Topic: Bitmain will be releasing S15 and T15 on 8/11/2018 - page 7. (Read 31485 times)

lets say  you do a sub net.

your main router is say  192.168.0.1-256

add a cheap  router  that is 192.168.1.1-256

so modem >>>> main router >>>>> cheap subnet router >>>>> attach 1 bad miner attach 1 pc  search with advanced ip or angry ip   see if it shows.

I had issues with early model s-9's  sound like this issue.

So  2 dead psu's
and 3 weird ip's out of 20 units correct?

To late now but the s15's in your case should have  been m10's

My 4 m10's = flawless  they do 59-60 watts  a th at  low setting about 20.5th

are you usa based?

Definitely have normal connectivity lights. Machines boot as normal, all lights normal, but cycle due to not having a network connection. When we plug them into the connections of working machines, they do not come up. All the miners on the network are static IP, but we have a DHCP range set aside on the router for new machines. We boot them DHCP and then assign static IP's.  Power cycling and resetting the machines does not help. Switching PDU, power cables, and NIC cables doesn't help. Units do not show up in our routers control panel of connected devices. I thought perhaps they had shipped us units that had already been assigned static IP's, but a factory reset would fix that. I though maybe the NIC's were being finicky with our gigabit switches, so I put them on a 10/100 switch and also plugged them directly into the router. No joy. Not sure what the problem is, but if the controllers aren't bad we'll figure it out.

Are you getting both blinky lights from the Ethernet ports?

One thing, which I assumed was because I swapped an Ethernet cable from an A4+ to the T15, is that when I powered up the T15 is it stole the IP address from the A4+. Given this may not be an isolated incident caused by my switch, are you using static IP’s or dhcp? If DHCP, you may want to move all the existing miners to static IP’s and see if the missing miners start showing up.

Our s15 adventure continues. This morning we shipped the (2) S15's that failed on 240v to the CA repair depot. This afternoon 10 more S15's arrived from BitMain. We put them on the 208v rack at the office. Of the 10, 7 work perfectly and 3 will not show up on our network despite all efforts and troubleshooting. Can't figure it out. So now I have 13 S15's running, 2 going back for PSU, and 3 that boot up but won't communicate with the network. I wish I could get ahold of a couple of spare PSU's and controllers for these guys. They aren't selling S15 parts yet, are they?

S15 - 28TH
220V / Ambient 17 degrees C

3 Days on Nicehash, so far so good.
https://i.imgur.com/TmyRbuK.jpg

Power Supply fans ruined sound improvement from lower miner fans speed.
Very high pitch noise.

thanks sarcheer...seems you could be right...ie 16nm chip

I was under the impression it was as well, however I cannot find any info on the Bitmain website which lends any fact to that impression. The blog/PR post from BM about miners featuring the new 7nm chipset only mentions the S15 and the T15. The efficiencies of the S15 and T15 are a significant jump from the S11. The S11 has 252 chips on 3 boards while the S15 and T15 have 288 on 4 boards and 180 on 3, respectively.

The above factors lead me to believe that the S11 is in fact using 16nm chips and was probably their "improved" S9 they intended to release had they not been experiencing such financial difficulty.

Edit: Link to blog post mentioned above:

https://blog.bitmain.com/en/the-7nm-antminer-s15-and-t15-are-now-available-for-purchase/

know this thread is relevant to S15/T15...but isnt the S11 based on newer chipset(7nm) as well?

I'm willing to guess their PSUs are optimal for chinese mains, which is 220v@50hz...

well so far one guy with 5 units

running 3 at 208 volt input  high speed = good to go

running 2 at 240.8 volt input high speed = dead psu's

not enough to say  gear is fucked up

but enough for all owners to watch units carefully  and make notes of the input voltage.

another guy  with 4 units all four at high speed.
2 run hot and shut down
2 run okay and don't shut down

It is new tech and bitmain has had new tech struggle during early batches.

fanatic26 usually has a lot of gear running.

maybe he can post  his s15 numbers  if he has them setup.

it's quite the opposite, high speed require more current therefore more resistance and less voltage.

the more "power" your miner "takes" the more drop in voltage you get

in more details.

assuming your voltage reading "unloaded" 240v.

if your miner needs more power , it asks for more current from the source (volt) , the more current that moves thought the wires, the more heat it generates, the more heat the more the resistance, the higher the resistance the less the volt.

so your miner needs 5 amps 240v might drop to only 239.
but if it needs 7 amps 240v might drop to 238.
if it needs only 1amp it might stay at 240.

all the above numbers are nothing even close to reality, but just an explanation, you could have your miner at full load and it may still fail to generate enough resistance that is capable of dropping  the voltage by anything measurable.


low voltage can damage your PSU , the psu has an ideal rated power of Volts * Amps it needs to keep it at x value
if the voltage drops it will try to cover up by drawing more amps (current) , and as explained above this will cause more heat , most of the times more heat than what your PSU can handle and that = psu damage.

but usually those psu come with voltage protection, which indeed is not perfect, but i am pretty certain it can handle at small drops and spikes.

Well I have gone as 247 very rarely. So if 240+ is an issue on both low setting and high setting I can’t use these that easily.  Lots of USA buyers will be complaining since lots of us go from 237 to 245.

Maybe it was just bad luck.  Years back the s7ln would have issues.  And the s4 or s2 had psu issues.

This will need to be watched for carefully.

If your voltage is too high I don’t think running at LPM would have made a difference, especially if these PSU’s are as sensitive as they are appearing to be. What’s strange is the previous APWs were advertised as having “overvoltage protection.” Where was that for your units? Really hope Bitmain didn’t scrimp on these PSU’s.

Link to info below in case you’re interested in removing your APW8 and giving it an inspection:

https://support.bitmain.com/hc/en-us/articles/360013898033-How-to-detach-APW8-PSU-from-DR5-S11-S15-T15

Yeah this likely it a spike above 243 is common in my area.  Even 245 can happen.

I almost never drop below 237.

And of course my 120 is often 121 or 122 and I almost never drop to 118.

I am guessing full power at 243 volts + equals psu death

We will need more confirmation of this to be sure.

We need a spare psu or 2 to test top limits.

This could be a real issue to USA miners as many of us get 238 to 245 for our 240 lines.

My new buildout with buysolar will be 40kwatt constant 240 but it could flux to 245.  This means no s15 or t15 on the 240 lines.

Although this guys warehouse has.

208
240
480

So if we do some s15 we will do it on the 208 . If someone can get spare psus we can test at 245 and see if there is a burn out at low power setting on the miner.

No, nothing. Units are totally dead. No Hash boards, fans, or lights on the controller. Re-power produces nothing. Leads me to believe that the PSU's failed due to the voltage. Really want to open one up and do some testing.

philipma1957 is probably right, the voltage from my 240 outlets here at the house probably varies by couple of volts up and down. When I test them I was reading 240.8v. Most appliances designed for 240v power have an outside operating range of 250v due to variance. Not these guys, apparently. Oh, was definitely running them full power mode. They were running about 28.4 when they were up. I should have switched the other unit to low power mode when the first one failed, but I wasn't thinking it was a voltage issue until they both failed.

Multimeter tells me my PDU is distributing 200v regardless of load, my panel is labeled 208v however. I haven’t had power problems (no dead psu’s) in the year I’ve had this panel under 24/7 load. Temps ranged from 65 to 95 over that time as ventilation and A/C changed.

I’ve only run my T15 in LPM for 3-4 hours since I received it on Tuesday and I’ve had no PSU problems. So, based on available info, we can preliminarily say that the APW8 tolerates the low end of its rating better than the high end. It will be interesting to see this range better defined as more units enter the field.

So my stats are 23.35 TH/s at 200v, 7.6 Amps

Please  let us know

high speed hash
low speed  hash
high speed watts
low speed watts
high speed volts in
low speed volts in

all temps of all boards on all settings

TIA

phil

I have ordered one S15 from Bitmain and I will do a review when I receive it.

So on your power meters you read 200 ?

I am asking since you could drop under 200 if you normally read 200.

The gear says 200-240.   So if 243 kills them. What would 195 do?

My guess is a shut down.
My next guess is   Low setting will be fine for s15 or t15.  At 190 to 250.

But. High speed you will need to be say 201 to 238.

I would ask for those with the gear to read volts suppiled and speed you run at.
Maybe we can get an idea if this is true.

I had asked blokforge to sell me one via cc.  He missed my pm and when he contacted me I had purchased 8 k in other gear.  So I will be needing to mine longer before I get one of these. As I will no longer purchase direct from bitmain due to trump tax.

I may get an s15 from a private sale in feb. Till then I won’t be testing or showing them.

It is important to see if the psu really needs to stay inside 200-240 to run high speed safely .

So those that have them please give us info.

Do the fans on the PSU spin when you try to re-power it up? More wondering if it is giving any indications that it is responding to input power.

The PSU on my T15 is not hot to the touch in normal mode in a ~75ºF room. I will check and confirm come Monday that this is still the case.