antminer U1 0.00Gh/s

[Sp00n1980]

so i have an issue with 2 of my antminer U1's, both have stopped hashing

in BFGminer i see  "AMU 9:       |  2.10/ 2.10/ 0.00Gh/s | A:   2 R: 0+0(none) HW:0/none" that's after 3 hours (though it can be after 24hrs the results are still the same)   and the other is exactly the same.

These are on a powered d-link DUB H7 hub, fed off a 350w PSU (so plenty of juice), over clocked to 0981, with a 120mm fan on them. Even if i move these ants onto there own hub, i still get nothing  

has anyone had similar, and is there a fix?

edit, both the heat sinks are tightened up correctly and there is no debris behind them etc

[1715155072]

1715155072

[1715155072]

1715155072

[1715155072]

1715155072

[1715155072]

1715155072

[nybbler905]

Closest I've had with that is a power issue on my powered hub... there wasn't power...

AntMiners tend to actually use more than the 500mA that most hubs try and put out per USB connection ( especially if you overclock, ty clock=x0981 ! )

Unplug the USB from the PC and if they are still lit, it's got power.  If the one in question goes out, not enough power.  My issue was a cheap regulator IN the hub ( bypassed by cutting the 5V line on the USB cable and using a D-Link wireless router's power supply for the hub ).

With the mod I use, you need to keep the hub unplugged until you actualy start the PC to prevent issues.

[Sp00n1980]

thanks for the reply 

i can assure you the hub is not the issue, as i have moved both my trouble some the ants on there own hub, moved the ants onto my PC etc etc, with exactly the same results  these U1's just don't seem to like life.

I have tried them at 0781 up to 0A81 (only briefly at 0A81, just to see if that made a difference), with the same results, its like they just don't want to run 

i guess these 2 could just be dud's 

[nybbler905]

I use Ubuntu ( Linux ) so if you don't have root ( which you must with the others to work ) it wouldn't.  With all the steps you seem to have covered, yea, looks like dead ants, you might want to contact the seller or go directly to the manufacturer to see if they can be RMA-ed.

If you have access to SMD rework stuff ( The not recomended way to look at it ) you may want to do a component by component check between a working unit and one of the dead ones in case it's as simple as missing a resistor, short, poorly added resistor for overclocking.  Shorting out the ( originally empty ) resistor spot kills both the ASIC and the voltage regulator as an example of what could cause dead AntMiners.  I tend to void warrenties myself and look at options to resolder stuff to make them work.  With the change over to the U2, it may be difficult to have it changed out for another ( and working ) U1 ( U2 has bigger heatsink ).

Good luck with it.

[Sp00n1980]

hows this for odd ..... one of my duff ones is now hashing .... instead of the 2Gh/s, it making 1Gh/s  i have no idea how it happened, but i have moved that rig today(but i never unplugged any ants), so that *may* have helped 

one off them i have voided the warranty (well both if you consider the minor over clock to 0981), i tried the resister upgrade, but it made no difference, on that one i was considering trying something similar to the "YLOD PS3 fix" (no clean flux and some "monitored" heat).

[nybbler905]

hows this for odd ..... one of my duff ones is now hashing .... instead of the 2Gh/s, it making 1Gh/s  i have no idea how it happened, but i have moved that rig today(but i never unplugged any ants), so that *may* have helped 

one off them i have voided the warranty (well both if you consider the minor over clock to 0981), i tried the resister upgrade, but it made no difference, on that one i was considering trying something similar to the "YLOD PS3 fix" (no clean flux and some "monitored" heat).

dumb as it sounds, with both CGMiner and BFGMiner you have to restart the miner for the ' new ' miner to use the clock settings.  I still recommend removing all but the 2 defunct ones from the system and trying just those 2, each on it's own USB chain or on different powered hubs.

I bought a cheap powered hub then found out it has the sideways mounts so I can only fit 4 miners on it AND they have power switches ( nice blue glow to show there is power going to the miners ).  I hit a switch while moving the hub and had to restart as when it re-detected the Ant, it was only getting 1GH/sec until I closed and reopened BFGMiner.  With the other one ' suddenly ' working, I'd definitly look at the cables and see if you can swap USB cables to see if it totaly clears things up.  Been there, Done that, got a USB B cable that only does power that came with a Pic18f4520 board  for powering it, they went cheap and didn't add the other 2 wires IN the cable but didn't remove the connectors at both ends so it looks good on the outside, just can't do more than power things with it.  Just saying, the cables may be adding to the issue.

