Butterfly Labs invests heavily in high speed production equipment

[michaelmclees]

But carrying the analogy, would Enigma say that what you're doing in your garage is simply not enough and you need to get a million dollars worth of tools and equipment from AMG?  That's the analogy.

I don't know if he's correct is my point.  Assuming that he is correct if the average BFL customer will settle for nothing less than 100% perfect, is he still correct if they are willing to restart their machines every week or so?

You can carry most analogies too far, and I think we can see the breaking point from here.

Walk into your average small dealership service department and see the equipment they have.  Then walk into a small but sustainable garage.  You'll see that people can get work done in both, but there's a lot less bodging in the dealership.  In the small garage, one employee has to wear multiple hats (hey, I do startups, that's how we roll) which decreases depth of knowledge and ability to foresee potential problems but minimises cost.  This can result in a bunch of "oh shit" moments where valuable lessons are learned that would have been avoided by spending hundreds of thousands or millions of dollars on a specialised piece of equipment and at least one trained tech.

I guess I just want to ask straight out at this point.  Will BFL be able to put out workable products without making the investments that Enigma is saying are required?

[1715142550]

1715142550

[1715142550]

1715142550

[Inaba]

Well... I'm having trouble with the question.  If the answer were no, why would we be doing it?  The answer must necessarily be yes.  Sure, we could turn out to be wildly out of our depth, but we are not basing our first batch of products on our new equipment (it's still being done by our current facility) and if something goes wildly wrong and our quality is not up to snuff going forward, we still have the current houses in place to manufacture our equipment, so it's not really a problem for the end user/customer in either event.

[michaelmclees]

If the answer were no, why would we be doing it?  The answer must necessarily be yes.

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

So the answer turns out to be... Enigma is full of it.

Response Enigma?

[Syke]

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

Yeah, but why? Every corner in Shenzhen has the team/equipment to do it quicker and cheaper.

[abeaulieu]

If the answer were no, why would we be doing it?  The answer must necessarily be yes.

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

So the answer turns out to be... Enigma is full of it.

Response Enigma?

Enigma's points are valid concerns. They may be slightly over-worrisome, but these are things that BFL should be thinking about. Of course their "lifetime warranty" now means that these are things that the customers have to worry less about .

[squeept]

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

Yeah, but why? Every corner in Shenzhen has the team/equipment to do it quicker and cheaper.

Because they take a month off right when these things will really ramp up in delivery. And overseas shipping and dealing with customs is a complete nightmare.

Also, I like that enigma ignored my response, quoted it, then became even more erect.

[PuertoLibre]

But carrying the analogy, would Enigma say that what you're doing in your garage is simply not enough and you need to get a million dollars worth of tools and equipment from AMG?  That's the analogy.

I don't know if he's correct is my point.  Assuming that he is correct if the average BFL customer will settle for nothing less than 100% perfect, is he still correct if they are willing to restart their machines every week or so?

I think Enigmas point is that there are many challenges and quality issues to work through.

He feels in his opinion that the commercial components they bought are inadequate to truly provide a high level of quality control and internal fabrication that BFL sought to aquire

All the other technical jargon supports his opinion. (much appreciated technical jargon!)

[PuertoLibre]

But carrying the analogy, would Enigma say that what you're doing in your garage is simply not enough and you need to get a million dollars worth of tools and equipment from AMG?  That's the analogy.

I don't know if he's correct is my point.  Assuming that he is correct if the average BFL customer will settle for nothing less than 100% perfect, is he still correct if they are willing to restart their machines every week or so?

You can carry most analogies too far, and I think we can see the breaking point from here.

Walk into your average small dealership service department and see the equipment they have.  Then walk into a small but sustainable garage.  You'll see that people can get work done in both, but there's a lot less bodging in the dealership.  In the small garage, one employee has to wear multiple hats (hey, I do startups, that's how we roll) which decreases depth of knowledge and ability to foresee potential problems but minimises cost.  This can result in a bunch of "oh shit" moments where valuable lessons are learned that would have been avoided by spending hundreds of thousands or millions of dollars on a specialised piece of equipment and at least one trained tech.

Thats what Enigma was aiming to put down in his points/posts.

[PuertoLibre]

If the answer were no, why would we be doing it?  The answer must necessarily be yes.

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

So the answer turns out to be... Enigma is full of it.

Response Enigma?

No, not even close....

Enigma mentioned plenty of real points that someone working at one of those fabrication plants would recognize as "an issue". He points out that more demanding hardware fabrication projects would need better hardware than those cited by BFL.

Assuming that BFL hired a good consultant that understood their current and future needs, there should be no problems. If the consultant didn't do a good job, then BFL just spent about 50k on inadequate equipment. (Thats Enigmas [professional] opinion)

Though Enigma does not know what techniques are employed to build the project that BFL is working with, so he does not ultimately know (for sure) if BFLs in-house equipment is adequate for the job. He only states that it would be a poor choice for most jobs hes worked on.

Again, it all depends on whether BFL purchased the right equipment for the project or if it will be inadequate for the task.

They (BFL) could be using very rudimentary fab techniques and therefore there is no issue. Or they could have a "Oh Sh*t!" or "Doh!" moment when they try to do the fab in-house. Hopefully the consultant or advice they were given was good advice...then there would be no problems.

------------------------

As BFL_Josh/Inaba said, if there is a problem it won't show up in the first batch (that is being done by a separate fab house). If they do it for the second batch and there are problems, then they will default to plan B and use a separate fab house if the equipment doesn't do what they hope it will do.

[Syke]

From my perspective, putting the emotional issues (including history) aside, it's completely feasible that a company could produce ASIC-based mining systems without the investments Enigma mentioned.  You or I could do it in our garages.

Yeah, but why? Every corner in Shenzhen has the team/equipment to do it quicker and cheaper.

Since my quote was lifted without context, I said it's feasible.  I didn't say it was profitable or even a good idea for a company to do.  Hell, look at all the Arduino kits that eight-year-olds can build.

Exactly. So why is BFL trying to go down that path? Why not let the pros do it?

[Keefe]

It all depends on relationships.  They may not have the face time in China to work out all the kinks.  They may have to work through an intermediary who drives up costs.  They may be unhappy with various stages, or all stages, of the work that is done overseas.

I have no idea why they've decided to take on specific parts of their assembly themselves, but we can all guess.  Maybe Chinese business ethics prevents them from sanding off all identifying marks on chips stolen from BTCFPGA's fab.

I've been fortunate to work in tech-heavy locations, but I've still had to travel to China to kick off relationships, projects, etc., to make sure things work out OK.  When we haven't had someone on the ground, we've gone back and forth with revised tech packs, killing time to market and margins.

ROFL!

No offence meant toward ngzhang and crew. It's just that I've done business with other Chinese who act like it's totally normal to recycle old computer parts and sell them as "new" as long as they're cleaned up enough to look new, and sometimes even if obviously used. I have a feeling that's quite common over there.

[ChipGeek]

A 6 foot long oven can't (literally, impossible) have enough zones to form a proper reflow profile for a mixed technology/mixed thermal mass PCBA.  We use 14 zone reflow ovens that are over 30 feet long, and there are times we wish we had bigger ones.  Longer ovens with more zones allow for better profiles on boards that have mixed thermal-mass areas.  Run an Oven Mole though a 6 foot oven with the thermocouples attached to different parts of the circuit board and you'll see how incredibly uneven the heating will be in a short oven.

I understand your point but consider this:  What if BFL is willing to sacrifice throughput?  Couldn't they park the boards in the middle and create whatever temperature profile they want? 

For production assembly of some products, we use an assembly house with a very long oven (I do not know how many feet it is).  For others, we use an assembly house with similar equipment to what it appears BFL has purchased.  For some in-house prototypes, we use a small commercial oven that is roughly the size of a household oven.  We place the board inside, set whatever temperature profile we want on the computer then wait for the cycle to finish. Granted this is VERY slow by normal production standards.  But if BFL was able to produce 50 to 100 boards per day, that would probably be enough.  Remember, BFL is not making Android smart phones here.  The total volume is not huge compared to industry standards.

[bobitza]

Exactly. So why is BFL trying to go down that path? Why not let the pros do it?

Because they will have some sort of control if something went wrong. Because if China decides it's the Year of the Dragon 1 month holiday, that means 1 month with no work being done. Etc.

And you can see how people are screaming if there are delays in deliveries.

[squid]

Exactly. So why is BFL trying to go down that path? Why not let the pros do it?

Because they will have some sort of control if something went wrong. Because if China decides it's the Year of the Dragon 1 month holiday, that means 1 month with no work being done. Etc.

And you can see how people are screaming if there are delays in deliveries.

Yea, I am sure China shuts everything down for 1 month. Makes sense.

[Tinua]

Exactly. So why is BFL trying to go down that path? Why not let the pros do it?

Because they will have some sort of control if something went wrong. Because if China decides it's the Year of the Dragon 1 month holiday, that means 1 month with no work being done. Etc.

And you can see how people are screaming if there are delays in deliveries.

Yea, I am sure China shuts everything down for 1 month. Makes sense.

During the Chinese New Year, most of the workers go home to their families, who live hundreds of miles away! During this time, many factories do not have much output! Thus, only large customers will be served!

[squeept]

Not only are they shut down for a full month, but they still accept orders, so when they start back up, they're bogged down for a long time.

[abeaulieu]

Exactly. So why is BFL trying to go down that path? Why not let the pros do it?

Because they will have some sort of control if something went wrong. Because if China decides it's the Year of the Dragon 1 month holiday, that means 1 month with no work being done. Etc.

And you can see how people are screaming if there are delays in deliveries.

Yea, I am sure China shuts everything down for 1 month. Makes sense.

During the Chinese New Year, most of the workers go home to their families, who live hundreds of miles away! During this time, many factories do not have much output! Thus, only large customers will be served!

This is what haunted BFL for the original single release and it probably what they're trying to hedge against now by acquiring some equipment for internal assembly.

[MrTeal]

Yeah, think of it like a couple week long Christmas. If you do much ordering from China, get used to hearing they're a two week lead time if they can get it out before New Year's. If not, it'll be a couple months.

As for BFL's high speed equipment, looking at their boards it seems there's a single small BGA, with the rest of the chips either being QFN or leaded. That makes the requirements for assembly and chip inspection much simpler than they would have even had for their FPGA singles.

[abeaulieu]

As for BFL's high speed equipment, looking at their boards it seems there's a single small BGA, with the rest of the chips either being QFN or leaded. That makes the requirements for assembly and chip inspection much simpler than they would have even had for their FPGA singles.

Good point.
That BGA might only exist in the MiniRig cards for communication purposes too. It might be a no-populate for the rest, who knows.

[squeept]

I almost lost my business the first time I experienced Chinese New Year.