Butterfly Labs

[jml]

user grnbrg has opened and dismantled his.   I think his name is bryan greenberg in winnipeg canada.   He has posted pics on butterfly labs site.

Can he then post something in this thread? And also this does not impact the ponzi potential. Only the shipping volume will determine if BFL is legitimate. Again, they can finance the manufacturing and shipping of previous units with new orders. As long as the queue is filled, they can continue to ship. If BFL can exhaust a large chunk of their queue while staying solvent, then they are not running a ponzi scheme. 

I will back off if units start shipping and simply wait and see.

Quoting from the site:
================
Update: I have exchanged some email with @BFL_Josh , and some things have been clarified.

My unit is probably closer to prototype than a production-quality review unit. In particular, one of the two chips in my miner is a dud, and non-functional. Production 5GH/s miners will all have 2 functional ASIC chips. Because mine has only one, it has been clocked higher, and is not quite capable of reaching 5 GH/s. But I'm happy with it.

Edit: I forgot to mention -- Shipping weight was 1.54 kilograms - about 3.4 pounds.
============

So basically BFL are selling dud ASIC's with 1 overclocked chip whilst the other chip is non functional. Hmmm, not good in terms of redundancy if the only chip fails!

[charleshoskinson]

Quoting from the site:
================
Update: I have exchanged some email with @BFL_Josh , and some things have been clarified.

My unit is probably closer to prototype than a production-quality review unit. In particular, one of the two chips in my miner is a dud, and non-functional. Production 5GH/s miners will all have 2 functional ASIC chips. Because mine has only one, it has been clocked higher, and is not quite capable of reaching 5 GH/s. But I'm happy with it.

Edit: I forgot to mention -- Shipping weight was 1.54 kilograms - about 3.4 pounds.
============

So basically BFL are selling dud ASIC's with 1 overclocked chip whilst the other chip is non functional. Hmmm, not good in terms of redundancy if the only chip fails!

There are many unanswered questions and BFL must deliver soon if they wish to maintain credibility. I will reserve judgement and wait for them to act. If they fail to act, then I will rally as many in the community who are willing and take the next step. This stuff is painful and difficult, but very necessary for a growing ecosystem.

[grue]

so the review prototypes that were sent out were for generating hype. NICE.

[charleshoskinson]

so the review prototypes that were sent out were for generating hype. NICE.

If more people believe BFL will actually ship, then they will have a spike in orders from people on the sidelines. The money can then be used to finance the original orders and continue the process. As long as the queue is filled, BFL stays in business. This is why we need an audit.

[jml]

Quoting from the site:
================
Update: I have exchanged some email with @BFL_Josh , and some things have been clarified.

My unit is probably closer to prototype than a production-quality review unit. In particular, one of the two chips in my miner is a dud, and non-functional. Production 5GH/s miners will all have 2 functional ASIC chips. Because mine has only one, it has been clocked higher, and is not quite capable of reaching 5 GH/s. But I'm happy with it.

Edit: I forgot to mention -- Shipping weight was 1.54 kilograms - about 3.4 pounds.
============

So basically BFL are selling dud ASIC's with 1 overclocked chip whilst the other chip is non functional. Hmmm, not good in terms of redundancy if the only chip fails!

There are many unanswered questions and BFL must deliver soon if they wish to maintain credibility. I will reserve judgement and wait for them to act. If they fail to act, then I will rally as many in the community who are willing and take the next step. This stuff is painful and difficult, but very necessary for a growing ecosystem.

I believe what grnberg said has answered a question in your proposed letter as it reads that the Jalapeno was not production quality. Now you need to ask is when was the prototype made and why are they shipping "prototype standard" ASIC miners?

[jml]

so the review prototypes that were sent out were for generating hype. NICE.

If more people believe BFL will actually ship, then they will have a spike in orders from people on the sidelines. The money can then be used to finance the original orders and continue the process. As long as the queue is filled, BFL stays in business. This is why we need an audit.

So then what happened to the so called FPGA financing? I thought this would already have generated profit to fund ASIC development.

[charleshoskinson]

I believe what grnberg said has answered a question in your proposed letter as it reads that the Jalapeno was not production quality. Now you need to ask is when was the prototype made and why are they shipping "prototype standard" ASIC miners?

His response gives credibility to the idea that BFL used preorder money to fund the development of the ASICs. This is a problem guys because they have an obligation to ship those preorders. Unless their margins are enormous, they will not be able to catch ahead and if there is a disruption in money inflow, BFL will not be able to purchase materials to manufacture and deliver the promised units.

An honest business would have done this with a kickstarter campaign and then we wouldn't be here today. But again, I do not have any hard evidence of these claims just a feeling  backed by the information I have received. I'm going to wait for the units to start shipping and see if the queue clears.

It is important to note that even if the product is poorly constructed, that does not denote a ponzi scheme. Just BFL's poor engineering skills. That is not a criminal act. If they knowingly developed a product off of orders and require a continuous influx of capital to service orders they have an obligation to deliver, then that would be a ponzi scheme.

So then what happened to the so called FPGA financing? I thought this would already have generated profit to fund ASIC development.

This question is why I asked about mining in my list of questions. I'd like to get an idea how much they raised in mining profit. We can instantly verify any claims of mining profit from the addresses they provide and sign from.

[jml]

So they have started to trickle prototype quality Jalaps, so what is going on with the rest of the product line? i.e Litte Single SC, Single SC, Minirig.

[charleshoskinson]

So they have started to trickle prototype quality Jalaps, so what is going on with the rest of the product line? i.e Litte Single SC, Single SC, Minirig.

I can't answer that question nor will I speculate. I'm just going to wait and inform the BTC community that I'm here to help in the event of a likely ponzi scheme.

[jml]

So they have started to trickle prototype quality Jalaps, so what is going on with the rest of the product line? i.e Litte Single SC, Single SC, Minirig.

I can't answer that question nor will I speculate. I'm just going to wait and inform the BTC community that I'm here to help in the event of a likely ponzi scheme.

That was just another question to add to your list.

[charleshoskinson]

That was just another question to add to your list.

Your question presupposes they are going to ship prototype ASICs to their customers. As it stands, it would unfair to ask a company to reveal trade secrets about their PCB and ASIC design. It however is fair game to ask when the RTM prototype was finalized  and what reliability testing they have performed. Also how does this scale? If a 5GH/s device is consuming 40 watts of power. How much power will the 1500 GH/s consume? With linear scaling that would be 60 kilowatts.

[jml]

That was just another question to add to your list.

Your question presupposes they are going to ship prototype ASICs to their customers. As it stands, it would unfair to ask a company to reveal trade secrets about their PCB and ASIC design. It however is fair game to ask when the RTM prototype was finalized  and what reliability testing they have performed. Also how does this scale? If a 5GH/s device is consuming 40 watts of power. How much power will the 1500 GH/s consume? With linear scaling that would be 60 kilowatts.

My question addresses the fate of the later models. My suspicion is that due to unforeseen increased exhaust heat, they will have to look at newer cases/passive heat dissipation, hence further delaying the shipping of those models. The jalap was intended to be a "coffee warmer" size ASIC and it is now the same size of the Bitforce SC.

[charleshoskinson]

My question addresses the fate of the later models. My suspicion is that due to unforeseen increased exhaust heat, they will have to look at newer cases/passive heat dissipation, hence further delaying the shipping of those models. The jalap was intended to be a "coffee warmer" size ASIC and it is now the same size of the Bitforce SC.

How many watts of power have to go through their rigs? Also have they done any reliability testing at all? These things will be on 24/7 any downtime is lost money.

[Logik]

My question addresses the fate of the later models. My suspicion is that due to unforeseen increased exhaust heat, they will have to look at newer cases/passive heat dissipation, hence further delaying the shipping of those models. The jalap was intended to be a "coffee warmer" size ASIC and it is now the same size of the Bitforce SC.

I follow the updates obsessively, if you're interested in a random stranger's opinion based on careful analysis of said updates:

There are 3 main products. Jalapeños, Single and Mini Rig. There are power issues in design of the ASIC chip itself.

- Jalapeños can ship with the existing chips, even with their power issues due to the low specs.
- The Bitforce Single SC is not feasible to construct with the original specs, without having new chips produced at the foundry (which will take a long time). Of course if BFL came out and said this then people would be extremely upset, so they will not say this directly or admit it to customers. What they will do is ship multiple products to make up for the difference. For example if you ordered a Single (that's 60GH/s) then they will in fact ship you 2x30GH/s at a higher power rate. That way they will say they are shipping "Single", but they are just playing around with the definition of their products so they can use the faulty chips instead waiting for the new ones.
- The MiniRig is DOA with current power specs. Not happening.

BFL are very clever in managing perceptions, especially with the power issues. They distract from the ASIC power problem (a very large problem) with information about PCB power problems (a small problem). For example you will hear again and again and PCB redesigns and PCB power components and firmware, etc, etc and this makes it seem they are fixing the power issues, but the real issue can't be fixed without new chips. Improving power aspects on the PCB just paints over this! It's kind of funny to watch.

[jml]

My question addresses the fate of the later models. My suspicion is that due to unforeseen increased exhaust heat, they will have to look at newer cases/passive heat dissipation, hence further delaying the shipping of those models. The jalap was intended to be a "coffee warmer" size ASIC and it is now the same size of the Bitforce SC.

How many watts of power have to go through their rigs? Also have they done any reliability testing at all? These things will be on 24/7 any downtime is lost money.

This was the same question I was asking a UK based company called cloudhashing.com. They state they bought 6 minirigs and their business model is entirely dependent on whether BFL delivers. They also have a SLA between them and a mining farm within the same city as BFL are (I already expressed that this kind of centralisation is not good). I asked mainly questions to the PR guy of cloudhashing surrounding reliability and fault tolerance as being an area I am well aware of. So basically their answer was that they have no MTBF or MTTF values or documentation and they depend exclusively on BFL parts to service the rigs if there is a fault and to repair them within 24hrs.

[charleshoskinson]

This was the same question I was asking a UK based company called cloudhashing.com. They state they bought 6 minirigs and their business model is entirely dependent on whether BFL delivers. They also have a SLA between them and a mining farm within the same city as BFL are (I already expressed that this kind of centralisation is not good). I asked mainly questions to the PR guy of cloudhashing surrounding reliability and fault tolerance as being an area I am well aware of. So basically their answer was that they have no MTBF or MTTF values or documentation and they depend exclusively on BFL parts to service the rigs if there is a fault and to repair them within 24hrs.

Did they just make a preorder fee or did they pay everything upfront?

[bitcoinminer]

My question addresses the fate of the later models. My suspicion is that due to unforeseen increased exhaust heat, they will have to look at newer cases/passive heat dissipation, hence further delaying the shipping of those models. The jalap was intended to be a "coffee warmer" size ASIC and it is now the same size of the Bitforce SC.

How many watts of power have to go through their rigs? Also have they done any reliability testing at all? These things will be on 24/7 any downtime is lost money.

This was the same question I was asking a UK based company called cloudhashing.com. They state they bought 6 minirigs and their business model is entirely dependent on whether BFL delivers. They also have a SLA between them and a mining farm within the same city as BFL are (I already expressed that this kind of centralisation is not good). I asked mainly questions to the PR guy of cloudhashing surrounding reliability and fault tolerance as being an area I am well aware of. So basically their answer was that they have no MTBF or MTTF values or documentation and they depend exclusively on BFL parts to service the rigs if there is a fault and to repair them within 24hrs.

This is stupid, and I hope it gets them goxxxed.  Why do people seem to think that if bitcoins are involved, they should throw common sense out the window?

"We're going to have a roller skating rink.  Our entire business is based on a delivery of skates, which we have entrusted to one company, who hasn't even sent us a sample yet".

Sounds retarded, doesn't it?

[charleshoskinson]

This is stupid, and I hope it gets them goxxxed.  Why do people seem to think that if bitcoins are involved, they should throw common sense out the window?

"We're going to have a roller skating rink.  Our entire business is based on a delivery of skates, which we have entrusted to one company, who hasn't even sent us a sample yet".

Sounds retarded, doesn't it?

Which is why I want this community to return things to commonsense land. BFL has a window to prove themselves. I could care less if they shipped something working to Dave. They took money for a product that is physically impossible to build with their ASIC design. 60KW rig? Could you imagine the cooling system? It's madness.

I mean honestly this comes to mind:

http://www.youtube.com/watch?v=g3ythpzsu18

[SgtSpike]

So they have started to trickle prototype quality Jalaps, so what is going on with the rest of the product line? i.e Litte Single SC, Single SC, Minirig.

The PCBs were designed to handle 60 watts of power draw.  While they sort of work with 180 watts of power draw at the little-single levels of hashing, it would not be in the interest of BFL to release such a beast.  They need to properly redesign the PCB to withstand the new anticipated power draw.  They can create Jalepeno units with the old boards because they are only drawing 30 watts - far below what the boards were originally meant to handle.  They are only creating a few of them because they believe the board redesign will also make the units more power efficient.  They just announced that they finished the board revisions, and should hopefully be receiving the new boards in late next week.  At that point, full-bore production of all the product lines can commence.

Make sense?

[charleshoskinson]

The PCBs were designed to handle 60 watts of power draw.  While they sort of work with 180 watts of power draw at the little-single levels of hashing, it would not be in the interest of BFL to release such a beast.  They need to properly redesign the PCB to withstand the new anticipated power draw.  They can create Jalepeno units with the old boards because they are only drawing 30 watts - far below what the boards were originally meant to handle.  They are only creating a few of them because they believe the board redesign will also make the units more power efficient.  They just announced that they finished the board revisions, and should hopefully be receiving the new boards in late next week.  At that point, full-bore production of all the product lines can commence.

Make sense?

What is my Hash per watt performance and how does this relate to the 1500 GH/s rig?