My record so far has been refusing to take in money for a product which does not exist. If someone wants to help fund development, with the full knowledge and understand that the product may never materialize and your money could be completely wasted, that's a bit different (and also something I've avoided thus far). But I have never, and likely will never, take in money pre-selling a product which I have not designed and successfully prototyped. If I've ever taken in money it's either been as a loan with repayment terms, or to buy materials to manufacture a batch of something already designed and tested and the investment was repaid in units of that thing once manufactured, according to the original terms of the sale.

And yes I understand that makes it hard for a broke guy to get anything done, but it also protects a customer base remarkably willing to put itself at risk of getting screwed by speculators from being screwed by me, accidentally or otherwise.

Also, the idea of using a blockchain for other things than the transfer of currency is not fundamentally bad. But I think leveraging an existing blockchain already filled to capacity with legitimate transactions, and designed specifically for processing legitimate transactions, for something other than processing legitimate transactions, is foolish. Repackaging something inherently not a bitcoin transfer to look like a bitcoin transfer so you can use the bitcoin blockchain for it seems kinda dumb. If you want to do something that isn't transferring bitcoins from one address to another, get your own blockchain.

Also anecdotal, but in the last month I've shipped out 20 custom PSU setups intended to replace failed Bitmain PSUs in S4+ units, which draw about the same power as the S7 and S9.

So, folks who know me on here know I have an interest in miner design. I have my ideas on how to do things and what is "good" and "not good", but plenty of other people also have their own ideas.

So I would like, if possible, to use this thread to facilitate civilised debate over possible features for a consumer-grade miner. Let's assume the machine sits in the spectrum of Avalon6 and S7 for general size and power consumption. Those attributes are fixed.

I'd like to see what the community consensus is about such items as integrated controllers (like Bitmain) versus external chainable controllers (like Avalon uses), or using purpose-designed cabling and protocols (lke Bitmain) versus a more generic bus (like ASICMiner's UART or Rockminer's USB) to connect. What do we want to see for power interfacing, or voltage control? Should hashboards have sub-controllers or be dumb and driven directly by the central controller? Stuff like that.

Partly I'm interested to see where my own ideas line up with the community at large, and partly I want to draw from the collective (and collectively overwhelming) variety of expertise present here, the end result of which should assist in designing an actual miner with the hopes that it's "the best". That part of the goal should be no surprise to anyone.

So, let's start today's discussion and start picking things apart.


I like the single-fan tube concept of the Avalon6, but would like to see heatsinks on both sides of the board. This isn't always possible, and with a string-topology miner can become dangerous. However, in designs like the A1 Dragon and AntMiner S3, having chip-side and PCB-side heatsinks allowed for efficient heat transfer without a lot of fan noise. Does the decreased Tca outweigh the risk of failure due to electrical short circuits through the PCB-side heatsink, and the extra milling requirements of at least one of the heatsinks to avoid contact with other components?

I like the daisy-chaining concept that Avalon used in the -4 and -6 machines. I'm not overly fond of the required USB dongle signal converter, and would prefer to keep operating requirements a bit simpler. Jstefanop mentioned some weeks ago an idea to make small (~20W) miners with USB connectivity, and each daisy-chainable. That would require a 2-output hub chip in each miner, which is definitely fun but I'm not sure if that's better than just using a hub. I think star topology afforded by a hub would be better (certainly more fault-tolerant) than a daisychain of miners. Does the availability and affordability of USB hubs and cabling, and the maintainability and fault-tolerance of a tree layout versus chaining, outweigh the potential software and hardware overhead (and physical connection reliability concerns) of using USB versus a more primitive protocol for miner interconnection?

Any chaining like this will only matter if the miner uses an external controller. This reduces overhead for a fleet of miners, and makes administration easier (with a single point of access). However, the controller becomes a single point of failure for what could be a fairly large operation, and distributed control software leveraging cgminer API functions already exists. Does the ease of administration, ease of replacement, and general reduction in equipment overhead for a single central controller outweigh the cost of single-point failure? Is it better to have a completely self-contained controller (like seen on Antminers, requiring only a network connection) in every miner unit?

If using an external controller, higher-level protocol or not, it makes sense to have an intermediary microcontroller on the hashboard (or, in the case of Avalon6, on a separate board that talks to the hashboards directly) which handles all upstream communication, and multiplexes chip-bound data, fan control and sensor reading. Would it make sense to have this integrated at the board level even if the miner has a self-contained controller? This increases board-level hardware complexity, but can also add features like semi-autonomous fan control in emergency situations when the controller software is exhibiting erratic behavior (for example, S5 overheats). It could also be responsible for implementing changes at the board's voltage regulators, in the case that adjustable core voltage is desired; you cannot convince me adjustable core voltage is not a mandatory feature, so something will need to do it.

So, lets get started.

You will not find a consumer-grade refrigerator capable of handling the heat put out by one of these miners, let alone several.

What a refrigerator does is pull heat from the air inside it and radiate it out into the room. So even if you had one that could keep up, you'd still have all that heat in your apartment. The refrigerator would be running constantly, and probably burn about as much electricity as the miner does (this point is up for debate and based on the efficiency of the chiller pump and such, but the point stands regardless) which would increase your power demands (and therefore operating cost) substantially, and possibly overtax your electrical infrastructure which can cause fires.

Short answer is, unless you have access to an industrial freezer it's not going to work. Of course, even if you did, condensation remains a problem. There's no good way to deal quietly with the heat density of an S7. Immersion cooling is possible, but I said "good way" and to do immersion cooling right tends to get fairly complex and expensive.

Okay, let me be more explicit. I am fundamentally opposed to overlaying projects on top of the Bitcoin blockchain that have nothing to do with the transfer of currency from one entity to another in exchange for goods or services. I would not desire to be an officer in an organization, nor would I desire to support an organization, seeking to usurp Bitcoin resources for other purposes. I don't know what DAO is and I don't care about smart contracts or whatever Ethereum is doing and this is literally the first time I've ever seen the phrase "colored coins protocol".

But you're welcome to provide details about just what the heck you're talking about, if you want to try and convince me otherwise.

Yeah, but if you try to run Avalon6 version of cgminer on a Bitmain product it's not going to work. Each miner has its own driver for cgminer core, and Bitmain's code hasn't really been that good in about two years. The hardware doesn't care what it's hashing as long as the data is fed in in the right format. Pool issues, vardiff issues, shitcoin issues, all that is software.

I changed the price to $420. I got to thinking about it, and the $430 I'd set was for a guy using PayPal. Their fees would end up eating about $20 off that, so adding $10 splits the fees evenly between me and the buyer. Course adding $10 also incentivises buyers to use BTC when possible, which is handy because Bitmain doesn't take PayPal and the coins for the order have to come from somewhere.

Looks like Round 2 units are still sitting in a DHL warehouse in Ohio, where they've been since Saturday night. I've already heard from them and given them updated customs info and expected to see the truck this morning but according to tracking they're still in Ohio. Sorry, batch 2 guys. I'll stick with UPS for shipping. The prices between the two weren't that different, and I wanted to test which had the better service. I've received express packages from China in under 30 hours (even account for time zone differences) via DHL before.

Pretty unlikely to be hardware-related, if it was on the S5. BM1384 have no problem working with vardiff 1.

I'm sorry, but you lost me. Somewhere between "Blockchain is about connecting people. money transfer transactions is just  sample" and "We can implement a cooperative DAO like protocol on top of Blockchain". I always thought the Bitcoin blockchain was about money transfer transactions alone, and trying to shoehorn something else on top of it (like 21's microtransaction-based services stuff they were touting last year) is a waste of time, space and energy. Or are you suggesting to start over with a new blockchain-based thing that ends up in some sort of competition with Bitcoin?

If you bought from someone that wasn't me, you should contact the seller directly when you have problems.

Sounds like probably an ASIC issue, if it's not submitting shares. If it's registering a berjillion errors, it's definitely an ASIC issue. if it's submitting shares but the light is stuck white, it's an issue with the blinker circuit.

Did we ever get an answer to that question? I know the going theories that he was dead in a hole, or hiding out on a private island.

Current cost is $420 at time of transaction. I'm not going to set a BTC price because every time I do it changes. Honestly I should have done that from the start. I can get one from Bitmain for $373 shipped, plus reshipping within the US and a bit of work and testing. I'm using some of the overhead to buy a spare miner or two in case something doesn't pass muster, which is always possible.

Round one has arrived and are currently running smoothly. Round 2 are in transit, hopefully to arrive tomorrow. I'm okay with a Round 3 if folks want to start chipping in. I'm starting them out at 620mV cold and running 450MHz (2TH, ~420W); any boards which aren't stable at 620 get bumped up to 630mV. Any boards not stable at 630mV get rejected. So far I've had two boards out of 16 not stable at 620mV, and one appears to be stable at 630mV.

If you specifically want a different setting (say, optimized for a specific hashrate or power consumption) let me know and I'll try for it.

All the miners are testing pointed to Eligius on the 1ColdS7 address so anyone can see how they're behaving.

To run 600MHz stable, you're probably going to want at least 640mV, which will cost you more like 900W. Still a fair sight better than stock. The best you could feasibly do at 620 is about 550MHz for 3.7TH at 800W but maybe that's the setpoint you're talking about.

Pictures are up showing where to plug in the programmer.

There is no way to do this with an Avalon6, just like you can't do it on a 162-chip S7.

Yes, that looks like what you need. If you get one with a ribbon cable it'll make hooking it up to your miner boards easier.

I did take some pictures when I worked up the first-round LNs today. I'll get those added to the explanation post.

Also, based on your TX dump and what I'm seeing in your kernel log (confirming with my own), I'd guess the "AB B0 0E D4" is the command for adjusting voltage, where the third byte (0E) is the volt-setting command. I'll test this later on a factory-firmware board.

So what, we should start the "buy Sidehack an S9 he can destroy without remorse" fund?

Right, I recall now you did already post that. Sorry.

Huh, interesting. Yeah, my S7LN was set to right around 10V (0x1B, which measured out as 10.16V on my DMM) which is definitely low for a stock S7 designed for 650 or 700MHz.

Interesting that it saves the value. Maybe they do some initial testing with their software-voltable firmware to bin the boards, and since the value they settle on is stored on the board they can then hook them up with the not-software-voltable firmware they give to customers.

Do you have a dump of the TX data that gets sent to the board when it lights up, that looks like what sets the voltage? It would be interesting to write a little program for the BBB to spoof that, but I'm not sure whatall they do for UART multiplexing on the IO board. And since their cgminer source hasn't been updated since the S5, which I don't know if it's compatible...

I did some math last week which put a 620mV/550MHz S7 (3.7TH/800W) bought for $300 at pretty much the same breakeven time as one S9, and you could buy seven of them for the same money.

I'm definitely going to try and find time in the near future to get back around to TypeZero prelim design work, so on the off-chance that I do get access to new chips I'll already have a good framework in place.

Round 1 machines are in and I've worked them up. So far of the twelve boards between six machines, 11 of them look to be stable at 650/620 and one needed to be stepped up to 660/630 to run 450MHz.
They're all cooking now. If I don't see any issues by about this time tomorrow they'll all get shipped out to Round 1 customers.

And another about 4% to PayPal. Set the price 1/7 higher than you want to actually get.