Good luck with the resistor upgrade.  They DO GET REALLY HOT so, if you have some old/dead ATX power supplies, a drill and a hack saw, there are aluminum heat sinks waiting to get cut down to the right size for Ants...   

[Sp00n1980]

dumb as it sounds, with both CGMiner and BFGMiner you have to restart the miner for the ' new ' miner to use the clock settings.  I still recommend removing all but the 2 defunct ones from the system and trying just those 2, each on it's own USB chain or on different powered hubs.

yeah, done that, still no joy  

I bought a cheap powered hub then found out it has the sideways mounts so I can only fit 4 miners on it AND they have power switches ( nice blue glow to show there is power going to the miners ).  I hit a switch while moving the hub and had to restart as when it re-detected the Ant, it was only getting 1GH/sec until I closed and reopened BFGMiner.  With the other one ' suddenly ' working, I'd definitly look at the cables and see if you can swap USB cables to see if it totaly clears things up.  Been there, Done that, got a USB B cable that only does power that came with a Pic18f4520 board  for powering it, they went cheap and didn't add the other 2 wires IN the cable but didn't remove the connectors at both ends so it looks good on the outside, just can't do more than power things with it.  Just saying, the cables may be adding to the issue.

they have been in there own d-link dub H7 (and i have swapped them about figuring it may be a bad hub), no joy ... straight on my PC, no joy 

Good luck with the resistor upgrade.  They DO GET REALLY HOT so, if you have some old/dead ATX power supplies, a drill and a hack saw, there are aluminum heat sinks waiting to get cut down to the right size for Ants...  

the resistor upgraded one isn't even getting warm ... i guess its dead  

[nybbler905]

There is dead, then there is dead... If you are using Linux, open a terminal window ( in Ubuntu it's control+alt+T ) and type lsusb with just the 2 assumed dead connected.  Usually you don't get the hubs listed.  If 1 shows, pull 1 and try again to see which shows up.  This will only show the ones that can still ' talk ' on the usb line and not anything else.  With the resistor overclock not even warming it up, it looks alot like SMD rework of the voltage regulator ( U4 by the resistors for the overclocking ) may be bad on them.  The dot indicating pin 1 should be closest to the USB connnector on it ( just in case in mass production on yours they got put in upside down ).  The power regulator is suseptable to transient current when soldering so a grounded tip soldering iron is recomended or hot air work station.  For what little cost those ICs are at places ( Mouser, etc ) it might be worth just checking that they are the correct orientation and no weird burn marks near them.  Also all the pins on my ants for U2 ( the ASIC ) are clean other than the solder and can be read with the USB connector DOWN.  Looking at the board, you will get what I mean.  you haven't said ( and I forgot to ask ) if the red LED goes on when you plug it in.

[Sp00n1980]

There is dead, then there is dead... If you are using Linux, open a terminal window ( in Ubuntu it's control+alt+T ) and type lsusb with just the 2 assumed dead connected.  Usually you don't get the hubs listed.  If 1 shows, pull 1 and try again to see which shows up.  This will only show the ones that can still ' talk ' on the usb line and not anything else.  With the resistor overclock not even warming it up, it looks alot like SMD rework of the voltage regulator ( U4 by the resistors for the overclocking ) may be bad on them.  The dot indicating pin 1 should be closest to the USB connnector on it ( just in case in mass production on yours they got put in upside down ).  The power regulator is suseptable to transient current when soldering so a grounded tip soldering iron is recomended or hot air work station.  For what little cost those ICs are at places ( Mouser, etc ) it might be worth just checking that they are the correct orientation and no weird burn marks near them.  Also all the pins on my ants for U2 ( the ASIC ) are clean other than the solder and can be read with the USB connector DOWN.  Looking at the board, you will get what I mean.  you haven't said ( and I forgot to ask ) if the red LED goes on when you plug it in.

chip is the right way round. the red LED is on and the green LED flashes (like my good ants) but in the miner window i get "AMU 0:       |  2.10/ 2.10/ 0.00Gh/s " ... no accepts no rejects and no hardware error's .... maybe its a ghost ant  dead and mining in spirit only 

good thinking, i am going see if i can get hold of the voltage regulators 

[nybbler905]

BFGMiner's been giving me issues lately, i've had that happen UNTIL it submits a share... but it's definitly at least 1 share per ant per min so it should show some hash rate...

Been working on an erupter and its now got issues, think I blew the buffer on it as it's acting that way now, the ant is a direct connect so that isn't likely.  If you have a USB camera you could use as a form of microscope ( I managed to find a used one for 10 bux used since they didn't know to pull the cap off the end ) it's darn handy for looking at those SMD parts.  It might be as simple as pull off the heat sink and the pad, put in a toaster oven for 4 min at 200 F and let cool before reaplying the pad then thermal paste then the heat sink.  BTW there are 2 green LEDs at the top corner, they both flash on found shares, lower one flashes on work done, no share.

Edited to add the following
i think my eyes were playing tricks on me, in the dark I only see 1 green flash..... but the other one looked like it flashed.... so which one flashes on the deadish one?

[Sp00n1980]

BFGMiner's been giving me issues lately, i've had that happen UNTIL it submits a share... but it's definitly at least 1 share per ant per min so it should show some hash rate...

Been working on an erupter and its now got issues, think I blew the buffer on it as it's acting that way now, the ant is a direct connect so that isn't likely.  If you have a USB camera you could use as a form of microscope ( I managed to find a used one for 10 bux used since they didn't know to pull the cap off the end ) it's darn handy for looking at those SMD parts.  It might be as simple as pull off the heat sink and the pad, put in a toaster oven for 4 min at 200 F and let cool before reaplying the pad then thermal paste then the heat sink.  BTW there are 2 green LEDs at the top corner, they both flash on found shares, lower one flashes on work done, no share.

Edited to add the following
i think my eyes were playing tricks on me, in the dark I only see 1 green flash..... but the other one looked like it flashed.... so which one flashes on the deadish one?

Just had to do a system reboot, at first neither of them flashed (0.00Gh/s on bfg miner), so i stopped the miner, and the (green led towards the) middle LED came on and stayed on, i then restarted the miner and it started hashing with the one toward the middle flashing (0.34Gh/s). every now and then the other green LED would flash though.

[nybbler905]

At this point, not sure what else to go on for getting it working other than keeping your eyes open for more Ants ( cheap and possibly dead ) so you can start swapping ICs and see what makes them stop working.  I ( for one ) can't exactly afford to drop off 2GHash/sec in hopes of getting a half dead Ant back up and running ( and gain that 2 GHash/sec back from the dead ).  Other than all that's been done by you and what I've pointed out for what to look at, the only thing left is quite literally the board and that's not likely.  If you have a reflow oven/pan/toaster I'd give it a good shot of flux and bake the unit at 200 F. for 4 min, cool then clean and see what happens.  Anything past that at this point is kicking a dead ant....

[Sp00n1980]

very true, i have ordered some voltage regulators, so i plan to change 1 out on the very dead ant ... whats the worst that could happen

[nybbler905]

had to move my USB mining to an old dual Coppermine rig ( yes, it does XP and barely ) in Ubuntu with a D-Link wireless's power adapter to the powered hub ( so it's got 2.5 A at 5V and doesn't get warm ) with 3 ants, 1 Erupter, 1 D-Link USB extender cable to a second erupter and thought it was a power issue that got the last in the chain not working.  It's a combination of the old hardware and screen update in BFGMiner!  I set it to 2 sec screen update and the erupter goes from solid green to hashing more often ( unless I'm browsing the net ).  from showing you are mining, D ( for display ) and drop the update to 2 to 10 seconds and it forces it to pay attention to all devices more often.  It would be interesting if this ' fixes ' the half dead ant.

[nybbler905]

very true, i have ordered some voltage regulators, so i plan to change 1 out on the very dead ant ... whats the worst that could happen

I've been digging in my old posts looking for unresolved issues and AntMiner U1 hacking issues and found this ... and to answer your question... a dead ASIC from soldering a ' 103 ' smd for R2 and finding one end also went to that cap close to those pads shorting part of the power so it's not even having any sign of life is the ' worst ' thing..

Yea, I admit I should have checked and solder wicked up the excess prior to plugging it in.  The red LED did flash 5 or so before nothing... total nothing..
I'm looking in to getting a fresh ant and doing a part by part swap one day to see if it's fixable.  In the mean time my ( averaged ) hashing 25 GHash/sec only during daylight hours is down to point tripple zero 3 to 7 per day on BTC so it's gonna take a while.... back to doing BTC faucets and dumping the proceeds on to either Cryptsy or Peerbet to have them deal with the ' dust ' aspect.