Author

Topic: [ANN] US/North American Bitfury sales NEW STOCK ***NOW SHIPPING*** - page 208. (Read 576785 times)

hero member
Activity: 574
Merit: 501
Put me down on the list of buyers for "reworkable" boards at a reasonable price.  I'd like 30, please..
hero member
Activity: 826
Merit: 500
Crypto Investor ;) @ Farmed Account Hunter
Crossing my fingers that 19 boards after my kit  are 5% under Dave's threshold. Lol
Just kidding.
I hope everyone gets theirs on the first shot.
vip
Activity: 472
Merit: 250
I'd buy 'bad' boards too :-)

Hah!  I was just talking to Niko about selling reworkable boards to "advanced hobbyists" out there.  We'll see what makes sense...
sr. member
Activity: 490
Merit: 255
I'd buy 'bad' boards too :-)

Sounds like a new product idea Dave... Mystery broken boards delivering between 0 and 20GH per board for a significant discount!  Hopefully available in august !!11! Wink
hero member
Activity: 826
Merit: 500
Crypto Investor ;) @ Farmed Account Hunter
Thanks for the updates, Dave!  We're all anxiously awaiting/anticipating the start of August shipping...   Just think, if you deliver by August 30th, you'll be the first ASIC manufacturer with available, substantial hashrate, product to have delivered early, ever.

Another first: ...not to have premined customer hardware before shipping.


Though we *will* test them and pull out bad boards.  Even with a 98% yield on the chips, it still means 30% of all boards have a chance of a bad chip Sad  Individual bad chips won't be much of a problem, but if a board has a big group of bad ones or one that doesn't pass SPI down the chain, it will be replaced by a good card before shipping.  Our hand-built full rig prototype has only 10 chips not showing up to the party, but there are several others hashing at subpar rates (and the thing is still producing over 400G).

To determine this we will test them for at least long enough to see where the high error rates and/or zero hashing chips are.  This is clearly visible within a few hours.  As we get more familiar with the boards we can tighten up our test time.  This will be a major bottleneck IMHO.

I know some of you will say, "Please don't test mine" and/or "Please send your test hashes to these pool creds", but I just can't support doing that, sorry!

1st dibs on buying "bad"  boards Smiley


Damn, I was thinking the same thing Grin

If they are August built and Dave reasonably prices them, then I am game. Smiley
The only way I can afford this stuff at this point. Lol
legendary
Activity: 910
Merit: 1000
I'd buy 'bad' boards too :-)
full member
Activity: 238
Merit: 100
Thanks for the updates, Dave!  We're all anxiously awaiting/anticipating the start of August shipping...   Just think, if you deliver by August 30th, you'll be the first ASIC manufacturer with available, substantial hashrate, product to have delivered early, ever.

Another first: ...not to have premined customer hardware before shipping.


Though we *will* test them and pull out bad boards.  Even with a 98% yield on the chips, it still means 30% of all boards have a chance of a bad chip Sad  Individual bad chips won't be much of a problem, but if a board has a big group of bad ones or one that doesn't pass SPI down the chain, it will be replaced by a good card before shipping.  Our hand-built full rig prototype has only 10 chips not showing up to the party, but there are several others hashing at subpar rates (and the thing is still producing over 400G).

To determine this we will test them for at least long enough to see where the high error rates and/or zero hashing chips are.  This is clearly visible within a few hours.  As we get more familiar with the boards we can tighten up our test time.  This will be a major bottleneck IMHO.

I know some of you will say, "Please don't test mine" and/or "Please send your test hashes to these pool creds", but I just can't support doing that, sorry!

1st dibs on buying "bad"  boards Smiley


Damn, I was thinking the same thing Grin
sr. member
Activity: 420
Merit: 250
   What happen if Hboard sold out on site and are you going to restock?
hero member
Activity: 826
Merit: 500
Crypto Investor ;) @ Farmed Account Hunter
Thanks for the updates, Dave!  We're all anxiously awaiting/anticipating the start of August shipping...   Just think, if you deliver by August 30th, you'll be the first ASIC manufacturer with available, substantial hashrate, product to have delivered early, ever.

Another first: ...not to have premined customer hardware before shipping.


Though we *will* test them and pull out bad boards.  Even with a 98% yield on the chips, it still means 30% of all boards have a chance of a bad chip Sad  Individual bad chips won't be much of a problem, but if a board has a big group of bad ones or one that doesn't pass SPI down the chain, it will be replaced by a good card before shipping.  Our hand-built full rig prototype has only 10 chips not showing up to the party, but there are several others hashing at subpar rates (and the thing is still producing over 400G).

To determine this we will test them for at least long enough to see where the high error rates and/or zero hashing chips are.  This is clearly visible within a few hours.  As we get more familiar with the boards we can tighten up our test time.  This will be a major bottleneck IMHO.

I know some of you will say, "Please don't test mine" and/or "Please send your test hashes to these pool creds", but I just can't support doing that, sorry!

1st dibs on buying "bad"  boards Smiley
vip
Activity: 472
Merit: 250
Thanks for the updates, Dave!  We're all anxiously awaiting/anticipating the start of August shipping...   Just think, if you deliver by August 30th, you'll be the first ASIC manufacturer with available, substantial hashrate, product to have delivered early, ever.

Another first: ...not to have premined customer hardware before shipping.


Though we *will* test them and pull out bad boards.  Even with a 98% yield on the chips, it still means 30% of all boards have a chance of a bad chip Sad  Individual bad chips won't be much of a problem, but if a board has a big group of bad ones or one that doesn't pass SPI down the chain, it will be replaced by a good card before shipping.  Our hand-built full rig prototype has only 10 chips not showing up to the party, but there are several others hashing at subpar rates (and the thing is still producing over 400G).

To determine this we will test them for at least long enough to see where the high error rates and/or zero hashing chips are.  This is clearly visible within a few hours.  As we get more familiar with the boards we can tighten up our test time.  This will be a major bottleneck IMHO.

I know some of you will say, "Please don't test mine" and/or "Please send your test hashes to these pool creds", but I just can't support doing that, sorry!
hero member
Activity: 574
Merit: 501
Thanks for the updates, Dave!  We're all anxiously awaiting/anticipating the start of August shipping...   Just think, if you deliver by August 30th, you'll be the first ASIC manufacturer with a substantial hashrate product to have delivered early, ever.
vip
Activity: 472
Merit: 250
If one were to add additional H boards - does this alter their place in the queue?  Thanks.

No, your original position is never affected by later ordering.  If we (ever) have enough quantity on hand to fulfill all the orders in the delivery month, we will combine shipping wherever possible.
vip
Activity: 472
Merit: 250
I'm leaning towards having the new M-board design made here in the states on a quick turn.  If I can't make that happen, I'll supply an appropriate cable so we don't have any confusion.

The first M-board was not made for ease of use or
something but to be as low-cost as possible for the
100TH. That's why the Keystone terminals were chosen,
they were the next cheapest after soldering bare wires
to the board:)

What you can do is take a bunch of the thin wires
coming from the PSU, bundle them together and
cut them. Then solder a thick, 2 mm^2 or so wire to
the bundle, connect that with a crimp terminal to
the Keystone connector. Leave the PSU wires as short
as possible and run the nice thick wire thru your rig.
Insulate the solder joints of course:)

intron

Thanks Intron - I have no problem with this for 100TH mine, but I think I want to be very careful with the rigs that ship to customers.  Does anyone know of a plastic safety cover for these Keystone terminals?  I'd like to get a bunch for the mine...

How many amps and volts will be required by the fully populated board? How much for M board and How much for each H board?

Tytus is still playing with software settings that can adjust clock rate (within fixed bounds).  I argue that achieving 400G computational hashrate meets spec.  He argues that we must deliver 400G nonce rate (excludes errors) and so has turned them up a bit.  This produces a 440G hashrate with 414G nonce rate (385G reported by Slushpool), but power at the wall is measured 10% higher at 330w (.825w per Gh/s total rig power).  Since 330w is after the PSU losses, actual board wattage is somewhat less than that.  I figure the board is pulling about 25A @ 12VDC.

I have no idea yet how well the chips will hold up as they are clocked higher.  What we do know is that the hotter they get the more power they pull, the more errors they throw.  However they like to run best in 45C range, so over cooling them doesn't help.  BitFury is running his chips at much higher clocks with heatsinks and fans and having no failures.  He also found a block already so we know these are lucky chips!




newbie
Activity: 40
Merit: 0
If one were to add additional H boards - does this alter their place in the queue?  Thanks.
vip
Activity: 472
Merit: 250
i have a question about paying with a credit card, can the payment be broken on 2 cards?
my highest limit is 7500 and i would like to order the 400gh.

Thanks

You can call your CC company and ask for a higher limit.

That's what I did, lol.

Though I haven't yet been able to pay as I haven't been contacted yet?? I hope that occurs soon because my local currency is tanking against us$ right now - so it's going to start really hurting soon.
Sorry Stack!  The credit card processing thing has become a major pain.   I'll have it running as soon as possible, then we will contact everyone who is waiting to pay.
hero member
Activity: 988
Merit: 1000
I'm leaning towards having the new M-board design made here in the states on a quick turn.  If I can't make that happen, I'll supply an appropriate cable so we don't have any confusion.

The first M-board was not made for ease of use or
something but to be as low-cost as possible for the
100TH. That's why the Keystone terminals were chosen,
they were the next cheapest after soldering bare wires
to the board:)

What you can do is take a bunch of the thin wires
coming from the PSU, bundle them together and
cut them. Then solder a thick, 2 mm^2 or so wire to
the bundle, connect that with a crimp terminal to
the Keystone connector. Leave the PSU wires as short
as possible and run the nice thick wire thru your rig.
Insulate the solder joints of course:)

intron

Thanks Intron - I have no problem with this for 100TH mine, but I think I want to be very careful with the rigs that ship to customers.  Does anyone know of a plastic safety cover for these Keystone terminals?  I'd like to get a bunch for the mine...

How many amps at 12 volts will be required by the fully populated board? How much for M board and How much for each H board?
sr. member
Activity: 420
Merit: 250
full member
Activity: 238
Merit: 100
i have a question about paying with a credit card, can the payment be broken on 2 cards?
my highest limit is 7500 and i would like to order the 400gh.

Thanks

You can call your CC company and ask for a higher limit.

That's what I did, lol.

Though I haven't yet been able to pay as I haven't been contacted yet?? I hope that occurs soon because my local currency is tanking against us$ right now - so it's going to start really hurting soon.
sr. member
Activity: 434
Merit: 250
Hey guys, trying to compile list of US/NA October orders, appreciate if you can contribute. link in signature.
Jump to: