ViaBTC has historically been a classic 4% PPS pool.  They dropped that to 2% for a few months after an issue as a "sorry" response.  If they are down to 1% now its likely because they are keeping a part of transaction fees.  

3.5% for BTC.COM as a PPS pool sharing full TX fees is not a bad deal - IF your need PPS's lower volatility.  

However, unless your a smaller miner (say a S9 or two or less), PPS will cost you more over PPLNS.  Its just a question of risk management (aka variance acceptance).  Slushpool at 2% is large enough you wouldn't feel that variance if measured in terms of months not days.  Kano at 0.9% is good if measured over quarters not weeks.  Ckpool is smaller than Kano, so has even larger variance, but has grown enough recently that your very likely going to find several blocks per difficulty period - which is big enough for those with a very long time horizon.

For those smaller miners "dust" becomes an issue - where your reward is less than the fee associated with sending it to you.  But even there, just select a pool like Slushpool that allows you to set a payout thresehold so your dust can accumulate to something reasonable before payout.

I concur.  Its working at spec speeds (presuming you bought a 13TH unit not a faster one).  Chances are the sensor is bad and reporting false information.  If it was really that hot it should have shut itself down.

Mine on, or open a ticket with Bitmain.

No real difference besides a day or three to convert BTC to USD and get it deposited into my bank.  Unless, of course, BTC had value independant of fiat.

My last months electric bill came in at $14,072.82, and that was with a temporary discount of $1,584 the co-op passed onto me.

Or perhaps consider that Kano has more experience with -ck code than you are, and has vastly more experience creating and running a fair pool.  Maybe, just maybe, consider that he is right and your too inexperienced to know it.

The problem goes back to the early days of the Internet when SMTP (Simple Mail Transfer Protocol) was defined.  The "to:" and "from:" fields are actually data in a mail message, not handshake stuff.  I use to fry peoples minds that didn't believe me by sending them e-mails that appeared to be from "Your Mother" or "GOD".  Thus its absolutely trivial to fake where an e-mail comes from as displayed by most mail programs.

Look at the example at https://en.wikipedia.org/wiki/Simple_Mail_Transfer_Protocol  and notice both the protocol "RCPT TO:" and "MAIL FROM:" that most mail checkers validate against for blantant fraud.  Then go down a bit and look at the message body, the stuff after the "DATA" protocol command.  The stuff there is what most mail programs display and its just data, you could put anything you like in there, including "techsupport@bitmain.com" or whatever.

I would be curious if someone that received the e-mail would look at the headers (an option turned off in most mail programs but that can be turned on) and see what the sending "MAIL FROM:" was?

So far, everything released that I'm aware of is ASICBoost Only.  Technically it would be possible to design a system which could go either way, but turning it off would hurt performance... (and thus advertising/sales impact)

S3 478GH, 28nm silicon, 366W = 1.306GH/W at the chip = 0.766 W/GH
S4 2000GH, 28nm silicon, 1450W =  1.279GH/W "at the wall" (maybe) ~= 0.725 W/Gh
S5 1155GH, 28nm silicon, 590W = 1.958GH/W = 0.511 W/GH (just goes to show what can be down within a chip 'size')
S7 4500GH, 16nm silicon, 1293W = 3.480GH/W at the chip = 0.287 W/GH - a 1.78X factor in power consumption improvement for a 1.75X decrease in die size - pretty linear.
S9 13500GH, 16nm silicon, 1323W = 10.20GH/W at the chip = 0.098 W/GH - a 2.93X improvement in power consumption via better chip design vs. the S7.

So I'm guessing a shinkage of die size can be estimated to make linear improvement in power consumption, perhaps.

Here is hoping this doesn't get deleted for being "off topic" in a speculation forum...

But many mid-size farm, like mine, have pre-wired expecting 20 amp 220V to be split between 2 devices, not one. 

Ah, hadn't spotted that thread - sorry.

No, I haven't changed them the wall warts - using the ones that came with the Pi 3s, all of which are rated at 2.5A.  Only running (3) AUCs, not (4) per controller as well.

Thanks for explaining your logic.  Alas, heat was not the issue - this happened on one of our cooler days.

Found it, compliments of Blokforge support asking me for screenshots...

My Avalons use Static IP addressing for the controllers.  My S9s use DHCP.  Ends up the DNS server tha Avalons were pointed too was having problems.

Woke up this morning to discover ALL my Avalons, both 821s and 841s, offline.  S9s on the same switch to the same pool are working fine.  These Avalons are spread between 3 controllers and 6 AUCs.  All have yellow lights on them.  I've tried rebooting the Pis several times - both by power cycling and soft reboots.

The cgminer API log is showing (with trivial variations):

[Firmware Version] => Avalon Firmware - 20180305
    luci: 62d814c
    cgminer: b5b497e
    cgminer-packages: 960e108
Socket connect failed => Connection refused

These machines have been working well for quite some time.

Any ideas?

Given that others report no problems running thousands of Avalons with full 20 member configurations, I'm curious what the basis of your recommendation is?  Also -2 is suppose to be less stable than -1 per Canaan, so am again confused.  Please explain.

Depends on the cost of your electricity.

Here in the mid-west USA I've had decent luck with Fed-Ex.  Their custom charges (for packages over $2500) are about half of what UPS rips me off for.

I believe that is incorrect.  Yes, setting a strong password would slow down web access to the miners, but the API would be broken into within minutes since the API userid/password are hard coded in the firmware.

Yeah, a single S9 should be fine... but not a S7 - at least per the previous posts on the subject.

I always pay with Bitcoin.  Fast, easy, and I use coins I mined with.  Twas a bit of a pain when they only accepted Bitcoin Cash, but now things are back to normal.  Well.  Kind of.  I've stopped buying S9s in favor of Avalons.  Now the trick to only do that when Avalon pricing is compariable.  Seems those windows are pretty short.

Highly suggest buying unistrut and just designing your own.  Home Depot has a similar brand for about $16 for a 10' length.  You can use normal 1/2" bolts, lock washers, and nuts and not have to buy their rip-off priced ones.    That is what I did and ended up with (4) 10' racks, (5) shelves each, spaced ideally for miners, that you could stand on it you wanted too.  Price wise, it was similar if not cheaper than anything else I found when you figured the cost/foot out.