Topic: Cointerra Hardware Support **Unofficial - page 12. (Read 57003 times)

There is not yet a class formed.  For one thing the purchase agreement includes wording that it must first go to binding arbitration, and they are in that process.  I am in the queue with the same law firm should that result in something positive.  If however arbitration totally breaks down then they can go to the courts and request the court for a class to be formed.  It will probably be a year or more before anything comes out of that, if it actually happens.

If they are successful in arbitration with the first client then they can cookie cutter deals based on that for the rest that they represent.  However I am still waiting for a refund from my Jan order and was told on July 8 that it would be processed on July 30.  On August 8 I was told that it was not yet processed.  Today I was told that there is no ETA but they assured me they are working as fast as they can citing how overwhelmed they are with refund requests.  And to think I still have multiple units that didnt hit the 2TH/s mark :/

The only thing I can think of is they used refund money from customers to build their cloud mining operation, so the store on broad street manhattan could have the miners in stock, etc.  They expected sales to explode and they didnt.  They expected profits to soar and it was more costly to have the units made and delivered.  They now do not have the cash to refund people and are struggling on that front hoping to mine their way out with gear they cant sell.  I could be wrong, but that is the feeling I get.  That insolvency is just around the corner. 

Hi all, been spending some time reading through this thread trying to see if anyone has a similar issue, and whilst I think people do, there doesn't seem to be a fix.

It all started with the top PSU showing an orange light on the right and green on the left, unit not mining at all.  When you turned on just the top one the fans weren't starting on the unit. 

The customer service at Cointerra is shocking (and rude), the issue was reported mere days after the 3 month warranty period, I'm in the UK so I'm sure it will cost more than the $350 non refundable fee to get it there (they won't even allow me to ship and return - they make a FORTUNE on their shipping rates I can tell you shipping myself on a corporate UPS account).  The breaking point was even if you do that, a 7 day warranty period on the repair; really?!  So I'm voting with my feet, I won't be buying anything else from them again and being an advisor to many BTC onboarders I'll ensure they don't get anymore new custom, in my mind they're not far off Butterfly labs now.  All that said, decided to self diagnose and try to fix.

The problem...

1) When opened main board had no lights (the night rider orange ones on the board).  Not even at startup and no fans.
2) Switched PSU's and ruled out that problem as the good PSU then became the bad.
3) Checked all connectors
4) Tried removing the black and white cables from the faulty board - nothing.
5) Swapped out the 2 boards checking all connections along the way.  No leaks found, no bad connections.
6) After swapping boards around we've worked out a board is at fault and as it was the primary board there was no mining.  Now the machine is mining on one board so that is an improvement.  Interestingly both PSU's need to be on or the unit won't stay online.  If you don't power up or power down the 2nd it goes offline (can't access over LAN etc)
7) Now the original problem of the orange light on the PSU is on the bottom one.
No lights on the faulty board at all, not even on startup.
9) Looking at the nightrider lights on the working board D42 and D45 don't light in sequence (on startup D42 does but not D45) so I assume this is why it's hashing at around 700. 
10) Only CTA0 displays on the admin screens (before we started I couldn't get to this menu as the faulty board was the main one).  On the front of the machine it blinks on Miner 2 as mining.
11) We've not amended or touched the coolant system yet, there appear to be no leaks.
12) We've tried resetting the boards, we've tried resetting the system etc.  Firmware is 0.7.6, I tried to install 0.6.32 but it loads but doesn't downgrade.

So.. a board seemingly without power, PSU's seemingly fine, cooling seemingly fine, everything follows the board.  Have meticulously checked the board and can't see anything obvious.  Any ideas?

Would really appreciate any advice, thanks all.

David

I have also filed a complaint with the Texas Attorney General https://www.texasattorneygeneral.gov/consumer/complain.shtml

Even-though the Attorney General can not seek restitution for you, I still felt that their unfair business practices needs to be brought to there attention!

I have purchased 3 units from CoinTerra and all 3 had to go through their lengthy RMA process. CoinTerra would only repair one at a time and the repairs have only lasted a month. Now all my miners are out of warranty and are running too hot again. I contacted CoinTerra and once they found the units to be out of warranty I receive no reply to my requests. Only confirms my suspicions that you would only get one repair out of them if you where lucky.

Even though there maybe some compensation from the Class Action Law Suit coming out of California it will only be related to the companies inability to hit the 2Th's mark and the excessive power consumption in conjunction with not being able too deliver the units in a timely fashion.

Their is nothing filed in the Courts to address the major losses of revenue due to CoinTerra's  lengthy RMA process. Which I found to be on a average of having a 6 weeks turn around until the unit is back in service and functioning normally. Nor for the time frame of the units operating even below the 1.6 Th's level till the RMA was recognized and CoinTerra took action to receive the units back.

As we all know time vs hashing power vs difficulty = major losses in revenue.

So, will CoinTerras consumers ever get what they are entitled to???
Posted from Bitcointa.lk - #7L0w4Gk6YGLaF4m6

For anyone who needs new PSU's for their cointerra rig.
I have two brand new powerone 80 plus platnium PSU's received directly from Cointerra.
Im selling them for 180 each or 320 for both.
I accept bitcoins or paypal
message me or reply if interested

Aight that's nice, thanks guys i learned something. Fingers crossed i get the job done okay !

thanks again .

All good waterblocks/coldplates will be copper, at least the part that touches the chip and what the water channels are cut into.  Copper transfers heat to the fluid better than aluminum, however radiators and mere heatsinks are aluminum because that helps to draw the heat out better. 

If you have an aluminum waterblock you have a really cheap one not to mention aftermarket because stock is copper.  Aluminum heatsinks are popular though which is probably why they mentioned aluminum.

The block is copper, the chip is a semiconductor with epoxy around it (see http://www.empf.org/empfasis/2007/June07/flip.html). You're good.

sorry to ask again :

 Coollaboratory Liquid Ultra v Coollaboratory Liquid Pro ?

'' After the overwhelming success of the Liquid Pro, the new Coollaboratory Liquid Ultra was developed. Of course the Liquid Ultra is also made of 100% metal, but it has an absolute new usability and offers an outstanding performance. Due to the paste-like texture the application is now very easy. ''

 - which sounds good because ive never done this kind of thing before, and i am unable to find the liquid pro in the country where i am.

just one thing scared me :

'' Please notice that the Liquid Ultra is not suitable with the usage of aluminum surfaces. ''

-- umm . so my fear is that the chip or cooling block has aluminum or something? - sorry if that makes some people laugh - i really have no idea about chip metals

thanks ~

touching on the PSUs a bit, one thing that may help is to reverse the fan.  There are two otherwise identical model PSUs made by power-one.  The PFE-1100-054RA and the PFE-1100-054NA, the R/N stands for reverse or normal airflow.  For some reason cointerrible got the reverse ones so cold is rear and hot is front even though the rest of the system is completely opposite.  Because these are otherwise identical and it does not make sense from a manufacturing standpoint to have two completely different layouts internally I suspect but have zero proof that the only difference is the direction of the fan.

The fan can either be flipped or possibly, maybe depending on stuff I couldnt see, the power wire reversed to reverse the flow of the fan.

I suspect that reversing the fan may actually be helpful, especially for boxes in a data center with a cold and hot side.  If people have exhaust strategies in a home environment this would probably be even more helpful since they are unlikely to have surplus airco capacity.

To remove the psu:

remove power to the machine
remove the 4 phillips screws holding the heatshield on and set aside
depress the grey tab and pull the psu, the connector may stick a bit but a firm yet gentle tug should dislodge it.
using a small screwdriver, tack or other item pry the 3-4 tabs on each side of the psu out just enough to let the  top cover slip off
using a small torx bit, smaller than T10 (which is the smallest I had so I never went this far) remove the screw
The top cover should be easily removed at this point and you can locate the fan.  Flip it around.

I got as far as being able to open the cover a bit to blow some air in and remove the dust that was in there but could not get to the fan so I could not  reverse it.  Any information on what size torx screw it is and especially if reversing the fan helps would be appreciated.

There very well may be a reason they got the reverse air psus, I just cant see it.   There is plenty of space for air to get in and it makes a whole lot more sense to suck the coldest air in rather than the hottest.  The heat shield would not even be required if the fans were not reversed.

The psu will shut down at about 115C on its own per power-one so if this turns out to be a bad idea you may have to wait for it to cool and undo whatever you did.  In  theory a cooler psu will output more stable voltages which in turn should result in better hashing.

my point was that it is clearly not LZMA compressed because its not 9000 yottabytes  There are also no other known signatures in the file indicating if it is anything it is not standard or well known.  It is more likely that it is completely uncompressed as there is little need.  And yes there are symbols and strings all over in there that indicate a lot of it is not compressed.

I have IDA pro and plan on using that although I have not yet reinstalled after I wiped  this system as part of an upgrade a while back.  That one generally does a better job than most at this type of thing.

onlinedisassembler does not appear to do armv7 which would explain some of the undefined instructions.  I would not rely on its output as it apparently lacks the proper support for this MCU.  It also appears to ignore the symbol information which can be quite helpful in knowing generally what a section of code does without having to parse every line and figure that out yourself.


BTW: email cointerra support and explain your woes, some of the newer firmwares deal better with usb resets than what is in 0.7.6.

It's clearly not compressed, there are printf strings all over the binary (e.g. "A%dC%d Pump bad %d"). That doesn't mean that the code itself doesn't do some decompression, but as you so perfectly put it, occams razor and all that. There's even a string "Set asic %d core %d to speed %d pll 0x%X". I stuck the firmware into this online disassembler; who knows if it's any good:

http://www.onlinedisassembler.com/odaweb/wzJdcS/0

I took a great interest in this firmware a few weeks back, but gave up. I think my non-working board is borked for some other reason, but the MCU _does_ communicate from time to time, so I held out hope. I really appreciate your writing that little essay, it's rekindled my interest. I just wish they would release the damn firmware code instead of us poking around like criminals.

I will add that cointerra charges about $350 just to look at the unit.  Once inspected they will tell you what they want to repair it.  The $350 is non-refundable but I believe does include shipping back to you.  

The warranty is only for 30 days on new sales 3 months on sales from long enough ago its out of warranty anyway.  You must use the original shipping container in order to send something back to them or you void the warranty.  They have so many ways that the warranty gets voided its almost as if they expect the units to fail.



I suggested the PFE drop in replacement because the power supply in the unit has an i2c bus and the boards might talk to the psu.  If you replace it with something totally different like an external you are likely going to have to do a lot more work to make that function properly.  I believe that the ps_load argument to cgminer is involved in talking to the psu over that bus but I have not researched that aspect enough.  I do know that cgminer just passes that argument to the CTA board and does not do anything internally with it, which hints that a firmware change would be required to change to a non-compatible psu.  The fact that power-one says
is helpful in assessing risk.  Specifically I do not think there is any as long as you stay in their PFE/PET series.

http://www.power-one.com/power-solutions/products/embedded-power/ac-dc-hot-swap-front-ends/platinum-efficiency/pfe-series/series


There might be some light shed on this if someone has a 2TH/s unit and can get the model number of the psu so that I can compare what all was changed.  If its also a power-one PFE/PET series then its likely that one firmware can deal with all of them trivially, if however its something totally different then that may mean the psu has less communication since the same firmware works on both and with only 55k of code at an average of 4 bytes per opcode excluding the symbols and other text you are looking at fewer than 10k instructions which makes it less likely that the firmware has 2 totally different psu control sections.  A totally different PSU would hint that the firmware does not do much with the psu.


Power-one does do cases of the power supplies, they are 10 packs.  There is a price discount at digikey, mouser, etc for 10 packs and some you can only buy in a case.  At 2 per unit a group buy would not take a large group before you start to see a price drop, but until someone swaps one and sees if it does anything I do not suspect that anyone will want to pony up the money just because.

The power efficiency graph on power-one says its most efficient at about 50% max capability.  Inefficiency is commonly presented as heat as well as slightly higher power bills.  Granted its only a few percent difference but a larger one might help keep things slightly cooler while providing enough of a surplus that the boards run better.  Just need a guinea pig to try it, preferably someone who can spare the $450 without compromising their ability to put food on the table for their family.

You could just swap 1 psu and see if that board performs better before ordering a 2nd.

I didnt see some of this posted elsewhere so I thought I would add to this thread.  Some may be mentioned elsewhere I did not search for everything I am putting in this post because of the size and scope of what I am contributing.  This post also turned out to be much larger than I had originally intended, sorry bout that for anyone who dislikes reading or pays per kb downloaded :/


I would have more info but have to spend more time on things that pay the bills.  I have no funding for this research so its kinda on an as available basis and currently I have only had about 30 minutes to look into it.  Should that change I can probably do a whole lot more.  As I get the time I am going to disassemble the board firmware and try to write compatible C code that would replace the closed proprietary cointerrible CTA firmware that is known to be buggy.  This is by no means a trivial task, disassembly has its own problems and is often a little more involved  than running IDA to get something that can be recompiled into a working software package.

If this info is helpful I do accept tips, should anyone want to fund my research I can make more time available.  1AhuDugRidP9HksF8s5WrDUziucmctXBNc


The PSU is a power-one PFE-1100-054RA ($350 or so at a few retailers).  This breaks down to 1100w max output, reverse airflow psu.  It seems to me that the psus are maxed out (and less efficient as a result).  The PET-1300-054NA would actually fit and according to power-one anything in the PFE or PET series is swappable.  This would give you a bit more available power which might solve some issues with the systems.  The 1300w one is a bit more at about $450 though.  It is my understanding that the DC/DC circuitry on the board as well as the psu is the reason the 1.6 TH/s units cant do 2TH/s so you may get a slight bump from swapping the psu.  It would be an expensive experiment for someone to try though.  I also like the NA rather than the RA models as they exhaust the heat rearward instead of to the front.  I am unsure exactly why cointerrible got reverse models, but the heat shield could be removed (you must remove it to swap the PSU) and for data center apps the hot/cold sides could be better maintained.

The mcu on the boards themselves is an arm.  specifically a TM4C1233H6PGEI.  The firmware loaded onto the board along with the tools to load the firmware is located in /opt/GSA_{VERSION}.dfu and at least some debugging symbols appear to be intact.  I have a few versions of board firmware that are not yet released specifically  from 0.7.53 CTA version 0.5.29 (2014-06-12) and from 0.7.46 CTA version 0.5.29 (2014-06-12).  See the end of this post for instructions on extracting the firmware so you dont have to ssh into your box to get it.

The waterblocks/coldplates are basically corsair H60s (which coolIT also was part of developing).  Any intel LGA 1150 (and others) compatible block should fit providing you have an X type mounting bracket with slots allowing for about 72mm spacing on the screws.  I would avoid any round screw hole and stick with slots like the cointerrible unit ships with to ensure that you can get the screws in exactly the right place.  This means people can experiment with alternate blocks, radiators and pumps to see what works best for them.  My quick measurements on a running unit (so I couldnt get the calipers in as good as I wanted) showed the exterior dimensions of the block is 2.5 inches or about 64mm and the screws center to center are 2.88 inches or about 72mm.  I know for a fact that I did not take good or exact measurements mostly because I was afraid of accidentally touching something with metal calipers and partly because I wanted approximate dimensions to see what would fit.  If you do replace the coldplate I strongly suggest that the part that touches the chips be made of copper and that the water channel be cut into that copper plate.  Copper allows for better heat transfer into the fluid than aluminum (while aluminum works better in the radiators to get the heat out).  The pump appears to use the standard 3 pin connector with the pins in the standard positions.


PSU datasheet http://www.power-one.com/sites/power-one.com/files/documents/power/datasheet/bcd.00012_ai_pfe1100-12-054xa.pdf
MCU datasheet http://www.ti.com/product/tm4c1233h6pge


The salvage price for various components in the cointerrible miner puts it at $1k or maybe a bit more if you have a good salvage guy.  Soon they will be selling well below this price point and just parting them out could be more profitable than mining with them


insturctions to extract firmware

First let me say that if you need these you probably shouldnt be modifying firmware for your device.  However sometimes all someone needs is a nudge in the right direction to notice and/or fix something that should have already been fixed.  There are plenty of things that can be fixed on these boxes.  The beaglebone is a ARM AM335x cpu so anything you build/add/replace must adhere to that platform.

The web stuff is php and one thing that would be easy to extend is the power stepping (and possibly adding the ps_load capability).  Cointerrible gives you 0-9 to step through, however the cgminer driver takes 0-255.  You appear to be able to do 25MHz steps if you can get finer grained control over this than the crude stepping they give you.  ps_load is 0-100 with a hardcoded default of 0.  I am unsure exactly what ps_load does exactly.  I was unable to observe any real device changes in the brief time I looked at things.

This at least opens the door for more people to fix many of the problems with the non-CTA firmware, and opens the door for a few who can work on the CTA firmware.  One is obviously harder than the other.  You will need mkimage from uboot to rebuild rdImage back so it can be flashed using the web interface.


In /opt you will see a few files.  Of interest is:
dfu-util and cointool which has symbolic links made for cointool-info and cointool-upgrade

dfu-util can reset the board (0 or 1)
cointool-info can query the board for its info - serial numbers, CTA version, hardware version, etc
cointool-upgrade can load the firmware of your choice, or even downgrade (default script loads anything that is not equal to the version on the board)
GSA_0.4.5.dfu is the firmware itself for the board.  

I have not yet looked for a recovery method in case you brick the board.  There should be jtag or something to allow you to load a minimal firmware that then lets you load a working one (jtag is slow++ and the MCU natively does usb and other faster stuff).  My guess is reviewing the data sheet on the MCU would yield that answer fairly quickly, but alas I have no time.  I have not even looked at the board for headers or contacts or silkscreening that would indicate where you can do this.  I am guessing there are contacts somewhere that directly connect to pins on the MCU that would enable a reflash for recovery but until this is resolved I personally would not reflash it with anything not from cointerrible.

If you try to be clever and use binwalk on the GSA firmware you will be disappointed.  It claims it contains a LZMA compressed data section and the uncompressed size is 9000 yottabytes.  This is obviously incorrect, which leads me to think that it is a direct image for the ARM MCU.  I have yet to further analyze the file to see if this is true, but occams razor and all that.  While its compiled size is only 55k, and it is ARM which generally have 4 byte opcodes usually with the same prefix that does not mean it will be fast, easy or trivial to reverse the firmware, then again maybe it will be.  I do know already that they use a PLL to set the asic chip frequency, so there is a potential however weak that one could play with that for overclocking.  The PLL is probably controlled via one of the 106 GPIO lines the MCU has.  There may also be some speedups that can happen during an audit of the code to see if they do anything in a silly way that is inefficient.

 Hi All,


I live in a country where it was hard to find Coollaboratory Liquid Pro

will Coollaboratory Liquid Ultra be okay ?? i found and bought 2 packs but just need a heads up on that

thanks in advance

I would like to thank all you good people for the good advice and technical support. I received my Innovative Cooling Diamond 24 Carat thermal paste yesterday and today I took my cointerra apart removing the cooling blocks and cleaning off all the shitty cointerra paste. I reapplied the Diamond 24 smearing a thin coat over the chips and then a bb sized drop in the middle. I reattached the heat sync blocks and fired the machine up. It is now running better than when cointerra shipped it to me. Temperatures have dropped from 120 plus to 50s and 60s. The crashing has completely gone away. I tell you it is bringing tears to my eyes. This thing has never run this good. Gentlemen my hat is off to you guys. Hip Hip hurray! 

Hi there. I am running a early Jan (hm, I mean late Feb) batch Terraminer. It ran fine the last months, but yesterday one board stopped hashing. The pump is listed as 0 pump speed.
Which alternative pump do you suggest? Are they easy to replace? I just saw single alternative pumps and not a double system like Cointerrra build.

Thanks for helping.

So, I have basically fixed my "1 core dead" issue by pump replacement + Liquid Pro.

These pumps are like shit, burn too easily.

Now I am left with the only problem, of "half board not receiving power" and the PSU is showing an orange little light.

I tried to pull off everything from the faulty board, even the power supply black / red cable.  PSU turns back to double green light.  Problem should therefore on the board.

Any successful story out there for fixing this orange light issue? 

If you have a Cointerra machine and you feel they haven't lived up to their promised service or performance, here's another link you can try to get some resolution.  You won't have to pay for it either:

http://www.bbb.org/central-texas/Business-Reviews/virtual-currency/cointerra-inc-in-austin-tx-1000104440/