I think the biggest market for a "100W USBx port" will be for device charging (e.g. phone, tablet, etc). While a higher power port would be great for USB stick miners, it won't drive the adoption.

For the last month or so, the speculation has been that Bitmain was replacing used S5's with "as yet unreleased" S7 miners. It could be that it was actually S5+ miners rather than S7's. While the S5+ has it's advantages, it's really a repackaging of the S5 technology, and nothing like the previous increases (e.g. S5 over S3).

If the actual replacement has been s5+ miners, then that suggests to me that the S7 will be even later than many had expected/hoped. Right now it's hard to see what will motivate Bitmain to actually release an S7, assuming it has actually got the ASIC work and other engineering completed. It's possible folks have been assuming Bitmain is further along than they actually are.

Just my $.02.

The picture above neglects to include the "hearing protection" (i.e. headphone type) required if you are in the vicinity of that monster. It also doesn't mention that you need about 8x the electrical power and heat output. Not really in the same league as the S3's mentioned in this thread.

The syntax you provided is Windows centric in nature. As a general guide, backslashes (i.e. \ ) are for Windows, and forward slashes (i.e. / ) is for Linux. USB mining devices will usually show up as /dev/ttyUSB devices. My limited experience with bfgminer under Linux didn't include a specific device references, it just "found" them.

I didn have to build bfgminer at the time, and make sure to include the proper device support for the type of miner. I would almost certainly use cgminer now under Linux.

While it's probably less efficient, I found the Rockminer R-box (i.e. the small 32GH "pod") to be very stable and reliable. I had 3 of them running for months without intervention at all. They shouldn't be too expensive.  

You might want to consider one of sidehack's Compac USB miner's when they are available. Less hash than a U3 though.

If it's that loud, you have lots more options for power supplies (e.g. noisy and efficient used server supplies).

That ignores the existing millions of coins as part of the "supply". I also have yet to see any cogent explanation of what constitutes Bitcoin "demand". The usual Supply and Demand stuff would be great, except I see no serious way to quantify either.

My interpretation of the above is that August 20th is the date for more details. That doesn't imply to me that they will actually ship in August. I expect we'll get specifications, and possibly pricing and more dates on August 20th.

I kinda get this feeling the various big ASIC vendors are waiting for the others to actually ship something, and then they'll respond. It would appear that SFARDS didn't trigger them, though maybe the S5+ is the response to SFARDS.

When you look carefully at the PCIe power connector spec, one of the +12V lines was allowed to be NC (Not Connected). If it was connected, it was required to be +12V. Hence to be absolutely safe, you can really only count two of the +12V pins as delivering power. Hence a more conservative estimate would be much closer to the 150W value for long term use (i.e. not the absolute maximum). They probably didn't want the connectors to discolor over time due to the heat.

Yes, I know that most all miner vendors assume you can get well over 200W per connector. Not everybody buys a power supply where the PSU manufacturer has connected all three +12V lines with 16AWG wire.

This also makes me wonder if the S5+ hashing boards will work with an S5 controller? It will be interesting if/when somebody that's interested gets their hands on one to start taking close-up pictures and measurements, both physical and electrical.

The current 7.7 TH miner make complete sense if you figure that they want to sell to big mining farms. It's got to be somewhat lest costly to manage N S5+ miners than 3N of something smaller. Not to mention some reduced cost for a single "bigger" controller compared to the existing S5.

It just isn't a "hobby" (used to be "home) miner. (Hobby term stolen from Phillip).

It looks like you could use 3 separate 120V power supplies, though they would need to be on 3 separate circuits (i.e. a breakers) to operate safely. Obviously a 240V setup would be much cleaner, although you will likely need more than 1 PSU since even a 2880W server PSU won't do it alone.

One thing to remember when you start calculating returns that are pretty far out (e.g. 11 months). The next "halving" of the block reward is projected for July of 2016. As with difficulty adjustments, it's block count based, so when it happens isn't precise in terms of date. Most folks don't remember that the block reward will then fall to 12.5 BTC/bloock, from the current 25 BTC/block.

Please put me down for -.51% to -.75%.

There's nothing to suggest, beyond speculation, that a BM1386 actually exists. The S5+ suggests that Bitmain has plenty of BM1384 parts for now.

You should probably investigate what you can do with the DHCP server within your router. Most routers will allow you to find the MAC (i.e. the hardwired Ethernet address, usually six 8-bit hex values separated by colons) associated with a TCP address (i.e. the usual 4 decimal values separated by periods). You might find the MAC address on a label on an S5, though maybe Bitmain didn't make it visible.

The only drawback to the process described above, is that it isn't guaranteed to persist across reboots of miners, and especially across a reboot of the router. The DHCP server in the router will assign a TCP address to the MAC address when it's requested. Usually it will re-assign the same to RCP address to a MAC address, but it isn't required to do so. In particular, the router when rebooted starts with an "empty" table and then assigns address as they are requested. They might match what happened the last time the router was power cycled, but it might not.

While it's tedious to work it all out, and write down stuff, you can probably have your router assign a specific IP address to a specific miner (i.e. MAC address). That way it won't move around on you if things get power cycled.

I also noticed with some IP Scanners you have explicitly request a "rescan" or "refresh" to observe a change of stuff if you power off a miner.

It's not trivial to manage forty-eight S5 miners!

Might be a while until my PC has that kind of port in it, and the corresponding coffee warmer.

While I think a better USB spec for power delivery is great, there may not be a lot of folks that implement it for maximum delivery. Particularly anytime soon.

One item I want to take issue with. What leads you to think that the "string design" is gone from the S5+? As I understand it, each blade has 48 chips, This could break down to a 3x16 layout, which would match the 3 PCIe power connectors on each blade. Maybe I missed something in the announcement, but I think the string design is alive and well for the S5+.

Just my picky two cents on a specific item above. I too think that this marks a later arrival of the S7 than folks may have expected.

One other thing I'll add that I expect you don't want to hear. If you are taking on any kind of debt for college education, do not spend more money than you would for a move and a pizza. BTC mining gear is really a risky proposition in terms of making money, and it's just as likely that you'll lose money as you will make money. Even if you don't lose money, you won't get rich. It's just not in the cards.

Sorry to be such a downer.

I would also dump any GPU miners, unless you wish to mine an alt-coin. They consume a huge amount of electricity for the GH that you get. You must have a ton of GPU's to get get 30+ GH of SHA-256 hashing.

It's OK to gamble, but why buy an expensive and slow ticket (i.e. the GPU's) to do so.