NanoFury Project - Open Source Design - page 3.

adib

sr. member

Activity: 368

Merit: 250

A fab would have the resources to build like 100 chip binning machine

but not even bitfury(guess from their calculations its not worthed...but they are screwing the endusers an retailers this way) itself tests the chips so its unlikely.

On a sidenote bitfury chips are cool but a pain to work with, all it takes is a chip to go nuts and it brings the whole chain down.

A1 seems kool but they are not so easy to work with.

TheRealSteve

hero member

Activity: 686

Merit: 500

FUN > ROI

Well, sure, if you're hand-soldering every board for a small run instead of having a bunch put together at a fab, then you might as well test-before-solder

adib

sr. member

Activity: 368

Merit: 250

Its worthed if you have the time, i soldered about 200 chips manually, and about 10 are partial duds and 2 are dead, stopping the whole chain from going.
10 duds is about 25ghs on the negative side if you give them 0.85v + even more on the rev2 chips.

Whats bothering me is why the rev2 chips dont ramp down if they do no work, an h-card with rev1 chips would go to about 8W while the same card would stay at 20W+ if there are rev2 chips on it.

Anyone know?

TheRealSteve

hero member

Activity: 686

Merit: 500

FUN > ROI

Take a BitFury board design, change layout for a ZIF socket ($$), and off you go? Wink

( I'd have to wonder if it's worth the bother in time and effort to actually bin them manually vs saying that hash rates may vary and just deal with complete duds on a case-by-case scenario. )

Taugeran

hero member

Activity: 658

Merit: 500

CCNA: There i fixed the internet.

Quote from: intron on June 06, 2014, 01:06:54 AM

Quote from: Taugeran on June 05, 2014, 06:49:59 PM

Quote from: intron on June 05, 2014, 08:36:04 AM

Quote from: adib on June 05, 2014, 07:10:53 AM

I belive its intron, if you search the alpha test thread it should be there.

Yes, it was the 'binfury' board:) We never
were able to get it functional, too busy
at that moment.

http://i.imgur.com/xCXaFsR.jpg

Willing to sell or OSHW?

It's not functional, so the board is not very
useful as it is now I guess.

sad panda

intron

sr. member

Activity: 427

Merit: 251

- electronics design|embedded software|verilog -

Quote from: Taugeran on June 05, 2014, 06:49:59 PM

Quote from: intron on June 05, 2014, 08:36:04 AM

Quote from: adib on June 05, 2014, 07:10:53 AM

I belive its intron, if you search the alpha test thread it should be there.

Yes, it was the 'binfury' board:) We never
were able to get it functional, too busy
at that moment.

http://i.imgur.com/xCXaFsR.jpg

Willing to sell or OSHW?

It's not functional, so the board is not very
useful as it is now I guess.

Taugeran

hero member

Activity: 658

Merit: 500

CCNA: There i fixed the internet.

Quote from: intron on June 05, 2014, 08:36:04 AM

Quote from: adib on June 05, 2014, 07:10:53 AM

I belive its intron, if you search the alpha test thread it should be there.

Yes, it was the 'binfury' board:) We never
were able to get it functional, too busy
at that moment.

Willing to sell or OSHW?

intron

sr. member

Activity: 427

Merit: 251

- electronics design|embedded software|verilog -

Quote from: adib on June 05, 2014, 07:10:53 AM

I belive its intron, if you search the alpha test thread it should be there.

Yes, it was the 'binfury' board:) We never
were able to get it functional, too busy
at that moment.

adib

sr. member

Activity: 368

Merit: 250

I belive its intron, if you search the alpha test thread it should be there.

Taugeran

hero member

Activity: 658

Merit: 500

CCNA: There i fixed the internet.

Quote from: pinkfloyd11 on June 05, 2014, 05:05:41 AM

I had use TTi power supply thinking that it was good.. Grin

Question:
there is a way to test bitfury chip before solder it on device?

Thanks

You mean for like functionality test. Not anything easy. One of the designers here in the forum came up with the idea of a bitfury chip binner to sort and categorize chips, as well as find dead/calamity ones but this was never seen to fruition. It still stands to reason he would license it

pinkfloyd11

newbie

Activity: 56

Merit: 0

I had use TTi power supply thinking that it was good.. Grin

Question:
there is a way to test bitfury chip before solder it on device?

Thanks

Taugeran

hero member

Activity: 658

Merit: 500

CCNA: There i fixed the internet.

Quote from: vs3 on May 30, 2014, 01:38:56 AM

Quote from: Taugeran on May 29, 2014, 12:51:11 PM

Quote from: pinkfloyd11 on May 29, 2014, 10:42:28 AM

Hello

my device work well, just exceed 1Gh/s

now which are the step for improve performance?

Thanks

Just a few notes

Why such an old version of bfgminer?
What oscillator bits are they running at? Type M to bring up management screen. Will show oscillator bits, at least in newer versions

Actually we chatted after he posted that image - seems his external power supply was the culprit (he saw voltage spikes of over 300mV). Adding a bunch of capacitors may help

What was he using to power it? (Jokingly,) a bike dynamo?

vs3

hero member

Activity: 622

Merit: 500

Quote from: Taugeran on May 29, 2014, 12:51:11 PM

Quote from: pinkfloyd11 on May 29, 2014, 10:42:28 AM

Hello

my device work well, just exceed 1Gh/s

now which are the step for improve performance?

Thanks

Just a few notes

Why such an old version of bfgminer?
What oscillator bits are they running at? Type M to bring up management screen. Will show oscillator bits, at least in newer versions

Actually we chatted after he posted that image - seems his external power supply was the culprit (he saw voltage spikes of over 300mV). Adding a bunch of capacitors may help

vs3

hero member

Activity: 622

Merit: 500

Quote from: pinkfloyd11 on May 26, 2014, 07:55:59 AM

Hello friends

this is instead the clock of non working device

Note that the signal go down but not raise 0 volt value

This means that on clock wire flow undesidered current and V DROP ON SOME RESISTANCE...

Have a suggest for found the problem?

Question1: the max core voltage must be MAX 0.95Volt also if ASIC is cooled very well?

Question2: IOREF not should be connect like here:

1) the clock picture is correct. That's exactly how it should look like. There are 3 voltage levels for the SCK:
SPI_0 = 0V
SPI_1 = 1.2V - during normal SPI operations, SCKOVR not active
SPI_1_SCKOVR = 1.6V - with SCKOVR active (the SPI signals will me muted/overridden in this case)

IOREF - can be anything you like. The only thing that matters is that it must be at least +/-50mV away from the 0 and 1 levels. Since your "1" is over 1V then connecting it to VDD (0.8V) is more than safe.

Taugeran

hero member

Activity: 658

Merit: 500

CCNA: There i fixed the internet.

Quote from: pinkfloyd11 on May 29, 2014, 10:42:28 AM

Hello

my device work well, just exceed 1Gh/s

now which are the step for improve performance?

Thanks

Just a few notes

Why such an old version of bfgminer?
What oscillator bits are they running at? Type M to bring up management screen. Will show oscillator bits, at least in newer versions

pinkfloyd11

newbie

Activity: 56

Merit: 0

Hello

my device work well, just exceed 1Gh/s
http://i.imgur.com/ZsuE6Bhl.png

now which are the step for improve performance?

Thanks

pinkfloyd11

newbie

Activity: 56

Merit: 0

Hello friends

this is instead the clock of non working device

http://i.imgur.com/9svlum8l.png
Note that the signal go down but not raise 0 volt value

This means that on clock wire flow undesidered current and V DROP ON SOME RESISTANCE...

Have a suggest for found the problem?

Question1: the max core voltage must be MAX 0.95Volt also if ASIC is cooled very well?

Question2: IOREF not should be connect like here:
http://i.imgur.com/gxNk4Rfl.png

vs3

hero member

Activity: 622

Merit: 500

Quote from: pinkfloyd11 on May 24, 2014, 05:23:55 AM

seem to work, really bad but seem to work.... Grin

http://imgur.com/nWwS8R2

Congrats!

And yes - that's how the clock is supposed to look like. When the reset/SCK_OVR is active it will be close to 1.8V and otherwise around 1.2-1.3V (the logic "1" level)

pinkfloyd11

newbie

Activity: 56

Merit: 0

seem to work, really bad but seem to work.... Grin

http://i.imgur.com/nWwS8R2.png?1

vs3

hero member

Activity: 622

Merit: 500

Quote from: pinkfloyd11 on May 22, 2014, 04:08:32 AM

Hello VS3
The schematics is the your.
The only difference is that I'm using external power supply for Vcore. However thanks to MCP2210 utility I was able to change product description instead that use fixID (look pag 21 photo).

http://i.imgur.com/1PSVrWk.png

In a while, where you will check with that kind of error?

Could you (or other that have working device)
made a sort of test of SPI using MCP2210 TERMINAL. This could help a lot also other people in debug error.

In any case BIG THANKS for your time

UPDATE: Vcore must be 0.8volt or could be also 0.6 - 0.7?

The reason why you were getting the NanoFury error was that you're most likely missing the SCK_OVR resistor(s). So, compare carefully the schematics and make sure you haven't missed some component or connection. Also, check that your components are wired and installed the way you expect.

Vcore can be as low as 0.6V and I'd recommend staying below 0.95V.

Topic: NanoFury Project - Open Source Design - page 3. (Read 75392 times)