Topic: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH - page 249. (Read 528055 times)

Even though my PSU has 12 PCI-Es, I specifically ran it on 7 to prove the point. As expected there were absolutely no problems for weeks on end, cables remain at ambient.


I'd much rather do a quick shake test when putting it on the rack than be constantly worried of lose fish paper being eaten by a fan and locking it up.

RMs initially suffered a derp in which the fan remained too quiet for too long, leading to the PSU reaching thermal shutdown before the fan kicked up and removed the heat.

just set fans to 60% 

 

Huh, got kinda quiet here....

Everyone must be playing with their S7's .... either that or the runaway military blimp in Maryland / Pennsilvania that is causing mayhem to the power grid ( and anything else strung by wire above ground like phone, CATV etc) has knocked everyone offline.

So, can we sue the government if the blimp knocks out the internet and the S7's burn up?  

Holy crap don`t do that, I just received my S7 today, it`s not even plugged in yet since i`m still at work and i was like " WHAT, the fucking this is already obsolete  " 

yes, and two other things fix the problem as well:
a. setting manual % for fan (in my experince 80% results in exactly the same fan speed as not setting anything at all). i have confirmed others observations that it results in fan continuously blowing at high speed EVEN when internet goes down.
b. update the software (someone had posted that it resulted in the same effect as #a)

however, having an outside fan is a good idea anyway.

OH WAIT!!!

The Maybe the S6 will finally make an appearance !!!   

I just scrapped a HUGE pile of old servers and ancient Chinese bitcoin mining rellics... I have fans up the ying and coming out the yang .... if anyone needs a few, let me know and I'll research the specs on these fans and post it along with a price ( cheap )....

WOW - HEART ATTACK!!!!   Whew,  if you only knew how much I have invested in the S7s right now.  Oh, you really got me there! 

Hmm .... I don't know what is more of an issue .... the unknowns of alien tech, or that nuclear generating station I'm building in my shed to power all my miners using the radioactive materials from smoke detectors....

 

 

 

It was posted that if the internet goes down it could over heat the miner due to the fans dropping down and the miner still hashing away. After reading that I now have a Delta fane about 3 inches in front of the intake fan blowing air into it as my internet provider is crap.

alien technology, man, straight from KIC 8462852 (what do you think is blocking that starlight-kiccoin mining operation)  

LOL, obviously I'm just messing with ya ... You guess is as good as anyone elses.... only Bitmain R&D knows.

Then again, I predict an S8 or an S9... lol... obviously

Bitmain is releasing a miner next week that will make the S7 obsolete. Its based on a 5.75nm process, 250TH/s in the same footprint as the S1/S3/S5/S7 ... and only uses 50watts at the wall....

  Does anyone have any information as to WHAT IS NEXT?  

Is there going to be another type of Antminer or are they just catching up on current orders? 

Thanks for all the replys that i have.

I read here that there was a problem with the internet going down.
I think that was the problem yesterday...

Since 8.00 AM from today, the miner is working as a bad ass with no worries at all.
I did not see wha was the firmware version, but i'll pass my eye on it.

It comes down to how many amps of 12VDC per Molex pin. the more amps being stuffed through each pin increases the heat generated in the pin, the pin to socket physical contact connection and at the solder connections.

Lets say the unit uses 100A of 12VDC. You have 3 Molex 6 pin connectors per hash board x 3 hash boards. In the molex connector 3 are positive DC, 3 are negative DC. I have no clue how many amps the controller uses so I'll just focus on the hash boards. ANALOGY : Picture the power supply as a water pump in a bucket. You turn on the pump, water flows through one hose,that is positive, it goes to a water cooling block for a cpu, the hot water exiting the water block enters another tube and goes back to the bucket, thats the negative.... so each positive and negative pin is a circuit and can only pump just so much through the circuit. What I'm getting at is, even though you use two pins, 1 positive and 1 negative, you are NOT adding the amperage capacity of the two pins together, both pins are the same "hose", basically, what goes through one end of the hose comes out the other.

A total of 9 Molex connectors, That would be 27 pins for positive and 27 pins for negative. 100amps of 12VDC divided by 27 equals 3.7A of 12vdc going through each pin.

Now if you reduce the amount of pins by not using ALL the molex connectors with your power supply, you will increase the current going through each pin. Using 2 Molex connectors per hash board equals 18 pins positive and 18 pins negative. 100A of 12VDC divided by 18 equals 5.5 amps going through each pin.


Each pin of the Molex Mini-Fit Jr. connector is rated for 9A maximum. So the pin and the mating connector would handle the current going through the connector if using just 2 connectors per hash board. BUT, how is the circuit board designed? Is the circuit board designed to handle the additional heat generated  by just using 2 connectors? I would not think so since adding additional layers of heat dissipating copper in the layered fiber board would cost money and we all know Bitmain is in it for the cashola.  My guess would be that if you use just 2 of the Molex connectors per hash board, you will find that the circuit board will start to discolor and or the Molex connector due to the excessive heat generated. You will also find that the voltage will drop under load at the hashboard causing inefficiency.

I am powering my S7 with 1-RM1000 and a RM850. I just left both PSU turned off, plugged all cables in and Ethernet cable then turned on the 850 first as it was just powering 1 hash board then turned on the 1000 as it's powering the other 2 boards and the controller. At first it looked to be about a 1TH short of hash so I did a soft reboot and it came right up to speed of 4.8+

I think that you are precisely correct when you say, "turning them on one at a time with the control board psu being the last one to turn on."  I have quite a few S7s now and have several powered by 3 power supplies and followed the power on procedure that you noted without any problems.

not a concern re overheating initially as it would take it a minute or so to situate itself on your network and then hashing will build up slowly (don't expect it to jump to 4.86 in a few seconds). Wait at least 1 hr before reaching any conclusion where your miner is operating as far as speed is concerned and if high errors are seen, try soft reboot in gui (I believe that it is in system pane).