|
The answer is no one, apparently. I couldn't find anyone in the list.
|
|
|
|
Thanks again both of you for helping with what most would assume was very elementary, and making me feel welcome instead of stupid. Much appreciated.
No problem, good luck. |
|
|
|
Reason why you need so many mid-priced PSUs is because each blade requires 3 PCIe connections and you CANT intermix them between boards, which mean an CX750W can only power 1 board.
A CX750* can power two hashing boards. |
|
|
|
Bitcoin is too important a technology to be rotted away by corporate interests.
Interestingly enough you don't have much problem with the corporate interests when they align with your own in the form of income. If you want to continue to discuss that you can do so in meta or a scam accusation. It doesn't matter who pays or doesn't pay me, I get accused of being biased in both and all directions at once which doesn't make any sense. Even now its being claimed that I am somehow biased in this situation when: 1. The claimant provided a picture of a mixture of nickel plated and non nickel plated heatsinks claiming that its not actually nickel plating but normal heatsinks from a charred and burnt miner Bitmain sent him. 2. A claim has been made that when he cleaned the heatsinks they changed color to match the others 3. The claimant made these claims during a period he was struggling to get Bitmain to agree to a warranty return. 4. The claimant posted these claims in every Bitmain thread in an effort to force Bitmain to accept his return in response to his silence. 5. The claimant now refuses to provide any proof what so ever to backup the claims, even with continued involvement in the thread, a camera and heatsinks which would "prove" the claims. That's some Katete level logic going on there if one is to conclude that is a bias.... |
|
|
|
Does anybody know approx how loud this will be compared to an Antminer S3?
S5/S7 loud at stock configuration, its an identically specified fan. Some notes from the wiki by the way, looks like we can't power fans from the PSU. *NOTICE* If the fan is broken or the fan's RPM cannot be measure,
the device will not start.
Controller:
Raspberry Pi (version B or B+)
Raspberrt Pi 2 is not supported yet.
Each RPi can control approximate 60 devices. |
|
|
|
The entire back panel of lights on one of my Avalon 4.1 is all lite up and blinking in unison. ERR, WARN, TXD etc. solid and pulsing in unison. I took the unit apart and nothing appears out of the ordinary. Any recommendations on trouble shooting?
I'd start with power. That arrangement isn't on the wiki so it might be continually resetting over and over again. |
|
|
|
Post behind someone trying to make them look unintelligent without any reasons at all. If Tupsu has been guilty of anything it has been due to a language issue
He made a ridiculous and damaging claim about a product which he is now unable to back up. I guess you two aren't used to being called out People like Tupsu are valuable to our community. He actually does things with equipment and posts about it.
Yes, by posting disparaging statements in every thread in order to 'expedite' his warranty claims. Maybe Tupsu should regurgitate the manufacturer specifications and clutter the forum with such.
He is more than welcome to do so, as are you and anyone else. In closing I would have to say for the people who love to run in riding the dog. Simply read back and see what he has done in this thread, and then, check out a couple of these.
As with those threads ( which were run by one person with 18 account accounts), I don't back down to shills and alts. Whether you think you're shilling for a pool, for an exchange, for Bitmain, Spondoolies *cough*, for fun or otherwise, I will call you out every single time. Bitcoin is too important a technology to be rotted away by corporate interests. |
|
|
|
Clean one then. If you're honestly trying to claim they're that color because they're dirty, clean one and show us.
I cleaned them all, not just one. Already when I got back to the rest of the two blades from repair. So you can show us how they changed color and the nickel plating magically disappeared, right?? But I'm not trying to prove anything.
If you're not willing to back up your serious allegations then don't make them. |
|
|
|
Still nothing useful as to proving a board was "burnt with flame and smoke". If you're saying its due to the smell then its probably from someone unglueing heatsinks off a dead board with a torch. If you're saying that the dark ones are a different color because they've been in a fire, that's even more ridiculous - its nickel plating.
The dark one are different color , because they are not cleaned and sooty. At this point I have to assume you're trolling or shilling, because you're a grown ass man with a technical profession claiming that the nickel coated heatsinks are actually "covered in soot". Nickel coated heatsinks are on the other side of the blade. Clean one then. If you're honestly trying to claim they're that color because they're dirty, clean one and show us. |
|
|
|
Run your own p2pool node - super easy & you'll help decentralize the network  P2pool isn't his own mining pool though. |
|
|
|
Still nothing useful as to proving a board was "burnt with flame and smoke". If you're saying its due to the smell then its probably from someone unglueing heatsinks off a dead board with a torch. If you're saying that the dark ones are a different color because they've been in a fire, that's even more ridiculous - its nickel plating.
The dark one are different color , because they are not cleaned and sooty. At this point I have to assume you're trolling or shilling, because you're a grown ass man with a technical profession claiming that the nickel coated heatsinks are actually "covered in soot". |
|
|
|
2) No tolerances on the efficiency
They've never had problems with tolerances or false claims, doubt It'd start now. 3) Blockc is not just the US distributor but the "global master distributor (ex China).
Why does that matter? |
|
|
|
no no no no no no. you could arc and kill the controller
Arc'ings not really a thing. One of the following things would happen. 1. The PSU turns off 2. The controller gets semi fried by an inconsistent 12V supply if there's no Vreg 3. Nothing happens and the controller turns on. Wow. that makes sense. it didn't make sense to me either. The instructions clearly said to power the hashboards first and the controller after. I have done this a few times but everything works. I hope I am ok
The instructions are designed for users using multiple PSUs on the same unit. Powering up the controller's PSU before the / some of the hashboards would lead to detection issues. Although we have worked out what Bitmain means in their instructions, they are far from clear and the OP's interpretation not an unreasonable one. This question has come up many times, since it first appeared on the S5+. It would be good if they could make them clearer. I agree, not up to me though. |
|
|
|
I may end up ordering one from the site. Price has a bit of a premium but they do take paypal, so some protection if they never ship me anything.
If I do I will post a review.
Would you be purchasing it mainly for reviewing, or to actually try and turn a profit? If you're looking for a profit, wouldn't purchasing an S7 make more sense? Right now you can purchase a 4.05 TH/s Antminer S7 for 4 BTC (higher hashrate, less cost than the Avalon 6) which would be arriving probably around Christmas time. The site you linked to has no mention of shipping dates. Plus, the S7 is more efficient than the Avalon 6, it seems like an S7 would be a better investment if you're looking to make some money. I believe what they are wanting is someone to confirm that the New Avalon will underclock in similar with the last model and if it does would run more efficient than the S7 and quieter. That's the intent of this thread from what I gather. Looks like I'll be ordering so I'll have a full clock speed vs power vs hashrate vs efficiency graph. |
|
|
|
I won't get into a big thing about it since I'm happily past it. But if you go through C1 thread you can verify all the issues I had mentioned in it.
I read every post in that thread, I know the issues the C1 had. I just don't like that they were misrepresented to be 90% failure when it was still a small single digit number. |
|
|
|
I have an Antminer S5, overclocked and draws 660w from the wall. Can I use a two dual PCI-E cable to power both boards? Or is 330W too much to put through a cable with 18AWG wire?
330W is fine through an 18AWG cable rail. [Official spec would be a minimum of 300W due to 2x 8 pin PCI-E]. The PCI-E cables I have are 8 pin at the PSU end, and 6 at the other. The CX750M I have on order only has two 8 pin PCI-E sockets and four 6 pin PCI-E sockets. All my PCI-E cables have 8 pin connectors at one end, so it appears the only option at the moment is to use two dual 6 pin PCI-E cables. Or can I buy some cables to make use of the 6 pin sockets on the PSU?
You seem to be confusing the PCI-E connectors and the modular connectors. CXMs have a fixed 8 pin connection that fits into the PSU. The CX750M has two such ports and two such cables. On the connector end of these cables are 2x 6 or 8 pin PCI-E connectors. Two of the pins are detachable depending on if you need a 6 or 8 pin connection, but don't worry about that. The 6 pin connections on your PSU are wired up for molex and sata cable rails, ignore them for mining. There is only one choice of cables and it will fit your requirements fine. A) it is a bronze it burns a lot of power.
Its actually about 86% efficient. B) the cable issue you picked up on
There are no cable issues here, he was asking for clarification because the layout is different to his temporary PSU. C) lots of better deals available.
Americans get spoilt by choice and price, there simply aren't other reasonably priced and non explosive PSUs in that price range in the UK (he is UK). The CX750M is as good as he's going to get. If it is too hard to return the cx750m plug in all four jacks and under clock to freq 275 run for a day check cables every hour or two. then bump to freq 287.5 run for a day check cables every hour or two.
There is no reason to do this, the S5 is well within specifications for an S5. but like I said if it is easy to return return it as it is a poor choice to mine a s-5 with.
It is most certainly not. |
|
|
|
no no no no no no. you could arc and kill the controller
Arc'ings not really a thing. One of the following things would happen. 1. The PSU turns off 2. The controller gets semi fried by an inconsistent 12V supply if there's no Vreg 3. Nothing happens and the controller turns on. Wow. that makes sense. it didn't make sense to me either. The instructions clearly said to power the hashboards first and the controller after. I have done this a few times but everything works. I hope I am ok
The instructions are designed for users using multiple PSUs on the same unit. Powering up the controller's PSU before the / some of the hashboards would lead to detection issues. |
|
|
|
There is an HW category strangely under Pool URLs that says 3. That can't be right can it?
That's the one. The first guy coded it in that position and it stuck. Is there some other formula I should be using to find the error rate? As far as I can tell, it should be HW/(diffA+diffR+HW)*100.
Off the top of my head it would be HW / Diff1 * 100 if you have Diff1, else HW / DiffA * 100 will give you a reasonable approximation. Do you not have a % box in the last row of DiffA calculating this though? |
|
|
|
It cannot, like i said, you cannot use different PSU for a single board. So to be more precise;
At an estimate 1200W at the wall, to give an idea;
-PSU #1 will need to draw 800W~ (2 boards)
-PSU #2 will need to draw 400W~ (1 board)
-PSU #2 would ideally power the controller as well (which power the fans), it need its own 6 pin, not its own PSU. Total 10 PCI-e pins.
Actual power draw of PSU#1 would be a bit lower, but overall a 750W PSU should probably not be asked to pull this much.
Last thing, if you get a corsair make sure the OEM is Seasonic or Superflower, they have some batches/series that used low quality OEM.
Personally i think i would go with a EVGA G2 1300 with this. Any 1300+ PSU made by Seasonic or Superflower behind whatever brand should do.
This is it, for using 2 PSU on a miner. As far as using a ATX PSU anyways.
I actually tested this exact configuration, it does work but its pretty terrible. 800W from the wall = around 720W load on the PSU which has a 825W peak load rating. However the limit of these PSUs is more the cooling actually, they don't fare well with high loadings and 35C+ ambients, so its more of a 750W peak rating when it gets warm. The other consideration is that the cables are 18AWG which means you're asking about 290W per cable rail. This is within the specification of the cable (2x 8 pin PCI-E) but they do get warm. The first PSU also requires 2x 2x molex to PCI-E adapters in order to make up the 6 it needs and to reduce PCI-E cable rail loads. |
|
|
|
The C1 was at the height of the anti Bitmain shilling so a lot of the "problems" were from a vocal minority. They did things like dry test all of batch one's pumps.... which running dry causes high RMA rate.
The context was a bad translation, and powering up an impeller for a second to check it works isn't the cause of failures. They did not test out all the wiring harness's the first batch mine if you plugged it in it started to spark....
A number of harnesses didn't work due to bad crimp connections which I had fixed very quickly. If a harness was cross wired then it wouldn't spark, the PSU would refuse to turn on due to short protection. And on top of that the RMA pumps one took one month and other over a month, both died within 2 day's of getting pump.
Pumps did have a high failure rate, I warned them about that way before release but they went with them anyway. My RMA pumps took 3 days to arrive and they provided a spare. |
|
|
|
|
Show Posts
