Pages:
Author

Topic: Bitmain introduces the Antminer S17 Pro, Antminer S17, and the Antminer T17 - page 17. (Read 11290 times)

member
Activity: 181
Merit: 53
You should definitely be running them on Turbo.

As long as your environment is a mining conducive one and there is no failure within the first few weeks you'll be fine with these long term.
newbie
Activity: 13
Merit: 7
UPDATE: I remembered I had a dual metered PDU that reports amps (in whole numbers) that I don't use that readout anymore.  So I tried one of the S17s on that so I could measure each cord separately.  It measured exactly 5 amps each on Turbo mode.  Exactly being that it only reports round numbers.  At my voltage that is around 1150 watts per cord.  But I'm sure it is 5.x not just 5 amps.  But either way, it seems to balance the load very close between the two cords.

I tested my 53TH/s pro on two separate banks on my PDU. Like yours my PDU only shows whole numbers and rounds up or down sometimes depending on the threshold. Together on the same bank both cords pull 11 amps on my 208v electric. When I plug each plug into their own bank my PDU’s show 5 amp draw per plug. I’m estimating it’s probably more around 5.5 amps so on 208v that would mean 1144 watts per plug. This is on normal mode so it looks like the PSU’s split the load evenly across the two plugs.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
Compelling argument. Thanks for elaborating! Smiley

It gets more complex when you are at 100 units. That is 66 dollars a day.

You could risk running at turbo until one breaks.

35 x 66 is about 2200. So every 35 days that extra profit pays for 1 unit.

So if 1 or 0 units break in 35 days it is a win or a push.

If  0 or 1 units break in 70 days it is a win.

This part of the you have to have money to make money.

100 units is about 200k worth of miners.

I know of a few people with 20 of these. They do normal speed with 4 cent power.

Down the road they will bump to turbo.  Since 20 earn them about 160 a day.  In 30 days they will earn 4800 that means 2 of the 20 are paid off.  They will know of the risks of turbo and will have at least paid off 2 units.
jr. member
Activity: 33
Merit: 11
So your example clearly shows 66 cent benefit on turbo  which means  it needs to run for 3300 days (3300 x .66 = 2178.00)  without breaking to reach a clear benefit for 1 unit.

You would risk running new unknown gear at this moment with zero knowledge of durability hoping that the 66 cent day profit will make it worth your while.

"at the moment with current coin prices" is the key  point of my statement.

so in a month if no one has broken gear on turbo you have a base line.  And you would lose about 20 bucks waiting for others to test the gear out.

or if prices spike hard and you make 32 a day vs 28 a day  so turbo gains 4 dollars a day .

but you could be correct so you earn the 66 cents and never break the gear you win the bet. you earn more your way.

my guess is run it on normal for a month and re assess what to do.

Compelling argument. Thanks for elaborating! Smiley
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
I'd like to see your calculation/thinking.  

By my profit model turbo mode makes sense currently.  But I am working with 5.4 cent electricity so that obviously will make a difference.  Simple Cryptocompare shows $8.50/day for Turbo, $7.84/day for Normal, $6.24/day for Low Power mode.  So Turbo currently makes the most profit.  I have wondered if the extra wear and tear is worth it and for that extra 8% profit and maybe not.  

As energy price goes up then it would flip to where normal would win out and then low power mode.

So your example clearly shows 66 cent benefit on turbo  which means  it needs to run for 3300 days (3300 x .66 = 2178.00)  without breaking to reach a clear benefit for 1 unit.

You would risk running new unknown gear at this moment with zero knowledge of durability hoping that the 66 cent day profit will make it worth your while.

"at the moment with current coin prices" is the key  point of my statement.

so in a month if no one has broken gear on turbo you have a base line.  And you would lose about 20 bucks waiting for others to test the gear out.

or if prices spike hard and you make 32 a day vs 28 a day  so turbo gains 4 dollars a day .

but you could be correct so you earn the 66 cents and never break the gear you win the bet. you earn more your way.

my guess is run it on normal for a month and re assess what to do.
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
... As energy price goes up then it would flip to where normal would win out and then low power mode.

You have to pay attention to Difficulty and BTC price long before you have to worry about energy price. You wont find your power bill changing every day (and no one pays it every day).

You will find Difficulty changing every 2 weeks and BTC price changing every day.
jr. member
Activity: 33
Merit: 11
I have metered smart switches for each circuit-PDU so the only way I could measure each cord would be to connect to two separate circuits which I'd rather not risk. Otherwise I'd need to measure with my amp-meter with a split cord which I don't have. If I can figure out how to safely do it I'll let you know.

UPDATE: I remembered I had a dual metered PDU that reports amps (in whole numbers) that I don't use that readout anymore.  So I tried one of the S17s on that so I could measure each cord separately.  It measured exactly 5 amps each on Turbo mode.  Exactly being that it only reports round numbers.  At my voltage that is around 1150 watts per cord.  But I'm sure it is 5.x not just 5 amps.  But either way, it seems to balance the load very close between the two cords.

... Turbo should never be used at the moment with current coin prices.

Normal speed should be used at the moment with current coin prices.

But I am sure some can create an example that shows turbo is smart to do at the moment.

I'd like to see your calculation/thinking. 

By my profit model turbo mode makes sense currently.  But I am working with 5.4 cent electricity so that obviously will make a difference.  Simple Cryptocompare shows $8.50/day for Turbo, $7.84/day for Normal, $6.24/day for Low Power mode.  So Turbo currently makes the most profit.  I have wondered if the extra wear and tear is worth it and for that extra 8% profit and maybe not. 

As energy price goes up then it would flip to where normal would win out and then low power mode.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
It is supposed to be based on the BM1397, correct?

https://chainbulletin.com/bitmain-announces-its-new-and-improved-7nm-mining-chip/

55J/TH seems low compared to what was initially advertised (30J/TH). Still incredibly impressive, and blowing everything else out of the water. But I wonder where this huge discrepancy is coming from.

you don't understand how they measure.

a single chip with test leads  can do 30J/TH

this gear is doing 50th at 2050 watts on = normal speed  that is 41J/TH

you are doing turbo numbers  to mine on turbo speeds gives you  a little more hash and burns power like mad.

so this gear does 41J/TH using the  fans and controller  on normal speed.  I suspect fans and controller are at least 100 watts

so it is  doing about 1950/50th = 39J/TH on normal speed

In what seems like a million years ago bit fury shows its chip in a single chip setup  it did 60J/TH  which when built into a  big machine did about 100J/TH

Turbo should never be used at the moment with current coin prices.

Normal speed should be used at the moment with current coin prices.

But I am sure some can create an example that shows turbo is smart to do at the moment.
member
Activity: 67
Merit: 31
Okay i found my answer by searching out new videos uploaded in the last few days

S17 Pro in Normal mode: power meter shows 1109 watts pulled from 1 plug, if it is in fact 50/50 that’s 2218ish watts so let’s say 42watt/T

Video below :



Antminer S17 Pro instruction and review
jr. member
Activity: 33
Merit: 11
it’s not definitive until someone tests it, depending on the amount of hash boards and the controller board it’s possible one plug pulls more than the other, bitmain isn’t exactly the best designers.

I have metered smart switches for each circuit-PDU so the only way I could measure each cord would be to connect to two separate circuits which I'd rather not risk. Otherwise I'd need to measure with my amp-meter with a split cord which I don't have. If I can figure out how to safely do it I'll let you know.
member
Activity: 238
Merit: 10
It is supposed to be based on the BM1397, correct?

https://chainbulletin.com/bitmain-announces-its-new-and-improved-7nm-mining-chip/

55J/TH seems low compared to what was initially advertised (30J/TH). Still incredibly impressive, and blowing everything else out of the water. But I wonder where this huge discrepancy is coming from.
newbie
Activity: 13
Merit: 7
could you test how much power each cord is pulling on normal and turbo mode?

I’ve heard from someone who’s contacted Bitmain about this that the PSU splits the load equally between the two cords. So it should be, based on his reported numbers, 1025 watts per cord in normal mode and 1250 watts in Turbo.

I have the 53 TH/s model and my metered PDU’s, which only measure in amps, are showing ~11 amps total in normal mode on 208v power at 53-54TH. They also only show whole numbers and sometimes round up, so it’s not super accurate. That would put me at around 2288 watts or 1144 watts per cord.
full member
Activity: 1022
Merit: 221
We are not retail.
People ask us for s9's all the time. We just started our escrow service as well. T17 yea looks like s15 for sure. Are you talking s11 or T15? either way Bitmain confirmed the 17 series is all they are producing for the moment.
legendary
Activity: 2254
Merit: 2419
EIN: 82-3893490
Next batch is back in stock, shipping in a week.

T17 is also out, 40th/s, 2200w, 55j/th. I think it uses S15 chips, just more of it.

https://shop.bitmain.com/product/detail?pid=00020190426171840402gq98giLA0699

i can dig that price on the T17 - still wish I could just get a bunch of T11's for cheap though - no tariff on those so quicker ROI for me. Now, if Kano pool could hit a few blocks before they go out of stock or if I could get someone to buy my S9's for 150-200 each I would be able to get some new mining gear.
sr. member
Activity: 610
Merit: 265
Next batch is back in stock, shipping in a week.

T17 is also out, 40th/s, 2200w, 55j/th. I think it uses S15 chips, just more of it.

https://shop.bitmain.com/product/detail?pid=00020190426171840402gq98giLA0699
member
Activity: 67
Merit: 31
Got my first S17 Pro (50TH model) up and running.  As smooth of a startup as typical with Antminers.  Just plugged in both power cords and viola.

I've only tried "normal" mode so far and I'm getting 51 TH on average and about 2050 Watts measured with my smart switch.  So right at 40 J/TH efficiency.  The noise seems similar to other antminers.  Not noticeably louder to me but then again I have 40 antminers running in my garage so I probably wouldn't notice.  And I haven't tried turbo mode.

Edit1: Turbo mode results: 57 TH, 2500 Watts - 43.8 J/TH.  Pretty impressive! And not really any louder that I can tell.

Edit2: Low power mode: 40 TH, 1540 Watts - 38.5 J/TH.

Below is a pic of it and the dashboard.  I'll be trying the other modes when I get a chance.

https://imgur.com/k6aMzDf
https://imgur.com/I2fRpwD

could you test how much power each cord is pulling on normal and turbo mode?
legendary
Activity: 2254
Merit: 2419
EIN: 82-3893490
that is also why i put switches on each circuit - I like to turn them on one at time, instead of all at once.
full member
Activity: 294
Merit: 100
What happens if you plug them with the wires energized? When you plug the first and then the second.. I don't think this should be done, but given how some people operate their mines...

Next experiment; Would it turn on if you used two 120v from different phases (would it get the 208v from the two hots?). Theoretically speaking, of course, don't try this at home, in the name of science.

Yea i want to know, since it has two plugs, what happen if you plug directly one cable first to power point (the psu wont turn on?) and then the second cable (then the psu turn on?), so no need PDU switches.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
What happens if you plug them with the wires energized? When you plug the first and then the second.. I don't think this should be done, but given how some people operate their mines...

Next experiment; Would it turn on if you used two 120v from different phases (would it get the 208v from the two hots?). Theoretically speaking, of course, don't try this at home, in the name of science.

I had an Italian stereo amp it could do that use two 120 volt circuits from separate legs of a USA circuit box ..

Back in the 90s maybe I could find a link to either speaker builder or stereophile.
Pages:
Jump to: