When Blockchain tells you what the hashing power share of each pool is, it is based on the rate of blocks found.  When your pool tells you it's own hash rate, it is based on the hash rate at which shares are submitted.  Therefore discrepancies will exist based on pool luck.  When your pool get's really lucky it will look like their hash rate is higher than it really is, and when it is unlucky, blockchain will make it appear the hash rate has dropped even if it has not.

Unfortunately that PSU is not ideal, and here's why:
http://www.newegg.com/Product/Product.aspx?Item=9SIA4CP1GG1829


Ideally, a single rail is preferred for powering miners.  Now you have to make sure you get each blade of each miner on it's own separate rail.  That seems difficult, as you have 6xPCI-E connectors, and from what you said and the pictures, it looks like it's 3 cables with two connectors per cable.  So that indicates that the CPU 4+4 and/or CPU 8pin connectors are *probably* on the 4th rail.  And I don't know if anyone makes adapters for those plugs, so you might have to find some way to get PCI-E connectors on those cables.

On the other hand, it does say each 12v rail is capable of 36A, and the total 12v power available is 912W.  That *should* mean  you can run an entire miner (both blades) from a single rail.  12*36 = 432W max.  The only problem is that means you're running it off of a single cable, but if your cables are quality enough, you might just be OK.  Just check them closely over the first few hours and make sure they don't get hot.  And don't overclock any miner that doesn't have all 4 PCI-E's attached (which would mean you'll be on two separate rails anyway).

With all of the built in protection that PSU has, and keeping an eye out to make sure the cables don't get hot, you should be fine.  But in the future, look for single rail PSU's.

Check your tracking # with your shipper.  It should be two packages, one shipment.  UPS actually assigns 3 different tracking numbers for parent and two children.  Regardless, most likely one package was lost and your shipper screwed up.  You should contact them and/or Bitmain ASAP.

When you swapped the cables from the controller blade, did the miner status change which chain showed to be hashing?  Did you also, at that time, confirm that the left hand side continued to warm while the right hand stayed cool?

If so, sounds like a bad blade, and I hope Bitmain would RMA just that blade, and allow you to at least keep the 200GH you do have mining.

At the end of the day, all pools pay BTC based on a formula used to calculate which and how many shares each user "owns" and what, in BTC, each of those shares is valued at.  I would assume since CEX operates the hardware, it is easier for them to just pay out an equal amout of BTC per GH that each user "owns".  Assigning actual hardware based variability by user would be pointless.

While I applaud your idea, and willingness to share with the community, you still have more to consider.

The problem is, there are other variables.  PSU quality/80Plus rating, 120v vs 240v, even the quality of power from your power company has some effect on the # of watts at the wall.  Not to mention, these S3's tend to vary all over the +/-5% manufacturer's acceptable tolerance range.  Much more so than the old S1's anyway.

But here, I'll share what I have.  2xB1 S3's clocked at 225mhz averaging 455GH each on a single EVGA 1300G2 (80Plus Gold) pulled about 720W from the wall.  Right in line with what most people are reporting for 225mhz on gold rated PSU's.

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.

While I agree you should not make rash decisions and jump out of your pool based on your payout dipping due to variance.  There does seem to be some suspicious patterns occuring.


Now that I see Ghash is also paying out less than expected I have to go "hmmm"...and do more research.

http://www.bitcoincharts.com/
Current network speed (calculated by rate of blocks turned in): 147.9PH

https://www.btcguild.com/index.php?page=pool_stats
BTCGuild reported speed (calculated by rate of shares hashed): 12.0PH

http://blockchain.info/pools
BTCGuild reported share of total network (calculated by rate of blocks turned in): 7%

149.9 * .07 = 10.4PH = 12PH - 1.6PH missing blocks

Now of course this can be explained by bad luck, but on how many pools can this bad luck occur.

Also note that the share of network blocks turned in is 9% for two individual unidentified addresses, and 12% completely unknown to blockchain.info.

Obviously there are private players out there, but are they 21% of the actual hashing capacity?

Last link: http://bitcoin.sipa.be/speed-lin-10k.png

A lot of blocks sure are being turned in now that we're in the last 300 blocks before the difficulty change.

With the currently available hardware, there is no possible way difficulty adjustments could be 20% for 6 months.  At that rate, even BitFury, MegabigPower, and CoinTerra would not be able to operate their gear at a profit.  Plus, at that rate they'd have to sell more and more BTC to cover operating/electricity costs, causing BTC price to tank, and making even more hardware unprofitable.  The difficulty is self adjusting, and if it got out of control like that, it could go down.

Now, when <20nm ASICS start joining the network at >5GH / Watt, things could accelerate again.

What pool, and what is the pool reporting about your mining / payout statictics?

Just look at the difference between your numbers:

you gain 400GH @ an extra cost of 300W.  That's still only 0.75W/GH, which, unless you have some crazy expensive electricity/hosting rates, is worth mining for a long time to come.

Maybe the cost of frame, controller, and shipping weight/volume is not negligable?

If you already have 2 circuits installed, then just buy the cables with the correct ends, and you should be good to go.  Honestly though, even at 3kW, you're not to go over 12-13A @240v, which is perfectly safe on a 20A circuit.

Find the plug that matches your recepticles:

http://smile.amazon.com/SF-Cable-NEMA-L6-20P-250V/dp/B009ZERV72

http://smile.amazon.com/6-20P-Power-Cord-Adapter-Foot/dp/B00H7JR636/

This one already has 2xC13, so you'd be good to go.
http://smile.amazon.com/L6-30P-C13-Splitter-Power-Cable/dp/B00BGAEBNU/

And of course you can easily get C13-C14 extension cables wherever (online...Don't expect to find them at Ace Hardware).  Though I would be sure to get those cables rated for 250v & at least 15-20A.

As to the W/Hash ratio, that is what matters "in the end" only so much as you are talking end of life of your hardware.  As RoadStress said after the difficulty is >70B.  Until then, maximize your hash rate as much as you safely can.

I have almost the exact same story.  Interestingly enough, most likely the electricity cost of running those GPU's in 2010/2011 could have been better invested in buying and holding bitcoin.  You could buy a BTC for $6 pretty easy back then.  My huge profit from GPU mining is mostly just from holding the BTC for that long.  Of course half of my gains got goxxed, because I wasn't paying attention.     But at least it was only half 

For My S3's on the miner config page, I just use: stratum.btcguild.com:3333

You'd need a cord that goes from whatever kind of recepticle you have (6-20p L6-20p or L6-30p would be common but not the only options) to C13.  However since it has two PSU's, you'd need a splitter, which might be difficult to find with enough amp rating, or you'd need a PDU.

LOL @ "launching rockets"

I haven't noticed anything on my B1's (July 21 Firmware), but I don't know if I would have noticed.  I'll keep an eye on them and my B5's after they arrive tomorrow, and let you know if I see anything.

Also a good idea, as it's what I have done with my own 4 miners.  Of course it does take a certain amount of understanding about the details of his own local network to get everything right.

I like the list also...Great startup!

I would would like to see availability details and pricing, where available.

Thanks!

When you change it to DHCP the S3 is getting a new IP address from your DHCP server (router).  That means you need to figure out what that IP address is by logging into your router and seeing what IP was assigned.  Would need to know more about your network to help more.