I have two Antminer S3 that are both hashing at 440GBs. They both are running at the stock frequency of 218.75 but one if the miners has 20% HW errors. I have read many people are lowering the frequency to 215 or 212.75 to lower power consumption. I was hoping maybe reducing the frequency might help my HW errors. However I cannot find where to lower the frequency. I have done scrypt mining, but this is my first sha256 mining. If someone could help explain to me how to lower the frequency I would greatly appreciate it.
Underclocking early without having a solid reason will only lower your earning potential. Are you sure you calculated your HW error % properly? Post an image of your Miner Stats page. I now have it underclocked so I cannot take a screenshot. But it had been running for almost two days. The accepts were 65,000 and the HW was in the 14,000. It has now been running over 2 hours on 212.5 and accepts are 3,774 and HW are 370. So the current HW is at 9%. The other miner which is still at 218.75 has 71,068 accepts and only 2,049 HW; only 2% HW. that's NOT how you calculate the HW error percentage. Why under-clocking causes inability to take a screen shot? You probably downshifted your machine(s) for no valid reason although people proposed to you to post your stats... Formula is: HW error %=HW / (diffA + diffR + HW) * 100 |
|
|
|
Preferably the unit that was bought with 10% discount ($4495).
Anybody? |
|
|
|
|
Questions to SPtech (Zvi) or anyone else at SPT or btctalk:
1. If Sp-30 is connected for 6 days to 240V, then switched to 120V for a day, then again 6 days on 240V with cycles repeating in perpetuum, how exactly it would affect it's "learned" behavior. Could this even be done?
2. Is throwing the off switch the best way to switch the miner off or it is better to be do so in gui?
|
|
|
|
As mentioned previously in this thread (although not quoted), those compose of perhaps 10% of the price, or approximately $40. Why wait around or hang your hopes on getting an upgrade kit to save $40? Not to mention an old S1 is worth about $150 or so on the open market, and you lose the ability to sell it if you upgrade.
That's your estimate. But why do you think that? Shipping alone must be non-negligible, as Bitmain ships super-fast from China. 4xS3 Blades, could easily fit in a single S3 sized box and weigh considerably less. Regardless, the real evidence is, Bitmain keeps saying they are going to do it. What reason could they possibly have to tell potential customers that an upgrade is coming if it is not economical for them to do so. Surely they don't want to prevent people from ordering entire S3's if they aren't planning to sell the upgrade. Because it is very obvious that some people who would otherwise purchase S3's are waiting for upgrade kits. Yes those are my estimates. What are yours? I know a little bit about manufacturing CCAs and production as that's what we do where I work. The bulk of the cost of the S1 or S3 is in the main boards. The controller cards are basically Raspberry Pis which can be bought retail for under $40, I'm sure Bitmains are less than that if sold separately. The aluminum frame is not pricey in mass production. You might have something with shipping. But here are the economic scenarios of where you'd be 2 months from now (ignore shipping times, assumed a wash for each scenario): Scenario A: Order new S3, sell old S1. 0.66BTC ~= $383. Sell S1 for $150 after S3 arrives. Net result: -$233, with double or better mining efficiency at current difficulty level for 2 months. Scenario B: Wait for an upgrade kit. Assume $250 cost w/ availability 1 month from now. No S1 salvage value. Net result: -$250, 1 month or longer using S1 at half the efficiency of S3. 1 month with S3 efficiency at future difficulty level(s). So clearly to make an upgrade work, you'd probably need to have an upgrade kit sold at less than $200 and be made available within a month to even come close to breaking even vs ordering an S3 now and selling your S1 when the S3 gets here when factoring everything in. I could be wrong, but I would be extremely surprised to see a $200 or even a $250 price on an upgrade kit in the near future when they are getting $383 for a full unit. Add in the headaches of support for helping some of us idiots tear apart our machines and do our own upgrades and the proposition doesn't seem to make much sense. Bottom line, there really is no reason to wait for an upgrade kit. Make your life easy and order a new S3 and sell your S1 when it arrives. Someone just bought S1 for $160 on ebay (0.27BTC, 0.25BTC after ebay fees!). I am surprised-it sells on ebay for almost the same as 1 month ago (maybe $10-20 $ less). People were selling S1 on BitcoinTalk for 0.22-0.23 a month ago-they were selling too cheap, obviously. Probably, low diff jumps played a role. |
|
|
|
You will get around 10% lower hashrate at 120V compared to 240V.
I get that (4.1Th at 110/120v), but would you use 2500, 2640 or 3000W at the wall to get there? I'm now getting 4178 GH/s on 122 V while using 2752 W (2*1376) at the wall for an efficiency of 0.659 J/GH. This is on 120V. Super, thanks, I just don't know about 122V. Typically, it is 115-118V, but i checked my house, and it is, indeed, 121.5-121.8V. |
|
|
|
You will get around 10% lower hashrate at 120V compared to 240V.
I get that (4.1Th at 110/120v), but would you use 2500, 2640 or 3000W at the wall to get there? |
|
|
|
What discount? And why would Bitmain charge less for blades use in an S1 upgrade vs an S3 purchase?
Maybe the cost of frame, controller, and shipping weight/volume is not negligable? As mentioned previously in this thread (although not quoted), those compose of perhaps 10% of the price, or approximately $40. Why wait around or hang your hopes on getting an upgrade kit to save $40? Not to mention an old S1 is worth about $150 or so on the open market, and you lose the ability to sell it if you upgrade. Yea, but since my S1s are all paid off and still making money, why be in a hurry to sell? It's like keeping a car for 20 years until you drive it to the junk yard - sometimes it's more cost effective than trading them in every few years. Plus, all the S1 guys can wait for you S3 pioneers to work out all the bugs. S1 uses 2-2.5 times more energy than S3 for the same hash plus pricing was good a few weeks back. I am limited to <3kw extra during summer, so it was a no-brainer. I loved the little fellas, but they just had to go. |
|
|
|
Disconnecting the PSU's power will reset the SP30s knowledge of the PSU's limit, and will cause this algorithm to start again. If you change the cooling conditions for your SP30, you should power cycle your PSUs so the SP30 learns its new limits.
I didn't explain myself correctly in mails: In order to cause recalibration, go to "settings" tab and set high PSU limit (1370W is over the top for 95% of PSUs). It will take it down again to it's possible maximum over the next day. Note that the hash rate is more expensive the higher you raise the watts. You can get ~4TH for 2300W, but the next 600W will only provide 0.5Th - much more 'expensive' hash-rate in terms of power. So I guess for 2500 the system will give about 4.2TH. If you pay 0.15$ per kWT, you still better running 4.5TH (100$ more). If you pay over 0.35c per kWT, you better run 4.2TH. Is the info above at 240V or same for 110/120V (for lower Th)? In other words, if I run Sp-30 on two 15A 110V circuits (ostensibly up to 2640W/24A from both at the wall), what would be the expected hash? do I need to do some mod to remove power limit? I would like this info so I can compare the economic benefit (or lack of it) for hosting vs home. |
|
|
|
I am having some trouble with an S2 unit that I have had running. It has been running perfectly at 1 TH/s since I got it, but after some installation of S3's last night, it is now showing problems.
The LCD on the unit says that it is hashing at 1 TH/s and it looks to be OK inside the unit. Everything is flashing normally and looking OK, but the pool is only showing about 100 GH/s hitting it from the unit. Does anyone know why the two might be showing this discrepancy?
Interesting, does S3 show 1300GH instead? Are both units on the same ethernet hub or connected independently? I had a freak occurrence when I unplugged S2 that was on the independent power strip and S3 that was nearby (and on a DIFFERENT power strip) switched off, then back on. This is weird, I know. I think that this is because S3 has some wi-fi capability which might have some interaction with the router or they somehow had some interaction through the ethernet hub. Anyway, it was spooky, but I double checked the fact that the units were on different power strips. Does that 900 mH go anywhere where you can detect it or is it just gone? I haven't been able to find it at all. Everything else connected to my pool is accounted for as each unit has a different worker name and everything else it running fine. (Besides one pesky S3 that I will have to spend some time trying to get running properly tonight). I can't imagine that one is causing problems for the other as there are 10 S3's, 5 or so S1's and 2 S2's running in parallel down there and this is the only issue. Perhaps, with a last miner you exceeded your downstream internet provider speed as you have so many. I assume that they are all on the same router. Are they all on one hub or several? If several, maybe try mixing them up differently. |
|
|
|
bumped it up to @237 clock and is stable last two hours hashing 465 - 470g/hs all asic 0's ..
i have one ant that shows all ooo for 2 days at 237.5, while hashing at 464-465, then it shows one X, but speed stays the same. My conclusion is that it does not show X properly from the beginning, but that is a malfunctioning asic from the start, otherwise the speed would be 478, not 464-465. Numbers fit, percentage wise: 31/32X478=463 yours might be different (truly fixed) |
|
|
|
I am having some trouble with an S2 unit that I have had running. It has been running perfectly at 1 TH/s since I got it, but after some installation of S3's last night, it is now showing problems.
The LCD on the unit says that it is hashing at 1 TH/s and it looks to be OK inside the unit. Everything is flashing normally and looking OK, but the pool is only showing about 100 GH/s hitting it from the unit. Does anyone know why the two might be showing this discrepancy?
Interesting, does S3 show 1300GH instead? Are both units on the same ethernet hub or connected independently? I had a freak occurrence when I unplugged S2 that was on the independent power strip and S3 that was nearby (and on a DIFFERENT power strip) switched off, then back on. This is weird, I know. I think that this is because S3 has some wi-fi capability which might have some interaction with the router or they somehow had some interaction through the ethernet hub. Anyway, it was spooky, but I double checked the fact that the units were on different power strips. Does that 900 mH go anywhere where you can detect it or is it just gone? |
|
|
|
I like to tell another thing, my 2 PCI-E Power cables to S3 were a little hot.
My S3 hashrate was not stable at the rated 441Ghs but was averaging at 420. I ordered thermal paste.
Before thermal plates were arriving, I tried to plug two additional PCI-E power cables to S3.
The Antminer S3 is running at stock clocks 218.75 and PCI-E power cables are cool now.
The main reason for this post is to report that the miners stabilised at 441. Earlier it was 420.
Two thing I'm pointing out,
1. The miners plugged in continuously for 4 or 5 days and seems like it was running-in and stabilised at-last at 441.
2. Plugging the all four connectors even if you are not overclocking will some way help you.
Yes, but did they stabilize because of a 4 days run-in or because of 4 vs 2? I am also in the middle of one experiment. A unit was fluctuating between 420 and 428 on 212.5 (~400 only on 218.75) on four connectors, now I switched to 2 connectors in front (first pair close to the ethernet port; connectors are from the same double cable from EVGA1300, incoming cable is warm (42C with a laser temp probe), temp of actual connectors is 31 and 29C-unit is at 429gh (after 18hr), so possible slight improvement. Conclusion: a possibility that 2 connectors are better than 4 (at least for some machines). Will report on latter stats and will try to go to 218.75, which it was NEVER able to properly do before (on 4 connectors). Another new unit I am "burning in"-checking it with just a single double cable from CX500M, connected to the "back" pair of connectors-incoming cable is at 42C, first connector is at 32C, second at 30C with ambient at 25-26C, speed 441.55GH at 218.75 for 20hr. Conclusion: CX500M is OK at regular speed, temp not that high, speed good |
|
|
|
|
Preferably the unit that was bought with 10% discount ($4495).
|
|
|
|
|
A long shot, but does anybody want to sell their Group buy Sp-30?
If yes, PM me.
|
|
|
|
One of my machines came pretty mangled up using DHL. Looks like it was scuffs from the factory as most of the damage looked like rubbing damage which wouldn't happen in shipping.
Just a heads up.
I am debating to send it back or keep it, looks to be hashing 430, while every other batch 5 is at 440+. Meh.
If it's hashing good, just bite the bullet or you'll end up with lost income. As for me, here we are on the 5th and still no movement on my 4 B5 units I paid 0.68 per unit for totaling 2.72 BTC which they said would ship early and he we now have B6 on the site for 0.66 per unit. Not really complaining since when I did buy, I bought in with a original Aug. 7th ship date. Just frustrated some that I haven't gotten early shipping per the one Bitmain post. i don't see any b6 for 0.66-did they took it off? |
|
|
|
Think it is safe to run one S2 and one S3 on a regular US 120v electrical socket? I know the safe limit is supposed to be somewhere around 1400 watts.
Could be pushing it=~13.27A (strictly speaking it should be no more than 12A). Just check if you have two circuits close together, then run S2 on one and S3(s) on another-that's what I did. |
|
|
|
About the payment process. I bought a couple of coins at CoinBase, can I use their wallet to pay for their site? how much should I donate to the transaction in order to make it faster?
Coinbase pays the miners fee themselves, so you don't need to put in any extra. However, it takes for coinbase up to 1hr sometimes to have the first confirmation, so your transaction would timed out initially, but they (bitmain) usually reconcile it later either by themselves or if you send them email with your transaction id. If time is of no particular importance, but you want a peace of mind, send your coinbase coin to blockchain.info wallet first and pay from there, paying minimal fee of 0.0001 btc or slightlly above. |
|
|
|
|
Collider, jtoomim, Roadsress or anyone with the early machines:
Please post your statistics and whether you achieved sustained ~4.5TH at 208-220V (please post your voltage) and what was the speed at 110/120V if applicable (I assume 4.1Th). Thanks.
|
|
|
|
I got an issue with one of my S3s. It does not hash at all. I can connect to it no problem. This one appears to have an issue where the unit does not hash at all. All the settings are identical to the other ones (minus the IP addresses which are statically assigned. I can connect to it from within my lan to get to the web interface. cgminer appears to be running but not hashing. Anyone run into this? It seems like it would be a quick fix. Thanks
Did you open it up and check the cables from the blades to the controller? You could try to reseat the cables as well. If that does not fix it the next step would be to swap out controllers to determine if it is the blade/s or controller. I did not cover the power supply verification as I am sure you have already been down that road. Please let us know what you find. I will give that a try. I checked on the IP conflict and that was not an issue. I statically assign everything in my network so I know thats not it. But just the same I checked just in case. I will open it up and check it. I will post back if that was the problem. Why after I statically assigned my ip address on my miner can i not pull the interface back up on the web browser? Did you set your subnet mask and dns settings correctly? Yeah I matched my IPv4, subnet mask, and default gateway 10.8.201.103 255.255.255.0 10.8.201.1 But, cant seem to reach the miner again not sure I understand. What do you mean "I matched my IPv4 address"? IP address of the miner has to be different (after the last .) from your computer/laptop and unique for your LAN (no other piece of equipment with the same IP). Also, you need a fourth lane-your DNS, which could be same as your router/getaway (10.8.201.1) Sorry, let me try to explain better. My network ip address is 10.8.201.103 so I had to temporarily change my computer to 192.168.1.99 (plugged the miner into computer and the interface ran) . When i changed to try and get the miner to run on my network ip (10.8.201.103) I was never able to get back to the miner interface without going into too much detail, go to sushi or dogie "S1 troubleshoot and setup" threads and read carefully the setup. What you describe has insuffient specifics, as to your PC (after changes), router, miner. Did you setup DNS? This line was completely missing from your explanation. And what is the current IP of PC and did you setup miner as DHCP or static? |
|
|
|
How can I safely power off the s3 antminer? Anyone have suggestions for the best way? Thanks.
the safest is to ssh (using terminal or putty) into the miner, login by "root" passwd, then issue the following press return/enter type <exit> return command line will return to Terminal/Putty miner will power down slowly in 3 sec wait 20-30 sec as fans decelerate before switching the PSU OFF what other people have jokingly posted is OK too, but would result in a bit higher temperature of the chips since fans would not slowly decelerate and upon immediate restart you could get a couple of ASIC errors, but you can remedy this by restarting cgminer once or twice next 1-2 min. |
|
|
|
|
Show Posts
