Then there is nothing to do but sending it back to bitmain. 

Make sure to swap the cables of the hashing boards to make sure it's not a cable / controller board issue. Furthermore you can try to set the frequency to something really low pure as a test. I.e. set it to 200Mhz and see if it get's detected. If it does you can decide to run it at a slightly lower frequency instead of sending it back to bitmain (which will probably cost you more loss...)

I haven't try this myself, but with the knowledge I have from the code and from the fact that you can connect a board to the 4th connector I am 99% sure this will work. Let me know if your run into an issue. I might be able to help you out

Harper86 and Cryptonica helped me out getting a backup of the A3 firmware. thanks for that guys!   As others already found out as well: The A3 and D3 are very similar. I even found out that the software inside the controller chips on the hashing boards is exactly the same.
Out of curiosity I tried to run the A3 firmware on the D3 and the A3 firmware actually detects the ASIC chips and tries to  start mining on the D3 hardware. Looking at the logging it's fine, but it doesn't accept the work as Bitmain did actually changed something on the ASIC chips I guess. I do think they used the D3 as a base and stripped out 10 of x11 algorithms.

For now the conclusion is that you can't convert the D3 to an A3.
It probably does need reprogramming of the ASIC chips on the hashing boards itself.

You need to have some linux knowledge: login with ssh as root/admin to your miner:

This guide helps you to make a copy:
https://help.ubuntu.com/community/BackupYourSystem/TAR

later you need to scp the tar / gz file to your windows pc. winscp can do that for example.
Don't worry to mess something up on your miner as the filesystem is loaded in RAM. So a reboot and you're back to the original state...

I hope to have the firmware earlier then that, but I will let it know in this topic. Thanks 

As there is no firmware available yet from bitmain I ask it over here:
If one that received an A3 already can make a copy of the filesystem (tar gz it and SCP) and can upload it to mega and share the link (PM me) I would be really thankful and give a 0.5 LTC or 0.1 DASH reward to the first one sharing this.

+1

Sorry for being a bit late at the party, but it can be that one of the chips on your hashing board is just a bad chip. In case you can get the board running with the HW errors I would say just ignore those few HW errors.
Increasing / decreasing the voltage might help you out.

Hi all,

As you might have noticed there is some discussion going on about the GPL license of the cgminer software that's used in my firmware.
I was not aware of this GPL license, but I am now, and I am quite sure the source should be published.
As I don't have enough time to do it now and want to do it properly I decided to temporary remove the download link to the modified firmware and lock the topic until this GPL issue is resolved. This will prevent that this topic get's locked / removed for ever and valuable information gets lost.
I hope I can resolve this issue as soon as possible. Sorry for the inconvenience,

ciao Blissz

ps. Please respect this and don't PM me for links to the firmware.

Hi all,

As you might have noticed there is some discussion going on about the GPL license of the cgminer software that's used in my firmware.
I was not aware of this GPL license, but I am now, and I am quite sure the source should be published.
As I don't have enough time to do it now and want to do it properly I decided to temporary remove the download link to the modified firmware and lock the topic until this GPL issue is resolved. This will prevent that this topic get's locked / removed for ever and valuable information gets lost.
I hope I can resolve this issue as soon as possible. Sorry for the inconvenience,

ciao Blissz

ps. Please respect this and don't PM me for links to the firmware.

Good to hear it works good! I don't know how to save the time as the memory is volatile, but I will put a warning to the daily fee mining that it should only be enabled when your miner works stable / doesn't reboot.

Looking good to me. Saves some heat and energy compared to stock. You might be able to lower the voltages of chain 2 and 4 a bit more.
daily dev fee should be a bit more efficient as there is less pools switches. It will mine ~21 minutes on the devfee pool each day. One day holds 1440 minutes so that's ~1.5% devfee mining.

great to hear it works good. I am not aware this 80 degree was an option so I accidently removed it. It's now enabled by default...


I am not aware this 85 degree was an option so I accidently removed it. It's now enabled by default...

That's strange indeed. Never seen it on my unit. Can you PM me a kernel log when this happened?

Hi All,

I wasn't aware of the GPL license, so I did some investigation last days. After reading the GPL licenses I can conclude that stangerm2 is right. So that means I do have to provide the modified sources as most of the current modifications are now build on top of a linked module in cgminer.

There are some consequences in releasing the source though:

- As far as I know, Bitmain is free to use these improvements in their current and their new products they will sell. I do hope they compensate these free goodies with a multi algorithm D3 firmware one day
- There will probably appear modified firmware binaries which can cause damaging your miners if the maker doesn't know what he's doing...
- I had plans to buy a S9 as there are a lot of requests and make a modified S9 version, but I might consider not to jump on that boat for now. (let's first see what will happen with the D3)

Legit alternatives instead of releasing the sourcecode

- I stop this project and take the binaries offline. I don't like this option, as the whole idea is to help improve the situations for the D3 owners.
- Temporary I take the binaries offline and I will write something from scratch / not linked to the sources of cgminer. This will take away the consequences, but a lot of time and effort (at least a month or two) will be put into something that will probably not lead to any improvements for you guys.

My proposal

- Although these consequences I will release source of a stable version (non BETA) that allow others to investigate the voltage / fan mods / fault check improvements I did. (devfee related code will be removed from the source). This allow others to look into the changes I did to make a binary for themselves without devfee.
BETA functionality will be left out of the source right now since it's not yet stable and will create dangerous situations when used in a wrong way. I might have to restrict the use of a BETA versions to a closed group instead of providing a public link to the BETA versions.

Happy mining

Ciao Blissz

Now it's 1 minute and 50 seconds each two hours mining on the devfee pools. It's actually switching between your pools / devfee pool. Have a look at  the kernel logging after some hours and you can see in detail what happened

You're not alone having this issue in the latest beta. I will investigate this

v1.02 :: 14/01/2018 :: L3+ recommended update

Download:
https://mega.nz/#!mkRWFKpB!J4pJLNgTheED1UkMLDHy8LiLMmsoKizkABCd_MbFslw

changelog v1.02 L3+ version 14/01/2018:
- adds power usage (Watt) indication per chain in the mining status overview. Really useful for if you don't own a wattage meter
- cleaned kernel logging when not using all the pools
- cleaned API code for faster miner status update speed.

Can you comment on the voltage setting?  I don't see it on the overview page.  Thanks!
[/quote]

You're still on v1.0 which had this issue. Please update to v1.01 or v1.02 when it comes out in a few minutes

if if falls below 92% of the expected hashrate for a few minutes it will reboot.