NanoFury Project - Open Source Design

[vs3]

Hello

just mounted mcp2210 3.3 voltage regulator and murata resonator.
...
[here MISO and MOSI are short for loopback test]

The chip was detect by Microchip Utility MCP2210: (...) but nf1_init fixID not work:

Yep, that seems to be the correct behavior.

You seem to be trying to use the MCP2210 on a non-NanoFury board (which also doesn't have the NanoFury schematic).
In this case the test does the correct thing and fails.
Here is the line that prints your message:
https://github.com/nanofury/NanoFury_Init/blob/master/nf1_init/nf_spidevc.cpp#L160

So the question is rather why do you expect this test to pass?

And also I'm curious as to what were you trying to achieve by doing all that work?
(as if you were just looking to try the "terminal" you could've shorted the two MCP2210 pins on your NanoFury board - as I mentioned earlier it will work that way even if you don't remove any components)

[1713997756]

1713997756

[1713997756]

1713997756

[1713997756]

1713997756

[pinkfloyd11]

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).

https://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?

[vs3]

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).

https://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.

[pinkfloyd11]

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

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

[vs3]

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

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]

Hello friends

this is instead the clock of non working device

https://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:
https://i.imgur.com/gxNk4Rfl.png

[pinkfloyd11]

Hello

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

now which are the step for improve performance?

Thanks

[Taugeran]

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

[vs3]

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.

[vs3]

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

[Taugeran]

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?

[pinkfloyd11]

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

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

Thanks

[Taugeran]

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

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

[adib]

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

[intron]

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.

[Taugeran]

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]

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.

https://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]

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.

https://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

[TheRealSteve]

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

( 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. )

[adib]

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?