Topic: [ANN] Bitfury is looking for alpha-testers of first chips! FREE MONEY HERE! - page 4. (Read 176664 times)
March 03, 2014, 07:59:22 AM
March 02, 2014, 05:42:13 PM
Put me in too, i have a few H-cards left and a Volcraft 3000 power meter to see how it behaves power wise. 
March 02, 2014, 05:25:35 PM
Would love to beta test
Have over 8000 pcs of rev1 IC up
Have over 8000 pcs of rev1 IC up
March 02, 2014, 12:03:20 PM
Is this 25GH per H-Card at €250?
March 02, 2014, 11:53:42 AM
The new bitfury chip will have the same pinout?
Our rev2 55nm chip is in the same 7mm punch-type QFN48 package with identical pinout to the rev1 55nm (summer 2013).Quote
There will be improvement in Ghash/s?
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
Yes, but this engineering batch will not be sold in big quantities. It's for evaluation purposes.
March 02, 2014, 10:42:30 AM
The new bitfury chip will have the same pinout?
Our rev2 55nm chip is in the same 7mm punch-type QFN48 package with identical pinout to the rev1 55nm (summer 2013).Quote
There will be improvement in Ghash/s?
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
Yes, but this engineering batch will not be sold in big quantities. It's for evaluation purposes.
Do yo have any price idea about next batches ?
March 02, 2014, 10:25:43 AM
The new bitfury chip will have the same pinout?
Our rev2 55nm chip is in the same 7mm punch-type QFN48 package with identical pinout to the rev1 55nm (summer 2013).Quote
There will be improvement in Ghash/s?
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
Yes, but this engineering batch will not be sold in big quantities. It's for evaluation purposes.
March 02, 2014, 09:03:36 AM
Is this 25GH per H-Card at €250?
March 02, 2014, 08:18:57 AM
The new bitfury chip will have the same pinout?
Our rev2 55nm chip is in the same 7mm punch-type QFN48 package with identical pinout to the rev1 55nm (summer 2013).Quote
There will be improvement in Ghash/s?
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
March 02, 2014, 07:08:52 AM
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
Can you add me to the list for samples please?
March 02, 2014, 06:10:35 AM
The new bitfury chip will have the same pinout?
Our rev2 55nm chip is in the same 7mm punch-type QFN48 package with identical pinout to the rev1 55nm (summer 2013).Quote
There will be improvement in Ghash/s?
Yes. Our preliminary tests show 25% increase in hashrate and 25% lower power consumption. We will receive new chips tomorrow and will have samples available to developers.
March 01, 2014, 04:10:54 AM
Hello
The new bitfury chip will have the same pinout?
There will be improvement in Ghash/s?
Thanks
February 25, 2014, 03:47:15 PM
After turning this:
into this: 
and this: 
into this: 
,
because there were no 0603 220k resistors in stock.
The result:
, 
, 
P.S. Thanks intron for the help and support
4 boards fixed and accepting chips
Schematic:
TOP VIEW
BOTTOM VIEW
NOTES:
- ONLY USE V2 CHAINMINER SOFTWARE AS THE SCHEMATIC IS BASED ON THAT, its actually a bank of the v2 m-board
- tried to keep air wires to a minimum (8 to be exact)
- LM2596 power module should be placed exactly where it is placed on the layout, otherwise youll have trouble in placing the H-cards, youll have some tolerance if you use flexible wires.
- Scale is 1:1 so i did not bother to place dimensions
- full resolution images here: https://www.dropbox.com/sh/yl0fov35cmv3r7p/gR4W5vdL6W
- pins on the slots are a bit tighter then they should be so youll have to bend the first 4 pins on each side inwards a bit (made them from the hip)
I have decided to stop making other version... aka the 8,12 or 16 slot versions although you can adapt as you see fit.
You can also connect 4 m-boards with 4 slots toghether like so:
full image here: https://www.dropbox.com/sh/yl0fov35cmv3r7p/gR4W5vdL6W
into this: and this: into this: , because there were no 0603 220k resistors in stock.
The result:
, , P.S. Thanks intron for the help and support
4 boards fixed and accepting chips
Schematic:
TOP VIEW
BOTTOM VIEW
NOTES:
- ONLY USE V2 CHAINMINER SOFTWARE AS THE SCHEMATIC IS BASED ON THAT, its actually a bank of the v2 m-board
- tried to keep air wires to a minimum (8 to be exact)
- LM2596 power module should be placed exactly where it is placed on the layout, otherwise youll have trouble in placing the H-cards, youll have some tolerance if you use flexible wires.
- Scale is 1:1 so i did not bother to place dimensions
- full resolution images here: https://www.dropbox.com/sh/yl0fov35cmv3r7p/gR4W5vdL6W
- pins on the slots are a bit tighter then they should be so youll have to bend the first 4 pins on each side inwards a bit (made them from the hip)
I have decided to stop making other version... aka the 8,12 or 16 slot versions although you can adapt as you see fit.
You can also connect 4 m-boards with 4 slots toghether like so:
full image here: https://www.dropbox.com/sh/yl0fov35cmv3r7p/gR4W5vdL6W
February 25, 2014, 06:35:20 AM
Update:
Got 4 chips, 3 working, 1 one i think i busted because i didnt check if the 1.8v line was ok on one of the H-cards(H-card still fixable, jut got to remove the connections as they lead to GND and make new ones on air wires ).
Out of the 5 H-cards repaired, 4 are confirmed to work.
Pictures and schematics doon.
Got 4 chips, 3 working, 1 one i think i busted because i didnt check if the 1.8v line was ok on one of the H-cards(H-card still fixable, jut got to remove the connections as they lead to GND and make new ones on air wires
).Out of the 5 H-cards repaired, 4 are confirmed to work.
Pictures and schematics doon.
Will send you some more chips to play with:)
February 25, 2014, 04:51:51 AM
Update:
Got 4 chips, 3 working, 1 one i think i busted because i didnt check if the 1.8v line was ok on one of the H-cards(H-card still fixable, jut got to remove the connections as they lead to GND and make new ones on air wires ).
Out of the 5 H-cards repaired, 4 are confirmed to work.
Pictures and schematics doon.
Got 4 chips, 3 working, 1 one i think i busted because i didnt check if the 1.8v line was ok on one of the H-cards(H-card still fixable, jut got to remove the connections as they lead to GND and make new ones on air wires
).Out of the 5 H-cards repaired, 4 are confirmed to work.
Pictures and schematics doon.
February 22, 2014, 07:07:48 AM
There is nothing to clearly associate a nonce with the work sent.
It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
The chip design is what it is and it can't be fixed for this, but the next chip (design)It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
Yes, there is a simple fix and without much CPU usage trade off.
P.S. I didn't mean to offend you and definitely do not want to start a flame war as I
https://bitcointalksearch.org/topic/optimal-firmwarehardware-design-for-mining-with-cgminer-294499
(yeah the lack of any temperature sensors anywhere sux also)
February 22, 2014, 06:54:16 AM
Sigh, people wasting their time reproducing the exact same hardware+firmware will get the exact same problem of HW errors.
Im using what i can afford and as i said im a DIY kindda guy, so i woulnt judge. and i admit i dont have much experience with those chips.
February 22, 2014, 06:22:10 AM
There is nothing to clearly associate a nonce with the work sent.
It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
The chip design is what it is and it can't be fixed for this, but the next chip (design)It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
Yes, there is a simple fix and without much CPU usage trade off.
P.S. I didn't mean to offend you and definitely do not want to start a flame war as I
February 22, 2014, 06:07:09 AM
Sigh, people wasting their time reproducing the exact same hardware+firmware will get the exact same problem of HW errors.
Other than the fact that I've reduced that as low as possible in my bab driver in cgminer, (with a CPU usage trade off) the problem is the design.
There is nothing to clearly associate a nonce with the work sent.
It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
End of problem - and my super smart nonce checking code can be removed
(yes that's a joke - there's nothing special about my nonce fix - it's just simple logic - oh and a few bugs from chainminer fixed)
P.S. KNK I don't care if you wish to reproduce my PM reply in public. Feel free to do so.
Edit: I've set the default limit in cgminer to 55 and on my BlackArrow board ... and it runs:
[Fast 0-15] => 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55
Other than the fact that I've reduced that as low as possible in my bab driver in cgminer, (with a CPU usage trade off) the problem is the design.
There is nothing to clearly associate a nonce with the work sent.
It's unbelievable how stupid that problem is and how simple it is to fix.
Fix in design is to send an ID with the work and send the ID with the nonce reply.
End of problem - and my super smart nonce checking code can be removed
(yes that's a joke - there's nothing special about my nonce fix - it's just simple logic - oh and a few bugs from chainminer fixed)
P.S. KNK I don't care if you wish to reproduce my PM reply in public. Feel free to do so.
Edit: I've set the default limit in cgminer to 55 and on my BlackArrow board ... and it runs:
[Fast 0-15] => 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55
February 16, 2014, 10:31:35 AM
An update on my m-board DIY project:
The schematic has been modified a little bit from the original design(on the fly).
The copper layer is now on the bottom, i am expecting the pieces needed to repair the power module on the H-cards next week.
Ill update the schematic also next week.
The H-card are empty, yes, im expecting some bitfury chips for testing, any donations(even bad chips)/trade/selling of chips are welcome.
Hope ill get them running.
As i understand one of these boards should be seen as 1 bank by the Rpi.
Keeping my fingers crossed and time to clean up those H-cards.
The schematic has been modified a little bit from the original design(on the fly).
The copper layer is now on the bottom, i am expecting the pieces needed to repair the power module on the H-cards next week.
Ill update the schematic also next week.
The H-card are empty, yes, im expecting some bitfury chips for testing, any donations(even bad chips)/trade/selling of chips are welcome.
Hope ill get them running.
As i understand one of these boards should be seen as 1 bank by the Rpi.
Keeping my fingers crossed and time to clean up those H-cards.
Jump to: