Topic: ANTMINER S3+ Discussion and Support Thread - page 147. (Read 710164 times)

Good luck with that...

Cheers,

ZiG

Is there a way to check the Batch number of an Antminer S3?  I bought 3 off of ebay, and am curious what batch(es) they are.  The seller did not know.

Coupons are on the site now. 0.035 BTC. Doesn't even cover shipping. Gimmie a f'ing break.

I don't think there is or will be much interest in these, unless you can use multiple at least up to the cost of shipping.

Nope, 1 coupon per miner, for as many coupons as you have. They get the price right around 0.66 BTC/ea with shipping, which resets the price back to Batch #7 pricing! I thought for a moment that the price was down to 0.53 BTC/ea for certain quantities, but it turns out I just clicked on the shipping method before their webpage had updated the $$/BTC figure. When I tried placing my 0.53 BTC/ea order, the net price was recalculated "correctly" with the current spot price and the total ended up at 0.65 BTC/ea.

For reference I ran a bunch of quantities through their cart to get a feel for how shipping is managed, it's not entirely consistent, and YMMV (I'm in the USA), but here's what I found (these are after coupon prices):

Quantity   Extended   BTC/ea
2      1.391      0.6955
3      2.043      0.681
4      2.616      0.654
5      3.274      0.6548
6      3.95      0.6583333333
7      4.61      0.6585714286
8      5.22      0.6525
9      5.87      0.6522222222
10      6.524      0.6524
11      7.193      0.6539090909
12      7.761      0.64675
14      9.055      0.6467857143
16      10.43      0.651875
18      11.723   0.6512777778
20      13.037   0.65185
22      14.355   0.6525
42      27.394   0.6522380952

And so there we are. An average price of 0.657 BTC each. Not exactly what I wanted to see. So for anyone who bought Batch #7, you basically mined -for free- for the past 66 days compared to anyone buying an S3+ today. Or, to take another way, if you're buying a Batch #11, you're giving Bitmain 66 days of mining time -for free-. The least they could do is offer free "shipping" to their cloud service, since the miners are already pretty much there, or give new customers a voucher for 100 days of the hashrate they're purchasing, of cloud mining (i.e. "Buy 2 S3+ and get a voucher good for 100 days of 900 GH on Hashnest!").

Does not work

anybody tried to use more than 1 coupon per each S3?

I got two unit from Batch 1, One rocking solid at 504GH and the other one running at stock freq and only at 424. He did run one month at 440 but i never understood why ... I tried reboot, shutdown/cooldown. I'm wondering if i should repaste it since it's not so successful around here.

well, 2 out of 2 and 1 out of one (mine) makes no rule, especially if we are talking about different batches. B1-3 had a "weaker" controller that resulted in some machines hashing at 410-415 GH. I had one of these and re-pasting it made absolutely no difference.

Every underperforming unit I have repasted (2 so far), have run better weeks after working on them.

they are saying 70amps on 12 votls or 840watts... so $50 for pump etc and $100+ for power supply + $100+ s/h

got my coupons!

For kicks, I threw the C1 into my spreadsheet, I get 1.048 BTC shipped to my mining facility in the USA. This does not include a PSU (which I don't include in my other calculations, which slightly inflates the relative cost of the S4 & S2) or the water pump. Assuming a customer buys the water pump Bitmain recommends (which you'd be dumb not to), it's $50 (let's call it 0.13 BTC today). The end result in BTC/TH is:

Block Erupter Tube (10x):   0.9875
Block Erupter Prisma (10x):   1.0642857143
Antminer S3+ (10x):      1.2682119205
Antminer S4 (1x):      1.8475
Antminer C1 (1x):      1.178
S2 – Used (1x):         0.9484897125

So yes, the C1 is cheaper than the S4, and even the S3, but still not as cheap as the Prisma, and requires a considerably more involved setup, with higher risk of failure. I might get one for the novelty, and as hashrates increase air-cooling may end up being insufficient to be competitive, but at the moment, air-cooled hardware is still more competitive on a BTC/GH ratio, but not on a GH/Watt ratio (but how much power do the pump and fans draw for that TH? That may tip it back in favor of the Prisma).

It's intriguing, but having seen Cointerras in action, I'm not entirely excited about the prospect.

Any word on which batches get coupons, and for how much? I haven't seen any applied to my account yet, but it's been a while since I ordered S3+s, maybe it's only for those who were charged shipping? Maybe too few people actually paid that price for their S3s, and so no-one is chiming in as being burned by not getting a coupon, or to have said they got it? All 10 of you who ordered Batch #10 got your coupons already eh? And none of you visit BitcoinTalk perhaps...

Anyway, the price is too stinking high still, but it's at least cheaper than...Batch 8 now. Ugh, Batch 8 was barely profitable, I can't justify this pricing. To be the same price as the Prisma Tube, S3s would have to be 0.48 BTC -this week- (shipping inclusive, or 0.38 BTC before shipping). I like the S3 better than the Prisma, but my purchasing decisions aren't driving by "like", they're driven by BTC/GH/Watt, and right now the S3 is just too high at 1.268 BTC/TH.

That is why I have been trying to stay away from the inital batches!!

The C1's however....  1th/s  1BTC seems fair, 800watts.... makes those s2 2th/s way too much money of course.

Bitmain,

  Where are the coupons? I see nothing.

only if the original job wasnt done properly and the only way to know that is by taking them apart to look. and remember to not use conductive paste!
When I did several upgrade kits I saw that the paste applied to the S1's wasnt really how I would apply it... basically it looks like they have a machine that sprays it by hand and they just spray it where ever in lines.

All this re-pasting is hugely overrated.
People re-paste, then switch miner on-it reports better numbers for a first 12 hours-and these fellows post giddily about it. They don't post that later it goes back to the same numbers as before or worse (due to possible reassembly errors). However, if you switch the miner off for a half an hour, it ALWAYS comes back slightly more perky than before, so essentially the re-pasting effect=placebo effect. I've done repasting with several compounds-and in no case there was any long term effect.
Granted, if the miner did not perform from the get go, then repasting might improve it, but it does not improve slightly underperforming miners-that is in most cases something else than paste.

Does repasting them really make a huge difference?

We bought a large lot of batch 7-9 and none of the units, except 1 dud out of 90, (Props to BMT for quickly responding for RMA) have issues running at 225. Some go to 250, and many in between.  Most run at 237.5 or 243.75.

Would repasting increase rates on the units that don't seem to like higher than 225 or 237.5 settings?

Seems like a lot of work with little return difference. Even at 225 the units give a solid 450-500 Hashrate.

Thanks,

Strato

You think that's bad ? You could be a batch 1 buyer that had 50% of his units unable to perform stably at 218.75mhz.

Have 3 units that can only run stably at 212.5mhz. Plan on getting around to re-pasting them in the near future to see if I can't improve things.

Ask me how annoyed I am. Especially when I emailed Bitmain early on in the S3 process and they assured me they would not be updating the S3's in the future to resolve the performance problems... And lo-and-behold the S3+ arrives a month later.

Don't get me wrong, I'm a huge fan of Bitmain, and more than 50% of my hash power comes from their devices, but the S3 launch did not go as well as the S1 launch, it appears the S4 launch (Don't own one) has been problematic, and given anecdotes I'm reading here, seems the community consensus is confirming my gut-feel: Bitmain is losing it's way

Funny that I'm thinking of circling back to AsicMiner products for my next miner purchase...