Pages:
Author

Topic: Antminer S9 Volt Rocket Ship firmware mod! ~Time to Max Out Your Bitmain PSU~ - page 12. (Read 32492 times)

jr. member
Activity: 306
Merit: 7
Hi again,

I tried the 880 on a S9 14TH that worked with the fixed FW 650.
Can you explain please what the differences are between 650 fixed FW and 880 and 870 FW? With 880 I dont see a big difference. But I used the mod on a machine that already did not show temps over 100° with fixed fw. So I am just cuious. Since I have a 1600W PSU, I dont want to overextent. It runs @ 16TH @ 1690W
member
Activity: 133
Merit: 15
~ Fortuna Favet Fortibus ~
I didn't understand exactly how to install.Can you make a video tutorial please?
sr. member
Activity: 800
Merit: 293
Created AutoTune to saved the planet! ~USA
Hi,
is it correct, that stock voltage of fixed frequency firmware is 940?

880 or 870
jr. member
Activity: 306
Merit: 7
Hi,
is it correct, that stock voltage of fixed frequency firmware is 940?
sr. member
Activity: 800
Merit: 293
Created AutoTune to saved the planet! ~USA
my question would still be Taserz. On GitHub I can see that there are several versions available for example: bmminer870, bmminer880, bmminer890, bmminer910, bmminer920, bmminer930, bmminer940, bmminer9v. What are the differences between them? Currently I'm using bmminer940 at 700mhz speed. Is that right?  Roll Eyes Smiley

The differences is the power draw. 8.7v 8.8v 8.9v etc etc

I would stick with 880 890 or 900 to play it safe.

I am sorry for not clarifying what I did there. All 3 of my Antminers were using default(fixed-frequency) firmware from Bitmain without taserz's edited bmminer.
However, I have now uploaded the 9v version of bmminer on to the first Antminer S9 13.5TH/s(IP 157)- I distinguish them by IP.
Here below you can find my results:

Non-modified 13.5TH/s with fixed frequency of 706 7h test: https://ibb.co/fQRBme
Taserz's 9Volt Rocket Ship bmminer9v 13.5TH/s with fixed frequency of 706 10min test: https://ibb.co/jNLbKz
                                                                                                              14h test: https://ibb.co/cCwVge

I am sharing this as a proof that it is actually working better with custom bmminer and I because I am just too happy about it(that is why I couldn't wait for more than 10min to test).
Will upload a picture with longer test time soon.  Cheesy Cheesy Cheesy Cheesy Cheesy
Thank you a lot @taserz

You are very welcome.

Also, it appears my s9i running at 19gh/s might be a bit sick Sad It appears one of the hashboards shit the bed. I am going to contact my hosting provider to power cycle it for me. Otherwise, we will see how undetected this mod is and if they will warranty it if I revert it to stock haha.
sr. member
Activity: 800
Merit: 293
Created AutoTune to saved the planet! ~USA
From what I've been noticing on my rigs, the RT asic number and the actual number relation is best when you can get all to stay within 100 of the expected. Like if it expects 5th even, and my actual is 4,990,900 or close on each board I'd leave it. Any more than that variation it's not getting enough power to it. The hw errors happen during low current situations I'm assuming from the power load, or during blocks that aren't sent to the pool in time and it has extra work rolled that leaks lower than difficulty.

Try to lower your frequency to keep the RT and Actual within 100 of each other. If you're on 8.9V go to 9v or 9.4v but remember it's more draw you can go higher if you have the amps!

I agree as this is what I have been realizing and trying to GR and average the same seems to be the key and if you have to little power you will get HW errors more often than the random ones.
member
Activity: 434
Merit: 30


i have them running for a few weeks now. So far so good.
both S9 13.5TH with APW5++ 1600W psu.
sr. member
Activity: 604
Merit: 416
From what I've been noticing on my rigs, the RT asic number and the actual number relation is best when you can get all to stay within 100 of the expected. Like if it expects 5th even, and my actual is 4,990,900 or close on each board I'd leave it. Any more than that variation it's not getting enough power to it. The hw errors happen during low current situations I'm assuming from the power load, or during blocks that aren't sent to the pool in time and it has extra work rolled that leaks lower than difficulty.

Try to lower your frequency to keep the RT and Actual within 100 of each other. If you're on 8.9V go to 9v or 9.4v but remember it's more draw you can go higher if you have the amps!

I am sorry for not clarifying what I did there. All 3 of my Antminers were using default(fixed-frequency) firmware from Bitmain without taserz's edited bmminer.
However, I have now uploaded the 9v version of bmminer on to the first Antminer S9 13.5TH/s(IP 157)- I distinguish them by IP.
Here below you can find my results:

Non-modified 13.5TH/s with fixed frequency of 706 7h test: https://ibb.co/fQRBme
Taserz's 9Volt Rocket Ship bmminer9v 13.5TH/s with fixed frequency of 706 10min test: https://ibb.co/jNLbKz
                                                                                                              14h test: https://ibb.co/cCwVge

Thank you a lot @taserz
newbie
Activity: 6
Merit: 0
my question would still be Taserz. On GitHub I can see that there are several versions available for example: bmminer870, bmminer880, bmminer890, bmminer910, bmminer920, bmminer930, bmminer940, bmminer9v. What are the differences between them? Currently I'm using bmminer940 at 700mhz speed. Is that right?  Roll Eyes Smiley
full member
Activity: 538
Merit: 175
Try to lower your frequency to keep the RT and Actual within 100 of each other. If you're on 8.9V go to 9v or 9.4v but remember it's more draw you can go higher if you have the amps!

By the Actual, are you referring to GH/s(avg) or GH/s(ideal) ?
newbie
Activity: 6
Merit: 0
I tried custom firmware. It works perfectly so far. Thank you so much for taserz! Smiley (Antminer S9i - 14TH\s)
jr. member
Activity: 59
Merit: 10
From what I've been noticing on my rigs, the RT asic number and the actual number relation is best when you can get all to stay within 100 of the expected. Like if it expects 5th even, and my actual is 4,990,900 or close on each board I'd leave it. Any more than that variation it's not getting enough power to it. The hw errors happen during low current situations I'm assuming from the power load, or during blocks that aren't sent to the pool in time and it has extra work rolled that leaks lower than difficulty.

Try to lower your frequency to keep the RT and Actual within 100 of each other. If you're on 8.9V go to 9v or 9.4v but remember it's more draw you can go higher if you have the amps!
sr. member
Activity: 604
Merit: 416
2x Bitmain. they normally give more than they are rated for.

Been using the 890V, absolutely awesome

the si one i tested for a week or so, topped it out at 15th on pool side  with virtually no errors. (712)

https://ibb.co/coQhpz

this is the 890. 13.5 th model. 120V, bitmain psu  the hw happen in groups im assuming its a stale situation when blocks change

That sounds awesome.
This is what I achieved with Bitmain's firmware and no additional changes except the frequency:

1st Antminer S9 13.5TH/s: https://ibb.co/bvOGhK
2nd Antminer S9 13.5TH/s: https://ibb.co/f0zBGe
Antminer S9i 14TH/s: https://ibb.co/gRd5be
jr. member
Activity: 59
Merit: 10
2x Bitmain. they normally give more than they are rated for.

Been using the 890V, absolutely awesome

the si one i tested for a week or so, topped it out at 15th on pool side  with virtually no errors. (712)

https://ibb.co/coQhpz

this is the 890. 13.5 th model. 120V, bitmain psu  the hw happen in groups im assuming its a stale situation when blocks change
sr. member
Activity: 604
Merit: 416
Hi, a little bit off-topic question, but what would you rather recommend, one server PSU of like 2kw+ or 2 Bitmain PSU 1600W for powering one Antminer S9/S9i overclocked?
sr. member
Activity: 446
Merit: 347
Sorry, they talk about a good time to troubleshoot a persian scammed by a firmware ... and I have recovered for my own tests! at my risk and perils Wink
full member
Activity: 538
Merit: 175
Ps: Taserz, you will notice that I do not allow myself to insult you, in spite of a BMMINER who completely burned my minor ... I just hope that my message will not be suppressed like all the others! and maybe one day it will be able to discuss it amicably !!! Wink

taserz clearly indicated at the top of the thread that "This is made for the 13.5th and 14th S9i." Nowhere at all does he mention anything about putting it onto the T9+. As far as I know you are the first to brick a T9+ with his S9 firmware mod.
sr. member
Activity: 446
Merit: 347
Sorry i edit my diag after most test !

Only chain 1 and 2 are lost but is reprograming the pic with wronk fw .....

I continued my recherche...
sr. member
Activity: 446
Merit: 347
teserz , still not have any news from DARVAL?

my T9+ still bricked because for Dollsmizer Scam firmware.

Hi, well ... I now have the same problem (thanks Taserz for BMMINER-880-T9 + in passing;))

After analysis, in fact the firmware did not burn your cards, when changing the BMMINER, this one in the kernel log did: "pic error, need restore, restoring"

In reality, the T9 + is composed of 3 channels of 18 asics per card,

After analysis, the damaged card:
-1 has a string of 18 unassigned asics
-2 string asics not assigned
-3 all the chains are lost not assign (the 3)

You can check my arguments by fading through the fan, you see red led lights, a red led corresponds to a chain of 18 asics ...

The problem now, now that we REALLY know the failure, we must successfully reprogram the "dsPIC33EP16GS202"!

Interestingly, when changing BMMINER, it was reprogrammed automatically!
                                       

Ps: Taserz, you will notice that I do not allow myself to insult you, in spite of a BMMINER who completely burned my minor ... I just hope that my message will not be suppressed like all the others! and maybe one day it will be able to discuss it amicably !!! Wink
member
Activity: 113
Merit: 31
Default web login: root
Default web password: root

Default SSH login: root
Default SSH password: admin

Default SCP login: root
Default SCP password: admin

Just like Taserz posted in the OP and responded with 2 posts above, no bugs involved, just a different default password.
Pages:
Jump to: