Topic: New Official AMT Thread - page 95. (Read 149472 times)

Well, possibly... Still, don't write up such a note then. Who are the attorneys they avidly keep refering to? Does anyone know?

Chances are they probably got barred from even doing that.

Yes, what IS the point of telling us to contact you when you have no intentions to answer any of us?  I emailed twice, once Monday and once Tuesday and nothing. Just the proverbial sound of crickets trying to get laid. At least someone is being **** here...

Well, that sounds like a good idea.  If AMT is unwilling to receive the miners at their address, then I am sure their lawyers would be more than thrilled to receive the miners at their address.   Is that how we are going to proceed to RMA all these broken miners?

I've never been able to get my miner hashing using the web interface. CG Miner never initialized when I would turn on the pi. I know because every time I'd go through the steps to kill cgminer it said no process was killed. So I'd cgminer -DT get a look at what was working and enter my pool info from there. Then I'd set my minimum difficulty via the web interface in the pool I was using.

From what I see just based on the small heatsinks of mine that have discolored is that it isn't just a issue with the chip in the 5 position.

On one of the boards that never hashed I have a completely discolored heatsink  in the number three position. On other board of mine that was hashing I also have that discoloration in the chip #3 position.
 issues that they have claimed to have faced ( which I do think is true, personally.)

So depending on the line it was worked on there could be at least from what was posted here a 2/5 failure rate if the boards were left running, your rig my vary depending on how many cards you got off a defective line. That is assuming only two lines had issues. Now that still doesn't speak to the issue of the other cards not being detected at all.

I did e-mail josh tonight and will be giving him a call tomorrow or the day after. We'll see if he can talk. If he says he can't I'll respect that and send any further communication to his lawyer.

I do believe that I'm going to send my equipment that needs to be RMAed to Josh care of his lawfirm and send it directly to their address with printed pictures I took of the damage. As I got a unit with manufacturing issues then I am due a new working unit and would be happy to talk to josh or his lawyers on how they would like to go about that.

Anyone know if the board will still initialize for operation, if just a single chip was removed?

Since I suspect this is alive enough to let it know that it shouldn't run, or it has just shorted from thermal-runoff, and just failed the initialization process. Thus, removing the short... or removing the "limiting failure broadcasting"... May allow the rest of the good chips to operate, until I can get it fixed, or just ignore it.

In my case, I assume it was the hardware for the most part. However, the software didn't quite know how to handle the "issue".

All except the issue with the starting, booting and resetting, which, for me, is still spotty.

There needs to be a longer delay between booting stages. It just pukes-out stage after stage, and obviously, some components are not actually ready, when it attempts to talk to them. (Namely the cards and the network.)

This, again, is where a multi-stage delayed-boot sequence would help. Allowing the devices to become powered-up and ready, before attempting to communicate with them.

The "reboot" in the settings, for the password-change and "target temps" (Not sure why those are both on the same page.), and also on the setup for the pools, is buggy as hell. (It is also annoying that it defaults to "balanced load" and not "failover". You do NOT want to run in balanced-mode, unless you are connecting to three separate servers, for some reason. You want failover, so when one pool fails, it attempts to go to the next one.) However, I do not change pools often, so that is just an annoying thing.

When changing the servers... "Save and reboot", does not actually reboot... it soft-boots... However, mine ends-up showing hashing like 12GHs and 45GHs per card after that... until I physically reboot by killing power to the PSU, since no-one has figured out how to turn off the unit unless you unplug the whole unit.

My biggest issue is that the miner does not connect to http connections for solomining. It demands "stratum+tcp" connections only. (Pools)

I was wrong before, about it hashing, but not reporting hashing... It does "get work", it just doesn't hash with it?!? Not sure how that got glitched. Pools are the worst thing to mine into, unless you have no choice, due to difficulty being high. However, coins I want to mine do not have high diffs, and pools are all in foreign-lands. (Too far, so lots of thrown-out share credits and valid work.)

If you could add some start-up delays between the start-up sequences... I am about 90% sure that will help resolve a lot of the issues where it simply gets "stuck" on the "Monitor" stage. Which is the last stage just before the network connects. (Or it is getting stuck on the network stage, and thus, does not refresh the screen, making you think it is stuck on the "Monitor" stage.)

Thanks ISAWHIM. On the software side I am currently compiling cgminer 4.3.3 for use. I will have a report likley tommorrow afternoon or later this evening depending on if it works on the first try. If it brings back the hardware, then its a huge win.

ISAWHIM, You think that the issues with the cards shutting down then are software, hardware or a combination of both?

Found the source of one issue on the card that fails to initialize...

Chip #5, (#1 being the closest to the power-connection), was not surface-flush. Thus, it overheated, tarnishing the copper heat-sink. Even with the card not operating (not hashing), it was glowing bright white and seared my finger.

Left unattended, with the rest of the cards running, it would surely have "popped", and heated to the point of destroying the bond with the copper-heat-sink.

This same exact chip location...


All the neighboring components above, were also abnormally hot... I assume feeding voltage into the unit, or attempting to try to limit the amperage or internal short-circuit on the chip.

The same chip that "popped" on RikKhaos's miner... But I assume, if he inspects it, he will find that the chip has a slight gap on one corner, lifted off the PCB. That non-flush mounting causes all the heat to distribute to the smaller heat-sink. Which, obviously, does not have the capacity to handle that level of heat dissipation. (Even if it did, it would have only prolonged the inevitable failure of that chip.)

In conclusion, the chips are not being set-up adequately. Either there is something like an obstruction, causing it to not pull-down in the reflow process, or there are contaminants in the reflow itself, or the reflow process is not getting complete. (Thus, being stopped before this specific component, or other components have had time to become "pulled-down", by the process of the liquefied surface-tension of the reflow solder.)

I will be pulling the heat-sink from this board, to use on another board, which may or may not work. This board will be sent-back to be reworked and tested. As I am not sure if the failure has caused other damage. Since it was heating-up, even without the card "hashing". No traces look seared or tarnished, so I am confident that a simple rework of that chip, to make it flush, would alleviate the issue. I would assume the initialization was not occurring, because of the internal detection of the chips temperature. Thus, causing the whole board to shut-down, out of protection.

I would rework it myself, however, I fear that the bond to the heat-sink may be compromised in the process. As a similar situation has been recorded.

I assume his heat-sink would not have fallen-off in normal operation, and thus, it was not the heat-sink causing the overheat, is was the overheating which caused the heat-sink to fall off. (In my opinion, looking at his photograph, and having experienced a nearly similar situation directly myself. Catching it before it had gotten to that point.)

Still, the V3.0.A series of the boards, otherwise seems robust and flawless, with the boards I have tested thus far.

I have prepared a specific report for this board, but will not RMA it, until I have tested the others, completely.

Still have only 3 cards up and running, at the moment... (Due to testing and lack of another free PSU at the moment.) They are operating at roughly 193GHs each, operating outside in 81F ambient air, without issues. Other than the one bad mounted chip, on that one card, mentioned above.

580GHs @ 760w total using only 6 fans for cooling the three cards inside a ducted cardboard shroud.
Rougly $2.92 a day, to run them, with a mining-reward of about $14.53 USD a day, leaving about $11.61 a day, for profit.
$11.61 a day * 30 = $348.30 USD a month. (Would be 2x that, if I had 6x cards running.)
That is just shy of 1 BTC a month. (2 BTC with 6x cards running.)

They likely got barred from making any replies at all.

AMT_miners,

 I tried contact you since last Friday by email and called you guys. no response at all.what is the point of this ask us send you email and call you. it's ridiculous.

Please response to my emails for options regarding my 2 miners.

Same here, I was able to mine with one rig for couple minutes then it went dead, the other rig had parts missing.

I agree at this point they could be getting out of this mess just by buying cheaper rigs and send to us with even higher speed, and I would be satisfied and not with the 2 RIGs that 1 worked for 1 hour and the other one has part missing. They are not very smart people, both of them!

And as I typed this message another card failed. I am down to 3 cards from 10......Hopefully this linux solution works on the already dead boards....

FYI I had no luck with bitmines image.  Doesn't mean it wont work...I just didn't spend enough time trying to make it work. Its just an updated version of the same yocto using a later version of cgminer. No screen or any of that stuff. I will be adding that to my image. Im adding screen and maybe some kind of admin interface like webmin. Not sure yet which route i am going with the web side if at all.

Yes but the difference is they can retaliate at you with legal action. Don't be that guy the news says AMT sues customer for leaking CEO's personal details. I mean its your call. I'm just saying, it could backfire on you is all.

had you changed the api look up in the web interface? or possibly disabled dhcp server which gives you an elaborate brick! trust me i did that 5 mins into seting up. click click click ohhhhh shizzzz! oh shiz what have i done! get the hammer out