We are sorry to hear that you are having problem.

If you could, please call +1-844-248-6246 and Option 2, so the tech may be able to assist you over the phone and teamviewer remote access.

No worries man. Thanks for testing out the crappy one and warning me. 

Hi there, the right side on my controller does not works, neither the right blade.
I not bought from the factory, instead ebay and probably one of the first batches.
Was working fine since July and this last sunday all of a sudden the right blade fail.
Tested both using another S3 that is working, left side and left blade are good.
Any chance Bitmain sell controllers and blades as a part?
Or simply better to buy another?

This is a though one, but we PM you.  Please contact the techs or the reps.  We should be able to assist you one way or the other.

+1-844-248-6246  or 844-BIT-MAIN  Option 2 or 0

Tough question... Would say.... start plan for the S3+ unit...    sorry  or may be miracle would happen but S3+ be the best bet!

they told me no in the US over the phone. but that doesn't mean they won't. btw i was offered to buy some parts in the US for S3 cost was a lot  but the tech told me parts won't be cheap.

Its saying NO but things can change any moment.  Its a wild wild BTC world.  Happy Hashing!

So I did a little experiment with one of my Antminer S3+ units (One of the better performing ones that is able to hold any frequency setting for long durations without any 'Xs.' I essentially tested out each of the frequency settings and compared them to the wattage numbers used at that setting with 110v power, then tested again at 230v.  The results of my testing are below:

https://i.imgur.com/3MZa9r0.jpg?1

In essence, what I found was somewhat obvious...that you get better Joules/Gh at 230V than you do at 110V, regardless of the frequency the miner is set to.  I also found that the 'sweet spot' of energy used per GH/s is around the 200 MHz mark (200 MHz @ 230v and 206.25 MHz @ 110v.)  I noted that in green, which incidently was the only reading that matched the 'published' efficiency numbers.  A few notes about my testing:

1. I left each of them at the specific MHz frequency for exactly 3 hours, then recorded the Wattage and Average hashrate at that time. I was off by a few minutes sometimes, but suffice it to say I left them for at least three hours. I did note that this number is more accurate the longer you wait...however, I didn't have so much time to play with this, so there may be some variation that I couldn't account for or the numbers may change more significantly over time.

2. I used the same Antminer for every setting, the same PS (Corsair CX750 with all four PCI-E cables plugged in) and the same Kil-a-watt unit for each test.  I was using the 8/11 firmware on the unit with the updated cgminer version (4.6.0-140908) posted by ckolivas.

3. The one major variable I couldn't control for was ambient air temperature, which was generally about 5C warmer in the 230V location than in the 110V location, plus during the day there were fluctuations. This likely would affect efficiency, though probably not by too great of a deal.

4. I used three variables which you don't see here...the price per kilowatt/hour ($0.10,) the current price of BTC in USD ($351.59,) and the amount of BTC/Gh/day based on difficulty (0.00001398 currently)

5. I included the power usage of the unit above when it is not hashing (noted as 0) just to note the overhead that the system takes up if it is on but not running.

Anyway, hope you can get some use out of this...at a minimum it taught me that it's not just a myth that 230V is more cost-efficient than 110V, but when it comes down to it, it's only more efficient by 1 or 2 cents a day...

Anyone is paying attention to unknown IPs by the antminer in addition to IP of pools setup on the miner.
I am not sure if this is related to this firmware upgrade but in my case there are one or two unknown ips connected at times that I cannot figure out.