Pages:
Author

Topic: Bitmain AntMiner U1 Tips & Tricks - page 27. (Read 106810 times)

newbie
Activity: 25
Merit: 0
January 11, 2014, 02:46:59 PM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?

Its probably the same stick with errors, the software allocates the table positions in sequence of power up, not port number. So each start up is different.

As an update, I went ahead and switched out all the miners to just random different ports, making sure each one is on a different port. Currently, I get no hardware errors on them, but one of them has only accepted 34 shares in 8.5 hours. I'm using eclipseMC and I'm not sure how to have the pool utilize my full potential as it's only displaying between 5-8 GH/s whenever I look at it's displayed hash rate.

New Update(Not that many of you care, but maybe someone's having a similar issue): I took out every stick, including block erupters, from my hub, and made sure to hit every port in the hub. It re-installed the silicon labs driver on some of them, so that leads me to believe the driver wasn't on all of them which is why I was getting so many issues.

Mixing devices in one session will continue to be difficult until the drivers are fully implemented.
Run one session for your erupters.
Then plug in your U1 sticks then start a second session for them.

nwoolls has done a great job with BFGminer but, until LukeJr builds then into the formal release, there may still be conflicts.

Currently everything is running smoothly after I did the whole unplug everything and plug them back in to make sure I hit every port
legendary
Activity: 4326
Merit: 8950
'The right to privacy matters'
January 11, 2014, 02:38:21 PM
these sticks will clock to 2 gh with a software adjustment and a 2.0 usb hub.

if you use a 3.0 hub they will also clock to 2 gh and the same software adjustment.

 if you want
2.2
2.4
 2.6  
  do all three below

 1) a 3.0 hub yes

2)a fan or some kind of extra cooling yes


3) and you need to remove a resistor and use a different resistor  for good oc  of 2.2-2.6 it is in an earlier post and in the ant miner notes.

 I have 5 different 3.0 hubs none of them allowed for good solid oc's at 2.2 or 2.4 .
hero member
Activity: 868
Merit: 1000
January 11, 2014, 02:27:40 PM
will we get a better hash rate if i use 3.0 port hub, but im pluggin it in a 2.0 port tho.

I get 1.97 out of 2.0 oc on the big 49 port  hubs they are 2.0 usb.

pretty much these sticks will do 35-40 ports out of 49 close to the oc of 1.8 or 2.0

with 37 sticks running on a 2.0 oc  i got 68.7gh out of a hoped for 74 gh over the  course of monday the 6th to fri the 10th I had a zombie float around pop in and out and a few sticks under preform.



same 37 sticks running on a 1.8 oc i get 66.2  gh out of a hoped for 66.2 gh over a  5 hour test from sat morn till now .   no zombies


power 68.7 gh was 112 watts

power 66.2 gh is 109-110  watts  

2 fans thermal take usb 2
1 psu a sea sonic 750 watt plat version possible to be 90%

 I will photo. note :
I run a ltc on a different cgminer
I run bfg miner and 14 ice fury nano
I run cgminer and the 37 ant miners


below 37 sticks set at a 1.8 oc

power the sticks use  109 watts for 66gh not to shabby


USB3 hubs will allow more power to a port, so are good if you are overclocking.

49 port USB2 hubs are ok for most devices.

bear in mind that if your psu is putting in 60 watts, you will probably get less than 40watts available at the ports.
legendary
Activity: 4326
Merit: 8950
'The right to privacy matters'
January 11, 2014, 02:06:42 PM
will we get a better hash rate if i use 3.0 port hub, but im pluggin it in a 2.0 port tho.

I get 1.97 out of 2.0 oc on the big 49 port  hubs they are 2.0 usb.

pretty much these sticks will do 35-40 ports out of 49 close to the oc of 1.8 or 2.0

with 37 sticks running on a 2.0 oc  i got 68.7gh out of a hoped for 74 gh over the  course of monday the 6th to fri the 10th I had a zombie float around pop in and out and a few sticks under preform.



same 37 sticks running on a 1.8 oc i get 66.2  gh out of a hoped for 66.2 gh over a  5 hour test from sat morn till now .   no zombies


power 68.7 gh was 112 watts

power 66.2 gh is 109-110  watts  

2 fans thermal take usb 2
1 psu a sea sonic 750 watt plat version possible to be 90%

 I will photo. note :
I run a ltc on a different cgminer
I run bfg miner and 14 ice fury nano
I run cgminer and the 37 ant miners


below 37 sticks set at a 1.8 oc



power the sticks use  109 watts for 66gh not to shabby


hero member
Activity: 868
Merit: 1000
January 11, 2014, 12:33:01 PM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?

Its probably the same stick with errors, the software allocates the table positions in sequence of power up, not port number. So each start up is different.

As an update, I went ahead and switched out all the miners to just random different ports, making sure each one is on a different port. Currently, I get no hardware errors on them, but one of them has only accepted 34 shares in 8.5 hours. I'm using eclipseMC and I'm not sure how to have the pool utilize my full potential as it's only displaying between 5-8 GH/s whenever I look at it's displayed hash rate.

New Update(Not that many of you care, but maybe someone's having a similar issue): I took out every stick, including block erupters, from my hub, and made sure to hit every port in the hub. It re-installed the silicon labs driver on some of them, so that leads me to believe the driver wasn't on all of them which is why I was getting so many issues.

Mixing devices in one session will continue to be difficult until the drivers are fully implemented.
Run one session for your erupters.
Then plug in your U1 sticks then start a second session for them.

nwoolls has done a great job with BFGminer but, until LukeJr builds then into the formal release, there may still be conflicts.
newbie
Activity: 25
Merit: 0
January 11, 2014, 12:26:57 PM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?

Its probably the same stick with errors, the software allocates the table positions in sequence of power up, not port number. So each start up is different.

As an update, I went ahead and switched out all the miners to just random different ports, making sure each one is on a different port. Currently, I get no hardware errors on them, but one of them has only accepted 34 shares in 8.5 hours. I'm using eclipseMC and I'm not sure how to have the pool utilize my full potential as it's only displaying between 5-8 GH/s whenever I look at it's displayed hash rate.

New Update(Not that many of you care, but maybe someone's having a similar issue): I took out every stick, including block erupters, from my hub, and made sure to hit every port in the hub. It re-installed the silicon labs driver on some of them, so that leads me to believe the driver wasn't on all of them which is why I was getting so many issues.
newbie
Activity: 28
Merit: 0
January 11, 2014, 11:42:51 AM
Running for 20+ minutes now and cgminer has been perfectly stable with 0%-1% CPU usage whenever I check (PiMiner consistently seems to take up 28%-30%, but I can live with that!)

Maybe your on to something here.

With my fresh built image ive just done (as per the post above) I still saw ~60% cpu usage by CGMiner.

Have just commented out that line in etc/init.d and rebooted and am instead running it with
Code:
sudo ./fractalbc/cgminer/cgminer --bmsc-options 115200:20 --bmsc-freq 0981 -o stratum.bitcoin.cz:3333 -u XXXXX-p XXXXX --api-listen --api-allow 0/0

And im seeing 1% usage. I wonder if its a specific option, or indeed the config file itself that's causing high CPU utilisation....

Gonna leave mine hashing overnight, if its still running tomorrow then il be 95% sure its the config file that's causing issues. Il slowly try removing lines and options until the CPU usage drops to these levels, or I cant remove anymore! I just prefer using the config file as it means the launch code is cleaner Smiley

Hate to throw a spanner in the works, but I've tried it with a config file too and that's been working perfectly for the past hour, so I'm kind of stumped now! (the config file was written from within cgminer while using the command line in my post above, so it could possibly be worth trying that to generate a fresh config file if you haven't done that already)

Mine is as follows if that's any help;
Code:
{
"pools" : [
        {
                "url" : "pool1",
                "user" : "user1",
                "pass" : "pass1"
        },
        {
                "url" : "pool2",
                "user" : "user2",
                "pass" : "pass2"
        }

]
,
"api-listen" : true,
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "120",
"hotplug" : "5",
"failover-only" : true,
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "60",
"shares" : "0",
"kernel-path" : "/usr/local/bin",
"api-allow" : "0/0",
"bmsc-options" : "115200:20",
"bmsc-freq" : "0881"
}

Hoe did yours go, mine ran until about 4:30AM before crashing, have since restarted the pi and was running all day, still is. CGminer CPU usage was/is a lot lower still but not 100% solid yet.
newbie
Activity: 25
Merit: 0
January 11, 2014, 10:50:11 AM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?

Its probably the same stick with errors, the software allocates the table positions in sequence of power up, not port number. So each start up is different.

As an update, I went ahead and switched out all the miners to just random different ports, making sure each one is on a different port. Currently, I get no hardware errors on them, but one of them has only accepted 34 shares in 8.5 hours. I'm using eclipseMC and I'm not sure how to have the pool utilize my full potential as it's only displaying between 5-8 GH/s whenever I look at it's displayed hash rate.
hero member
Activity: 868
Merit: 1000
January 11, 2014, 09:42:37 AM
I have 4 U1's but I consistently have problems getting them to hash even at a consistent 1.6 Gh/s.  I've tried OC'ing but the hash rate just drops to around 1 Gh/s or below.  Do I need to cool them or just a better hub?  I've got one attached to my laptop's USB port and 3 on a Belkin 7 port, powered hub.

I'm using the newer BFG Miner posted on this thread.

Any help is appreciated.
some dude in this thread had  a similar prob , think he was using the old bfg tho , he couldnt get the cgminer to work the first time, but ended up getting it to work and problem solve for him.

I've tried to find a newer version of the BFG Miner that supports the U1's but everyone I've tried has a compile date of the 9th leading me to believe there's only one version currently out?

I've restart my U1's a few times, trying different port configurations on my laptop and I've managed to get them stable for 2.5 hours.  Unfortunately the speeds aren't what they really should be - even when overclocked to 2 Gh/s.  I'm hoping this is just a power issue, I have a TechNet 10-port hub on it's way with a 48W rated PSU so when that arrives I expect the U1's to perform as they should. 

This image is with the '--set-device antminer:freq=0981' command which just highlights what the problem is.
 

Oh and before I forget, I was having BFG Miner 'load balance' between 3 separate pools but it seems to have stabilized with the 'balance' option between two instead.  Not sure if this is down to the power, BFG Miner or a mixture of the two but I've yet to test them using only one pool since I'm afraid that whatever luck has gotten the U1's to hash near their rated speed may run out and I may not get these speeds again until my new hub arrives.
Better off using only one pool, maybe with a fail over onto another pool.
Each time your miner switches pool , the work in the system is wasted, hence the stale shares
newbie
Activity: 14
Merit: 0
January 11, 2014, 08:46:31 AM
I have 4 U1's but I consistently have problems getting them to hash even at a consistent 1.6 Gh/s.  I've tried OC'ing but the hash rate just drops to around 1 Gh/s or below.  Do I need to cool them or just a better hub?  I've got one attached to my laptop's USB port and 3 on a Belkin 7 port, powered hub.

I'm using the newer BFG Miner posted on this thread.

Any help is appreciated.
some dude in this thread had  a similar prob , think he was using the old bfg tho , he couldnt get the cgminer to work the first time, but ended up getting it to work and problem solve for him.

I've tried to find a newer version of the BFG Miner that supports the U1's but everyone I've tried has a compile date of the 9th leading me to believe there's only one version currently out?

I've restart my U1's a few times, trying different port configurations on my laptop and I've managed to get them stable for 2.5 hours.  Unfortunately the speeds aren't what they really should be - even when overclocked to 2 Gh/s.  I'm hoping this is just a power issue, I have a TechNet 10-port hub on it's way with a 48W rated PSU so when that arrives I expect the U1's to perform as they should. 

This image is with the '--set-device antminer:freq=0981' command which just highlights what the problem is.
http://i42.tinypic.com/2dl99h5.jpg 

Oh and before I forget, I was having BFG Miner 'load balance' between 3 separate pools but it seems to have stabilized with the 'balance' option between two instead.  Not sure if this is down to the power, BFG Miner or a mixture of the two but I've yet to test them using only one pool since I'm afraid that whatever luck has gotten the U1's to hash near their rated speed may run out and I may not get these speeds again until my new hub arrives.
sr. member
Activity: 434
Merit: 250
January 11, 2014, 08:18:43 AM
nah,  the hub is band new. I will  return it  on monday:)

BTW

I overclocked one of my 4 miners with 10K and 2.2K resistors.  and  noticed that not only that  Antminer performs better (stable at 2.2GHS) but  overall performance of  all miners is increased.

Is there any explanation for this?
Do all miners work independently form each other, or do they split the work like multicore CPU's?
hero member
Activity: 868
Merit: 1000
January 11, 2014, 08:00:26 AM
you are right, i connected miners directly to my PC, and  everything goes ok.
But i doubt  my USB-ports have more power then USB3.0 Hub with 4A output. Most probably it  just sucks.
If your hub has been driving erupters for a while, then the power supply may be getting tired.
sr. member
Activity: 434
Merit: 250
January 11, 2014, 07:03:41 AM
you are right, i connected miners directly to my PC, and  everything goes ok.
But i doubt  my USB-ports have more power then USB3.0 Hub with 4A output. Most probably it  just sucks.
hero member
Activity: 868
Merit: 1000
January 11, 2014, 06:53:20 AM
Dammit i get some kind of interference. When i run all 3 Sticks from same HUB one Miner tends to clok up and  producing mass of HW. If i run only 2, then the problem is not present.

I think the problem is with those  ZADIG drivers.

I tried to sue nwools bfg miner, but as i start the .exe  it just dont detect anything. I deinstalled the zadig drivers
Try another hub, your power may be low.
sr. member
Activity: 434
Merit: 250
January 11, 2014, 06:48:35 AM
Dammit i get some kind of interference. When i run all 3 Sticks from same HUB one Miner tends to clok up and  producing mass of HW. If i run only 2, then the problem is not present.

I think the problem is with those  ZADIG drivers.

I tried to sue nwools bfg miner, but as i start the .exe  it just dont detect anything. I deinstalled the zadig drivers
hero member
Activity: 868
Merit: 1000
January 11, 2014, 04:02:06 AM
BTW
Running at 2 GHS  at chip Temp.  just 40 C°  (room temp 21 C°)
Passive cooling FTW=)



This is a really nice piece of creative engineering  Grin

Room heating that pays for itself.  Roll Eyes
hero member
Activity: 868
Merit: 1000
January 11, 2014, 04:00:56 AM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?

Its probably the same stick with errors, the software allocates the table positions in sequence of power up, not port number. So each start up is different.
newbie
Activity: 25
Merit: 0
January 11, 2014, 02:12:43 AM
I'm using the new bfgminer that was posted here(Thank you forever, btw) and one of my miners is getting a bunch of HW errors. The problem is it's always a different one whenever I cycle the mining software. This happened with both cgminer and BFGMiner. Does anyone have any ideas on this problem?
hero member
Activity: 854
Merit: 500
einc.io
January 10, 2014, 10:23:34 PM
BTW
Running at 2 GHS  at chip Temp.  just 40 C°  (room temp 21 C°)
Passive cooling FTW=)



This is a really nice piece of creative engineering  Grin
legendary
Activity: 924
Merit: 1000
Think. Positive. Thoughts.
January 10, 2014, 09:55:00 PM
Ok guys. I tried to overclock Antminer u1  according to  the official   guide.
I replaced  resistors  to 10K and 4.7K, to set the voltage near 1.2V ( I am sure the exact R Value is not necessary)
As  result I couldn't set any operating frequency  with  "nonce timeout" Error. Undecided

Then I replaced the R1 to 3.5K, that allowed me to  use  Antminer at 225 MHZ (all frequencies over it caused nonce timeout).
Instead  of getting 1.8 GHS I got  about 2.3 GHS but  too much  HW, so I lowered  R1 to  1.5K.
 That allowed me to run Antminer at max. 250MHZ, but getting  again about 2,3 GHS  with acceptable HW rate.Still  the miner got pretty hot, so i replaced the resistors back to the original state.
I also tried to vary   Timeout parameter   by setting it to higher values(from 20 to 100), that allowed me to start  Antminer at R1=3.5K R and 250MHZ, but still it acted VERY weird. The  "work" LED   was permanently "ON" and the miner delivered only hardware errors.


Can anyone explain my WTF? Who the hell posted this PDF with overclock instruction? What am i doing wrong?

PS
The test U1 was  powered through USB3.0 Hub  with max output current of 4A.

USB 3 was acting weird for me w/these. It sounds like cooling might be an issue for you. You can leave these heatsinks on and add another heatsink to them for extra cooling.... or just a large piece of metal. Mind showing pics of the resistors soldered? I have the big resistors but I don't think I can get them on such a small lead.
Pages:
Jump to: