Braiins OS & Braiins OS+ custom ASIC firmware: optimize performance & efficiency

[balla69]

I've set this before:

Code:

cd braiins-os_am1-s9_ssh_VERSION
virtualenv --python=/usr/bin/python3 .env
source .env/bin/activate
pip install -r ./requirements.txt

after:

Code:

python3 restore2factory.py Antminer-S9-all-201812051512-autofreq-user-Update2UBI-NF.tar.gz 172.16.0.134
Extracting backup tarball...
Invalid backup tarball!

It's invalid backup....

This restore accept only the backup make before from braiins ?

[1713924155]

1713924155

[1713924155]

1713924155

[1713924155]

1713924155

[1713924155]

1713924155

[1713924155]

1713924155

[Artemis3]

This restore accept only the backup make before from braiins ?

Yes. But there is a method to restore the firmware manually.

[Apprentice]

is any of you having issues with fan speeds?

some miners are like running at fan full speed every 1-2 minutes for like 10s (continuous)

[RadekG2]

is any of you having issues with fan speeds?

some miners are like running at fan full speed every 1-2 minutes for like 10s (continuous)

Yes, this problem exist under light load or cold ambient temperature. Next release will probably fix it.

I've installed this via the web portal on one of my R4's and am having problems getting it to hash. It will submit 3-4 shares before stopping. It appears to be linked to fan speed as when the fans drop from start up, around 3k rpm, the miner stops hashing. For some reason the fans drop as low as 800 rpm (never seen them that low) and my guess is thats causing a failure mode to kick in like it does on the s9's when people disconnect the fans to try and oil cool. As there is no way to control fan speeds through BOS at the moment, I can't test this theory. I'm in contact with Braiins support and will keep the thread update with how this issue gets resolved.

Yes, the problem is that it expect TWO FANs. To overcome it, you can connect RPM pin from FAN1 with FAN2 on control board. Just connect pin with yellow wire with the same position at FAN2 position. Always do this when disconnected power - next pin is 12V which will permanently damage control board when connected to RPM yellow pin. (I did it.)

Next release should fix this problem.

[Apprentice]

similar behavior with the S9, after few seconds the miner will stop hashing if there is a fan error.

[Artemis3]

Yes, the problem is that it expect TWO FANs. To overcome it, you can connect RPM pin from FAN1 with FAN2 on control board. Just connect pin with yellow wire with the same position at FAN2 position. Always do this when disconnected power - next pin is 12V which will permanently damage control board when connected to RPM yellow pin. (I did it.)

For the time being you are not supposed to use Braiins OS with the R4 at all.

[sarcheer]

I've set this before:

Code:

cd braiins-os_am1-s9_ssh_VERSION
virtualenv --python=/usr/bin/python3 .env
source .env/bin/activate
pip install -r ./requirements.txt

after:

Code:

python3 restore2factory.py Antminer-S9-all-201812051512-autofreq-user-Update2UBI-NF.tar.gz 172.16.0.134
Extracting backup tarball...
Invalid backup tarball!

It's invalid backup....

You need to run:

Code:

python3 restore2factory.py path/to/backup hostname --factory-image path/to/Antminer-S9-all-201812051512-autofreq-user-Update2UBI-NF.tar.gz

you may need to use an older factory firmware as I can’t remember whether the newest LPM firmwares are partial updates as opposed to full firmware images.

Here are the instructions for if you do not have a backup from when you upgraded to bOS

1.) create empty directory backup
2.) create text file mac (without any file extension)
3.) write one line with MAC address to the file mac (you can obtain your MAC address from web interface Network -> Interfaces -> MAC-Address)
4.) compress the file mac with tar using gzip compression and store it to the backup directory with a name config.tar.gz
5.) create a file called uEnv.txt in the back up directory containing the following information:

Code:

recovery_mtdparts=mtdparts=pl35x-nand:32m(BOOT.bin-env-dts-kernel),144m(angstram-rootfs),80m(upgrade-rootfs)

This is from the walkthrough given to me by Braiins support.

[Apprentice]

another question about chip temperature? is it now correct or estimation?

I like the BOS cause the fans are running low thus the miners are quiet. but need to make sure the chi temps are correct.

and regarding voltages, I know this was asked before but probably it can be summarized as BOS is showing recommended voltage based on the frequency:

shouldn't I increase the voltage when increasing the frequency?

Does any have a "optimal" values (Frequency/voltage) they can share?

Thanks!


is anyone having issues with sudden "miner death"?
This is the first time it happens, I have 10 S9 running on BOS Cobalt and 4 S9 on Bitmain's, all 10 BOS S9 stopped mining at the same time (red lights flashing, can't connect via LAN); the other 4 S9 that are connected to the same switch are working fine.

Fixed by recycling the power.

[Artemis3]

is anyone having issues with sudden "miner death"?

Are you overclocking or using the auto settings?

[Apprentice]

Are you overclocking or using the auto settings?

Max frequency is 655

[sarcheer]

Are you certain they have unique MAC addresses? I think as part of the bOS upgrade it changes the MAC addresses if you don't specify them.

[Artemis3]

Max frequency is 655

If you put that frequency manually in the cgminer settings, try unsetting it by unticking the left buttons, and see if with the auto/default frequencies remains stable.

