Bitmain's Released Antminer S9, World's First 16nm Miner Ready to Order

[Biodom]

Are S9 i/o connectors the same as S7 (18pin?). I can't tell as my S9s are at the hosting site.
Somehow, there are connectors on sale at bitmain's part shop, but they don't mention S9.

[HagssFIN]

Are S9 i/o connectors the same as S7 (18pin?). I can't tell as my S9s are at the hosting site.
Somehow, there are connectors on sale at bitmain's part shop, but they don't mention S9.

Yeah connectors are the same.
See these pictures (copyright dogiecoin.com):
S7: https://c2.staticflickr.com/6/5711/21733063766_1932e7d84b_o.jpg

S9: https://c2.staticflickr.com/8/7350/27791304401_90b0eccce8_o.jpg


So are you talking about this 18-pin cable at Bitmaintech.com?
https://shop.bitmain.com/productDetail.htm?pid=00020160105105043457FPP8X92I0616

[philipma1957]

May be a stupid question but has anyone changed the fans for quieter ones on the S9? How does it perform temp wise and noise wise?

not an option worth doing.  gear tends to overheat.

if noise is an issue   you can attempt sound proofing methods.

I have to go to the solar array today.  I will photo how I use this material


https://www.amazon.com/Car-Insulation-Deadener-Automotive-Lightweight/dp/B016QUS18U/ref=sr_1_3?


with this shelfing


http://www.target.com/p/adjustable-3-tier-wide-wire-shelving-chrome-room-essentials/-/A-17136152?lnk=abtest_searchpdp_3

Thanks Philip, I'm intrigued by the thinner soundproofing method. Only reason I ask is that I'm considering an S9 just incase they don't bring out another batch of the R4 as I need to get a more efficient unit now. And as you've guessed it's a sound issue  

it's dropped again after 13 hours

And the gui for the miner looks like all 3 boards are still hashing?  Based on the actual hashrate it looks like one of the boards stops working.  If the miner gui says everything is working normally and shows a hashrate for all 3 boards then I would be inclined to suspect it was a controller problem and not a hashboard problem.

Are the ambient temps ok? My S7LN would drop a board if it goes below 18c ambient.

thread is up on sound proofing tunnel /tent

https://bitcointalk.org/index.php?topic=1650756.0

[numnutz2009]

ok so some more info after i got a reply from bitmain...


1. these miners have an auto freq checker that checks every freq when the miner turns on to run the board at the highest posible freq which causes huge fluctuations in board temps.

2. bitmain confirmed when i said before....they say it can take up to 40 minutes before the miner starts hashing. this is because every time the miner reboots it has to run through the freq checker again! (i honestly think this is to use lesser quality boards with higher quality boards. one board like mine will run harder and much hotter which isnt good for that board)

3. no fan control. instead it uses an auto fan control feature that fluctuates the fan speed based on the temps of the boards. this sounds great but like i said one of my miners has a board that runs almost 100 degrees at the chip which none of my other miners come close to that so the flans constantly speed up to 5krpm when the temps hit 100 degrees then drop back down to 2400 rpm when the temps go below what looks like 95 degrees at the chip. this happens over and over with no end and sounds like a kid pullin up at the light next to u revin his mommy and daddys dodge charger just a little slower. i told them i want fan controls back in the firmware.

4. take screen shots after the miner shows 40+ minutes of runtime before sending them to bitmain. i had no idea i had the wait the better part of an hour for the miner to finally start so i emailed them after waiting and seeing socket connection errors which was a false error...it was just that the miner was testing the frequencies.

5. you can see if your miner is testing the frequencies by waiting for the miner to fully boot up and putting you hand in the air coming out. when testing frequencies the temps of the air will go warm then cold then warm then cold....it does this for like 40 minutes so if you notice the temp changes wait 40 minutes at least before assuming your miner is dead! orrrr sell you so called "dead" miner to me at a significant discount . lol jk but this part is pretty important.


this new auto freq is dangerous. i run my s9's with 2 pcie cables per board. the cables are not warm to the touch at all when using 2 because they are 16awg wires and i dont overclock the miner at all. it is easy to figure out how many to use...you just take the wattage (i have a meter) and divide it by 3 ( yes the controllers power isnt separated from the wattage just in case) and each 16awg cable if well made can handle something like 300w each but i try not to push it past 200 and change so my wires dont run hot at all. esentially what this new firmware does is make the miner less efficient because some boards are being overclocked to higher freqs making the math above useless because each board wont use about the same amount of power....the usage will fluctuate alot between the 3 boards. this is retardedddddddd.

[Biodom]

Are S9 i/o connectors the same as S7 (18pin?). I can't tell as my S9s are at the hosting site.
Somehow, there are connectors on sale at bitmain's part shop, but they don't mention S9.

Yeah connectors are the same.
See these pictures (copyright dogiecoin.com):
S7: https://c2.staticflickr.com/6/5711/21733063766_1932e7d84b_o.jpg

S9: https://c2.staticflickr.com/8/7350/27791304401_90b0eccce8_o.jpg


So are you talking about this 18-pin cable at Bitmaintech.com?
https://shop.bitmain.com/productDetail.htm?pid=00020160105105043457FPP8X92I0616

yeah, those, thanks.

[tbonetony]

ok so some more info after i got a reply from bitmain...


1. these miners have an auto freq checker that checks every freq when the miner turns on to run the board at the highest posible freq which causes huge fluctuations in board temps.

2. bitmain confirmed when i said before....they say it can take up to 40 minutes before the miner starts hashing. this is because every time the miner reboots it has to run through the freq checker again! (i honestly think this is to use lesser quality boards with higher quality boards. one board like mine will run harder and much hotter which isnt good for that board)

3. no fan control. instead it uses an auto fan control feature that fluctuates the fan speed based on the temps of the boards. this sounds great but like i said one of my miners has a board that runs almost 100 degrees at the chip which none of my other miners come close to that so the flans constantly speed up to 5krpm when the temps hit 100 degrees then drop back down to 2400 rpm when the temps go below what looks like 95 degrees at the chip. this happens over and over with no end and sounds like a kid pullin up at the light next to u revin his mommy and daddys dodge charger just a little slower. i told them i want fan controls back in the firmware.

4. take screen shots after the miner shows 40+ minutes of runtime before sending them to bitmain. i had no idea i had the wait the better part of an hour for the miner to finally start so i emailed them after waiting and seeing socket connection errors which was a false error...it was just that the miner was testing the frequencies.

5. you can see if your miner is testing the frequencies by waiting for the miner to fully boot up and putting you hand in the air coming out. when testing frequencies the temps of the air will go warm then cold then warm then cold....it does this for like 40 minutes so if you notice the temp changes wait 40 minutes at least before assuming your miner is dead! orrrr sell you so called "dead" miner to me at a significant discount . lol jk but this part is pretty important.


this new auto freq is dangerous. i run my s9's with 2 pcie cables per board. the cables are not warm to the touch at all when using 2 because they are 16awg wires and i dont overclock the miner at all. it is easy to figure out how many to use...you just take the wattage (i have a meter) and divide it by 3 ( yes the controllers power isnt separated from the wattage just in case) and each 16awg cable if well made can handle something like 300w each but i try not to push it past 200 and change so my wires dont run hot at all. esentially what this new firmware does is make the miner less efficient because some boards are being overclocked to higher freqs making the math above useless because each board wont use about the same amount of power....the usage will fluctuate alot between the 3 boards. this is retardedddddddd.

Thanks for sharing this great information. I learned some of it and it does echo with some of my observations with S9.

[mamiakimo]

i lost internet for 3 minutes and after return all miners return to hash with the same hash rate except B17 return to hash with 9.8TH/s


after restart it returned to hash with full hash rate

[mamiakimo]

now after 5 hours i lost all hash rate

[yslyung]

did anyone notice that the chip temp "magically" becomes lower ?

it used to be & also a rule of thumb, PCB temp +/- 30 = CHIP temp ?

eg (post above on chain #6): PCB temp 81 + 30 = 111 (CHIP temp)  but in this "new magical" FW version 81 PCB temp/ 98 CHIP temp

maybe those who owns the earlier versions before batch 16 & 17 can do a comparison ?

hmmm, IDEAL hashrate & Auto Tune, but why it takes 20 or 30 or even up to 40 mins to start to mine ? it "really" has to be doing something "magical" !

was expecting it to be more advanced instead of taking a or a few steps backwards.

now u see me now u don't but just pay me (BMT) ! LoL

sad, really sad, bring back the simple & super reliable S3's to S10/11/12/13/14/15 .... with OPEN SOURCE to the PUBLIC ! you can just chuck it at almost anywhere & run it be it high ambient temps, dust or no dust, not that SUPER LOUD jet engine.

really sad to see the newer it is the worse it becomes.

ah yeah the superb pricing too . . .

just my 2 cents, if the buying STOPS the KILLING STOPS too.

[gt_addict]

now after 5 hours i lost all hash rate

You've got a bad board. That HW figure is crazy. The "auto tune" hasn't worked for that board. Contact bitmain and tell them the error rate is high and due to there new "improved " firmware you can't change the freq to get it to run any better.

[mamiakimo]

i asked them to give me an old firmware to test and there support need more than 24 hours to answer every time!

[d57heinz]

did anyone notice that the chip temp "magically" becomes lower ?

it used to be & also a rule of thumb, PCB temp +/- 30 = CHIP temp ?

eg (post above on chain #6): PCB temp 81 + 30 = 111 (CHIP temp)  but in this "new magical" FW version 81 PCB temp/ 98 CHIP temp

maybe those who owns the earlier versions before batch 16 & 17 can do a comparison ?

hmmm, IDEAL hashrate & Auto Tune, but why it takes 20 or 30 or even up to 40 mins to start to mine ? it "really" has to be doing something "magical" !

was expecting it to be more advanced instead of taking a or a few steps backwards.

now u see me now u don't but just pay me (BMT) ! LoL

sad, really sad, bring back the simple & super reliable S3's to S10/11/12/13/14/15 .... with OPEN SOURCE to the PUBLIC ! you can just chuck it at almost anywhere & run it be it high ambient temps, dust or no dust, not that SUPER LOUD jet engine.

really sad to see the newer it is the worse it becomes.

ah yeah the superb pricing too . . .

just my 2 cents, if the buying STOPS the KILLING STOPS too.

My thoughts exactly. And what scares me most is most of the network is built on this hw. What does this say for Bitcoin reliability.  Bitmain is becoming more and more an Achilles heal more than helping secure the network.  As time passes more are starting to see the clusterfuck that is bitmain tech. Really sad as I met Yoshi back at hashers united in 2014 in vegas.  Was good dude it seemed. Had some nice eye candy with him to sell s4 to all us nerds.  Got quite the kick outta that.  Remember that yoshi hehehe.  Anyway it is sad to see them walk backwards.  It really could be so much better and it shows we're willing to pay for that by what we paid for the initial cost of s9. Compared to what it drop they could put that into reliability instead of profits. But meh that's me thinking of that ideal world again.

Best regards

[mamiakimo]

i went to boot loop now
every 20 minutes and before the machine start it restart by self

[AriesIV10]

now after 5 hours i lost all hash rate

What software is this?

[NotFuzzyWarm]

What software is this?

If you are referring to the monitor software -- it is Awesome Miner from Awesomeminer.com.
I also use it and gotta say - it's great and well worth the 1-time cost when controlling >4 miners. 4 or less is free!

[NotFuzzyWarm]

i went to boot loop now
every 20 minutes and before the machine start it restart by self

As others said - contact Bitmain to start the RMA process.
For now -- and document all of this with screen shots --,
At the bad hashboard, swap the data cable with a hashboard next to it.

Restart the miner and take screen shot. Did the problem chain move?
If it did the hashboard is bad.
If it did NOT then either the cable or controller is bad.
Odds are it's the board.

Once the problem is identified I'd disconnect that 1 board/cable so the remaining 2 can work properly. DO NOT BREAK THE WARRANTY SEAL UNTIL BITMAIN SAYS YOU CAN!

**TIP** unplugging the data cable can be -- difficult. You need to squeeze the long tab very firmly to disengage the hooks, then rock the connector back and forth as you pull.

[mamiakimo]

i went to boot loop now
every 20 minutes and before the machine start it restart by self

As others said - contact Bitmain to start the RMA process.
For now -- and document all of this with screen shots --,
At the bad hashboard, swap the data cable with a hashboard next to it.

Restart the miner and take screen shot. Did the problem chain move?
If it did the hashboard is bad.
If it did NOT then either the cable or controller is bad.
Odds are it's the board.

Once the problem is identified I disconnect that 1 board/cable so the remaining 2 can work properly. DO NOT BREAK THE WARRANTY SEAL UNTIL BITMAIN SAYS YOU CAN!

**TIP** unplugging the data cable can be -- difficult. You need to squeeze the long tab very firmly to disengage the hooks, then rock the connector back and forth as you pull.

i replaced the psu to another one used on B600 and they both work correctly now!
i hope this issue wouldn't return

[NotFuzzyWarm]

<snip>

i replaced the psu to another one used on B600 and they both work correctly now!
i hope this issue wouldn't return

Cool.
Query: what PSU's are you using? I've pretty much have mostly used Bitmains 1,600W ones and some are getting very long in the tooth... At least one that was powering a s9 b3 had to be swapped out with a bit newer one (previously used on a s7). Miner seems more stable now although even at worst dropouts were maybe 1 every5-6 days.

[mamiakimo]

<snip>

i replaced the psu to another one used on B600 and they both work correctly now!
i hope this issue wouldn't return

Cool.
Query: what PSU's are you using? I've pretty much have mostly used Bitmains 1,600W ones and some are getting very long in the tooth... At least one that was powering a s9 b3 had to be swapped out with a bit newer one (previously used on a s7). Miner seems more stable now although even at worst dropouts were maybe 1 every5-6 days.

PSU: EVGA 1600G2
i think i found the problem
i have unstable internet connection
i am losing it about 5-10 times a day for 1 minute or 2
all my miners haven't problem with that except B17
some times it can handle this disconnect but some other times it losing hash rate and not return to full hash rate until i restart it
so it's the new firmware issue i guess
i will ask the support to give me an old firmware until they fix this issue

[NotFuzzyWarm]

Since you are using Awesome miner to monitor/control, try editing the Miner Offline rule to use 'Start Miner" Instead of "Restart Miner. I just started trying that and so far it seems to give the miner a kick in the pants without doing a full restart/warm boot.