GekkoScience 2Pac/Compac BM1384 Stickminer Official Support Thread

[marousi]

BTW, the green light on my 2pac became white, yet it is mining at 300mhz, go figure.

[1713407598]

1713407598

[1713407598]

1713407598

[1713407598]

1713407598

[mdtrav]

I just received my 2pac and I cannot get it to work correctly.

I consistently get "share above target" very few of my shares are getting accepted. Same issue on windows and linux machines.

I've tried at 100mhz and up to 200mhz. No change it the issue. I also use a powered hub, this is the only device on the hub currently.

http://puu.sh/zg3yh/59ce44f73d.png

Any help getting it to work would be appreciated. Thanks!

What I use to start it.

./cgminer -o stratum+tcp://eu1.miningpool.shop:3333 -u 5xcZLmxBTU6whZWFCRc3qxVU8svZJQQTMD -p c=PLTCR --gekko-2pac-freq 150 --verbose

[zachj00]

I just received my 2pac and I cannot get it to work correctly.

I consistently get "share above target" very few of my shares are getting accepted. Same issue on windows and linux machines.

I've tried at 100mhz and up to 200mhz. No change it the issue. I also use a powered hub, this is the only device on the hub currently.



Any help getting it to work would be appreciated. Thanks!

What I use to start it.

./cgminer -o stratum+tcp://eu1.miningpool.shop:3333 -u 5xcZLmxBTU6whZWFCRc3qxVU8svZJQQTMD -p c=PLTCR --gekko-2pac-freq 150 --verbose

Try going to your pools webite and adjusting the mining difficulty to 128. Most pools adjust difficulty automatically but I've never heard of the pool you're using.

[mdtrav]

I just received my 2pac and I cannot get it to work correctly.

I consistently get "share above target" very few of my shares are getting accepted. Same issue on windows and linux machines.

I've tried at 100mhz and up to 200mhz. No change it the issue. I also use a powered hub, this is the only device on the hub currently.

http://puu.sh/zg3yh/59ce44f73d.png

Any help getting it to work would be appreciated. Thanks!

What I use to start it.

./cgminer -o stratum+tcp://eu1.miningpool.shop:3333 -u 5xcZLmxBTU6whZWFCRc3qxVU8svZJQQTMD -p c=PLTCR --gekko-2pac-freq 150 --verbose

Try going to your pools webite and adjusting the mining difficulty to 128. Most pools adjust difficulty automatically but I've never heard of the pool you're using.

Thanks for the reply.
I set the diff to 128 and it did not change anything. I've tried it with several pools with no luck.

[sidehack]

Leave it on one pool for 24 hours and see what happens. Check the pool stats and see if anything's getting through.

[Nut110]

Thanks @sidehack for your great learning sesson about reading forum!  
When I read the post of neowhere on this page 78 - the problem seems to be still there! And your answer "try adjusting the voltage a little bit" sounds not like the final solution....So why I should read page 23...37...56...?

I read page one, found issue depening links and get not the right answer... Also I googled my problem and hit some pages, here in the forum, in the middle between page 1 and 78... Problem is only get more questions, but no answers... after 2 replies, the forum changing in a totally different theme. Thats the problem of this one-thread - al lot of themes and issues in one chain.... So I need first of all 2 weeks exploring the forum....

But back to my ZOMBIE issue. I made the following steps and seems to get it solved. I tested it only with my two 2Pac's, named #1 and #2 and get no ZOMBIE stats at the moment.
1. Plugin #1 on power supply, only with the customized soldered power lines.
2. Measuring the ASIC-voltage at the two capacitances and adjusting it to 0.68V and 0.7V (slight diffenrence of 0.01..0.02V still present). --> voltage level recommended for 200..250Hz in the manual.
3. Plug off #1 and plugin #2 on power supply (seems to be importand that every stick is alone on the power supply --> if not, one gets into zombi (asymmetric asic voltage)).
4. Measuring the ASIC-voltage on #2 at the two capacitances and adjusting it to the same level and difference as #1 = 0.68V and 0.7V (slight diffenrence of 0.01..0.02V still present). --> seems, both stick should use the same asic voltage levels, despite there should independent from each other

After this symmetric adjustment of both sticks, I made the following observation:
5. starting cgminer without any plugged 2Pac.
6. When I plugin #1 to the power and then to usb-hub (data connection), this miner starts working (LED flashes).
7. when I add the second stick #2 to the power and to the usb-hub, one of the sticks getting into zombie, most of the time the first plugged stick gets into zombie, second starts mining --> seems to be a voltage volatilitiy during plugin or something....ASIC-voltages of the zombie sticks gets immediatly asymmetric...the first working stick seems to get a problem, when the second is plugged to the power supply.

So I made the connection in the following way:
8. power off power supply and usb-hub (are on the same power connection).
9. connect both sticks to the power connection and to the data connection (usb-hub).
10. power on power supply, so sticks starting at the same moment, also the connection via usb-hub
11. both sticks are getting recognized and starting work with running cgminer.
At the moment this setup runs without any ZOMBIE stats at 200MHz (tested only for the last night and some hours today).

But there is still a problem - the Raspberry gets freezed sometimes (the whole Raspberry, not only cgminer).
I read the link at page 1 about kernel debug mode. But not modified and tested at the moment.

Hello,

My 1st message here in this post.

I am putting my hands on 3 2-pac running PC-stick Windows 10 now.
Simply conclude the status now that I am still getting zombie 1 out of 3 in every half a day (every time I come to observe every morning and night before going to bed, so not sure how long would it take before 1/3 zombie happens).

Here is my dirty jobs.
1. Similar to what you are doing = open the hub and solder the power rails + use adjustable 5V-7V 60A switching PSU.
2. I cut the red power pin of the hub that connects to my PC-stick in order to prevent the voltage from PSU fed back to my PC-stick. Yes I burnt one of my previous PC-stick and all HDMI ports at my flat plasma TV from uncut power hub. Now I connect my new PC-stick to the TV via HDMI-VGA adapter + VGA port.
3. The core voltage is now at 0.685V for 200MHz.
4. Previously I even added the super capacitor 250F (serial 2.7V 500F x 2 with the balance voltage circuit) to clamp the hub power rail too, but seems like it is not relevant. Zombie is still.
5. I also increase the USB voltage too. I measure from the port and sometimes also from the 2-pac pins right at the USB connector (leftmost and rightmost, right ?). Very early I used the other USB 3.0 hub and found that the voltage was 4.8V something and not stable at all. So comes later the mod of this USB 2.0 hub. Now not only the vcore voltage, I ramp up the USB voltage to as it has been obvious that the setup yield more stable. But zombie is still. Now the USB voltage is up to 5.75V at the USB connector already. The good is...... my new PC-stick and flat TV are still safe not getting burnt.......
6. With vcore 0.685V + USB voltage 5.75V the zombie is still 1 out of 3. I come and check twice a day so not sure how long would it take until 1/3 happens.
7. Super capacitor is not used anymore as it will limit the USB voltage up to 5.35V only.
8. Start-up is easier by increasing USB voltage. Early time I had to put so much frustrating efforts to start the 3 running simultaneously. Got me mad. But with > 5.5V USB voltage start-up is a lot easier. Plugging in one-by-one even still fail to start all sometimes but a lot easier.
9. Right during my post here I got 1 zombie. Different stick so should not relevant to the position of the ports. Now I put the sticks at 1st, 3rd, and 5th ports counted from the near side to the PSU connector.
10. Of course I put a big fan for all 3.

Next will try increasing vcore up to 0.7V @200MHz. Let see if the zombie would be gone.

My 4th 2-pack is on the way to make me more mad....... Oh ............

Not recommend to do this USB voltage if you are not prepared to loss your PC, Raspberry, screen/monitor/TV, etc.
My 1/2 cent.

https://www.facebook.com/photo.php?fbid=1903782269652749&set=p.1903782269652749&type=3&theater

[wavelengthsf]

Thanks for the reply.
I set the diff to 128 and it did not change anything. I've tried it with several pools with no luck.

Are you mining a SHA-256 coin? Looking at your screen shot, that's definitely not Bitcoin (or even BCash) - can it be mined with a SHA 256 ASIC?

[coinop99]

I just received my 2pac and I cannot get it to work correctly.

