Hacking KNC Titan / Jupiter / Neptune miners back to life. Why not?

[lightfoot]

Tarkin: The first one was a Neptune, it had the smaller dies. I've only taken apart one Titan, this is it.

Another note: DO NOT BURN THE CONNECTORS. If the ground burns out first the cube will try to find ground through the 10 pin ribbon cable. This will either blow up the cube's die, the controller, or every cube on the controller. I've pretty much got the failure mode mapped out: If the 12v side lifts first the cube shuts down. If the main ground lifts first all of the other cube dies basically have ground dragged through them causing a serious failure.

[1715133185]

1715133185

[1715133185]

1715133185

[1715133185]

1715133185

[1715133185]

1715133185

[1715133185]

1715133185

[lightfoot]

So an update: I tried the same trick with a chip that was not bent like a pretzel and the glue would not loosen. Problem is if you push too hard you will rip the plastic underlay that carries all the signals. Too much heat and you lift the balls under the die.

Hm.

[lightfoot]

In the meantime I worked on the unit with the lid off. Die 0 seems to be the problem, and sure enough it's shorted according to the output of the dc-dc supply. And the board had 0 ohms on the pins 4,6. So I heated that die to super hot then let it cool down.

The resistances changed. Both on the 4-6 and the dc-dc. Still way too low, but they did change, first evidence a specific die is apparently the cause. Heated it again, back to shorted both places.

So I pulled the die. Real royal bitch. Under it the solder was smeared and there was a bronze ball, which is caused by a really high overtemp. Moral, chip overheated, melted the solder under the die (not the chip) and blew up the board.

[coinut]

was thinking that you could use MEK (Methyl ethyl ketone) in a syringe and inject it all round the sides of the chip to try and get it to penetrate the glue, we use this stuff at work and it will loosen up most bonding.

anyway thought I would add this, It might help for spreader removal.   

[Searing]

Ok now we are talking  Article on Water Cooling a Titan

http://cryptomining-blog.com/tag/knc-titan-cooling/

the cooler (i think this is the one) on Amazon below

http://www.amazon.com/Silverstone-Durable-High-Performance-Adjustable-TD02-E/dp/B00U8IS8F8/ref=sr_1_1/176-2500252-4794240?ie=UTF8&qid=1463481063&sr=8-1&keywords=tundra+td02


I MIGHT try this on my cube with only ONE working die....assuming any mods I do will bring none of the others back


er maybe not ...still too scary ...shudder (ping.....pop Titan Cube Sprinkler System)

anyway for those braver then myself for your consideration

[Longsnowsm]

Ok now we are talking  Article on Water Cooling a Titan

http://cryptomining-blog.com/tag/knc-titan-cooling/

the cooler (i think this is the one) on Amazon below

http://www.amazon.com/Silverstone-Durable-High-Performance-Adjustable-TD02-E/dp/B00U8IS8F8/ref=sr_1_1/176-2500252-4794240?ie=UTF8&qid=1463481063&sr=8-1&keywords=tundra+td02


I MIGHT try this on my cube with only ONE working die....assuming any mods I do will bring none of the others back


er maybe not ...still too scary ...shudder (ping.....pop Titan Cube Sprinkler System)

anyway for those braver then myself for your consideration

This water cooling project looks like it would be a lot of fun.  If you had a Jupiter case laying around it would be kinda cool to retro the Titan's into the Jupiter case which would give you the room to setup the radiators and plumbing for the cooling.  That is good to know that the Titan board has a standard 115X mount for coolers.  That would be very doable I think if you happened to have a case laying around that you could configure for the new cooling options.  That article he didn't show it mounted up in a case, but I don't think it would be too hard to do.

[lightfoot]

Nice. I ran a Neptune for awhile with a Corsair H100 water block secured to the top, it worked very well for a few months keeping the chip cool until the water block sprung a leak. Then all hell broke loose :-)

Leak detection is a big problem: The stupid things don't have temp sensing in the chip dies and little to no thermal mass, as a result the chip can overheat very quickly before software manages to shut things down. One of the leading causes of Titan Die death is when the die inside that chip carrier overheats, as soon as that happens the solder BGA grid under the chip melts, extrudes as solder balls, and either shorts the +.8v supply lines (if you are very lucky) or shorts the signal lines (which are shared by all dies) which kills the whole board.

So you might want to continue running that big heat sink thing: It actually cools very well and as long as the compound they used (which is crap) is secure you should be in good shape. Just remember that if you have an air gap on a corner of the chip, that die will extrude, melt, and fail.

Meantime I tried to source the displays and failed (got the wrong part) for the blown display boards. Anyone know what the right DigiKey or Mouser part number is?

[Searing]

Nice. I ran a Neptune for awhile with a Corsair H100 water block secured to the top, it worked very well for a few months keeping the chip cool until the water block sprung a leak. Then all hell broke loose :-)

Leak detection is a big problem: The stupid things don't have temp sensing in the chip dies and little to no thermal mass, as a result the chip can overheat very quickly before software manages to shut things down. One of the leading causes of Titan Die death is when the die inside that chip carrier overheats, as soon as that happens the solder BGA grid under the chip melts, extrudes as solder balls, and either shorts the +.8v supply lines (if you are very lucky) or shorts the signal lines (which are shared by all dies) which kills the whole board.

So you might want to continue running that big heat sink thing: It actually cools very well and as long as the compound they used (which is crap) is secure you should be in good shape. Just remember that if you have an air gap on a corner of the chip, that die will extrude, melt, and fail.

Meantime I tried to source the displays and failed (got the wrong part) for the blown display boards. Anyone know what the right DigiKey or Mouser part number is?

Well I have a 1 working die cube....so of course I threw like $50 at it (dumb) the long spiky copper heatsinks and the a15-ippc 3000 rpm fan (33% louder then the 2000 rpm ones on
Titans I was told.....er and 33% more wind ..too)

So>>>>>>>>> assuming <<<<<<<<<<<<<<<<<< always a bad idea to assume with kNC stuff

That does NOT work ...and it still is a 1 die working cube only unit (currently at 24 on putty.....they run great all alone at 325 on adv page lol )

Anway if it does not work the above. I MAY just get one of these water coolers for the fun of seeing how fast I can get one die to hash......(I need a life)

HOPEFULLY with the repaste/heatsinks/3000 rpm fan I can get another die working out of this cube. Never will make my money back likely..but it is like an 'itch' to try

[Spike53]

Lightfoot, have you traced out pin 6 of the 10 pin connector on a Titan cube?  I am reading 39 ohms on cube here.  I thought this was a good working cube as i was solid on all four dies before I blew the controller board with another cube.  I am afraid to fire it up as I only have one working controller until I get others fixed.  Have additional tools to arrive over weekend and then can work on the controllers.  Gonna try replacing the FPGA chip again.  Will follow your advice on replacing that.  Got a preheat table.  What you think?

[mmfiore]

I just wanted to report on a problem that I was having with a titan cube. It was purchased in new condition as the person/company that bought it never used the Titans.
Unfortunately when I received the single cube two of the dies, the top two dies did not work properly. I opened the cube and found that the heat sync popped loose because KNC must have over tightened the screw on the end and the base popped loose that holds the screw to the circuit board. I bolted the end back down by getting parts at the hardware store.

Before I had fixed it the symptoms were the top die would hash for about 10 seconds and overheat and shutdown. The second die would immediately shutdown. I could run the top die at a 100mhz and it would run steady but still running a little hot. After properly clamping down the single screw end of the heat synce the top die would run at 300 mhz no heating problem. The second die runs at 175 mhz setting but is using more current than it should be using. The cube amazingly still runs at over 70 MHZ hashing rate even though second die is under clocked. I opened the cube and found that the heat sync popped loose because they over tightened the screw on the end and the base popped loose that holds the screw. I bolted the end back down by getting parts at the hardware store. I got the cube hashing at full power. Thanks Chris for the help.

I have another cube with the same problem that I never could get fully hashing as the two top dies are shut off because the heat sync is not bolted down properly. I will work on that one today and see if I can get that one fully operational. I figured out away to bolt the heat sync down by buying slightly longer screws and using a washer and lock washer and nut to bolt down the front end of the heat sync properly. That is my mechanical solution.

[lightfoot]

No problem. That front screw spot is not very strong, the torque from the heat sink coupled with the slight bending needed to put down the sides to get to the screw can cause it to come loose. Then there is no pressure on the die.

The second die probably has a bad power supply. Note I have seen a number of people try to "repair" the supplies it is very difficult to remove them without a lot of bottom heat, air, and patience. Otherwise you wind up pulling up the traces which can wreck the cube. Run it at 150mhz (half of 300) on the remaining supply and feel good knowing that it is keeping the overall power draw at a reasonable level.

C

[lightfoot]

Something else I am noticing: Have a reference Titan plugged into a controller here. Just noticed the FPGA is getting hot. Pulled the Titan, put nothing else on the controller, still hot. Odd. I think it's going to blow the FPGA soon, no biggie as I can fix it, but is this a symptom of some sort?

Anyone else have a hot FPGA? I've seen a few repairs come in where people put heat sinks on the thing, it's not supposed to run warm. Something else is happpening.

C

[jacobmayes94]

I had a Neptune whos controller blew, the light would flash orange not white when it came on, the FPGA felt hot, i mean hot enough to burn your finger. the second controller the guy sent did the same thing, chip got roasting hot, it would see the cubes but they would not start hashing.

[lightfoot]

I had a Neptune whos controller blew, the light would flash orange not white when it came on, the FPGA felt hot, i mean hot enough to burn your finger. the second controller the guy sent did the same thing, chip got roasting hot, it would see the cubes but they would not start hashing.

I wonder why this happens. Note I unplugged both the cube and the display so that is not it. There's something going on in the FPGA.

Maybe it's the lack of driver buffers, maybe the chips just suck. Regardless it's odd.

I'll order 4 more FPGAs, replace the two bad ones I have here and fix two other boards someone sent in to me. Need to get this stuff done :-)

[Searing]

I had a Neptune whos controller blew, the light would flash orange not white when it came on, the FPGA felt hot, i mean hot enough to burn your finger. the second controller the guy sent did the same thing, chip got roasting hot, it would see the cubes but they would not start hashing.

I wonder why this happens. Note I unplugged both the cube and the display so that is not it. There's something going on in the FPGA.

Maybe it's the lack of driver buffers, maybe the chips just suck. Regardless it's odd.

I'll order 4 more FPGAs, replace the two bad ones I have here and fix two other boards someone sent in to me. Need to get this stuff done :-)

Yeah it is a LOT of work being a Tech Mage....busy busy Everyone is in AWE of your massive Tech Powers (that look like Magic) but no 'you' time

[jacobmayes94]

In my case the chip would still be hot, even when the cubes were unplugged. Incidentally I had a junk Neptune that belonged to a friend of mine, if you connect the ribbon cable BACKWARDS on the cube but correct on the board, as he did, (its not keyed on the cubes) it will then begin the yellow flash thing, but not only that, the chip would roast even when cubes were disconnected. At this point, it won't see the cubes. The yellow flash thing is either a warning from the LED, or the white LED being overdriven to a point where its yellow?

[Searing]

Something else I am noticing: Have a reference Titan plugged into a controller here. Just noticed the FPGA is getting hot. Pulled the Titan, put nothing else on the controller, still hot. Odd. I think it's going to blow the FPGA soon, no biggie as I can fix it, but is this a symptom of some sort?

Anyone else have a hot FPGA? I've seen a few repairs come in where people put heat sinks on the thing, it's not supposed to run warm. Something else is happpening.

C

Hmmm...did not someone on this thread or the swedish thread say the raspberry pi can start to go and do this? I"m not sure. But swap the pi out with another raspberry pi
B+ 512mb would be a quick test imho

again you probably have already done this

[lightfoot]

Hi lightfoot,

I just discovered your thread, so much good info here, reading thru it now.

I picked up another rig, that has 1 bad cube. None of the other cubes will run when the bad cube is plugged into the controller, however the other 4 run great when it's not plugged in. Everything looks brand new

The fan powers up but controller won't recognize it. I replaced the electrolytic capacitor but that didn't help. Can you take a look at it when you get a chance?

thx

I could, however I can guess what's wrong with it: When the chips get too hot from running too fast with a bad heatsink connection the solder under the dies inside the chips starts to heat up. Reason is the heat starts going down INTO the board instead of up and out via the tower heatsink.

As a result (and since the connection to the board is through a interconnecting junction at the bottom of the chip) that solder gets hot. When it gets hot enough to melt it extrudes solder balls around the die and kills the die. Sometimes it just takes out the .8v supply lines (most of the bga pins are this) however sometimes it hits either the die control lines or (way worse) the die power lines to the hotel circuits on the die.

That line is powered by the FPGA board's TMS chip because KNC was (efficient, cheap, something) and didn't put in a small dc-dc charge pump. However since it is shared by all dies a short will basically kill the cube and short the line to all of the other cubes through the 10 pin. Result=sucks.

The other way this blows up is if the cube has a bad power plug from running too hard with only one supply. If the gound lifts before the +12, the cube will try to draw power from the ground line on the 10 pin, then will try to draw ground *through* the signal lines and the hotel power supply (because 1.8v is closer to ground than 12v is). This blows out the chip, the controller (sometimes) and makes a large mess.

C

[fanatic26]

I used a dewalt heat gun to reflow hundreds of cointerra boards back to life, its worth a shot.

[lightfoot]

It could work, can you tell if the pin 4,6 on the 10 pin connectors are shorted to ground?

C