Just an update for anyone who finds this thread..

It's been about a month and both fans are still doing great.  I'm really glad I decided to fix these myself.

These 5850 Xtremes were the right card at the right price at the right time.. They came out during the Bitcoin bubble last year and for what? $140 each? With 1440 Stream Processing Units they are hard to beat.  I like them better than my 5970s for thermal reasons.. I ended up underclocking the 5970s and overclocking the 5850 Xtremes and I end up with about the same output for $140 / core instead of $300+ (the average / core price I paid for the 5970s)

This other fan on eBay has the same fin structure but the wrong mounting and only 3 wire power connector:
New ASUS HD 4870 GTS 250 Video Card Replacement 75mm fan YD128015EL

Here is one which has similar mountings but with 43mm hole spacing and it has a 4 wire power connector.. who knows if the pinout is the same though..
75mm VGA Video Card Fan Replacement 43mm

The inability to find a replacement part is what pushed me to go the repair route..

It looks like it would fit.. The listing specifies 72x10 and the original fan is listed as 80x15 on the Polish website which seems to have them: Wentylator FD8015H12S 80x15 12V

I'm not sure about the mounting placement and I don't have a ruler in mm to measure one.. if you are trying to replace one you should be able to do the measurement yourself (please post your results)

12 hours and still going strong.  It looks like the Tri-Flow did the trick.  I plan to reapply at the 24 hour mark as abracadabra suggested.

My issue with the "too energetic" propellant involves the hole being too small for the tube to enter so I ended up with most of it just spraying all over the place.  Next time I will drill 3/32 holes so I don't have that problem.

My Tri-Flow does not say anything about Teflon.. In fact I don't see any current Tri-Flow products which mention Teflon.  They have something called P.T.F.E. which is considered a hazardous substance.

Here is the Material Safety Data Sheet which lists the contents: [MSDS]

Here is the product info: [Tri-Flow]

[UPDATE] P.T.F.E. (Polytetrafluoroethylene) IS Teflon.  Teflon is a brand name owned by DuPont.  They must have gotten tired of paying to use the name.

Ok..

So I found Tri-Flow at REI in the bicycle section.. Unfortunately, there isn’t an REI very close to my house but they are open to 8pm so I drove out and bought it.

Tri-Flow is pretty cool, it is rated to 475F and the cards will shut the system down at less than half that temp so it seems like a good match. It’s a 4oz Aerosol spray can.. The tube which attaches to the sprayer is too large for the 1/16” hole I drilled (the tube seems to be about 3/32”) and the propellant was way too energetic so I filled a needle oiler and used that to apply the lubricant.





The fans move way more freely than they ever have so I reinstalled them and they are running great (so far at least) with no kick starting required.

I plan to run the fans overnight at 100% and then set them to Automatic tomorrow.

I did apply some force to attempt to pop the hub off but it got to the point where it felt like it would break and I went the hole drilling path..  Can you give a little more detail?  I have 9 of these cards so I expect I'll be doing this again at some point.

Thanks,
Tim

Wow is Tri-Flow hard to find..

I went to like 4 different auto parts stores and a big box hardware store and no dice..

Granger and a local bike shop both seem to have it but neither will be open until Monday..

As of this morning both fans have failed.

The theories I'm working on are:

1. Grease was too thick/hole too small so not enough actually got where I wanted it

2. Hole was in the wrong place.

3. Debris or damage to the sub-surface layer is creating additional drag.

4. Both fans are DOA and not resuscitatable.

1 and 2 I can work with.. 3 I can’t do anything without breaking the fan open and I’m not confident that these are able to be reassembled in working order.  And 4 doesn’t really have a good solution given the lack of replacement parts.

Plan of attack: (fails cascade to next letter)
A.   Buy some Tri-Flow as suggested by abracadabra and apply to existing holes.
B.   Drill new larger holes at the 8mm-10mm offset suggested by SlaveInDebt and apply SuperLube
C.   Tri-Flow on the new holes
D.   
E.   Profit!?

Thanks and I agree.. I only ran it for about a minute as a test and to get the pict.  I ended up getting the Sapphire stickers to stay on (for now).. if they end up falling off later I may have to clean the fan tops to remove the grease remnants and use a new sticker (or tape).   

All right.. It’s done  and so far it looks like Great Success!..

I unfortunately, I didn't read my PMs first because SlaveInDebt got back to me and said he drilled 8mm-10mm from the center.. I was closer to 3mm.. but hey.. whatever works right?

Drilling..  I went about 6mm down.




It seems like I drilled through a top level and a bit into the level below.  I can tell because when insert a needle into the hole and rotate the fan it catches a bit where the hole was initially drilled.

Next I tried to inject as much lubercant into the hole as I could.  I inserted a needle and rotated the fan to try to spread the grease around.  Repeat x2..

Fan one complete:




Zoom! 100%




Fan 2 complete: (more confident this time)




It was a little difficult to get the stickers to stay on.  After wiping the excess grease off of the top of the fans the stickers didn't like staying put.. I got them on but I'll be watching to be sure they don't fall off later.

The second fan didn't want to start at first so I had to kick start it with my finger.. after a few tries it started up and is running solid at 100% like the other.

Back home again:




I'm planning to run them both at 100% overnight and then set them back on automatic sometime tomorrow..

I'll post a status update this weekend..

Please post picts if you do 

The guy who drilled from the bottom said it stopped working again.. SlaveInDebt said he drilled from the top and his still works months later.

I actually removed both stickers and the bottom one has a PCB visible through the gaps.. I didn't feel like drilling through that was a good idea.  Other fans which have a grease reservoir cap seem to have that located on the top.

Thanks for the reply.. good idea..

I wasn't going to use a drill at all..

I have those quick swap drill bits with the hex bases.. I was going to use a 1/16th inch bit and simply spin it in my fingers while pushing down on the end.  That way I can control it better and avoid drilling into something important.

Ok.. I'm going to give this a try.

I have 2 failed 5850 Xtreme fans and I don't really want to deal with the RMA process for a discontinued item..

Given SlaveInDebt's success I am going to attempt his method of drilling through the top..




For the people asking how to take one of these apart..

Remove the 3 screws under the fan:




Unplug the power connector:




Push the power cable under the heat sink while pulling on the fan to give yourself some room to work..








Here is the top with the sticker removed:




I'm planning to drill just outside the ring around the center..

Here is the lubricant I'm planning to use:




Does anyone have any last minute suggestions?

-Tim

There are some problems with your scheme.

1.   Development time is expensive.  
2.   Implementing all these extra features would take a significant amount of time.
3.   Optional pay if you want it add on features are generally not widely adopted.
4.   People who are paying extra for something expect immediate support if there are problems.
5.   Users who are interested in the features but do not want to pay may become dissatisfied.

Basically, eleuthria wouldn’t make his money back and would end up pissing everyone off at the same time.

These sorts of services are better offered by third parties.  That way they can market to a larger number of users across multiple pools.  Also, because they are focused on a their service and not a whole bunch of other things like running a Bitcoin pool they can provide better / faster service to their clients.

It's showing my workers at 84051157.07 MH/s..  I hope I'm getting paid for all these hashes 

There are a few issues with your plan..

Your routers may have a technical limitation. The amount of storage space available for the router firmware is very limited.  There may not be enough storage to accommodate the routing software, mining software and whatever temporary storage is required for routing and mining.

Currently your routers do not run at 100% CPU.  Changing them to 24/7 100% operation would consume more electricity.. likely enough to eliminate any profits (assuming you are paying the bill).  Also, most routers are not designed to run with such high constant load and generally do not have sufficient cooling solutions installed.  You are likely to reduce the operating life of your routers significantly.

You are talking about writing your own custom firmware which would both enable all current router functions as well as some sort of idle cycle detection which would launch the mining activity.  This would be a significant amount of effort.  I suspect you are thinking of deploying this to a significant number of routers to achieve a multiplier to your 0.20 / year Bitcoin bounty.  You might require multiple firmware packages to accommodate different router models and hardware revisions.  The amount of time required to develop, test and maintain all this firmware will be very significant.  

You are likely to see an increase in network issues.  The combined router / miner will not have the same routing efficiency as a dedicated router.  Overburdened and overheating hardware will likely cause intermittent failures. If these routers are located at your work these failures could result in discovery which could cost you your job.

So.. if a few Bitcoins / year is worth hundreds of hours of programming work and the possibility of losing your job then this sounds like a great idea.

I'd take this a step further.  Bitcoin mining is something I am doing anyway.  The heat is a byproduct.  The fact that I can use this byproduct to reduce (or eliminate) my heating costs is where the "free heat" comes into play.  It does not matter that the bitcoin miners cost money to run as long as the heat is a unintended byproduct it is "free".

Yes! This is what I'm banking on.  Right now I have electricity costs for Bitcoin mining and my miners generate heat. This is a constant.  I also have a natural gas forced air furnace.  Because the Bitcoin mining waste heat will reduce my need for natural gas heat I will save money on my gas bill which changes my Bitcoin mining profit equation.

The point isn’t that there aren’t more efficient ways to generate heat.  The point is you have heat that is a free byproduct of an activity you were already engaged in. This free waste heat reduces your need to purchase additional heat at whatever that would cost.  Nobody is suggesting that anyone purchase Bitcoin mining rigs to replace their high efficiency heaters.