DIY PCB with AVALON: "The Quarter Stick" - Needs Help!

[loshia]

Dude,
Are you going to publish board design somewehe?
So we can make PCB?
10X

addressing your questions:

[tigerfree]

i dont think it will be open source

[loshia]

i dont think it will be open source

Why? DIY - Do it Yourself - that is exactly what i want to do
If you buy ready to go product where is DIY Part?

I really do hope that NG will publish PCB layout of Avalon + Controler + Spartan bitstream as he did for Icarus and lancelot
Buying a product is something different than DIY dude.
The joy and satisfaction for me personally will be great. And i do think i can contribute with some coding for it i can not guarantiee but i will do my best

Allten, Zefir, burnin pls comment

10X

PS: I do still think that 90+% of the guys will go for assembled and ready to mine board. Publishing the layout + code will do no harm it will bring more benefits only.
Making a git repository will give the opportunity to all of us to contribute - everyone according to it's skills 

[zefir]

Folks and miners,

we are still in the planning phase, so please be patient.

Obviously this whole idea depends on whether existing mining ASIC manufacturers are willing and able to sell their chips. So far, Avalon is the only party that announced support for it, but for weeks I am still waiting to hear back BitSyncom's response to my inquiry.

Aside from this hard dependency, here are the basic points so far:

Funding
We have enough coins to fund volume chip orders, as soon as they are available. No need for kickstarter or other types of public offerings planned. On demand, we will act as distributor for those ASICs, i.e. everyone will be able to order in smaller quantities.

Open Source
Yes, there will be at least one fully open source (HW and SW) project provided for the DIY folks of us. Since assembly requires equipment most DIY won't have at home, this might end up being an academic issue. The best take could be to offer fully assembled boards and leave integration to the miners (stacking, thermal design, housing, etc.). This is yet TBD, and since some are working full-time on this and are aiming at serious business, we still need to explore potential monetizing options.


Feel free to continue posting comments and suggestions, but as said above, as long as there is no confirmed commitment to offer ASIC chips to the public, don't be too enthusiastic in that.


Cheers,
zefir

[Bicknellski]

+1 Zefir...

Can't wait... none of us can. This is a super option hope we see this DIY ASIC happen. It would put a HUGE dent into BFL I think.

Could we raise money, kickstarter or indiegogo, then design and fab the ASIC chip? Something using older tech?

[DarkPunk]

Zefir, I'm glad to hear you plan on reselling the chips in smaller quantities.  I'm sitting on the edge of my chair waiting to be able to buy a few.

[loshia]

+1 Zefir...

Can't wait... none of us can. This is a super option hope we see this DIY ASIC happen. It would put a HUGE dent into BFL I think.

Could we raise money, kickstarter or indiegogo, then design and fab the ASIC chip? Something using older tech?

This is job to be done from few only And is not easy at all
We can dream however...

[DPoS]

I think we have enough info to set up a b2b website and take preorders from BFL, Dragon, Swedish miners, etc.

[loshia]

I think we have enough info to set up a b2b website and take preorders from BFL, Dragon, Swedish miners, etc.

What is your point dude b2b?
Is this a joke or?

[tigerfree]

i dont think it will be open source

Why? DIY - Do it Yourself - that is exactly what i want to do
If you buy ready to go product where is DIY Part?

I really do hope that NG will publish PCB layout of Avalon + Controler + Spartan bitstream as he did for Icarus and lancelot
Buying a product is something different than DIY dude.
The joy and satisfaction for me personally will be great. And i do think i can contribute with some coding for it i can not guarantiee but i will do my best

Allten, Zefir, burnin pls comment

10X

PS: I do still think that 90+% of the guys will go for assembled and ready to mine board. Publishing the layout + code will do no harm it will bring more benefits only.
Making a git repository will give the opportunity to all of us to contribute - everyone according to it's skills 

i wish they release  it and we can make our miner it would be great for Bitcoin community 

[zefir]

Could we raise money, kickstarter or indiegogo, then design and fab the ASIC chip? Something using older tech?

Observing the freaking-out of miners driven by under-supply of ASIC mining rig (peaking in lucky lottery winners selling their $1.5k device for $20k), and based on the fact that ASICMINER developed their chips with NRE of less than $200k, one might conclude that developing ones own Bitcoin ASIC is the most obvious next step.

At least I concluded recently, but after discussing with some experienced insiders, it only supposedly sounds easy. The two core issues are

Time to Market
There is an absolute minimum development time of ~6 months to be considered. That is, no matter which expertise you hire and how hard you try to negotiate with fabs - starting our mining ASIC today we would have chips in our hands for Christmas. Quite some time for numerous 'what ifs' and surprises in the Bitcoin world.

Expertise and Luck
Having only two successful ASIC manufacturers out of four taking the challenge indicates that designing a mining chip is far from being trivial. There are rants over BFL all over the forum for being incompetent idiots failing to provide what they promised long ago. Being more objective, there is no doubt they had the highest skilled experts available money can buy (and they had tons of money) - and yet they failed to meet expectations. That's because luck is an immanent component with the development of such uncommon ASICs. Those engineers in this forum believing to work error free, go cast the first stone.


tl;dr: starting a new mining ASIC from scratch is already too late and overall maybe not the best idea.

Avalon initially planned to develop only the chips and let others do the productizing. Taking the opportunity to monetize the current shortage of mining rig supply to dictate prices and recover NRE fast is a legitimate approach. But as soon as a second manufacturer enters the competition, they will gradually be forced to concentrate on their core business - and that's where we should build on top of. There are only two manufacturers of CPUs for desktop PCs worldwide, that's enough for everybody to have a computer at home, as it is enough to keep up a healthy competition with continuous technologal advances.


I expect this scenario to be reached in Bitcoin world not later than by the end of this year. Avalon might still have one or two successful batches to sell, but then things will go to normal again. That is, to what we had at the beginning with CPUs/GPUs: whoever likes to support the Bitcoin network buys and attaches rig right away. No pre-order limbo and no pre-sale lotteries.

At least thats the plan

[DPoS]

I think we have enough info to set up a b2b website and take preorders from BFL, Dragon, Swedish miners, etc.

What is your point dude b2b?
Is this a joke or?

yah.. business2business we sell ASIC to the scammers...   so i guess it is s2s

it would be good to scam them back

[loshia]

I think we have enough info to set up a b2b website and take preorders from BFL, Dragon, Swedish miners, etc.

What is your point dude b2b?
Is this a joke or?

yah.. business2business we sell ASIC to the scammers...   so i guess it is s2s

it would be good to scam them back

+1 i did not get the joke dude
Best:)

[SebastianJu]

Hm... sounds like this turns into a business... i wonder if the previous plans of $25 including 3 to 4 chips will still remain or if it goes into another direction. At least i would like if the chips are sold cheap, or with a groupbuy, the plans to create the miner are open source and the same goes for the software. So that everyone can buy the parts and create.
That some will create a business with it is clear. They will build them and sell them. Thats unavoidable.

I only wanted to mention that i would like to see that at the end you dont only have a miner that has to be bought completely built from a new established, profit oriented, company and thats the only way to get such a miner.

It only sounded a bit like it could go into that direction...

[zefir]

Hm... sounds like this turns into a business... i wonder if the previous plans of $25 including 3 to 4 chips will still remain or if it goes into another direction. At least i would like if the chips are sold cheap, or with a groupbuy, the plans to create the miner are open source and the same goes for the software. So that everyone can buy the parts and create.
That some will create a business with it is clear. They will build them and sell them. Thats unavoidable.

I only wanted to mention that i would like to see that at the end you dont only have a miner that has to be bought completely built from a new established, profit oriented, company and thats the only way to get such a miner.

It only sounded a bit like it could go into that direction...

First off: business and open source do not contradict. If you ever have the chance to visit an open source centric conference (try LinuxCon when you're nearby), go there and listen to the big guys (IBM, Google, RedHat, etc.) how they are making tons of money by giving things away for free. It is the services that you are charging for, not the technology.

Therefore: no, plans did not change. What we have in mind as business is services and support to be built on top of the open source design. Effectively, the DIY guy will be able to get all SW and HW design documents to build his own boards - but with no service, no support, no warranty. Whereas the typical customer will have to pay a premium and get ready to mine boards, including warranty, service ans support.


What we need to think of is how to prevent the inherent tendency for centralization in Bitcoin mining. Assume we open source all design documents; to build your miner you need to manufacture your PCB, source the material, and finally find an EMS partner to assemble it (which alone introduces ramp-up costs of several k€). You won't do this for 2 boards, not for 10, and also not for 100. It starts to get interesting with 1000 boards, which ultimately leads to centralization of mining power.

That is basically the challenge we are facing: how can we provide an open source approach but limit centralization tendency.

Feel free to post your ideas.

[SebastianJu]

Hm... sounds like this turns into a business... i wonder if the previous plans of $25 including 3 to 4 chips will still remain or if it goes into another direction. At least i would like if the chips are sold cheap, or with a groupbuy, the plans to create the miner are open source and the same goes for the software. So that everyone can buy the parts and create.
That some will create a business with it is clear. They will build them and sell them. Thats unavoidable.

I only wanted to mention that i would like to see that at the end you dont only have a miner that has to be bought completely built from a new established, profit oriented, company and thats the only way to get such a miner.

It only sounded a bit like it could go into that direction...

First off: business and open source do not contradict. If you ever have the chance to visit an open source centric conference (try LinuxCon when you're nearby), go there and listen to the big guys (IBM, Google, RedHat, etc.) how they are making tons of money by giving things away for free. It is the services that you are charging for, not the technology.

Therefore: no, plans did not change. What we have in mind as business is services and support to be built on top of the open source design. Effectively, the DIY guy will be able to get all SW and HW design documents to build his own boards - but with no service, no support, no warranty. Whereas the typical customer will have to pay a premium and get ready to mine boards, including warranty, service ans support.


What we need to think of is how to prevent the inherent tendency for centralization in Bitcoin mining. Assume we open source all design documents; to build your miner you need to manufacture your PCB, source the material, and finally find an EMS partner to assemble it (which alone introduces ramp-up costs of several k€). You won't do this for 2 boards, not for 10, and also not for 100. It starts to get interesting with 1000 boards, which ultimately leads to centralization of mining power.

That is basically the challenge we are facing: how can we provide an open source approach but limit centralization tendency.

Feel free to post your ideas.

Ok, sounds good. I awaited such services anyway but if it can be built on ones own its fine.

I think it will be hard to stop centralization. People could open different accounts here and order them. I think the only way to prevent it a bit would be to store the receiving address and let it automatically check. I guess it will be hard, or at least some work to come up with different addresses to receive more chips.

On the other hand... if the design will be ready and open source i bet there are guys out there that can afford 10000 chips on its own or as a group buy. They could create the miners and could mine.

Regarding the PCB... will it be that hard? I mean cant it be done with some standard home DIY PCB-Sets? I mean has the PCB to be manufactured in a fab? That would be difficult then.
Till now i thought you only need to buy the parts, download the software and start creating.

[jimmy3dita]

Hi everyone, I'm following you with much attention from Italy, and I'd like to spam the "principles" of our Arduino:
- it's fully open source
- you can clone/make/copy and improve the project as much as you want
- it's plenty of "local" shops selling the small boards, the PICs, and a whole bunch of plug&play accessories

This approach, together with a strong "DYI" philosophy, in my opinion can cut costs and time to market (think about assembly, logistics, expedition) while keeping speculators (if they aren't fainted all today) at the right distance.

This "legit" approach can be also a winner if we move on platforms like Kickstarter for fundraising, trying to get the basic machinery to jumpstart the project.

One last word about assembling/chips/etc: if I'm right we are talking about BGA components, nothing that can be done with home-made circuit boards and some soldering... nothing to do with my proposal

[segfault88]

I'm Excited. Looks good so far, I'll be watching with keen interest! I am an electronics DIY'er, so I wouldn't mind doing the reflow/soldering myself even.

[tbd]

First off: business and open source do not contradict. If you ever have the chance to visit an open source centric conference (try LinuxCon when you're nearby), go there and listen to the big guys (IBM, Google, RedHat, etc.) how they are making tons of money by giving things away for free. It is the services that you are charging for, not the technology.

Therefore: no, plans did not change. What we have in mind as business is services and support to be built on top of the open source design. Effectively, the DIY guy will be able to get all SW and HW design documents to build his own boards - but with no service, no support, no warranty. Whereas the typical customer will have to pay a premium and get ready to mine boards, including warranty, service ans support.


What we need to think of is how to prevent the inherent tendency for centralization in Bitcoin mining. Assume we open source all design documents; to build your miner you need to manufacture your PCB, source the material, and finally find an EMS partner to assemble it (which alone introduces ramp-up costs of several k€). You won't do this for 2 boards, not for 10, and also not for 100. It starts to get interesting with 1000 boards, which ultimately leads to centralization of mining power.

That is basically the challenge we are facing: how can we provide an open source approach but limit centralization tendency.

Feel free to post your ideas.

As long as the design documents sufficient to build the devices are freely available, then anyone can manufacture the devices.  That removes the centralization problem -- just like with Linux (RedHat, Ubuntu, Debian, etc.).

I believe your cost estimates are high.  Firms in the US can do smaller batches without fixed costs of several thousand dollars to get going.

[Remember remember the 5th of November]

How much will it cost to buy the chip(if it ever goes on sale) + the parts you would need?