[PLOG] Mac G5 --> PC Conversion

[angelkiller]

nice hand job

Ummm, thanks. (wait, that makes it worse!)

I made a bracket that bends around the front of the PSU & gets screwed into new holes in the floor.

You could move your radiator down to the floor of the case. Extend a piece of plexi or flat stock or whatever to span from the upper lip or top of the radiator across the motherboard to the rear & wedge in between the bottom of the PCI bump-in & the rear fan grill assembly. You could put your pump/res on that span of material. If you are really careful with the sizing, you could wedge your radiator against the front of the case. and the span would be wedged between the lip & the hose connections.

Or screw in a piece of 90° bent flat stock, screw it into the drive plate separator and mount the res/pump on that hanging material.

What's flat stock? Sounds like some kind of metal sheet..?

Anyway, I was thinking of using 90 degree brackets to attach the radiator to the side of the case. But I may mount it from the top/bottom. Depends on what they have at the hardware store. I think we have the same idea about how to mount the stuff.


My res came yesterday. It's a tad smaller than I thought but it has a pretty simple mounting system, it just needs 4 screws on a flat surface. However, there are 5 outlets on the res and the guy I bought it from only sent 2 stop fittings. So, yeah.. I need another one.

Question: How do I screw things into the case? I assume I need metal screws. But do I just drill it in? I'm worried a screw will go through the side panel. Or should I avoid mounting things on the side panel and stick the the top & bottom of the case?

Question 2: I noticed in your G5 mod, you have some custom metal pieces. Specifically, I was looking at the one on the mobo's IO section. The IO section doesn't extend flush against the back of the case, (as you know) it sits even with the expansion slots, which are recessed into the case. So if I cut a hole in the rear panel, there will be a short distance from the back of the case to the IO panel. What materials did you use to make that piece that surrounds the IO ports and how did you make it?

Thanks for all the replies people!

 

[distemper]

What's flat stock? Sounds like some kind of metal sheet..?

sheet: very thin, usually <= 1/32
plate: usually thicker than 1/32
bar: flat material, but often of limited width (1/2 to 4 inches)
stock: any piece of material (plastic, metal...)

For example, the flat stock I used to make my zone separators was 1/10" 6063 plate aluminum (6063 is the specific allow of aluminum). Your flat stock might be Lexan. Or the side of a kitty litter pail. Don't laugh, there is some of that in my case holding the fans together.

Question: How do I screw things into the case? I assume I need metal screws. But do I just drill it in? I'm worried a screw will go through the side panel. Or should I avoid mounting things on the side panel and stick the the top & bottom of the case?

You are definitely right about top/bottom/hanging & NOT drilling into the side. If you are on a mission to avoid that I suppose you could epoxy an a nut - but you should know that what JB Weld (I think all epoxys) have a huge weakness: shear strength. While pulling a JB Weld-ed part off is difficult, its much easier to push it off to the side. Apple has very shallow screws, and when they were drilled & tapped, the bits were flat bottom. Standard drill bits are tapered, not flat. Also, there are a great number of the screws because each screw is so shallow and has little binding force. Similarly hanging a motherboard off the side subjects the JB Weld to shear stresses - but over 9 points (and a graphics card screwed in @ 2 places) its manageable.

Question 2: What materials did you use to make that piece that surrounds the IO ports and how did you make it?

Here is a pic of the I/O plate that I made. It isn't epoxied in - it is just wedged between the case & the screwed in motherboard.


The offset angle aluminum & aluminum sheet were both purchased @ HomeDepot.

I used a piece of aluminum offset angle, measured how long the I/O section was, how tall - meaning how high off the motherboard (this is a very important measurement - too high and it hits your fan grill assembly, but you should be fine). I then practiced on a scrap piece: I made a cut on one edge with a jigsaw and bent it into a right angle and noted how tight the radius I could make. I bent it over another piece of angle stock & hammered it to the tightest radius I could (and its easy to bend off-angle and make a multi-dimensional nightmare). The material was too long (extended too far into the case) so I had to use an angle grinder and grind the depth down. Sorry, I don't remember what the measurement was. And if you know what a b*tch it was to get my Mugen2 cooler in you would understand why I'm not offering to measure it. Finally, I notched out the fan grill rounds. Originally I thought I would epoxy the flange (the part of the angle stock that will press against the inside of the rear case mesh) onto the back, and drill holes to "mask" its presence. Now I realize it just doesn't bother me. The end is wedged into forming a rectangle

On a piece of paper I traced the I/O shield openings (from the stock I/O shield) & taped the paper onto the piece of aluminum. I used a jigsaw to cut every opening - they don't look great, but they sure look better than what my drill bit did to the round openings (twisting metal where there were other openings). Then I cut the sheet with some allowance on 2 sides to bend to form a lip which fits over the outside of the frame I made above. I sandwiched the sheet between wood & C-clamps. Placing a piece of angle stock next to the sheet, I used a hammer to start pushing the sheet into an angle.

You could also:
• For the frame of the I/O recess, you could use just a piece of flat aluminum bar. Con: subject to move in the 2 directions where the I/O shield doesn't have a flange.
• I also had an idea of using only sheet metal (no angle stock). Cons: making 2 90° bends within 1/2" seemed too complicated for me, but I don't know what you have available.

 

[angelkiller]

Ok, I've got some mounting issues. I still don't know how to attach everything.

I just got some aluminum strips (for lack of a better word). My plan is to bend the pieces into a C shape, the top of the C will attach on the top of the case, and the bottom of the C will attach on the bottom of the case. I've also decided that the res will be mounted on the top panel and the pump will just sit on the floor for the sake of simplicity. So my latest plan is shown here:



Now my issue is that I need to attach thin pieces of metal to other thin pieces of metal. If you look at this diagram, the arrows show where attachments need to be made. Normally, I'd just use a screw and nut. But in every case here, clearance is extremely limited. So I can't have a screw sticking out. The optical drive sits less than 1cm above the top separator and the bottom of the case is just that. I don't want screws sticking out of the bottom of the case. (screw heads are not desirable either, but acceptable though) A simple solution would be to get really short screws. However, the shortest screw I could find was still 3/8in. Another solution would be put the screw in so the heads are on the outside. That's better, but then the screws are showing on the inside... And plus that's not possible on the radiator, a screw must go into it. So, there must be a better solution. I thought about JB Weld, but I wanted this to be removable. Is there any better way to attach two thin pieces of metal?

I hope that all made sense. Thanks.

 

[DanishDevil]

As far as thin metal to thin metal, get a rivet gun (they're cheap). They aren't super easily removable though.

 

[phanbuey]

what about using standard computer case screws? they are much shorter than 3/8". They are 1/4" AFAIK. You can pick up a kit of them at just about any electronics store.

your other option, if you have a dremel or wire cutters is to cut the screw after the mount, when you remove it, it will tap correctly and be usable again at the desired length.

you can cut all the screws at the same spot, if you cant easily reach other spots with the dreme/wc's. You can even do it with two pieces of scrap metal that are the same thickness as w/e u are trying to mount so you don't risk scratching the case (best option IMO).

 

[Cuzza]

this is excelent, finally someone putting a mac to good use....

Keep up the good work. And dont give up on the spud gun, they are endless hours of frustration for minimal reward, but still great fun.

 

[angelkiller]

Alright, small update. I said earlier that I'd polish that seperator plate. Well I did. It's the same as lapping a CPU, just you don't have to worry about being 100% flat. In the end, it's not a perfect mirror like a CPU, but it's a dull shine, which is exactly what I wanted.



Rivets and cutting the screw. 2 really good ideas. I think I'm going to cut the screws so I don't have to buy yet another tool I'll never use again. But can't you drill rivets out? (I think I've done that before.) IIRC, there is no trace that the rivet was there after you drill it out. So that could work. As long as it wasn't permanent. (eg JB Weld, etc)

 

[angelkiller]

New stuff: DS3L Pencil Mod to reduce/minimize Vdroop and LEDs on the reservoir.



I made the brackets for the radiator, but I really messed up the screw holes. It's hard to see, but the bracket holes don't line up with the screw holes on the rad. So I'm going to redo those now. Hopefully I'll finish tonight. Pics of the bad one:



Ok, new brackets made. They look the same as the old ones, but the screws actually fit. I went out and made the best investment as of recent: a center punch. My holes line up this time!

 

[angelkiller]

I thought I'd give some updates on my plans. I don't really have too much to go. I'd say I'm about halfway there. I still need to mount the radiator, do something about the PSU and mount the motherboard. Yay! Not much left. The rad should get mounted tomorrow.

For the PSU, I think I'm set on taking it out of its casing. Why? Because the PSU PCB has holes in the corners that I can use to mount it to the floor. I would get 2 80mm fans to blow across the PSU (front to rear of case) for cooling. Then I would cut all the cables I'm not using. And I want to shorten the cables that I will be using, but shortening the ATX connector seems like alot of work. I don't know if I should take the pins out, shorten them and resolder the pins back to the wires or if I should just cut a section out of the wire and solder the two ends together. What kind of impact will this have electrically? Do you think my mobo will still be stable with a modified ATX cable and 12v cable? I only need 3 cables, the ATX one, the 4pin 12v one and one with 3 molex connectors. That last cable will be routed under the motherboard to the top chamber, where it will connect to the Optical drive, the 2 80mm fans, the HDD's (adapter) and the res LEDs. The two front 120mm fans and the rear 80mm fans will be wired with its twin and powered/controlled by the motherboard. (CPU and SYS2 FAN headers)

Another thing I'm unsure of at the moment: the optical drive switch. Remember my plan? I wanted to use IR LEDs and a detector, so when you place your hand in front of the drive, (light will be reflected back) and the drive will open. I bought an IR Emitter and Dector from Radio Shack, but I have no idea how to use them in a circuit. I think the emitter is just a plain IR LED and the detector will have to act as a momentary switch. (maybe wrong terminology? ) Anyway, I'm not sure exactly how that works. I know that the IR LED needs a constant source... I was thinking of the 5v or 12v on the Molex connector on theback of the drive. Here is a pic of the tech specs. Annnny help there would be awesome.

 

[p_o_s_pc]

could you wire the IR LED into the 5vSB on the PSU? If so that would be the constant source.Its on even when the computer is off

 

[angelkiller]

could you wire the IR LED into the 5vSB on the PSU? If so that would be the constant source.Its on even when the computer is off

Oh, the LED doesn't need to be on when the computer is off. I meant constant source as in constant when the computer is running. I was planning on wiring that to the 5v molex connector on the back of the drive.

 

[p_o_s_pc]

Oh, the LED doesn't need to be on when the computer is off. I meant constant source as in constant when the computer is running. I was planning on wiring that to the 5v molex connector on the back of the drive.

meme meyou I slapped me because i read it wrong i slapped you for not making it clear for someone that hasn't slept in 2days

 

[Water Drop]

Awesome mod man, I want to do this to build a "Hackintosh", but simply do not have the funds. Would love to do it some day though. Great work. Not exactly a new mod, I've seen it done multiple times before, but yours is a little different then the others I've seen. I like your water cooling plan, should work well with the "cheese grater" air intakes on the G5 case.

 

[angelkiller]

meme meyou I slapped me because i read it wrong i slapped you for not making it clear for someone that hasn't slept in 2days

Fixed, just for you.

Awesome mod man, I want to do this to build a "Hackintosh", but simply do not have the funds. Would love to do it some day though. Great work. Not exactly a new mod, I've seen it done multiple times before, but yours is a little different then the others I've seen. I like your water cooling plan, should work well with the "cheese grater" air intakes on the G5 case.

Thanks. Yeah, I have 18 bookmarks of other people doing this mod too. Oh well. It's still way cool though.

 

[Cuzza]

The emitter is a diode. You can read the reverse voltage is 5V. Ignore that, that is the voltage where the diode stops acting as a diode and will conduct in the reverse direction. Forward voltage reads 1.3V typical, so aim for feeding it that much voltage and you should be ok.

 

[angelkiller]

The emitter is a diode. You can read the reverse voltage is 5V. Ignore that, that is the voltage where the diode stops acting as a diode and will conduct in the reverse direction. Forward voltage reads 1.3V typical, so aim for feeding it that much voltage and you should be ok.

Ok, that makes sense. The reverse voltage is a cool property. It's a diode until a certain voltage then it conducts... that probably has some cool applications.

I used a LED calculator (because I always screw up the math) and I got the following result:

It says that the power dissipated by the resistor is a problem. At .6W, it is. I know I can just get a 1W resistor, but why is the resistor dissapatting so much power? In other LED/resistor setups that I've done, I don't remember them dealing with so much power. Can I use 3 10ohm resistors to spread the power out?

 

[Cuzza]

Hey that's a handy tool. I can do the math but I always forget and it takes me ages. I generally get it right but that is faster, nice.

If you are worried about the heat dissipation by the resistor, power it off 3.3V instead. EDIT: Then you can use a 12-ohm 1/2W resistor

 

[angelkiller]

Hey that's a handy tool. I can do the math but I always forget and it takes me ages. I generally get it right but that is faster, nice.

If you are worried about the heat dissipation by the resistor, power it off 3.3V instead. EDIT: Then you can use a 12-ohm 1/2W resistor

But getting 3.3v up to the front of the case will be kinda hard. The only 3.3v source I know of is off of the ATX connector. I'd rather just get a 1W resistor.

Any one have any ideas about how to use the IR detector?

 

[Cuzza]

It's not hard, just tap into the wire where it leaves the PSU, one wire is easy to route to the front of the case.

EDIT: I am also working on the detector problem, we'll figure this out my friend.

 

[Cuzza]

Right, making progress here.

The detector is a phototransistor. look here for how to use them http://www.fairchildsemi.com/an/AN/AN-3005.pdf or if you want to get complicated http://vorlon.case.edu/~flm/eecs245/Datasheets/Sharp photodevices.pdf

I think this is the one you want to be using:

Here you will connect Vcc and Vout across the switch on the drive. And GND to ground obviously.

Now you want this to operate in switching mode so VCC < RL × IC

we know IC is 50mA (from your specsheet)

VCC you will have to measure across the switch on your drive. I tried on one of mine it was 2.4V, that's probably what you can expect.

So follow my algebra:

VCC < RL × IC
RL > VCC / IC

RL > 2.4 / 0.05 = 48Ω

So use any resistor over 48Ω. I would go substantially over to be sure.

 

[p_o_s_pc]

Fixed, just for you.

Thank you it did help. I think the idea of the 5v on the molex should work just fine

EDIT: from the other post it looks as if that you would have to use a resister so never mind i am not good with this stuff when i have slept.

 

[angelkiller]

Right, making progress here.

The detector is a phototransistor. ....

Awesome! I think that actually makes sense to me. The second link is over my head though.

When you say the resistance needs to be substantially over 48 ohms, how much is substantially? 75 ohms? 100? 1000?

Thanks a bunch, I'll try and set this up when I get home.

 

[Ahhzz]

Actually, just as a quick note, a standard diode "Forward biases" at +.5 v, causing it to pass current. So, anything over .5v gets passed thru the diode. Basically, all you need for that to be active is a half a volt applied to it. An LED can have a different bias voltage, dependent on a lot of things, like color of the LED (White takes like over 3v). Basically, you want to make sure you have enough to push it over, and if it's too dim, you'll have to bump the voltage a little.

 

[angelkiller]

Actually, just as a quick note, a standard diode "Forward biases" at +.5 v, causing it to pass current. So, anything over .5v gets passed thru the diode. Basically, all you need for that to be active is a half a volt applied to it. An LED can have a different bias voltage, dependent on a lot of things, like color of the LED (White takes like over 3v). Basically, you want to make sure you have enough to push it over, and if it's too dim, you'll have to bump the voltage a little.

Ok, everything you said makes sense... but I don't get what you're saying. What do you want me to do? (sorry)


Ok, so I got the optical drive connected to a spare machine. It gets dected by Win7 RC just fine and everything looks fine and dandy. Then I try to eject the tray. Motor noise. No tray. Now I understand why the drive had been opened before and stripped of its mounting screws. The damn thing doesn't work. How disappointing. I was really looking forward to trying the IR thing out.

So does anyone have any ideas about how to fix this? When I hit the eject button all I hear is this motor running, but no movement whatsoever. I've taken the bottom cover off and when I hit eject, nothing moves. I hear that something is moving, but I don't see anything. I'll probably have to get another drive. Luckily this doesn't affect anything because the mounting is a standard. But damn. I was really excited.

The voltage from one side of the switch to ground on that drive was 3.19v. So I need to use a resistor thats 64 or greater ohms, right? (assuming I use that drive)

I'm going to see if I can figure out what's wrong and/or fix it. If not I'll have a go at mounting the rad in the case.

 

[Frick]

Are there some pinions or gearwheels or whatever you call them missing? If you find out the problem you can salvage parts from some other drive to fix this one.

EDIT: Oh, this mod looks really good. Keep it up!