I will get a new batt for my multimeter and do some measureing for you guys.
My S7 is running stable at 4.8 at 600. Power at wall 1180. Using a RM1000 and a RM650
Topic: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH - page 340. (Read 527816 times)
September 30, 2015, 04:23:09 AM
September 30, 2015, 03:44:02 AM
I think all my problems were due to power off power on when I went from 2 to 1 machines. A simple software system reboot really looks like a winning solution.
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
I've tried that at stock freq and to no avail, still lots of HW and lower hash.. I'm curious if your unit holds its current levels for a day or two after soft reboot. I'm going to try each of my miners individually on my 2880 tomorrow and report back
thank you. I am still doing well over 4800 - 1200 watts and 0.0110 errors vs 0.1100 errors.
My cables are providing 11.94 watts with draw taken into account
I am able to measure 8 of the 9 cables. if you can find a two headed pcie cable and plug in one head. then us the other head to measure the draw.
I will photo what I mean later today.
September 30, 2015, 01:38:47 AM
I think all my problems were due to power off power on when I went from 2 to 1 machines. A simple software system reboot really looks like a winning solution.
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
I've tried that at stock freq and to no avail, still lots of HW and lower hash.. I'm curious if your unit holds its current levels for a day or two after soft reboot. I'm going to try each of my miners individually on my 2880 tomorrow and report back
September 30, 2015, 01:32:33 AM
Looks like Bitmain temporarily stopped selling everything except for Pacmic V3 on their website.
September 30, 2015, 12:17:52 AM
Thanks for letting some of us sleep tonight while awaiting our s7's, you rock!
Ufo
Ufo
September 30, 2015, 12:07:37 AM
I think all my problems were due to power off power on when I went from 2 to 1 machines. A simple software system reboot really looks like a winning solution.
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
https://bitcointalksearch.org/topic/my-bitmaintech-s-7-underpreforms-at-freq-600-freq-550-not-freq-500-fixed-1196671
September 29, 2015, 11:25:36 PM
1 of my S7s is exactly the same. I was getting 4300th/s at 600. I dropped it to 593 and got about 4400. As a test I have been running at 575 and it is a steady 4400th/s.
What PSU are you using? Are you using the server one bitmain sells, APW3-12-1600-B2?
I am using 3 Antec 750W Platinum PSUs. Plenty of Power.
September 29, 2015, 11:19:00 PM
[/quote]
Actually it is... 600MHz / 4860GH/s * 575MHz = 4657GH/s
So guess what... They are running with 575MHz instead of 600MHz -> I bet my left hand on it.
The price is quite good, I'm in for a batch 2 now. Does anyone have an S7 coupon to sell? Thanks.
[/quote]
I think you are right, the stock power usage on batch is lower than batch 3.
Batch 2: Power Consumption: 1160 W
Batch 3: Power Consumption: 1210 W
September 29, 2015, 11:17:45 PM
I'm suppose to get mine tomorrow or the next day (2). I will probably test them with 2 Corsair RM1000. (My IBM 2880w is not ready yet).
I will keep you posted.
I will keep you posted.
September 29, 2015, 11:02:34 PM
You know, if Bitmain is going to sit on our orders for a week during their national holiday, the least they could do is give us a web cam in their data center so we can watch our S7's make money for them 
A "convenient" burn-in test
And let's not kid ourselves, we're being sold units that have been used by bitmain for AT LEAST 1 month. They know full well that some of their earliest built rigs had these issues. This isn't a last minute "oh shit" on their part as they were packing them up for shipment. They're playing dumb, and you've got to be pretty dumb yourself to believe anything they say when it comes to their fuck ups. They have a proven history of lying to their customers, and even screwing them intentionally. Voltage-gate with the S4 is a prime example.
September 29, 2015, 11:00:30 PM
well the evga 1600 p2 at 240 volts is a 1600 watt psu.
I need to do really good testing on this. As I suspect the issue is complex.
Ie if you are pointing at a pool under mild ddos you could get f'd up rates.
When I ran the 2 the room was hot 89 f so I ran the 2 at freq 493 500 and 506 and I got good numbers.
Now back to 1 unit I went to 600 and was under 4500 I went to 593 and was at 4350.
I will meter the cables now at freq 550.
I get 11.96
I have 9 cables every one is 2 plugged so I have 18 plugs
8 go into the blades 1 goes into a blade and the controller.
So I can measure 8 of the 9 cables they all measure 11.96 or 11.97 volts this is after the draw as I plug in the first pcie jack.
unplugged pcie they measure 12.07 so my voltage sag is 12.07 to 11.96 this is completely acceptable numbers.
But if I am using a 1600 watt psu feeding it 240 volts using 9 cables anyone with lessor psu's should get a bigger drop.
I will post my 550 screenshot in a bit.
I need to do really good testing on this. As I suspect the issue is complex.
Ie if you are pointing at a pool under mild ddos you could get f'd up rates.
When I ran the 2 the room was hot 89 f so I ran the 2 at freq 493 500 and 506 and I got good numbers.
Now back to 1 unit I went to 600 and was under 4500 I went to 593 and was at 4350.
I will meter the cables now at freq 550.
I get 11.96
I have 9 cables every one is 2 plugged so I have 18 plugs
8 go into the blades 1 goes into a blade and the controller.
So I can measure 8 of the 9 cables they all measure 11.96 or 11.97 volts this is after the draw as I plug in the first pcie jack.
unplugged pcie they measure 12.07 so my voltage sag is 12.07 to 11.96 this is completely acceptable numbers.
But if I am using a 1600 watt psu feeding it 240 volts using 9 cables anyone with lessor psu's should get a bigger drop.
I will post my 550 screenshot in a bit.
September 29, 2015, 10:54:02 PM
You know, if Bitmain is going to sit on our orders for a week during their national holiday, the least they could do is give us a web cam in their data center so we can watch our S7's make money for them
A "convenient" burn-in test
September 29, 2015, 10:49:20 PM
You know, if Bitmain is going to sit on our orders for a week during their national holiday, the least they could do is give us a web cam in their data center so we can watch our S7's make money for them
September 29, 2015, 10:45:22 PM
That 1500w comment from bitmain is total bullshit. I'm certain -droid-'s test tomorrow will prove that.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
Has anyone with a sagging miner checked the output on their psu for any variance on the 12v rail with a multimeter?
Does anyone have some that are working and others that are not to compare the rev on the hashing board?
Ufo
Would be nice if a few people could verify that. Please make sure you take the 12v reading on a PCI-e connector that is plugged to a blade and not an unused wire as there is a voltage drop that needs to be accounted for across the loaded wires.
September 29, 2015, 10:25:00 PM
That 1500w comment from bitmain is total bullshit. I'm certain -droid-'s test tomorrow will prove that.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
Has anyone with a sagging miner checked the output on their psu for any variance on the 12v rail with a multimeter?
Does anyone have some that are working and others that are not to compare the rev on the hashing board?
Ufo
September 29, 2015, 10:18:02 PM
That 1500w comment from bitmain is total bullshit. I'm certain -droid-'s test tomorrow will prove that.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
Maybe these sub par machines have lower quality chips? Perhaps a slight voltage bump would help some of the ones running on ATX PSUs which supply as close to 12v as possible? Most server PSUs including the ever popular 2880w and DPS2000 will put out 12.2v.
September 29, 2015, 10:14:41 PM
1 of my S7s is exactly the same. I was getting 4300th/s at 600. I dropped it to 593 and got about 4400. As a test I have been running at 575 and it is a steady 4400th/s.
What PSU are you using? Are you using the server one bitmain sells, APW3-12-1600-B2?
September 29, 2015, 09:51:59 PM
my miner came quickly, but it runs slowly.
i can not get past 4600 and errors are .1100
i can not get past 4600 and errors are .1100
sorry to hear this, but I thought that you got a couple.
is second similar to the first one?
yes both were able to do 4750, 4775 = 4625 but after a week burn in they sagged a bit.
I find that freq at 500 is better so frankly 500/600 = 83.3 % this is 16.7% short .
I paid about 7.8btc so I should get about 1.3 btc refund.
Now they will argue that at 600 I get say 4400 th which is 91.67% or about 8.33% short which is about a .65 btc refund.
1 of my S7s is exactly the same. I was getting 4300th/s at 600. I dropped it to 593 and got about 4400. As a test I have been running at 575 and it is a steady 4400th/s.
September 29, 2015, 09:50:15 PM
Anyone running the APW3-12-1600-B2 server psu with the s7 that is experiencing the sagging? Those are the ones I bought for my s7's since this was a new build.
September 29, 2015, 09:48:31 PM
my miner came quickly, but it runs slowly.
i can not get past 4600 and errors are .1100
i can not get past 4600 and errors are .1100
Sound pretty subpar, like the first batch of S3's iirc and not all that different with S5 batch one? Maybe you're illegible for a partial refund? Kinda disappointing that Bitmain keep repeating the same error with First batches.
Like this S5 Batch one?.....
Jump to: