Topic: Bitmain Announces the Antminer R4 and APW5 Power Supply, Designed for Silence - page 40. (Read 84827 times)

Would be nice to figure it out, as I said I am not quite sure if it is diverted hash or a dying board or something in the middle of that.

re Philipma: I've never had any hints of your issue at the farm. Ja your setup works but I'm talking about getting PROOF of skimming. As in just where that boards hashing is going if not to your assigned pool where the other boards are pointed.

Get a cheap  router DSL

Do this modem ----- original router ------ cheap router.   The phantom s9s with a different password like rootfuckyouhacker.

If all boards pop back up I would be suspecting some diversion.

Although it could be that is the first sign of a board dying and the board will die in a week or so..

I am pretty much ready to go a bit nuts over this gear.

But I now have a second dead board and it came from the mystery drop out miner.

There is one more miner that was dropping a powered board it works now but my gut tells me it will die and I will end up with three dead boards out of six miners.


philipma1957.4           9.227k   3s   15,511   138,573,887   11.92THs   4m 36s   0.340%   0.026%   11.24THs-----down clock ---Stable
philipma1957.two2   9.499k   10s   14,150   137,968,161   10.64THs   4m 36s   0.164%   0.026%   11.21THs-----down clock ---Stable

philipma1957.6           7.911k   1s   12,721   112,717,709   8.74THs   4m 36s   0.192%   0.021%   10.08THs-----down clock----shaky
philipma1957.1           6.442k   3s   14,932   96,191,944   8.22THs   4m 36s   0.107%   0.018%    7.58THs------down clock----dead board




philipma1957.one   5.622k   4hrs   11,358   60,543,030   0GHs   4m 36s   0.142%   0.012%   0GHs ____off line testing pretty sure dead board


pattern seems to be mystery hashing  all board used then a vanished board
can be restored behind second router
but now  one seems dead
the other shaky.

and of course the one bad one  never did  this it just went xxxxx  then 0 then zip powerless

My guess is there are stages of death for the boards not a mystery hacker.

Oh no reply on second board as to where to send it.

and no info on board one which they have had since weds.

I am not a huge player  but 85,000usd  in direct business since December 2013 should be worth something.

I never thought I would be forced back to gpu mining but I now run 20 gpus and 470 mh  mostly due to the warranty service of bitmaintech.

What kills me is I want an r4 but 1400 and 90 days for  a warranty is no deal for me.

Yeah, a wider node is better for stability since differences in voltage based on difference in current draw through the chips is averaged out. Worst part of a string is the weakest link issue where the worst-performing chip sets the voltage for the whole thing. That's why there's so much bottom-end performance variance in undervolted boards.

Which, one of these days I need to get ahold of some S9 enough to work on a good snap-in undervolt solution. I suspect it'd also work on this R4 just as well.

Dinna know that. Good to learn something new
Makes sense in that since Vcore si divided between all the 3-chip nodes any differences in voltage between nodes is averaged out nicely. Guess I never just did the maths: Assuming a Vcore  supply of 10v/21=476mv per node. Right in line with expected chip Vcore range...

Not really, no. Each board has 21 nodes of 3 chips. Comms go from chip to chip at the same node, then from the third chip of one node to the first chip of the node above in a continuous data path of all 63 chips. You take out one chip, you take out the whole board. On S1 and S3 you could take out one chip and all chips before it would keep hashing, but you know if you interrupt the current in a string topology the whole thing shuts down.

That's what I wanted to know. Only whole board is tasked. I'm on Skype with Opieum2 and the CompSi Master on the issue of some folks having -- odd -- hashrate dropouts eg Phillpma. As in a whole board seeming to be feeding a shadow pool.

That aside, I though electrically the chips were powered in several paralleled banks of series strings. No?

Not sure what you mean. The whole card is an individual string, and each card is individually addressed at the control board.

Why ? Even with a good warranty, you are just better off buying and holding Bitcoin.

This product is a loser on every conceivable level for the consumer. Being in the R* product line, this thing can't even act as a router...

I mean, for fucks sake, they are still using individual chip heat-sinks. Their best products were always the ones with the solid block of metal that covered the entire board.

All in the name of gross profits.

re: cost of milling. Um, CNC mills... Almost dirt cheap in China to buy and operate.

How about IO power on board for coms? Query: Can individual strings be addressed or does it have to be the whole card?

Other reasons include cooling both sides without having to mill around components, and cooling both sides without risking shorting across the different node-level ground potentials.

Both the dead boards in my care have string power.

Can't blame ya  for being hesitant and doing more GPU stuff. May consider it myself. Knock on wood, so far all my s9's are running beautifully though I dropped the B1 from 650mHz to 606 to keep chip temp around 95C.

As for the chips being more unstable/fragile....
(cough) been bringing up that point since BitFury's smoke about their foray into 16nm...

As fof Bob's comment about a mono sink on top: A single extrusion milled flat on the bottom then relief milled for any clearance issues has to be cheaper than all those lil' sinks and the labor involved with them. The only reason I can see why BM didn't go that route is concern about spots with different amount of thermal expansion between the PCB/chips/topsink could stress the solder connections. For that -- don't use thermal epoxy. Use a soft non-hardening conformal thermal putty. And NOT paste. It hardens under the heat.

OT but query: Do the strings still have power on your dead boards?

I didn't buy one, because I doesn't make any sense when it comes to ROI. I just tried to put it to the shopping cart and see the total price.

APW5 in the top of the page they write 5KG
and in the Parameters tables 2.5KG

what is the correct weight?

I was ready to order a unit at $999 + shipping.  At this price I will continue waiting for a purchase that I can pretend has the potential to make sense.

You say that like its some new phenomenon.   Supply and demand it literally is as simple as that...

I would pay them the 1400 if they had a good warranty.

But based on my 6 s-9's   having now 2 dead boards  an 11% fail rate on in 80 days the other in one day I simply can't fork out the 1400.

These are smaller chips then ever and I am beginning to think they will have a very high fail rate.

Wow bitmain. Almost 400 days to break even on this sucker with 0.10 USD power cost. Going to pass on this one.

You guys actually bought this thing at $1400???  No wonder Bitmain can charge what ever they like...miners will buy anything from them no matter what the cost.

@Bitmain, change the price of the R4 to $4000...you'll still be able to sell them like hotcakes!

Hmmm, well that is true. I tried now as well and I get $0 for shipping with all logistics companies.