Searing
Copper Member
Legendary
Offline
Activity: 2898
Merit: 1465
Clueless!
|
|
February 13, 2016, 07:13:01 AM |
|
I recently bought titan on eBay. One cube got damaged Durring transport. Heatsink lose (screw holding it got lose) which i think is common. I managed to put it back with some DIY. Was hashing fine. Then unplugged it and now its dead and on top of it psu shuts down when connected to that cube. Any idea why? How to fix it?
the below may help (you may have already seen it) swedish guy videos see swedish one first (better video) and english one 2nd (not as good on details) I assume this is what you mean by 'heatsink and paste" stuff https://www.youtube.com/watch?v=O-RBI4am-wYhttps://www.youtube.com/watch?v=-G_zcJ0vbOwon a side note many have said on the old knc forums that issues with lousy paste job of thermal paste on the cubes on 1st couple months titans caused or contributed to dead die on cube issues on the 1st batch anyway you are probably referring to some other techy thing ...for what the above is worth as a repeat on the thread
|
Old Style Legacy Plug & Play BBS System. Get it from www.synchro.net. Updated 1/1/2021. It also works with Windows 10 and likely 11 and allows 16 bit DOS game doors on the same Win 10 Machine in Multi-Node! Five Minute Install! Look it over it uninstalls just as fast, if you simply want to look it over. Freeware! Full BBS System! It is a frigging hoot!:)
|
|
|
psy112
|
|
February 13, 2016, 10:24:11 AM |
|
I looked through the swedish video, looks pretty simple and I have the exact same issue with higher heat on my 1-2 chips. The question/wondering is, I'm in the lawsuit again KNC and I'm not sure If I'm allowed to alter the machine at all? Anyone knows? I recently bought titan on eBay. One cube got damaged Durring transport. Heatsink lose (screw holding it got lose) which i think is common. I managed to put it back with some DIY. Was hashing fine. Then unplugged it and now its dead and on top of it psu shuts down when connected to that cube. Any idea why? How to fix it?
the below may help (you may have already seen it) swedish guy videos see swedish one first (better video) and english one 2nd (not as good on details) I assume this is what you mean by 'heatsink and paste" stuff https://www.youtube.com/watch?v=O-RBI4am-wYhttps://www.youtube.com/watch?v=-G_zcJ0vbOwon a side note many have said on the old knc forums that issues with lousy paste job of thermal paste on the cubes on 1st couple months titans caused or contributed to dead die on cube issues on the 1st batch anyway you are probably referring to some other techy thing ...for what the above is worth as a repeat on the thread
|
|
|
|
wolverine5pl
|
|
February 13, 2016, 10:27:28 AM |
|
I recently bought titan on eBay. One cube got damaged Durring transport. Heatsink lose (screw holding it got lose) which i think is common. I managed to put it back with some DIY. Was hashing fine. Then unplugged it and now its dead and on top of it psu shuts down when connected to that cube. Any idea why? How to fix it?
Um. What speed were you hashing at? And um... How did you put the heat sink back on? And um.... can you take a picture of the board? I run a repair service for sad Titans, but I'm wondering what happened to this one. when it arrived after trying to fix it: did run for a while fine. when i turned it off and started later what happens is PSU shuts down. and psu is ok. tried 3 different.
|
|
|
|
|
The Avenger
|
|
February 13, 2016, 10:51:19 AM |
|
|
"I am not The Avenger" 1AthxGvreWbkmtTXed6EQfjXMXXdSG7dD6
|
|
|
psy112
|
|
February 13, 2016, 11:34:10 AM |
|
strange because I got this from my lawyer "Finally, the date for the main hearing is set to 16, 17 and 18 (half day) Mars 2016."
|
|
|
|
The Avenger
|
|
February 13, 2016, 12:37:47 PM |
|
strange because I got this from my lawyer "Finally, the date for the main hearing is set to 16, 17 and 18 (half day) Mars 2016."
The first line reads "Nacka District Court rejects class action against KnCMiner", so I assumed it failed? But it's hard to tell from that translated article what happened. But if you say prosecutions against kfc are still going ahead, that's good news. Hopefully they'll have their day in court before they all murder each other
|
"I am not The Avenger" 1AthxGvreWbkmtTXed6EQfjXMXXdSG7dD6
|
|
|
wolverine5pl
|
|
February 13, 2016, 12:48:58 PM |
|
Pictures of board:
|
|
|
|
Maxumark
|
|
February 13, 2016, 03:49:51 PM Last edit: February 13, 2016, 04:38:20 PM by Maxumark |
|
Any pointers on repairing burnt power connectors on cubes ?
Here is the best, quickest, and easiest way to deal with a burnt connection in my opinion. It also allows about 3/16in additional contact between the power connector and the pins. Sorry its a bit of a hack. and you must absolutely be sure to put the connector on clip side up, ground wires up since the will not be a receiver connector on the board anymore. The burnt connector will be brittle, and easy to break up with a pair of side cutters. Do it in small pieces until just the very bottom is left. Then pry up gently with a screwdriver to dislodge the 2 plastic retainers and it will slide right off. Then take a piece of 220 grit sand paper to clean up the pins. Clean everything up and carefully push the power connector on the bare pins while using a flashlight to make sure its on right. It will now go further on the pins making a better connection.
|
Mining LTC and other alts since 2014 when I thought I missed the BTC train.
|
|
|
lightfoot
Legendary
Offline
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
|
|
February 13, 2016, 04:01:05 PM |
|
Do you have a pic of it? Do you know what pins you're looking at, just don't know which is which?
It's tiny, getting a picture even with a 6x loupe on the camera is a job. These guys have a picture of one of the parts, but no description on what it is: http://www.aliexpress.com/item/Transistor-IC-chip-SMD-6-pin-IC-THR-SOT23-6/32240467148.htmlLooking at the grainy picture I have so far of the part it could be a 2 as the first letter.
|
|
|
|
psy112
|
|
February 13, 2016, 04:09:11 PM |
|
strange because I got this from my lawyer "Finally, the date for the main hearing is set to 16, 17 and 18 (half day) Mars 2016."
The first line reads "Nacka District Court rejects class action against KnCMiner", so I assumed it failed? But it's hard to tell from that translated article what happened. But if you say prosecutions against kfc are still going ahead, that's good news. Hopefully they'll have their day in court before they all murder each other I read the original article, since I'm swedish. But it said the same thing. Have sent the article to my lawyer, so have to wait and see what he says about it.
|
|
|
|
GenTarkin
Legendary
Offline
Activity: 2450
Merit: 1002
|
|
February 13, 2016, 06:32:09 PM |
|
Pictures of board: HOLY SHIT those 2 dies got way way way to fucking hot. Most likely fried the package substrate or something along those lines. If its not working, that would be why. I really wish KNC would have designed the dies with ondie thermistors. So we knew what the fuck the actual temps of the dies are...
|
|
|
|
lightfoot
Legendary
Offline
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
|
|
February 13, 2016, 07:02:13 PM |
|
Pictures of board: HOLY SHIT those 2 dies got way way way to fucking hot. Most likely fried the package substrate or something along those lines. If its not working, that would be why. I really wish KNC would have designed the dies with ondie thermistors. So we knew what the fuck the actual temps of the dies are... Yeah I was thinking that. When I say I have seen burned titans this is what I mean... *However* (and here is the magic however button) the chip itself doesn't have any of the +12v lines going to it. Therefore it should not be crowbarring the supply. It could be a blown FET on the power supplies, or a blown cap that's not showing up, but it's not just the chip. The chip has the .7 volt supplies going to it, some SPI lines one for each quadrant and the jury is still out as to what else it does.
|
|
|
|
2112
Legendary
Offline
Activity: 2128
Merit: 1073
|
|
February 13, 2016, 07:04:24 PM |
|
I really wish KNC would have designed the dies with ondie thermistors. So we knew what the fuck the actual temps of the dies are...
Thermistor wouldn't work without calibration which would be additional expense in testing. But some simple circuit with diodes or BJTs using reverse polarized junctions could have sensible accuracy without additional calibration.
|
|
|
|
wolverine5pl
|
|
February 13, 2016, 07:06:45 PM |
|
Pictures of board: HOLY SHIT those 2 dies got way way way to fucking hot. Most likely fried the package substrate or something along those lines. If its not working, that would be why. I really wish KNC would have designed the dies with ondie thermistors. So we knew what the fuck the actual temps of the dies are... Design of the cubes is terrible starting from screws holding everything. I guess when I put it back thermal paste wasnt spread properly. dont have luck with titans and ebay. Cant wait for A4 miners
|
|
|
|
GenTarkin
Legendary
Offline
Activity: 2450
Merit: 1002
|
|
February 13, 2016, 07:56:27 PM |
|
Pictures of board: HOLY SHIT those 2 dies got way way way to fucking hot. Most likely fried the package substrate or something along those lines. If its not working, that would be why. I really wish KNC would have designed the dies with ondie thermistors. So we knew what the fuck the actual temps of the dies are... Yeah I was thinking that. When I say I have seen burned titans this is what I mean... *However* (and here is the magic however button) the chip itself doesn't have any of the +12v lines going to it. Therefore it should not be crowbarring the supply. It could be a blown FET on the power supplies, or a blown cap that's not showing up, but it's not just the chip. The chip has the .7 volt supplies going to it, some SPI lines one for each quadrant and the jury is still out as to what else it does. Yeah, of course the 12v aint going directly to the dies =P ... the DCDC's convert the 12v to .7v for the dies(chips as u call them). Oh and, did u see my question about the SPI data u obtained? How did u get that? Im trying to figure out commands to run to get data from SPI but I cant =/ I used spi-test and cant figure out how to "talk to the dies"
|
|
|
|
teslaman
Member
Offline
Activity: 65
Merit: 15
|
|
February 14, 2016, 12:53:01 AM |
|
yeah thanks glad i tripped over the knc ip address again
wondering if you or others know...or have better google skills then me
Sorry I haven't replied sooner, some personal stuff has nearly destroyed me, so I haven't felt like coming around much at all these days. Anyways...yeah I have WinHTTrack sitting around too, but haven't used it much. As far as I know, it's really only good for cloning a site? I think it's near impossible to auto-backup that Wayback Machine archive with existing tools, since you can't really auto-follow each link with them. Whenever I do it these days, I just save each page in whatever browser I'm using. It's painstaking and time consuming. Maybe others know a better way. As far as merging those archives, I think you would have to go through the HTML and re-connect all of the links that are now different. Added bonus would be to delete all of the links that weren't archived. One quite old batch tool that I like using for changing links is SearchAndReplace V2.5.1.0. If someone knows a better one, I'm all ears.
|
|
|
|
lightfoot
Legendary
Offline
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
|
|
February 14, 2016, 12:59:44 AM |
|
Yeah, of course the 12v aint going directly to the dies =P ... the DCDC's convert the 12v to .7v for the dies(chips as u call them).
Well, you never know. They have to power the hashing engines and the computer that collects the engines on each die and formats it for the FPGA. That takes more than .7 volts, so either they have a charge pump on the chip or they have something on this stupid board that does it. Oh and, did u see my question about the SPI data u obtained? How did u get that? Im trying to figure out commands to run to get data from SPI but I cant =/ I used spi-test and cant figure out how to "talk to the dies"
I cheated. Using the driver code in bfgminer trip to diagnostic mode and save all the output to the disk. Question: Just include the driver code in your own routine and push/pull info from the die that way. Now what are those SOT-6 chips?
|
|
|
|
GenTarkin
Legendary
Offline
Activity: 2450
Merit: 1002
|
|
February 14, 2016, 01:32:46 AM |
|
Yeah, of course the 12v aint going directly to the dies =P ... the DCDC's convert the 12v to .7v for the dies(chips as u call them).
Well, you never know. They have to power the hashing engines and the computer that collects the engines on each die and formats it for the FPGA. That takes more than .7 volts, so either they have a charge pump on the chip or they have something on this stupid board that does it. Oh and, did u see my question about the SPI data u obtained? How did u get that? Im trying to figure out commands to run to get data from SPI but I cant =/ I used spi-test and cant figure out how to "talk to the dies"
I cheated. Using the driver code in bfgminer trip to diagnostic mode and save all the output to the disk. Question: Just include the driver code in your own routine and push/pull info from the die that way.
no idea how to do that =( Would implementing disabling / enabling of the dies be hard in the bfgminer driver? cuz that code is unfinished =/ ...
|
|
|
|
edgar
Legendary
Offline
Activity: 1848
Merit: 1001
|
|
February 14, 2016, 05:33:56 AM Last edit: February 14, 2016, 05:59:00 AM by edgar |
|
|
|
|
|
|