1: Heat sink design. It is one massive piece of metal and retains as much heat as it releases.

2: The controller shuts off at 45c yet its inside a metal shell sitting on top of a giant heatsink radiating 75c heat. This means over about 30c-32c ambient the controller shuts down due to heat issues.

3: They radiate some kind of infared heat that means when you put a bunch of them close together they tend to overheat each other even if the ambient temperature is not that high.


Also please take into account that I am talking about how they act in an industrial environment, not some random home user that thinks running 10 or 20 is a large enough sample size to call them great. The scale of the operation and the amount of heat generated will cause these flaws to show up whereas someone just running a few at home will never see the issue. I have had these machines tested and they shut down well before their rated max operating temperature. In the lab environment at 85f they have serious problems running versus the s7 that was able to run in 110f ambient without shutting down in the same test.environment.

I have some S7s that do not like to hash at the full 700. Turn em down to 693 and see if that kickstarts it. If that doesnt work just flash the latest firmware. I flashed 90 machines yesterday with the 5/23 firmware with zero issues.

Having a strange issue that I cannot figure out.

I have a mix of S7, S9, and Avalon6 units. All of the S7s at various times lose about 20% performance. This happens across the board to ALL s7s I have regardless of what pool they are running on. I have tested on Kano, antpool, f2pool, etc. All of the S9's as well as the Avalons do not lose speed at all. The issue is limited to only the S7, which I have a mix of batch 1 through batch 17 and all of them do the same thing. I have different machines with different firmware so it seems firmware independant as well. Any input would be appreciated.

If you have hard rebooted it a few times and it still comes back, you just have a bad chip. No way to fix it.

What is it exactly that causes the boards to drop out at the lower voltage? Is it a hardware error issue or is it something else? I have been flashing 660/690 because its what i have found to be universally stable but it really doesnt get them down into the power efficiency level that I want.

Lets hope the A7 fixes the giant flaws with the A6...namely the massive cooling problems the A6 has

1028 is two codes.

loop0fail: there is a failure in the hashing board itself

pga0fail: this is a power failure...either something in the power stage of the board went bad or you are not supplying the board with the proper amount of power. (this power issue will cause the loop fail error code as well since the board cant power up to actually proceed with the POST)

I experienced a few A6's that only hashed at 3.1th with a 12.0v psu. Plugging in a server PSU putting out 12.26v got the same machine hashing at 3.6t Its not about the wattage its about the voltage being supplied to the boards. There is no internal regulation so you can overclock in this manner to get back your lost speed. I would not recommend going over 12.3v as a number of people fried their boards that way on this forum.

Also about the bricked controller....thats exceptionally rare on the Avalons. i was able to recover every controller I worked on by simply reflashing a good MM to the controller (outside of one that popped a cap of course)

There are two marked pinholes on the controller to power it up on its own without being connected to an Avalon board. I would try and power them up in this manner and then flash a known working MM to it. I am almost 100% sure it will recover the dead controller.

1028 is probably a dead board or there is a power fault in the PSU. If you look in between the two PCEI plugs there should be two green LEDs on whenever the board is powered. If you see this and still get the error your board is pretty much dead.

FYI to check on any avalon status codes go here:  https://canaan.io/downloads/querystatus.html

Loop 0 is the right hand board looking at it from the power plug side. Loop 1 is on the left.

Look at that speed, i'm gonna be rich!

Yea that probably wasnt the best unit to show off with. Most are more like this one as far as the error rate goes

For those looking to mass mod without having to experiment with every single board to find the most stable number I have done some testing.

Running later batch S7s at 660/690 with frequency at 600 they are doing on average 960w @ 3.95TH
I also removed the back fan and flashed the S7-F1 firmware to save even more power. Temps are better with 1 fan modded version than a stock version at my location.

Its almost like you are reading another topic because you keep coming back with these nonsensical responses. Where did I say a damn thing about GPU mining...especially GPU mining bitcoin?

Kindly pull your head out of your ass and read the whole thread before you just throw shit at the wall that makes no sense. I mean seriously read my fucking posts.

So they saw they were about to become insolvent and stopped operations before that happened. Is there anything you sheep dont call a scam on this board? Just because YOU do not understand how business works you should not be running around calling everything in the world a scam.

Well my dear Watson, considering my response was to the previous post stating coins are moving away from PoW and me not actually asking when it would happen, I think you should feel a bit silly about your comment.

Thanks for playing!

As with almost everyone on this forum you refuse to look at mining from a business perspective. I just laid out in very simple terms the basics of what you need to make it profitable. The OP was asking about making a career out of it, he wasnt asking people that are butthurt that they cant make their fortune running half a dozen miners in their garage to tell him how he cant.



P.S. Bitcoin is the only coin that matters in the long run....when is it moving away from PoW exactly?

You guys use the word scam so often and in the wrong context most all of the time.

Its like half this forums active users just run around screaming scam at everything they dont like, its ridiculous.

Does Minera average the temperatures from all 3 boards for its readout? I have a machine with 2 bad temperature sensors and minera reports it running at 20c  when its actually around 67-68c

I am not sure you understand 21 Incs business model. It has nothing to do with mining other than to use them as a token for microtransactions. They have zero skin in the actual mining game....

Its crazy to think so many people pumped money into this thing without thinking of the problems inherent in the DAOs design.


http://hackingdistributed.com/2016/05/27/dao-call-for-moratorium/