Efudd Z-Series Fuddware 2.3 -Z11/Z11e/Z11j/Z9/Mini

[fabry1999]

What is the advantage on the next version for Z9 mini compared the original firmware ?

[efudd]

Hello Fabry1999,

That is an excellent question. I have attempted to address that in the original post, as well as on http://releases.broked.net. I'll copy the note from there to here for posterity below. In short, the Z9 mini value add is about bringing additional features like the per-hashboard overclocking feature to the mini shortly after they are made available on the Z9. This means that voltage control, fan/temperature curve control, etc. will be made available to the mini to provide you with additional control in the management of your unit(s). This has come about upon request by multiple mini owners (I was reluctant to even look at the mini), with the best path forward being to bring feature-set parity.

Just to be clear, this version is not officially supported on the mini as of this post -- that support will be coming in the next couple of days. That said, I know others have used my Z9 firmware on the minis with success, so you are welcome to try (I know that there is an additional failure condition that will retry and continue due to not finding enough ASICs, which is why I am stating that it is not currently supported).

"NOTE: I have been asked by many individuals for firmware on the Z9 mini units. I will be carrying the features for the Z9 forward to the mini units in the near future. The goal for the mini units is consistency in features and unification of the firmware across platforms as many of the features I have planned will work across all units."

Thank you,

Jason

[jewbsv]

I tried your firmware on Z9 mini, then I decided to return the old firmware, but your inscription still remains. How to remove it?
http://prntscr.com/l9z5xj

[efudd]

I tried your firmware on Z9 mini, then I decided to return the old firmware, but your inscription still remains. How to remove it?
http://prntscr.com/l9z5xj

Then you did not downgrade with a full version and I honestly have no idea how you could have possibly gotten to that state. I will try again with the factory firmware on my own in a few minutes just to verify, however.

http://releases.broked.net/antminer-z9-mini-201805262047-500m.tar.gz

That is the original batch1 factory firmware.

I know for a fact that the factory firmware flashes cleanly

I tested very possible combination of upgrade/downgrade/with and without saving configuration.

Please flash the factory image again, uncheck 'save configuration'.

Thank you,

Jason

[jewbsv]

Please flash the factory image again, uncheck 'save configuration'.

I did that. I also tried 3 firmwares including V9, you see the result.

[efudd]

Please flash the factory image again, uncheck 'save configuration'.

[/quote]

I did that. I also tried 3 firmwares including V9, you see the result.
[/quote]

Except it's simply impossible to get into that state unless your flashes are not completing properly.....

Let me try to explain:

Inside the .tar.gz, there is actually another image file.

That image, en whole, is flashed to NAND memory.
It is not flashed in part.

Code:

if [ -e uramdisk.image.gz ]; then
    md5=`md5sum uramdisk.image.gz | awk {'print $1'}`
    md5_r=`cat md5_info`
    if [ $md5 == $md5_r ];then
                flash_erase /dev/mtd1 0x0 0x100 >/dev/null 2>&1
                nandwrite -p -s 0x0 /dev/mtd1 uramdisk.image.gz >/dev/null 2>&1
                if [ -e /dev/mtd4 ]; then
                        flash_erase /dev/mtd4 0x0 0x100 >/dev/null 2>&1
                        nandwrite -p -s 0x0 /dev/mtd4 uramdisk.image.gz >/dev/null 2>&1
                fi
        else
                echo $md5 > /config/md5_error
                echo $md5_r >> /config/md5_error
                echo "Error md5! $md5 $md5_r" >> /tmp/upgrade_result
        fi
fi

This code says "completely erase the partition /dev/mtd1 and copy uramdisk.image.gz there", followed by "if there is a 2nd partition, copy it there as well", and "if the md5 check does not match, do nothing.".

No where in here is there anything that says "Copy part of the files out of an image and put them in the filesystem."

The entire image is embedded in the one file, uramdisk.image.gz -- the web pages, the cgminer, everything.

It is literally impossible to get in the state you portray unless you have a failed flash.

If you'd like to give me access to login to your machine via teamviewer, I'll be happy to take a look.

Jason

[efudd]

Please flash the factory image again, uncheck 'save configuration'.

I did that. I also tried 3 firmwares including V9, you see the result.
[/quote]

Except it's simply impossible to get into that state unless your flashes are not completing properly.....

Let me try to explain:

Inside the .tar.gz, there is actually another image file.

That image, en whole, is flashed to NAND memory.
It is not flashed in part.

Code:

if [ -e uramdisk.image.gz ]; then
    md5=`md5sum uramdisk.image.gz | awk {'print $1'}`
    md5_r=`cat md5_info`
    if [ $md5 == $md5_r ];then
                flash_erase /dev/mtd1 0x0 0x100 >/dev/null 2>&1
                nandwrite -p -s 0x0 /dev/mtd1 uramdisk.image.gz >/dev/null 2>&1
                if [ -e /dev/mtd4 ]; then
                        flash_erase /dev/mtd4 0x0 0x100 >/dev/null 2>&1
                        nandwrite -p -s 0x0 /dev/mtd4 uramdisk.image.gz >/dev/null 2>&1
                fi
        else
                echo $md5 > /config/md5_error
                echo $md5_r >> /config/md5_error
                echo "Error md5! $md5 $md5_r" >> /tmp/upgrade_result
        fi
fi

This code says "completely erase the partition /dev/mtd1 and copy uramdisk.image.gz there", followed by "if there is a 2nd partition, copy it there as well", and "if the md5 check does not match, do nothing.".

No where in here is there anything that says "Copy part of the files out of an image and put them in the filesystem."

The entire image is embedded in the one file, uramdisk.image.gz -- the web pages, the cgminer, everything.

It is literally impossible to get in the state you portray unless you have a failed flash.

If you'd like to give me access to login to your machine via teamviewer, I'll be happy to take a look.

Jason

[/quote]

This is now resolved -- local browser cache needed to be cleared to read the new html pages after changing firmware.

Jason

[jewbsv]

Efudd Thank you, sorry to trouble you. 
We are looking forward to your firmware for Z9 mini.

[efudd]

Folk,

I just put a version 2.0b up at the same download link. If you are already using the 2.0, you are fine where you are, but are welcome to use an updated image. It corrects an issue with the reporting of frequencies in the web interface after a reboot. The frequencies are still correctly applied to the hardware, however. Once you select and submit frequencies again, it will correct the web interface.

Thank you,

Jason

[chipless]

I tried it on my z9 with 4 hash boards and the 4th don't show in the web gui so I cant control it unless I change the global to what I want board 4 running at then set the other three manually in the web gui

[efudd]

I tried it on my z9 with 4 hash boards and the 4th don't show in the web gui so I cant control it unless I change the global to what I want board 4 running at then set the other three manually in the web gui

chipless, yeah, that is a reasonable way to handle that condition. I had not tested that since I only have 3 hash boards. Updating their web interface can be done, but it is ugly 'code'.

jason

[chipless]

I tried it on my z9 with 4 hash boards and the 4th don't show in the web gui so I cant control it unless I change the global to what I want board 4 running at then set the other three manually in the web gui

chipless, yeah, that is a reasonable way to handle that condition. I had not tested that since I only have 3 hash boards. Updating their web interface can be done, but it is ugly 'code'.

jason

I may take a stab at the web gui to add it but for now the other way is easiest to handle it. I did coding for over 30yrs and just got burned out on it. So far it seems ok once I get it all dialed in

[efudd]

Folk,

I made a mistake in the "2.0b" image that I have since corrected. However, it looks like there were approximately 8 downloads during that time window. If you are running 2.0b and you are having an issue with cgminer "restarting for no reason", please re-download the updated image and apply it.

I apologize for this mistake and will strive to not try to rush a fix out the door next time around.

Thank you,

Jason

[chipless]

A feature that would help my wife at night would be:

On network loss the fans don't max out
On server loss same issue with fans
On reboot/restart same issue with fans
oh did I mention any help on the fans would be great...LOL

I am being sarcastic...I get yelled at about 30 times a week from the wife
about the fans when I am playing around with it. Other then that the rebooting
has stopped.

I do have a question about the cgminer recompile, in all other versions factory and mods
I had cards randomly show x for an asic then disappear later back to normal in your version
I don't seem to have that issue with the same settings was this an api issue that you fixed or
something else might have changed that issue

Thanks

[FrozenChip]

Could you please add one more feature related to fans?
When Customize the fan speed percentage is ticked and set to 0%, all fan detection/monitoring should be disabled or at least ignored.
A little bit of background, so because of the noise, people are trying to use no standard fans with fan simulators.
Fan detection on Z9 controller gets burned easily by messing up middle pins on fan connector and controller becomes unusable, throwing something like no fan detected, everything else on controller seems to be working fine. So by adding this feature, there will be no more need to purchase fan simulators/spoofers, and everyone that has burned fan detection on controller, will be able to use that controller again.
 https://image.ibb.co/mv7XTA/Fan0.jpg

[efudd]

Could you please add one more feature related to fans?
When Customize the fan speed percentage is ticked and set to 0%, all fan detection/monitoring should be disabled or at least ignored.
A little bit of background, so because of the noise, people are trying to use no standard fans with fan simulators.
Fan detection on Z9 controller gets burned easily by messing up middle pins on fan connector and controller becomes unusable, throwing something like no fan detected, everything else on controller seems to be working fine. So by adding this feature, there will be no more need to purchase fan simulators/spoofers, and everyone that has burned fan detection on controller, will be able to use that controller again.
 

This sounds reasonable and something I can look at in the initial fan work.

Jason

[efudd]

A feature that would help my wife at night would be:

On network loss the fans don't max out
On server loss same issue with fans
On reboot/restart same issue with fans
oh did I mention any help on the fans would be great...LOL

I am being sarcastic...I get yelled at about 30 times a week from the wife
about the fans when I am playing around with it. Other then that the rebooting
has stopped.

I do have a question about the cgminer recompile, in all other versions factory and mods
I had cards randomly show x for an asic then disappear later back to normal in your version
I don't seem to have that issue with the same settings was this an api issue that you fixed or
something else might have changed that issue

Thanks

Definitely heard on the fans....

As far as cgminer goes, I want to be clear that this is not a recompile. I wish it were as this would be 1000 times easier. As far as ASICs disappearing, that I cannot explain for you at this time. I have not personally seen individual ASICs "x" out and would consider that a problem, I think... at least without any other knowledge. I have not changed anything related to that, at least. I definitely would not expect that to be an 'api issue' regardless of the firmware, but rather a real indication of something to investigate (chip at too high of a frequency, perhaps?).

Thank you,

Jason

[efudd]

Folk,

I've corrected an issue with the "developer fee" running too often on the Z9 in version 2.0c -- it is a recommended upgrade for any Z9 user.

In addition, the 140minute/5minute duration has been change to 280minute/10minute.

http://releases.broked.net/Z9_Unlocked__hashboard_control__2.0c.tar.gz

Thank you,

Jason

[efudd]

Folk,

I should have the mini variant out soon. I have the majority of it ready, I just need to get it tested.

Thank you!

Jason

[dohckiller]

i have batch 2, would it matter or the FW will work as batch 1 z9's ?