I consistently get "share above target" very few of my shares are getting accepted. Same issue on windows and linux machines.

I've tried at 100mhz and up to 200mhz. No change it the issue. I also use a powered hub, this is the only device on the hub currently.

http://puu.sh/zg3yh/59ce44f73d.png

Any help getting it to work would be appreciated. Thanks!

What I use to start it.

./cgminer -o stratum+tcp://eu1.miningpool.shop:3333 -u 5xcZLmxBTU6whZWFCRc3qxVU8svZJQQTMD -p c=PLTCR --gekko-2pac-freq 150 --verbose

I think the pool itself is suspect. I've tried your command with my 2pacs which are perfectly fine, but get the same 'above target' errors on that pool.
I suggest to try with one of the major btc pools like slush or ckpool.

[edc2k12]

I bought two 2pac usb stick and got them in December. Everything was working great until this week. (I did not change any settings or hardware changes).

We recently have been having colder weather and an earthquake. No power outages. One stick continues to work perfectly but the other has stopped working with both the blue and green lights continuously on. Any suggestions on fixing the non working stick?

I have tried the sticks on Windows 10, 8.1, Ubuntu and Centos. One does not work...

[sidehack]

Check the top of this page. Solid white light usually means one chip is dead. Check with who you bought them from about warranty options. If that's not good enough, contact the manufacturer *ahem* about getting it fixed.

[edc2k12]

Check the top of this page. Solid white light usually means one chip is dead. Check with who you bought them from about warranty options. If that's not good enough, contact the manufacturer *ahem* about getting it fixed.

Thanks for letting me know. I will check where I got them from and email the seller. Mind you, it might be more expensive to send it to them then to get a new one.

[ftguy]

I bought recently 2 sticks and when I running them at freq:100 , one of the stick is hashing at  21.5gh/s average with cgminer , is that normal...it seems the freq = freq *2

[marousi]

The hashrate is not constant, they bounce all over the hashrate range.

[ftguy]

I bought recently 2 sticks and when I running them at freq:100 , one of the stick is hashing at  21.5gh/s average with cgminer , is that normal...it seems the freq = freq *2

I fixed the issue, it was simply a problem with the USB connection.

[marousi]

I have 8 2pacs, and have optimized each one individually, and yet when they all run on 4 60 watt powered anker hubs, some go to zombie, even ones that ran perfectly for 48 hours alone. A power issue? Is the PC supplying enough power?

[sidehack]

Recall all the advice that's been given to you already.

[marousi]

and i have followed that advice, but still have problems.

[zachj00]

It's your Hub. You need more power to run them.  I can only run 3 2pacs up to 175 MHz per 60W USB hub.  try running 3 on it with 175 MHz.

[marousi]

i've been running 2 2pacs per 60 watt hub at 300mhz. think I need to reduce to 250mhz or so

[nQuant]

ck has released a new version of cgminer this week, does this build include the latest code or the version prior?

Yes, it includes all the latest updates from ck's repository.
Including some related to system libusb that has sent my windows build example back to the drawing board. 
   
The entry page to my repository should say "This branch is N commits ahead of ckolivas:master. ".
If it's ever "commits behind" just let me know.

vh, I was wondering if these 2pac checks against difficulty ? Datasheet says 4 bytes from byte position 36 to 39 set as ticket mask for difficulty.
But gekko driver code seems only setting ticket_mask at 39th byte position. What extactly is this ticket_mask?

I tried setting Network difficulty as below and seemed it work but later found its not always returning true nonce for every sample worked example I was using.
I think these chips keep iterating from 0 to max nonce and return every nonce which result in hash with first 4 bytes with 0. Is that true ?

Code:

_data[36] = (byte)((difficulty >> 24) & 0xff);
_data[37] = (byte)((difficulty >> 16) & 0xff);
_data[38] = (byte)((difficulty >> 8) & 0xff);
_data[39] = (byte)(difficulty & 0xff);

Actually I have written small POC kind of app in c# to test and wondering how to control difficulty its checking against.
Also it seems you are not using Start nonce offset, in my test I found any value above 1 was not working.
How to set the SNO properly so that device only starts the work from particular nonce instead of 0.

Appreciate if you could help me here. Thanks !!!