Author

Topic: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH - page 237. (Read 527809 times)

sr. member
Activity: 277
Merit: 250
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

That's over 350W DC for 1 cable.  The connectors/pins are only rated for a maxijum of 324W, and those ratings aren't overly conservative given the demanding environments mining puts on them (vibration + high ambient temps).  Plug more cables into them, like now...

Edit: Plus it violates Bitmain's ridiculous warranty terms...

I'm doing 8 total, 2 each to outside boards and 3 to center plus controller. That way I can use 2 dell 750 PSU with 8 cables. Using high quality 16awg cables and breakout boards. No issues. Been monitoring temps of connectors and cables. Avg connect temp is 79F. Miner runs at 55-59c and ambient temp is 69-72F.

That would be 175w per cable on the boards with 2 plugged in. Consumer PSU may not handle this since cables are lower gauge and will heat up more.
The wire gauge on the bitmain 1600 psu's is pretty teenie.
Maybe that is why the "three cable rule" is pushed.
My corsair rm1000 modular cables are pretty beefy. I use two per board with the daisy chain from one of them, filling all connectors.
The board next to the controller socket gets the daisy chain from the second cable.

I'm not sure what the wire gauge number is on the Bitmain power supplies, someone will have to see if there is some printing on the wires and post the AWG size.

BUT, for reference, when I made the cables for the DPS-1200FB power supplies I have, I used 20ga wire.... its what I had laying around that would fit the pins I have.... Calculating the load per wire,  I divided the maximum amperage the power supply could produce across the 6 Molex plugs, the amperage per pin/wire was 5.5 amp.... WELL within the specification for the wire size and insulation type using an NEC wire chart.

YES, insulation type is also a factor. You can have anything from varnish coated wires, plastic insulated, silicone insulated, or even teflon insulated. Some insulation melts at lower temps, so if you have a Molex connector that heating up because of old age or you have been plugging and unplugging a lot and you have cheap ass plastic insulated wires, more likely you will have a meltdown from insulation melting then wires shorting out than if you were running a smaller wire gauge, same amperage, and teflon insulation.

You might think silicone insulated wires will handle the heat... somewhat. Silicone insulation is used when maximum flexability is required, or is in a higher heat environment, or is carrying higher voltage (think 1000V or more). Silicone insulation is not very good when its in direct contact with a high temperature material such as a wire which has a contact that is micro arcing causing the wire to heat up massively.... it will melt with relative ease.


legendary
Activity: 3892
Merit: 4331
Did anybody try to put S7 in the attic instead of regular floors, basement  or garage? We don't have basements here, typically.
What would be a worry?
Sound should not be a problem, not sure about heat, although with fall coming (it is still balmy 75F right now), maybe it will be OK?
full member
Activity: 120
Merit: 100
is there going to be a batch 4?
legendary
Activity: 1218
Merit: 1003
That is consistent with my experience.  I have had many new cables/connectors under high loads that stay nice and cool, but a few months later they start getting hot.  I have also noticed variance even between samples of high quality cables.  After many negative experiences with KncMiner Neptunes that only have one PCIe, I plan to use all three on my S7s for backup and redundancy.  Burning my house down to save a few $'s on cables is not worth it for me.

Today I got the smell of smoke from my Corsair CX750M

It was connected with one S7 board and S7 controller.
legendary
Activity: 3892
Merit: 4331
Hello

Have a question.

What should I consider when chosing a mining pool?. I have a S7 B2 on its way to my house. It should be delivered on thursday and this is my first miner.

I have been reading about the different methods each pool work (PPS, PPLNS, DGM, SMPPS, etc...) and also we can find several pools on each method. Which method and which pool do you think work better and why? If two given pools work on PPS for example, how is that possible that one of them could have their luck on "-8%" (just saying something), I thought statiscally both of them should be "the same thing" (of course according to their each mining total power)

Do they have any trick I should take care of?

Is there anyway to measure efficiency of each mining pool without having to test each pool for several hours or days?

Thank you

Re luck and mining pool.
Below is the great writeup from a known pool observing expert:
http://organofcorti.blogspot.com/2015/07/faq-bitcoin-mining-and-luck-statistic.html
Any miner should read it, in my opinion.

basically two points:
1. any pool would have longer bad luck than good luck
2. from #1 it follows that you will underperform if you chase luck from pool to pool because you would never know when a period of bad luck stops and good luck begins and vice versa.

jr. member
Activity: 35
Merit: 1
That is consistent with my experience.  I have had many new cables/connectors under high loads that stay nice and cool, but a few months later they start getting hot.  I have also noticed variance even between samples of high quality cables.  After many negative experiences with KncMiner Neptunes that only have one PCIe, I plan to use all three on my S7s for backup and redundancy.  Burning my house down to save a few $'s on cables is not worth it for me.
newbie
Activity: 15
Merit: 0
Hello

Have a question.

What should I consider when chosing a mining pool?. I have a S7 B2 on its way to my house. It should be delivered on thursday and this is my first miner.

I have been reading about the different methods each pool work (PPS, PPLNS, DGM, SMPPS, etc...) and also we can find several pools on each method. Which method and which pool do you think work better and why? If two given pools work on PPS for example, how is that possible that one of them could have their luck on "-8%" (just saying something), I thought statiscally both of them should be "the same thing" (of course according to their each mining total power)

Do they have any trick I should take care of?

Is there anyway to measure efficiency of each mining pool without having to test each pool for several hours or days?

Thank you
sr. member
Activity: 277
Merit: 250
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

wow that's gutsy. cables not getting warm?

I'd also be worried about the connectors, sure they'll be overspec'ed but not infinitely.

As I mentioned when someone asked about running with just 2 connectors powering each board ... The Molex Mini-Fit Jr pins are rated for 9 amp maximum. HOWEVER, its not a continuous rating. The friction fit of the pins also introduces a resistance, the older the plug on the power supply, the looser the pin to pin contact thus increasing heat produced and potential "micro" arcing which then DRAMATICALLY increases pin temp. The inserting and removing of the plug from devices "stretches" and bends the contacts, the first time a plug is used, its the strongest contact, but as heat is produced due to load, the springyness of the contact material diminishes and the connector will loosen up on its own with time.

I can post some pictures of some of the Molex Mini-Fit Jr connectors from an Avalon that was running stock clock that just had enough of running continuous moderate amperage - not near the maximum-  through the pins.... nice n chared and had a small fire which delaminated the circuit board and required extensive modifications to the board to get it running again.

Unless you are made of money and can afford a $1600 doorstop... go right ahead.... Also, make sure your renters / home owners insurance is up to date and would cover such an instance.... its only a matter of time when taking that kind of short cut.
full member
Activity: 175
Merit: 101
cryptominer.ca


Well there you go guys!  It soulds like a pretty good possibility for a host site...

EDIT:  I am having a bit of an issue reaching the http://cryptominer.ca website...
EDIT 2:  I did finally get past the certificate issue...

Hi!

Sorry... I thought I fixed the issue this weekend but that wasn't the case. It should be fixed now.

The website is very minimalistic, but it's in my todo list to improve the company image. I'm currently busy building the backend for now..
legendary
Activity: 1722
Merit: 1032
Carl, aka Sonny :)
The last quote I got from Great North Data was $60/KW but you had to have a minimum requirement of 30KW in equipment.  I wanted to send 20 S5s to them but that was only 11-12KW and they would not take it.

Hi,

I'm based in Canada too and I do hosting. My rates are (Taxes included) 74.73 CAD which is currently ~ 57 USD / kW. I don't have any minimum and I'm currently working on a new website where customers will be able to manage their miners. There is no minimum needed.. (Why would there be...??)

My website is currently http://cryptominer.ca .

Just PM/send me an email if you are interested.

Well there you go guys!  It soulds like a pretty good possibility for a host site...

EDIT:  I am having a bit of an issue reaching the http://cryptominer.ca website...
EDIT 2:  I did finally get past the certificate issue...
legendary
Activity: 1456
Merit: 1000
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

wow that's gutsy. cables not getting warm?

I'd also be worried about the connectors, sure they'll be overspec'ed but not infinitely.

If you do it put up a fire alarm near it seriously.  One is asking for trouble.  I imagine it will get pretty hot, and your talking about possible burning with 1.

Get a better PSU and use more then 1 per board, is best advice.
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

wow that's gutsy. cables not getting warm?

I'd also be worried about the connectors, sure they'll be overspec'ed but not infinitely.
full member
Activity: 180
Merit: 100
Running against nicehash.
S5's would every so often crash the os against nicehash as it was switching rentals.
My remote pdu allowed me to cycle the power and bring them back up.
The s5's are gone now.

Just now my first s7 had the os crash with nicehash.
Did the remote pdu power recycle and it is back hashing away.
More of an impact when 5th are taken out rather than the 1th.

Now I only have a single S7 but haven't had this issue before.  Up 2200 minutes still with nicehash.
Only one once with 17 running against nicehash.
Can't even ping the miner. HAVE to power reset.
Pretty stable usually.
Payout pretty good on MRR?

MRR can payout good.  Depends on what you set your rental rate at, I tend to set mine a bit higher then 100% and have had good rentals, but not much lately with the increase in hash network-wide and high diff.  I like MRR too because I point all my miners to it and it proxies my connections so switching pools is a simple mouse click and all my miners switch together rather than having to log into each one and manually changing the settings.

Can you use their proxy to switch all your miners to any pool? Even if not renting? Sounds like an easy way to flip pool all at once on your own gear and pools.

That's exactly why I use them.  You can set up 5 pools and you can turn off your rentals if you don't want them to be rented (or just price them really high so if they are rented, you're making bank).  All your miners are set up to MRR through their web ui, then on the MRR page, you simply select which pool you want them to mine while not being rented.  And if you have a lot of miners, that saves a lot of time.

But I have to add that they all mine the same pool, so you can't have 3 mining on one pool, 5 on another...all 8 would be on the same pool.
hero member
Activity: 770
Merit: 523
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

That's over 350W DC for 1 cable.  The connectors/pins are only rated for a maxijum of 324W, and those ratings aren't overly conservative given the demanding environments mining puts on them (vibration + high ambient temps).  Plug more cables into them, like now...

Edit: Plus it violates Bitmain's ridiculous warranty terms...

I'm doing 8 total, 2 each to outside boards and 3 to center plus controller. That way I can use 2 dell 750 PSU with 8 cables. Using high quality 16awg cables and breakout boards. No issues. Been monitoring temps of connectors and cables. Avg connect temp is 79F. Miner runs at 55-59c and ambient temp is 69-72F.

That would be 175w per cable on the boards with 2 plugged in. Consumer PSU may not handle this since cables are lower gauge and will heat up more.
The wire gauge on the bitmain 1600 psu's is pretty teenie.
Maybe that is why the "three cable rule" is pushed.
My corsair rm1000 modular cables are pretty beefy. I use two per board with the daisy chain from one of them, filling all connectors.
The board next to the controller socket gets the daisy chain from the second cable.
hero member
Activity: 575
Merit: 500
Running against nicehash.
S5's would every so often crash the os against nicehash as it was switching rentals.
My remote pdu allowed me to cycle the power and bring them back up.
The s5's are gone now.

Just now my first s7 had the os crash with nicehash.
Did the remote pdu power recycle and it is back hashing away.
More of an impact when 5th are taken out rather than the 1th.

Now I only have a single S7 but haven't had this issue before.  Up 2200 minutes still with nicehash.
Only one once with 17 running against nicehash.
Can't even ping the miner. HAVE to power reset.
Pretty stable usually.
Payout pretty good on MRR?

MRR can payout good.  Depends on what you set your rental rate at, I tend to set mine a bit higher then 100% and have had good rentals, but not much lately with the increase in hash network-wide and high diff.  I like MRR too because I point all my miners to it and it proxies my connections so switching pools is a simple mouse click and all my miners switch together rather than having to log into each one and manually changing the settings.

Can you use their proxy to switch all your miners to any pool? Even if not renting? Sounds like an easy way to flip pool all at once on your own gear and pools.
full member
Activity: 180
Merit: 100
Running against nicehash.
S5's would every so often crash the os against nicehash as it was switching rentals.
My remote pdu allowed me to cycle the power and bring them back up.
The s5's are gone now.

Just now my first s7 had the os crash with nicehash.
Did the remote pdu power recycle and it is back hashing away.
More of an impact when 5th are taken out rather than the 1th.

Now I only have a single S7 but haven't had this issue before.  Up 2200 minutes still with nicehash.
Only one once with 17 running against nicehash.
Can't even ping the miner. HAVE to power reset.
Pretty stable usually.
Payout pretty good on MRR?

MRR can payout good.  Depends on what you set your rental rate at, I tend to set mine a bit higher then 100% and have had good rentals, but not much lately with the increase in hash network-wide and high diff.  I like MRR too because I point all my miners to it and it proxies my connections so switching pools is a simple mouse click and all my miners switch together rather than having to log into each one and manually changing the settings.
hero member
Activity: 575
Merit: 500
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

That's over 350W DC for 1 cable.  The connectors/pins are only rated for a maxijum of 324W, and those ratings aren't overly conservative given the demanding environments mining puts on them (vibration + high ambient temps).  Plug more cables into them, like now...

Edit: Plus it violates Bitmain's ridiculous warranty terms...

I'm doing 8 total, 2 each to outside boards and 3 to center plus controller. That way I can use 2 dell 750 PSU with 8 cables. Using high quality 16awg cables and breakout boards. No issues. Been monitoring temps of connectors and cables. Avg connect temp is 79F. Miner runs at 55-59c and ambient temp is 69-72F.

That would be 175w per cable on the boards with 2 plugged in. Consumer PSU may not handle this since cables are lower gauge and will heat up more.
hero member
Activity: 770
Merit: 523
Running against nicehash.
S5's would every so often crash the os against nicehash as it was switching rentals.
My remote pdu allowed me to cycle the power and bring them back up.
The s5's are gone now.

Just now my first s7 had the os crash with nicehash.
Did the remote pdu power recycle and it is back hashing away.
More of an impact when 5th are taken out rather than the 1th.

Now I only have a single S7 but haven't had this issue before.  Up 2200 minutes still with nicehash.
Only one once with 17 running against nicehash.
Can't even ping the miner. HAVE to power reset.
Pretty stable usually.
Payout pretty good on MRR?
That's the problem I had, luck would go in the toilet on regular pools and I had 45 s5's. Got tired of switching.
I have used putty in a batch file to shift pools and reboot on  s5's.
I have used a proxy server (the java one on ubuntu) but I seem to get a lower payout with the proxy.
Need more time to play.
full member
Activity: 180
Merit: 100
Running against nicehash.
S5's would every so often crash the os against nicehash as it was switching rentals.
My remote pdu allowed me to cycle the power and bring them back up.
The s5's are gone now.

Just now my first s7 had the os crash with nicehash.
Did the remote pdu power recycle and it is back hashing away.
More of an impact when 5th are taken out rather than the 1th.

Now I only have a single S7 but haven't had this issue before.  Up 2200 minutes still with nicehash.

Why are you running on nicehash?  I can't figure out why people still are.  Seems like negative amounts each time I look at site.

I think you would make more off a standard pool.  But I could be wrong on this.  It's currently -8 percent compared to mining BTC.

Considering I usually mine on slush and their luck is down, I'm making more on nicehash than I would be putting my hash at slush.  I move back and forth though.  I'll probably test my luck on slush again because good days there are nice on the wallet.  nicehash is just a steady stream even though it's not ideal.  I much prefer renting my hash out on MRR though.
legendary
Activity: 1022
Merit: 1003
Anyone powered his S7 using only 1 PCi-E 6 per board ?
I have powered them with only 1 cable per board for the last 2 days and they were working fine
Just today I read somewhere that it actually needs all 3 cables per board.

That's over 350W DC for 1 cable.  The connectors/pins are only rated for a maxijum of 324W, and those ratings aren't overly conservative given the demanding environments mining puts on them (vibration + high ambient temps).  Plug more cables into them, like now...

Edit: Plus it violates Bitmain's ridiculous warranty terms...
Jump to: