Just come back to bitcoin mining. I am going to setup a mini farm, thinking about sp20 or Bitmain S5 However, can anyone tell me what is sp20 delivery time? Noted that S5 will delivery on 27/12 . Thanks.
I wouldn'y buy S5 if i were you, you'll have troubles with it, i can assure you https://bitcointalk.org/index.php?topic=903102.msg9922620#msg9922620I dont think its fair to jump the gun on claiming theres problems with the S5. That said, the SP20 is proven as reliable and efficient
|
|
|
^1 in a chain failing will almost certainly have no effect - the chain is only new in the way it divides voltage across chips, and the chip design should still act as a load even if the chip is disabled. Unlikely to be a major failure
^2 the s5 could more easily be undervolted by adding additional chips of some other form of voltage drop to the series circuit. If bitmain was smart they would incorporate some sort of toggle/dummy-load or make an easy way to plug a 3-chip module onto the board to effectively knock down the voltage/chip to improve efficiency down the road
|
|
|
Another thing that really worries me is the chained design. Does it mean if one chip fails, the whole board fails?
So instead of getting 00XX0X00 we get -----------------?
Please tell me I'm wrong.
Hi, there will be two kinds of possible fails. Both are possible. But string design makes the PCB very simple, which left far more less possible reasons for the hardware to go wrong. That is what we have seen in the S3++, a private batch which use chained design with BM1382 chip. Thank you for your reply. But like you said, there is still a chance of failure. With the current S4 for example. Even if one chip is completely dead. I can still use it. But with the chained design if one chip goes silent, that's it. End of story. I like the power efficiency but your price tag.. and the chance of having one miner out of action just because of 1 single chip... I don't know, it's nagging me in my head... If one chip fails it shouldn't affect the rest. The chip schematic has multiple Vin and Vout pins, and the only way to knock out the entire chain would be disrupting the voltage drop across chips. The s1 and the s3 both used chains for the data already, chaining power is the only new thing.
|
|
|
also - will the S5 have a full-frame case again like the s3? None of the pictures show a top cover, or even a bottom cover. I assume its to show off the innards, but if a lid isnt actually included the entire dynamic is changed somewhat.
ps: The Sp20 has all the connections at the front, and could be stacked into a 5x5 grid unlike the antminers, which can only stack side-by-side (becomes an issue if you have limited space or 15+ units)
|
|
|
$450 seems little high
with shipping and customs tax on top. and poor English support. The new SP is far more efficient and what Asicminer has coming. Underclocked, i'm getting < 0.51w/gh (at the wall) with my SP20 already, and at 1.1TH/s... I see no reason to buy S5 since the SP20 gives that at low clock, and then I have space to move up if I want to. +1. The SP20 can be clocked to match the S5, and can easily be clocked higher or lower via the UI, not by having to find a reliable&efficient 9V or 6-10V adjustable source. (friedcat did a similar thing with the prisma - saying if you can feed it 11V its more efficient - but in reality thats easier said than done) If you assume an MOQ:2 as before, and shipping of ~$100/unit like for the S3 you get two approximate options; 3x SP20 = $1900 shipped, or $630/unit 3x S5 = $1350+shipping (~$300) = $1650, or $550/unit or 15xSP20 = $8000 shipped = $533 each 20x S5 = $8400 (no info on shipping) = $420 each (plus ~$50 shipping?) TBH its not bad pricing on both sides, but the spondoolies have FAR more flexibility to adjust for higher hashrate or better efficiency, while the S5 is stuck based on the input voltage.
|
|
|
would love to see some S5 boards to fit on s1 frames
|
|
|
"Hash Rate: 2,000GH/s guaranteed" - hm, wondering how Bitmain fulfills this "guarantee" if the miners show a lower hashrate?
I can tell you from experience it's not guaranteed. If you're using it on a pool or coin that does frequent work restarts, like p2pool, you won't get 2,000 GH/s. Just like the S2, it does not like regular work restarts. M Hm, my average is 1.8863636364 TH/s over two weeks. Tried on several pools. So it is 0.1136363636 below guaranteed hashrate. With 6 S4, this means that I loose about 20 TH/s per month. So if there would be a guarantee, they should reimburse me for this :-) thats 94.3% of spec - only about 5% under spec, and you could likely increase frequency slightly to get a full 2TH
|
|
|
PROS: - same great design - terrific efficiency in new chip - good efficiency gains by running chips in series without dc/dc regulators - at lower voltage these chips are perhaps the most efficient on the market right now (spondoolies can reach about the same values when drastically underclocked)
CONS: - Price: shipping still is not included, and compared with offers like the spondoolies 3xSP20 microfarm (which can underclock to around 1TH/550W per unit) its not an obvious winner -underclocking: chip voltage cannot be adjusted via settings like with the S3+/S4. instead, a lower supply voltage is necessary, which is not easily achieved without buying a dedicated 9V, 250W PSU per antminer in a few months from now.
it looks nice, but right now its not competitive enough with the SP20 units which have been dropping in price crazily (down to $0.31/GH at full speed, or ~0.45/GH at more efficient clockrates). Hopefully we see an option with better undervolting abilities or a S6 at 4TH/1.2kW
|
|
|
So will we get S5 orders going b4 Xsmas?
My guess is after Christmas. It is a guess so take it worth a grain of salt. right now the SP20 or SP31 are clear winners - Bitmain's prices have been stagnant for over a month while spondoolies has now come down to similar $/GH with about 20-40% better efficiency. The same spondoolies chip will easily keep pace with Bitmain down to 0.35W/GH, perhaps further
|
|
|
Dogie, besides the new crisp pictures in your review, what new features come with SP35 and how do they relate to the earlier versions SP31 and SP30? Why should one buy a SP35?
here's my understanding: SP31 SP353000W @ wall 3650W @ wall 4900GH 5500GH $2,595 $2,795 tldr; $200 more gets you 600GH/650W extra mining power. So the added benefits are fairly minimal - you're paying for better PSUs mostly.
|
|
|
^+1 its been a great week.
My XPY experience is similar: during the timespan that would have netted around 0.75BTC ive made about 2.4BTC on nicehash and directly mined 12XPY (about 0.3BTC)
when paycoin ends, Ill drop back to bitcoin mining and its likely the difficulty wont scale up quickly enough to go positive for the next jump. my money is still on a -1% change this period, which alone is a great christmas ROI bonus
paycoin looks like its not doing too hot though. the 'magic $20+ face-value coins' are still only $6 each on the free market which means GAW will have scammed a lot of investors, or be forced to throw a lot of money into buying back at $20 as they promised to
XPY rebounded nicely to ~$10. I have now sold off the last of my Peecoin. Now we just need someone to come up with another SHA scamcoin that gets all of the clueless fools excited again. I heard Titcoin was pretty sweet. well dang, it did. Looking at some charts theres a 50%+ upswing in the last 24hrs. shouldve bought up a bitcoin worth while they were rock bottom, and sold before they next hit rock bottom
|
|
|
very interesting Looks like its a redesigned 28nm as I expected was likely. Probably a larger number of lower-frequency cores per chip. S5 should be advertised as 1TH/380W or so, and an S6 could probably achieve 4TH/1.2kW
|
|
|
^+1 its been a great week.
My XPY experience is similar: during the timespan that would have netted around 0.75BTC ive made about 2.4BTC on nicehash and directly mined 12XPY (about 0.3BTC)
when paycoin ends, Ill drop back to bitcoin mining and its likely the difficulty wont scale up quickly enough to go positive for the next jump. my money is still on a -1% change this period, which alone is a great christmas ROI bonus
paycoin looks like its not doing too hot though. the 'magic $20+ face-value coins' are still only $6 each on the free market which means GAW will have scammed a lot of investors, or be forced to throw a lot of money into buying back at $20 as they promised to
|
|
|
BitFury has become a well run business. It has emerged as the only company that went from "amateur hour" to professional business. Look at the C-Level team; look at their Board; look at their investors! the new chip coming out any time will be game changer and thats on 28 nm (!!!) Can you imagine what they ll do on 16 GAME OVER.. Question how do others keep BitFury contain so there is no great deal of centralization ? Bitfury are hoping for 0.2 J/Gh for their next gen ASICs, AM already have 0.2 J/Gh ASIC samples and are expecting mass production to begin in February. Do you have a source for when Bitfury's next gen ASICs be will available? I've heard nothing about that. At the moment though, BitFury are a joke. Look at their new cloudmining offering - Bit-x: Compare that to AMHash's last dividend of 0.00000788 BTC. Bitfury is MASSIVE, they went the private route but have been excellently successful. Thier first-gen design was terrific, and they very quickly (with gen1 chips) discovered the benefits of running 15(?) chips in series on a 12V capacitated source, eliminating all the efficiency looses of a DC/DC regulator component. They then redesigned a 2nd-gen 55nm chip that was about 15% more powerful at the same power draw, to achieve better that 1w/Gh. They deployed over 50MW/50PH of hashrate already, in larger private entities. Their next chip can easily achieve 0.2-0.3w/GH is moving to the 28nm node realises the potential 4x efficiency gains, if not more if they further develop the chip.
|
|
|
When is it that our chariots turn back into pumpkins?
Saturday, noon EST (about 17.5 hours from now). that date confuses me somewhat, as from what ive read the distribution of POW speed will fall well short of the coins designated for the POW phase. Either POW phase will last serveral extra days or the last block mined tomorrow will have a lot of XPY in it As far as I know - neither. They will not mine all coins, it is time-limited, not amount-limited. ie: paycoin is a bullshit altcoin where 12 million are pre-mined for investors, and less than 0.4 million will be mined during this PoW period. Its no wonder these "$20+ coins" are selling for only $6 right now
|
|
|
When is it that our chariots turn back into pumpkins?
Saturday, noon EST (about 17.5 hours from now). that date confuses me somewhat, as from what ive read the distribution of POW speed will fall well short of the coins designated for the POW phase. Either POW phase will last serveral extra days or the last block mined tomorrow will have a lot of XPY in it
|
|
|
You Bitcoiners ready for some moon? Everyone aboard and strapped in?
ive been aboard and buckled in too long it seems. I have to sell about $3k worth of coins soon to cover some $$$ costs but would hate to have to sell coins at this low a price.
|
|
|
even a big guy will have second thoughts on dumping 1ph in s-3's for 1.5ph in sp20's.
I perfectly timed going from s-3 to sp20. More luck then any other reason, but it does feel good.
I know what you mean. There been times when ive looked back at hardware purchases and been sad, but more often i look back and realise i got in the right moment. I bought both S1 and S3 units at the beginning of thier lifecycles, and they retained value far longer than i expected. When i bought the S3 units I expected them to underclock around janruary and turn of in may. Now it seems more like March/August respectively. (that said, I sold most of mine to get a trio of SP31 units and pay my $CAD power bills) - Its exciting to see that people can get 1TH/400W from the SP20. An Sp31 could achieve 3.75TH/1.5kW, perhaps even better than that eventually, using similar settingsReally? I would be interested in running SP30 at 1440w (one 15A circuit) at some point. Is there a thread for underclocking SP30 or I can pretty much follow SP20 settings? not sure 100% of SP30 specs, but yes, you would follow similar instructions. limit the maximum voltage and start voltages by about 0.05V each and see how that works out. The other way is to change the PSU draw limits, reducing them by 100-200W each.
|
|
|
Just curious, if this is a legit mining company, why are you posting your stuff in Games and Rounds? I got burned by PBmining unfortunately (I should have known better)
Any reason you are posting it here?
It seems this tread was moved here by forum moderators. But I'm not sure. its the gambling subforum, for the reasons mabsark wrote.
|
|
|
The second set of chips get hit with warmer air, but not +40C hotter.
Ok, my intake is about 40c, and exhaust is about 80c, so the back chips are getting +20c air. It significant enough that they are downclocked 200 MHz as a result of the hot air from the front chips. why is your intake temperature so high? thats the problem. Ideal intake is about 15-30C and ideal outlet is 65-75C
|
|
|
|