New Official AMT Thread

[mrpark]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

[1714134754]

1714134754

[1714134754]

1714134754

[NotFuzzyWarm]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

[mrpark]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

I think it's the heat sinks causing the ASICS to slow down. They put a tiny square much smaller than the heatsink for attachment. I laid an aluminum heatsink and it works much better. I could feel the heat sinks were not greased right because they were not that hot. What a mess...

[opieum2]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

If ALL 8 chips and all cores are working you should see 8/255 in cgminer....with this number you should see between 190-220Ghs. But if you are seeing MHs then its a VERY low number and something else. I suggest pointing it at btcguild just to get a consistent result for the moment until you are sure that the problem is not the pool setting. It likely is the card, but just to rule out any pool related issues.

[opieum2]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

I think it's the heat sinks causing the ASICS to slow down. They put a tiny square much smaller than the heatsink for attachment. I laid an aluminum heatsink and it works much better. I could feel the heat sinks were not greased right because they were not that hot. What a mess...

Yea ISA covered that. In some cases the assemblers put the heatsinks on apparently in a way that left airpockets...not a good thing. I have been mostly quiet as I have been working on the software side of things. Lotta reading on these issues. Its not just AMT having them. They just got hit the worst by it due to all the other problems with assembly.

[opieum2]

@AMT

Not sure if you online tonight, but hopefully you can provide me with some info on the particulars regarding the driver (preferably the original C files used to complile your version of cgminer). I am trying to get an updated cgminer working and I have some of the driver stuff I would like to test. However the A1 driver changes that were made for AMT's setup I would need to know the specific changes to the driver-SPI-bitmine-A1.c....or any other cgminer/bitmine related driver file for this to work. Also the difference between the AMT and the bitmine backplane as first and foremost thats not being detected. Once I get the backplane detected, the rest I can start isolating the reset issue. If you can post or email me the files I can figure out where the issue is. Will save me alot of guess work to get the backplane charged up. I expect I might have some insight on being able to force a reset on the cards as well.

I have been going over Zefir's work, Trying to figure out where there would be changes for the AMT variant. Seeing as the original work is GPL this might not be a bad idea to post publicly or post on github for people to work on and improve. At any rate I am inching closer to a possible test but I need to see the actual files used in the cgminer build you guys have to make this work. Thanks

[mrpark]

If you put your finger on a heatsink, and it's cold, you know it wasn't done right.

[FrictionlessCoin]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

I think it's the heat sinks causing the ASICS to slow down. They put a tiny square much smaller than the heatsink for attachment. I laid an aluminum heatsink and it works much better. I could feel the heat sinks were not greased right because they were not that hot. What a mess...

Do you refer to the big heatsink?

[AMT_miners]

I am unfamiliar with the "grey market", could you elaborate?

<snip>
 There have even been cases of chips with no actual die in them!
<snip>

I thought the chips were ordered from Bitmine, and through them, from Innosilicon. I'm unclear how the manufacturer would have any part of how the chips were presented. I don't know a whole lot about how ASIC chips work but it seems like that would be very hard to miss and take effort to ignore. Then again, I'm no engineer.

Chips are not the problem, if they were, then they would not work when removing them and remounting them on a different board/design etc. Chips came from bitmine, and they are not not the problem. 

[AMT_miners]

@AMT

Not sure if you online tonight, but hopefully you can provide me with some info on the particulars regarding the driver (preferably the original C files used to complile your version of cgminer). I am trying to get an updated cgminer working and I have some of the driver stuff I would like to test. However the A1 driver changes that were made for AMT's setup I would need to know the specific changes to the driver-SPI-bitmine-A1.c....or any other cgminer/bitmine related driver file for this to work. Also the difference between the AMT and the bitmine backplane as first and foremost thats not being detected. Once I get the backplane detected, the rest I can start isolating the reset issue. If you can post or email me the files I can figure out where the issue is. Will save me alot of guess work to get the backplane charged up. I expect I might have some insight on being able to force a reset on the cards as well.

I have been going over Zefir's work, Trying to figure out where there would be changes for the AMT variant. Seeing as the original work is GPL this might not be a bad idea to post publicly or post on github for people to work on and improve. At any rate I am inching closer to a possible test but I need to see the actual files used in the cgminer build you guys have to make this work. Thanks

Will send to your email today. Posting it publicly is not a good idea, because of the difference in hardware. When using bitmine's original image to run the PI the default voltage was outputting .9 and above. This was due to hardware miscalculations and components with improper values. Our setup script was designed to modify voltages based on the batch of boards which came off the line. If someone with an imporper component setup use for example

"amt-setup dpot 0x5a"  would result in a voltage above .98 resulting and a potential heat issue. This is why we've hesitated to release that option to the public and its why this was simply a script for us to workaround and configure what given to us off the line.

We don't think we changed the drivers at all from bitmines original image for the 1.2 SDcards. For the 2.4, we did rewrite the drivers in order to redress the backplanes so that two setups could work together with on extended 3 port gpio cable.

[AMT_miners]

Question: I have 2 boards hashing at 380Mh is this correct for 2 boards? I guess all ASICS are working on the 2 boards. Any way to check individual ASICs?

Opium/ISA would know for sure but sounds a little low. Think 200ghs or more per board is expected. Depends on what they set the clock and such at but I say sounds like a safe setting or maybe a chip has a few bad cores.

Default clock setting should be 16000:750000:4000   essentially under-clocked, or "properly clocked" based on the component setup we have.

[AMT_miners]

<snip>

Here is my issue with this complaint about poor parts.

AMT promised delivery on February.

If the boards were coming out of the manufacturer and testing bad in January, they should have informed customers of the problem and like any legit supplier, refunded anyone who requested.

Now we are here 5 months later,  nothing but bad shipments if any shipment at all.  

The point is this, do not ask people to pre-order if you cannot guarantee that you can deliver.  If you can't deliver then it is your responsibility to refund your customer.  It is like any other product,  if I order 500 barrels of oil to be delivered in February for a price of $100 and you can't do that because of someone bombing your pipeline, then you refund the $100 because you never delivered product.

Has AMT delivered working product?  No.   So we all deserve a refund.  Their manufacturing issues (if even remotely true) is their problem.  No sympathy from me.  They should have purchased insurance prior to embarking in this risky endeavor.  

By and large no disagreement from me there aside from the boards timeline. Even on that someone should have at least said, 'Um folks... Oops...'  Not that anyone probably read it but on AMT's site under News they did mention on Feb.10  that multiple versions of boards may be needed. https://advancedminers.com/pcb-prototype-testing-diagnostics/ Why was that not brought here where folks see it?

Gotta love the closer on that page:

Note from the swiss: “Several prototype runs of the boards will be be necessary, along with fine tuning and optimizations (especially on the power supply blocks and cooling) carried out by our team along with the highly specialized engineers in switzerland.” This can be a quick or lengthy process, only time can tell

Especially agree on the pre-order mess. Still baffles me that the practice is allowed at all.... This endeavor as well as those by BFL, HashFast, TerraMiner and others should have been properly financed 100% by their own sources. Not unwitting Investomers.

Funny thing is as consumers you guys HAD ALL the power and you were throwing it away faster than they could collect it...  You guys made the pre-order model the easiest fastest solution for these companies to raise money (and what they raised was liabilities and not capital lol) so most of them used it.  Funny thing is once you guys handed over all the money the power switched form you to them.  We all have seen how they have handled the power you gave them, all cherries and ice cream before they got in your pants but then ignore became SoP once they blew their load.

There is a reason why the real world doesn't use pre-orders the reward simply rarely outweighs the risk.  The companies named above prove this point beyond a doubt.  Why do you think very few if any VC's were willing to take all the risk on NRE of BTC mining ASIC, because they knew the risk was so high that the reward could never come close to making it worth while for them...

The investomers of all of those companies got swindled by snake oil salesmen, the last resort is basically use the law to the best of your ability to bring restitution. Good luck guys I hope you get as much back as you can but it is going to be a LONG haul from here likely.

It is for that reason alone as to why we have not accepted new orders in the last few months. We're in the process of changing our business model completely, which will only be an immediate delivery option upon receipt of payment.

[NotFuzzyWarm]

It is for that reason alone as to why we have not accepted new orders in the last few months. We're in the process of changing our business model completely, which will only be an immediate delivery option upon receipt of payment.

Excellent. Only once you get a handle on actual production times for stable miners should you even think about doing capacity-induced order delays. Those of course would need to be kept within FTC guidelines and fulfilled as they state.

That includes the need to stop/slow down orders as needed until production capacity can be increased. It is a bitch but has to be done. Our company has turned down 3 large orders so far this year because we are booked into next year.

[Flying Hellfish]

It is for that reason alone as to why we have not accepted new orders in the last few months.

This is definitely the smartest move (if true) I have seen you guys make publicly.  You have probably saved yourselves more shit than you can imagine by simply doing that.  I assume your council has advised you that in case you lose one, some or all of current and future legal problems that it will be much more believable and easier to ask for leniency because you made a big mistake and once you learned you were making a big mistake you took immediate actions to correct the mistake.  Had you kept taking orders no one would have believed you made a mistake and just "got in over your heads" and you would have lost the ability to plead for leniency based on ignorance.

Little things like this makes the difference (should you lose) between hostile judges/juries and "sympathetic" judges/juries.   Not hard to see which one is better for the defendant if they lose...

It doesn't change what happened and won't get you off the hook for what's done already but it certainly helps minimize the damage and does nothing but give you another option to argue IF you lose cases and need to ask for leniency at some point.

[opieum2]

At the end of the day though hopefully they can make us all whole. There are those of us who can help that are trying to. In the end its to make it better for everyone. I for one would like to see ONE US company come out of this better. While I know they have made some massive fuck ups and I am pretty pissed personally about it, I can get past it if they are making a genuine effort to make things right. For comparison BFL has not done shit to make things right. They have delayed and worse, are now taking that same money and buying up personal things with it. At least seeing AMT trying to restructure properly and hopefully helping the people here who were wronged be righted, that might make for better things down the road for them.

Lotta haters no doubt. But everyone has fucked up and some spectacularly so. But there are those who deserve at least a second shot at getting it right. This isnt baseball tho. A second shot is really all they get. I would be interested in hearing this new plan. And if it involves in stock hardware, let's hope its something worthwhile. I'm still moving forward with my own plans but hope to get these miners working. FYI down to a single working card now. The other one gave up this afternoon. So 1/10 working cards now.

Im still looking at how to reinitialize the cards. The backplane init info is what I need to make this work as its not being detected on either system with my firmware. I am going to attempt to port over your bitmine files you guys have and see and do some code merging. They are in the firmware so I can modify them and port them over to my work in progress.

[Flying Hellfish]

It is for that reason alone as to why we have not accepted new orders in the last few months. We're in the process of changing our business model completely, which will only be an immediate delivery option upon receipt of payment.

Excellent. Only once you get a handle on actual production times for stable miners should you even think about doing capacity-induced order delays. Those of course would need to be kept within FTC guidelines and fulfilled as they state.

That includes the need to stop/slow down orders as needed until production capacity can be increased. It is a bitch but has to be done. Our company has turned down 3 large orders so far this year because we are booked into next year.

If they make it to that point they should strongly consider using escrow services at the start.  Their reputation is pretty shot IMO and this would help instil confidence in the consumer.  They should be proving to the customer they have changed and the customer should have the chance to give them a second shot at NO RISK to the customer.  They want a "last chance" they should bloody well earn it and start taking the risk with THEIR money!

I wonder if this new BM has them using existing monies to build miners to ship from stock. IF that's the case then nothing has changed really.  Until they dump a bunch of their own money (or investor money) to clean up the shit pile they're in, make things right for original investomers AND build new stock miners they are still just running the same old model.  IMO they need capital to make the switch, I hope they do make the switch but am very leery to be honest.

Righting a sinking ship is usually a very difficult task.  Based on all we have seen publicly from the company I am not personally convinced they have the ability to right the ship.  They need CAPITAL, hard wok, dedication, smarts and LUCK to pull this one out of the fire.

I hope I am wrong and I wish these guys the best of luck turning this thing around.  But ask yourself this, if you were a venture capitalist would you risk your investment on this company?  If they go outside of FF investments (friends and family) then they are going to go through a due diligence phase and the law suits and liabilities (ie unfulfilled orders or refunds) are going to be a huge red flag.  I personally think a lot of investors would look at them as WAY to high risk and that the rewards could never match the risk level (you can't sell more than 100% of you company...)

Fully capitalized entities go out of business everyday in the USA, even more under cap'd ones go under.  Very few of them turn it around.

[rik_khaos]

you can't sell more than 100% of you company

I've seen companies sell off even with debt (I'd consider that a exception to selling all of your company) unless of course you ment sell all of it have and have remaining equity.

If the can get the board issue turned around and the company sold I imagine they would clear house in terms of management and rebrand. They could just become BMT -Bitcoin Mining Technologies.

Then they could have a sweet (onion sauce) cross promotion with Subway. Or maybe as hot as the miners get they could have a subway promotion where you order a BMT with the creamy sriracha sauce and you are entered to win a miner.

But seriously if they do get it turned around, step one is to get properly working miners shipped before another order is ever taken. That includes miners to everyone who is currently sitting at home with their paper weight. And what ships should be what was advertised, case psu and all unless requested otherwise.

What they do have going for them is they don't appear to have the high profile with the media like BFL, Hash Fast and the like even though they do share a reputation. That could make them a decent candidate to take on a partner to raise capital.

[rik_khaos]

No one mention the fact that the thermal tape catches fire or they'll end up deleting it next time they are on.

[VinCeCream]

No one mention the fact that the thermal tape catches fire or they'll end up deleting it next time they are on.

Damn.. haven't heard about it but I'm glad to know it. Has anyone been hurt ?

[rik_khaos]

No one mention the fact that the thermal tape catches fire or they'll end up deleting it next time they are on.

Damn.. haven't heard about it but I'm glad to know it. Has anyone been hurt ?

Nope. Just  burned the chip out as the heatsink fell off. Less life threatening and more a disappointment in performance.

This could well be an issue caused by their manufacture, I won't lay the blame directly on them. Just make sure not to put your miner on gas soaked rags or anything and you should be fine.