliers and channel locks, but if I did have a bench vice I would just bend stuff over and over untill I ended up with something that fits. but like they say, nothing is hard, IF you have the right tools (And I do not).

I can't even fit the stock hard Drive cage AND a 120x2 mm radiator in my stacker!
 
Indeed, having the right tools is definitely key!

The PCB software I used is called Eagle. There is (for me at least) quite a learn curve to it, but after getting used to it, it works really well and is very powerful.
 
The next step in this front panel sub-project is to cut a hole in the custom PCB for the Logisys connector to pass. After drilling a few holes, I ended up passing a coping saw through the hole to rough it out...
1670.jpg


...and then cleaned it up with a file...
1690.jpg


The dimensions came from a printout of the layout. I realized after starting to mark it that those dimensions should've been to the HOLE instead of the connector :rolleyes:... I hate it when I have to do math in my head! :p:D
1680.jpg


I took some pics of the first test fit. Looks like the hole at least line up, so that's good...
1700.jpg


1720.jpg


Here's how it looks from the top. Next thing to do is cut some standoffs...
1710.jpg


I printed another drawing with dimensions for the standoff heights, and then carefully cut them to length...
1730.jpg


Here are the standoffs in place over the drawing. With those large toggle switches in there, the original header on the Logisys daughter board isn't long enough to reach the connector (check the gap circled in pencil). That's the reason for that wire-wrap header: it's going to replace the header on the daughter board...
1740.jpg


I wish I had more, but it's all I have for now. There's a few more things to do, then the front panel should be done:
  1. Find a 1" screw that'll thread into the original holes at the back of the Logisys panel.
  2. Swap the header on the daughter board for the new one.
  3. Solder the rest of the PCB components to height.
  4. Make the left and right sides.
  5. Test
  6. Assemble!
 
Update: bezel came today! They did an excellent job with it. Everything seems to fit as planned so far. I plan to post another update with it hopefully fully assembled and lit up this weekend. Here's some pics. :D

1750.jpg


1751.jpg


1760.jpg


1770.jpg


1780.jpg


1790.jpg


1800.jpg


1810.jpg


1820.jpg
 
Thanks Craig! It's just a little guy compared to others I've seen ;). FPE owes you another one! :D
 
Just one sick looking front panel there... :)
I want to order a full faceplate(5.25 x 9) to go with my stacker 832.
How do you plan to mount that plate? I am interested to know because I will soon do the same.
 
To k00lance, it'll mount to the case with these...

1830.jpg


Greg at a local machine shop made them up for me over the weekend. They fit perfectly and his fee was unbeatable. All I did was send him a dimensioned drawing, and he had them ready the next day. Thanks Greg! :)

Here's how it looks now assembled and fully soldered...
1850.jpg


The header on the daughter board is replaced, but I still have to find the 2-56 x 1.25" screws to mount it. I was hoping to find them at a local hardware store, but no luck...
1860.jpg


Btw, the LEDs were "diffused" using a Dremel and a wire brush bit...
1840.jpg


The next step was to start building the wires. I went with a ribbon cable to make the connection between the LED headers on the RAID card and the front panel. Having the ribbon cable makes it a lot easier to build compared to crimping connectors to individual wires, but even then, it was still quiet a tedious job!...
1880.jpg


1890.jpg


Here's the light switches lit up...
1870.jpg


And then the moment of truth came!... Is it actually going to work? Yup! :D (almost :rolleyes: ) Out of the 24 drive leds (12 activity and 12 vault), 7 of them didn't work. I was concerned at first that I had fried a component or two during the soldering, but after poking around with the tester for a bit, I discovered it was just a few of the legs didn't fully solder to their pads (it's no surprise ;)). After fixing those, all was good! It's a little hard to watch because of the quality, but anyway here's a video taken from my cell phone while it boots...
 
It's interesting how the hard drives are accessed in a row like that :) , very cool indeed.
 
SUBSCRIBED!

That panel is all kinds of awesome...
 
Bah! Lost a few updates. I deserve it for posting during the database move and not backing it up! :rolleyes: Anyway, this update is only a summary of those other lost updates (rip)...

First, I finished off the custom front panel with some 2-56 threaded rod to assemble the daughter board...
1900.jpg


Front panel DONE!! :D (except for the cables)...
1910.jpg


I'm going to flip the PSU so it's fan-down. In that case, I'm going to need some holes in the mobo tray to pass the PSU cables...
1930.jpg


1940.jpg


1950.jpg


Holes completed...
1960.jpg


A new bracket template to flip the PSU:
1920.jpg


Formed and cut...
2010.jpg


Painted and mouted...
2020.jpg


Bottom vent cover painted...
1980.jpg


and mounted to the bottom to allow the PSU fan to breath better and to allow the pump to sit lower (more on that)...
1990.jpg


Placed some nylon spacers to take the place of the vent cover to support the PSU...
2000.jpg


1970.jpg


PSU installed...
2030.jpg


PSU power cables now run behind the mobo tray...
2040.jpg


:D
 
Over the weekend I built my new rig in a CM Stacker orig, so this build is wayy more appealing now. I assembled it to test the higher-performance parts but will disassemble it to WC once all the parts come in.

I'm seriously considering painting the inside of the case a matte black, any recommendations?

As well, are you going to use the USB ports from the stock front panel? If you are you might find the cables a little too short to span accross the case to the bottom left of the mobo. Yet as I'm typing this I'm remembering the rad in the top bay meaning if you keep the panel it wont be there :rolleyes: blahblah blah

COOL!
 
oneil said:
Yet as I'm typing this I'm remembering the rad in the top bay meaning if you keep the panel it wont be there :rolleyes: blahblah blahCOOL!
Click to expand...

Yes, I'm using the stock I/O panel. And, you're right: it'll get mounted about 5 bays down from the top. From there, the USB cable reaches the header without a problem. I'm going to route it from the panel, out the hole in the side of the bay, behind the mobo tray, and then through the hole I cut there.

oneil said:
I'm seriously considering painting the inside of the case a matte black, any recommendations?
Click to expand...

If I was going to paint another case, I'd start with etching primer which is something I didn't use. The paint job I did is ok (it's functional and doesn't flake), but it could be better. For example, the clear coat I used scratches too easily (needs to be harder). Also, I've since read somewhere that among the spray-can paint, Rustoleum is better than Krylon. Other than that, I don't have any solid recommendations, but there's lot of guides out there on the subject, I'm sure.
 
Recovery update summary #2:

I ordered the 1/2" threaded pump housing from Laing (bottom left)...
PumpHousing.jpg


Having the threaded inlet, a PVC union can be used to connect the res...:
2060.jpg


But that big nut on there didn't fit in with the original pump mount, so two more brackets fixed the problem by spreading the legs further apart:
2070.jpg


Painted the res hanger while I was at it...
2080.jpg


Pump mount with pump looks like this now...
2100.jpg


2090.jpg


The StarWars shot... :p
2110.jpg


Assembled...
2120.jpg


and fit into the case...
2140.jpg


2130.jpg


The outlet is actually a straight thread, which made finding a fitting a problem. Finally, Lowes had CPVC fittings which fit perfectly...
2150.jpg


2160.jpg

It's a very tight fit...
2170.jpg


2180.jpg


Painted up and ready...
2190.jpg


That's it for the lost updates. I've done a lot with the project since a few days ago, so the next update will be new. Thanks for looking! :)
 
The setup and ideas are looking great. Sometime by the end of the year I'll be starting a project of my own that could be similar to this :) .
 
Your work is absolutely amazing, I can't wait to see the final product!

/subscribe
 
You must log in or register to reply here.
Back
Top