Thanks for sharing the photos and story.  And to give everyone that warm fuzzy feeling inside, they have repaid my loan!

I have been repaid in full! And it seems he sent an extra BTC?

https://blockchain.info/address/13M5iwY1dzVT2Hq4aC1CcsXxbepg7wKzPw

I am ecstatic that I helped keep the museum intact.

https://blockchain.info/tx/a53280605ca1afc9dd63b519b435c66e50cfdab2409079b5178b0c83f66e57c7

I seem to take months off these forums as I get busy IRL, I just happened to be looking in the Goods section today to buy something to recoup recent electricity costs and saw the museum items up for sale.  This perplexed me, so I read through all the threads and their Facebook public pages.  I have loaned the remainder of what they need to get their wedding and honeymoon funded and to keep the museum together.  I myself have a mini museum of my 5770's, MMQ, Avalon B3, and Bitfury, so I know what that feeling is all about.

I think I figured it out.

You only get +1 if the person who left you feedback is in your trust circle.  So for example, I gave you positive feedback, and now I see you as +1 (because I trust myself).  But you still view yourself as +0 because all the feedback are from untrusted sources.

So you can see in "Trust settings" that you need to get positive feedback from someone in this 2 depth network of trust in order for most people to see your "+1"

Bought 6 plex, fast and easy, thanks!

I have two of these: http://www.newegg.com/Product/Product.aspx?Item=N82E16835146003

You are looking for high CFM.  The Noctua is 115.5 m^3/h which translates to 68 CFM.  The Corsair is slightly worse at 62.74 CFM and noisier.  In comparison, the NZXT I am using have 166 CFM.  If you have a Frys nearby, they sell the NZXT for $15

Just wanted to add my positive experience, since forums tend to get filled with negative experiences:

I received my full kit order on Nov 5 and put it to work as soon as I could.  I encountered many problems as others have described on here with boards cycling to 0, even with 2x200mm 166cfm fans moving air through.  I kept rearranging boards to minimize the SPI errors from dragging other boards down, until I was left with 8 cycling boards and 1 EOL board (only 11/16 chips detected).  I do believe there are still 4 buses, since the EOL board only brought down slots 6 and 7 with it when it was in 5 at the beginning.  I was only getting somewhere between 200 - 300 GH/s as boards kept cycling.

My heatsinks finally arrived on Nov 6, and I put them on the back of all the regulators except the EOL board.  This stabilized most of the boards, leaving only 2 cycling boards and 1 EOL board.  This stabilized my hash rate to about 31-33GH/s per board.  That was roughly 450GH/s with 2 boards cycling on and off and the EOL at about 22GH/s.  It dropped to a bit less than 400GH/s during the day when the room temp got warm, when 2 other boards started cycling for a total of 4 cycling boards.

On Nov 8, I finally found time to put heatsinks behind all the chips except the EOL board, and this stabilized all the cycling boards.  Each board was getting 33-35GH/s and I was getting a stable 530GH/s.  On Nov 9, I moved it to its permanent home where the temp is 10F less than my testing area, and I have been getting a stable 580GH/s (36-38GH/s per board, 25GH/s EOL board)

I think the following quote from punin is exactly why so many are having problems in this thread in the past few days, these cards are going to be unstable unless you have adequate cooling or turn down the pot setting.  I needed heatsinks at both the regulator and chips in order to get all the boards stable (some were stable with only air cooling, some needed heatsink at the regulator, and some needed heatsinks at the chips)

For a batch 3 avalon, I used the following parts:

http://www.homedepot.com/p/Master-Flow-14-in-x-6-in-to-8-in-Universal-Register-Box-with-Flange-RBF14X6X8/100154377

Obviously, find a register box that would fit your needs.  This 14x6 barely fit the batch 3 after I straightened out the flanges (tear apart welds and unfold) to act as an air funnel that the avalon sits in.

http://www.ehydroponics.com/ecoplus-inline-duct-booster.html

Choose the correct size booster to fit the other side of the register.  I connected the register box directly to the booster, and used duct tape to seal it.  I didn't seal the register box to the avalon, so I figured a booster would help the exhaust go the right way, instead of out the unsealed part.

http://www.ehydroponics.com/deflecto-metallic-duct-4-x-25.html

Choose the correct size duct to match the booster.  I simply used duct tape and some bungee cords to exhaust the hot air outside.  I chose this instead of the insulated kind because it would be easier to work with (insulation needs respirators and special washing of clothing after).

These are the parts I used because they were all available in B&M stores, as opposed to having to order online and wait.  I had just gotten my avalon 1.5 days before and the uninsulated room I had put it in was already 10F warmer than outside and still rising.  After this, the room is at most 2F hotter than outside.  I debated using it as intake, but I worried about taking outside air directly without a filter / water drain.  As an exhaust, the walls act as a big filter.

DH(%) is the stat to look at.  It is calculated as HW / (Diff1Shares + HW).  Given that --avalon-auto tries to keep it around 1.5%, I'd say yours is doing better.

I'd love one, thanks in advance!

It looks like http://www.amazon.com/Acer-S211HL-21-5-Inch-Widescreen-Ultra-Slim/dp/B003U4IROY since I have 2 sitting on my desk right now

I'll bid 3 BTC

There are other "ASIC" Intellectual Property out there already that can do the same computation.  However, those IPs are optimized for low area and low power, so that they can sell it to someone who wants to add hashing capability to their ASIC.


This is the key point.  IP out there are meant to do a few hashes.  So the tech sheets I've seen from IP providers calculate a hash in 66 clock cycles.  Furthermore, because they are minimizing area used, they reuse as much as possible during these 66 clock cycles, meaning you can only calculate one hash every 66 clock cycles.  Assuming you are running it at 660MHz, then you only get 5MH/s per IP core (bitcoin hash is SHA256(SHA256(x))).

But to the mobile phone ASIC designers, they are happy to license it so that they can calculate a hash in 100ns, compared to however long it would take to put it through the MIPS or ARM core, or having to implement one themselves.

In comparison, the initial open source FPGA design unrolled the SHA256 cores (if the FPGA can fit it), which means it computes SHA256 in one clock cycle and a bitcoin hash in 2 clock cycles.  It uses up a lot of area, but can roughly get MHz -> MH/s (because of pipelining, the first core can start calculating the next hash while the second one finishes)

I was actually researching the same thing recently after seeing 80 PLUS having better efficiency on 230V

An electric dryer / stove socket would likely be NEMA 14 in newer homes, where there are two 120V lines with 180 degree phase difference, creating 240V.  Older homes would use NEMA 10.

I found these adapters which are meant for electric vehicle charging.  Then you could use a rackmount cable to connect to the PSU.

What's been stopping me is that the 2% gain in efficiency is only worth it if you are drawing a lot of power, having to put the machine in a specific room or pay an electrician to wire another socket to the room you want.

Bought and redeemed, thanks!

+1 for Graet, just used him for my first BTC purchase.  Quick and easy.

I preordered the bASIC because of these reasons, from most important to least:

• All the good words about cablepair's transparency in the forums
• They were the only one with batch 1 pre-orders remaining when I discovered ASICs
• BTC was continuously getting weaker in October and I wanted to lock in a higher price for my BTC
• If I was scammed, I would just forget about bitcoin completely, since the BTC I paid with were from mining on a $200 initial investment

I should have refunded earlier when the transparency stopped around December, but I was blinded by the "place in line" concept.  Hindsight is always 20/20.

I finally asked for a refund on his drunken post (I got it last Friday).  I will still buy a bASIC if it is realized, but not a pre-order anymore.

I do not blame them for refunding USD equivalent instead of the BTC I paid.  I lost the bet that BTC prices would keep declining.  It was entirely up to cablepair whether to risk holding the BTC or convert them to USD immediately, and is completely unrelated to the customer.  I certainly would not expect them to ask me for more BTC if BTC kept on declining.  Likewise, I did not expect them to refund me the original BTC amount if BTC strengthened.  (I guess we will never know if he would have refunded more BTC if it had kept on declining.)

Personally, I learned about BTC a year ago and started mining with a small $200 purchase of GPUs.  I mostly read posts and seldom posted.  Then I just got busy and stopped following it for a while (I even switched from pool mining to solo mining so I wouldn't have to deal with pools shutting down or needing miner upgrades).  I forgot about it until one day when I needed to reboot a computer and saw I had found a block solo mining.  I decided to look into BTC again, and saw that ASICs were being developed.

So I looked into it a bit, pre-ordered one, and have mostly been reading posts and seldom posting.

In the DD-WRT solution, you are simply using the USB drive as a storage device.  Initiating connections to other nodes, downloading / uploading the blockchain, and verifying that blocks satisfy the hashes are performed by the Broadcom (MIPS) CPU in the router.  DD-WRT is simply an OS with many router features that the Broadcom CPUs can understand, similar to how Mac / Windows / Linux are OSes for the Intel / AMD CPUs.  Usually, you can take Linux source code and compile with MIPS as the target architecture, then run it in DD-WRT.


Simply plugging in a USB storage device with the blockchain on it does nothing.  You need the CPU to process commands to and from that blockchain.  It sounds like you have two choices: 1. Rewrite the code to reduce memory usage and compile for MIPS, or 2. Have an external CPU like the Raspberry Pi or a phone to do the CPU work.

Here is an elaboration on choice 1 based on my work on rooting and getting apcupsd running on a NAS running busybox:

• Root your router, once you get shell access at busybox, you do not have to install DD-WRT but can just use the default OS already loaded
• Install Ubuntu somewhere and compile bitcoin for MIPS
• Try to run it with your shell access
• Make changes in code and recompile to reduce memory usage

Yay! A replacement for my BTCFPGA advertisement sig.


Actually, it only works if you filed for non-delivery 45 days or less after payment


Beyond that 45 days, you will need to do a chargeback on your credit card.