[Velles2003]

thanks for your help
i have successfully restored factory firmware image by using these commands from windows
i'll put the line here so that it may help another

Code:

C:\python\braiins-os>python restore2factory.py backup/ 192.168.2.244 --factory-image backup/Antminer-S9-all-201711171757-autofreq-user-Update2UBI-NF.tar.gz

install python
make a folder called backup
put the factory image in it, along with two more files: uEnv.txt and config.tar.gz (that contains mac adress)

You need to run:

Code:

python3 restore2factory.py path/to/backup hostname --factory-image path/to/Antminer-S9-all-201812051512-autofreq-user-Update2UBI-NF.tar.gz

you may need to use an older factory firmware as I can’t remember whether the newest LPM firmwares are partial updates as opposed to full firmware images.

Here are the instructions for if you do not have a backup from when you upgraded to bOS

1.) create empty directory backup
2.) create text file mac (without any file extension)
3.) write one line with MAC address to the file mac (you can obtain your MAC address from web interface Network -> Interfaces -> MAC-Address)
4.) compress the file mac with tar using gzip compression and store it to the backup directory with a name config.tar.gz
5.) create a file called uEnv.txt in the back up directory containing the following information:

Code:

recovery_mtdparts=mtdparts=pl35x-nand:32m(BOOT.bin-env-dts-kernel),144m(angstram-rootfs),80m(upgrade-rootfs)

This is from the walkthrough given to me by Braiins support.

[sarcheer]

So glad I could be of help and that you've successfully returned to factory! Hopefully future releases will go better.

Braiins support mentioned that they're working to do away with the dependency on backups to return to factory firmware in future releases.

[RadekG2]

Yes, the problem is that it expect TWO FANs. To overcome it, you can connect RPM pin from FAN1 with FAN2 on control board. Just connect pin with yellow wire with the same position at FAN2 position. Always do this when disconnected power - next pin is 12V which will permanently damage control board when connected to RPM yellow pin. (I did it.)

For the time being you are not supposed to use Braiins OS with the R4 at all.

Oh sorry, I forgot one much easier workaround. Just simply add any 3pin or 4pin CPU fan and leave it on the top of the miner running freely.

[Apprentice]

speaking about fan speeds? what I love about BOS is that fan speeds are lower than Bitmain's thus making less noise.
but is it OK for the miner to run at higher temperatures? my miners are not exceeding 82
thanks.

[wilkbtc]

Hello .. the braiin os is cool but the problem i run into is:  when i am at my mini farm 30 miners i get API connection stalled and then cant connect to miner ...

i mean i can still go mess with the other settings and pool side i am connected but it is very annoying because i cannot monitor the overclock settings so it makes it very hard to fine tune the miners

However when i bring 1 machine home and test run it there everything works very well and miners run optimal ....

anyone else got that issue ? ...

feels like it has something to do with my switch but all my miners that are not on brain os  dont have that issue

[Artemis3]

Hello .. the braiin os is cool but the problem i run into is:  when i am at my mini farm 30 miners i get API connection stalled and then cant connect to miner ...

i mean i can still go mess with the other settings and pool side i am connected but it is very annoying because i cannot monitor the overclock settings so it makes it very hard to fine tune the miners

However when i bring 1 machine home and test run it there everything works very well and miners run optimal ....

anyone else got that issue ? ...

feels like it has something to do with my switch but all my miners that are not on brain os  dont have that issue

I have not encountered that problem. Seems to me you have an issue with the way you connect to the miners. Is SSH working ok? Can you connect remotely using SSH instead of the web interface?

[philipma1957]

952 watts doing 10500  set at freq 500 and 8.7 v

I can do 90 watts a th with bitmain firmware.

at the moment this is doing 90 watts a th.

I will play with it. to see if I go to 85 watts a

1002 watts doing 11050 set at freq and 8.7volt

which is 90 watts a th.

[Shad0wSmurf]

Hello .. the braiin os is cool but the problem i run into is:  when i am at my mini farm 30 miners i get API connection stalled and then cant connect to miner ...

i mean i can still go mess with the other settings and pool side i am connected but it is very annoying because i cannot monitor the overclock settings so it makes it very hard to fine tune the miners

However when i bring 1 machine home and test run it there everything works very well and miners run optimal ....

anyone else got that issue ? ...

feels like it has something to do with my switch but all my miners that are not on brain os  dont have that issue

Make sure you're able to use unsecured port 80. That's what's uhttp is looking for access to. My one computer won't allow me to port forward from off network due to some security issues. Idk, I just made an exception in the tcp forwards to allow all connections to that ip. Idk what it was but try to allow all connections to 443,4028,80,8080

See what works

952 watts doing 10500  set at freq 500 and 8.7 v

I can do 90 watts a th with bitmain firmware.

at the moment this is doing 90 watts a th.

I will play with it. to see if I go to 85 watts a

1002 watts doing 11050 set at freq and 8.7volt

which is 90 watts a th.

The new bitmain firmware sets your boards to (roughly) 450 frequency and 8.3 volts. So, I'd assume you could do 500 at 8.3 or 8.4 and get better watt/TH