Butterfly Labs - Bitforce Single and Mini Rig Box

[despoiler]

What heatsink compound have you guys been using as a redo on the singles

Personally, I despise thermal grease.  I plan to replace the grease in my mini rigs with phase change material (PCM).

Some info:
http://www.electronics-cooling.com/2002/05/understanding-phase-change-materials

Where I buy it:
http://www.opentip.com/Electronics/Startech-Heatsink-Thermal-Pads-Pack-p-1075383.html

Also on Newegg:
http://www.newegg.com/Product/Product.aspx?Item=N82E16835230030

Experience has led me to love PCM for CPU and GPU cooling.  Never tried on an FPGA, but I'm definitely going to.

Thanks,but I have grown VERY fond of Ceramique for CPU's & GPU's,for the last 6-8 years,but for these FPGA's I think it's a little too thick.

I'm mainly curious about what folks have actually used for the FPGA's.

So far MX-4,any others??

Please keep the suggestions coming,Thanks

Most people use MX-2.  It's pretty much the price to performance king of thermal pastes.  I think it's about $4-5 cheaper than MX-4. 

[e21]

I used Thermaltake TG-1, lowered temps by about 3-4*, stopped my single from throttling 

[Unacceptable]

I used Thermaltake TG-1, lowered temps by about 3-4*, stopped my single from throttling  

Thanks for the great suggestions  

MX-4

MX-2

Thermaltake TG-1

Keep em coming.Always good to have choices  

Again,I would like to know what you have used on any FPGA.

[bitpop]

http://www.newegg.com/Product/Product.aspx?Item=N82E16835100007

[rjk]

Tuniq TX-4 is discontinued, but you can still get it a few places. It has the best thermal conductivity in real-world applications, but is expensive. Tuniq TX-2 is almost as good, and much less expensive.

[Cablez]

I use the prolimatech pk-1 for most everything.

For troublesome chips I use the IC Diamond compound.  Considering the tops of the chips are sanded one would not have to worry about the diamond scratching things.

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

[BTC-engineer]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

[DeathAndTaxes]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Just make sure the plate is copper and relatively thick (1/8"?).  You are going to have two junctions with high thermal resistence (chip & plate then plate & heatsink).  Using a material with good conductivity and sufficient mass will improve heat transfer.

[BFL-Engineer]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

[BTCurious]

I recently won gigavps' BFL Single raffle, and the single arrived yesterday. It was working less than satisfactory though. It's pretty hot here this week, ~26°C (78°F), that may have something to do with it. The stock firmware throttled down very quickly, when the reported temperature reached around 50°C, and afterwards only doing something between 500 and 700 Mhash/sec. It would also give nearly only errors on the light diagnostic.

A lower firmware seemed to help: 816 only sometimes seemed to throttle down, and could complete a light diagnostic most of the times.

Anyway, after reading some stuff, it seemed like the thermal paste might be an issue, so today I replaced the thermal paste with a Cooler Master paste. I took some pictures and put them in an album.

Here are some highlights:


It's a revision 3.


Sadly, the fancy new cooling block does not actually cover one of the aluminium plates fully.


The sanding on the chips themselves doesn't look too pretty either…


I blobified the uncovered part, hoping it would help.

And the result is:
It didn't help. The behavior is the same as before. For no particular reason though, I tried out the fastest firmware available. This actually gets me good rates O.o
So I don't really have any issue anymore, and am just sharing the album

[BFL]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

It's also worth pointing out that the silicon under the heat spreader is much smaller than the heat spreader itself.  The heat sink covers the full area of the heat generating silicon's footprint, so there's no real value in trying to cover the end of the heat spreader which has no heat producing component underneath.

Here's an example (not our chip)  

[spiccioli]

It's also worth pointing out that the silicon under the heat spreader is much smaller than the heat spreader itself.  The heat sink covers the full area of the heat generating silicon's footprint, so there's no real value in trying to cover the end of the heat spreader which has no heat producing component underneath.

Here's an example (not our chip)  

So chips are not sanded, it's the heat spreader what we see when someone removes the heatsink...

spiccioli.

[rjk]

So chips are not sanded, it's the heat spreader what we see when someone removes the heatsink...

spiccioli.

The heatspreader is an integral part of the chip packaging, they don't apply it themselves.

[e21]

I recently won gigavps' BFL Single raffle, and the single arrived yesterday. It was working less than satisfactory though. It's pretty hot here this week, ~26°C (78°F), that may have something to do with it. The stock firmware throttled down very quickly, when the reported temperature reached around 50°C, and afterwards only doing something between 500 and 700 Mhash/sec. It would also give nearly only errors on the light diagnostic.

A lower firmware seemed to help: 816 only sometimes seemed to throttle down, and could complete a light diagnostic most of the times.

Anyway, after reading some stuff, it seemed like the thermal paste might be an issue, so today I replaced the thermal paste with a Cooler Master paste. I took some pictures and put them in an album.

Here are some highlights:


It's a revision 3.


Sadly, the fancy new cooling block does not actually cover one of the aluminium plates fully.


The sanding on the chips themselves doesn't look too pretty either…


I blobified the uncovered part, hoping it would help.

And the result is:
It didn't help. The behavior is the same as before. For no particular reason though, I tried out the fastest firmware available. This actually gets me good rates O.o
So I don't really have any issue anymore, and am just sharing the album

That paste doesn't look much better than the stock that the single ships with. I would try something higher-grade like IC Diamond, Thermaltake TG-1, or Arctic MX-4 maybe.. I used TG-1 and it dropped my temps by a solid 3*C

[e21]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

I would agree here, don't add anything extra or your temps will just go up.

The heatsink itself could be better for sure, the aluminum part could be solid copper for start; and I noticed it is slightly short as well, however the copper pipes seem to touch the surface of the chip where the aluminum part cuts off, so that's something I suppose.

[BFL]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

I would agree here, don't add anything extra or your temps will just go up.

The heatsink itself could be better for sure, the aluminum part could be solid copper for start; and I noticed it is slightly short as well, however the copper pipes seem to touch the surface of the chip where the aluminum part cuts off, so that's something I suppose.

Again...  to clarify...   (trying to save unnecessary efforts among users), the chip itself is smaller than the aluminum heat spreader you see.  There is little purpose in trying to extend the heatsink out to the edge of the heat spreader because there's no heat being generated at that point...  There is only air there.  See picture above for visual reference.

The reason the units all have a slightly different throttling point is because the chips are full wafer supply.  There is a natural variance which will frustrate someone trying to 'fix' what they assume is an abnormality. 

[Inspector 2211]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

I would agree here, don't add anything extra or your temps will just go up.

The heatsink itself could be better for sure, the aluminum part could be solid copper for start; and I noticed it is slightly short as well, however the copper pipes seem to touch the surface of the chip where the aluminum part cuts off, so that's something I suppose.

Again...  to clarify...   (trying to save unnecessary efforts among users), the chip itself is smaller than the aluminum heat spreader you see.  There is no purpose in trying to extend the heatsink out to the edge of the heat spreader because there is no heat being generated under the edge...  There is only air there.  See picture above for visual reference.

From a technical perspective, I beg to disagree.

There's a thermal resistance RT1 from the die to the heat spreader, and a thermal resistance RT2 from the heat spreader to the cooler. While RT1 cannot be changed, it is a good thing to minimize RT2. One way of minimizing RT2 is by removing the heat from the whole area of the heat spreader.

[BFL]

Has anyone else noticed how the heatpipe cooler on the rev3 doesn't really cover the second FPGA? It is short by about 5cm. The center of the chip is covered but one whole side is left open.

I noticed this also. If this heatpipe is really custom made the job wasn't done good.

I'm thinking about putting a metal plate between the FPGA's and the heatpipe. This metal plate should cover the full area of both FPGA's and transfer it to the heatpipe. Maybe this could improve the cooling which is obviously mainly limited by the only parted covered FPGA.

Has someone already done this?

Adding any additional substance will only increase thermal resistance and makes cooling less efficient.
The aluminium plate which the heatsink sits on, is a by itself a heat-spreader directly attached to the die inside.


Regards,
BF Labs Inc.

I would agree here, don't add anything extra or your temps will just go up.

The heatsink itself could be better for sure, the aluminum part could be solid copper for start; and I noticed it is slightly short as well, however the copper pipes seem to touch the surface of the chip where the aluminum part cuts off, so that's something I suppose.

Again...  to clarify...   (trying to save unnecessary efforts among users), the chip itself is smaller than the aluminum heat spreader you see.  There is no purpose in trying to extend the heatsink out to the edge of the heat spreader because there is no heat being generated under the edge...  There is only air there.  See picture above for visual reference.

From a technical perspective, I beg to disagree.

There's a thermal resistance RT1 from the die to the heat spreader, and a thermal resistance RT2 from the heat spreader to the cooler. While RT1 cannot be changed, it is a good thing to minimize RT2. One way of minimizing RT2 is by removing the heat from the whole area of the heat spreader.

If you could choose full or partial coverage of the overhang...  certainly full would be an easy choice.  I'm just pointing out that the practical value of doing so in this case is nil.   This is the result of our thermal simulations in designing the heatsink.  The reason the few mm on the edge of the spreader are left uncovered is for package fitting. 

[malevolent]

If you could choose full or partial coverage of the overhang...  certainly full would be an easy choice.  I'm just pointing out that the practical value of doing so in this case is nil.   This is the result of our thermal simulations in designing the heatsink.  The reason the few mm on the edge of the spreader are left uncovered is for package fitting. 

Not that again, please...

[Cablez]

I get what you are saying and I have to agree as its only 5mm in this case. The only reason I brought it up was I noticed it on one unit that had very poor thermal properties compared to its brethren.

Upon examining more closely I would have to say that it is probably the FPGAs co-planarity that might be the culprit. I may end up removing the pushpins and using bolts to secure the heatsink (gently of course).