 Show Posts
|
|
Pages: « 1 [2] 3 »
|
I think you are worrying to much about the heat. You realize that 70F is only 21C. Here in Texas last summer I was running a bunch of GPU miners. They run at 70C which is 158F. The hot Texas summer day of 115F (46.1) air was enough to cool the miners down.
This year I have a bunch of S1 Antminers and we have already had high 95+F days so far. The S1's did fine with just some outside air being blown into the room with the miners and the hot air leaving the room. In the main mining room with the building closed up, the room was at 115F (46.1C). I opened the doors and windows to get the air circulating and the room dropped to near outside ambient temp in no time.
I don't see that much of a problem here when it gets to summer temps at 115F (46.1C), just keep the air moving with outside air and have several replacement miner fans on hand as fans fail when you don't want them to. If you cannot create the air exchange from outside with the inside, then you need to rethink you setup. You need air flow from outside to inside to outside.
This reminds me of the late 80's and early 90's when small and medium businesses where having their computers networked. I could not count how many times I was called in to troubleshoot network problems only to find a server or router closet. That was the dumbest idea I seen done over and over by inexperienced so called "tech's". Putting electronic equipment that produces heat into a closed off, normally not air cooled, closet. You have to have air exchanging to keep the equipment running. Same with the miners.
Yes, 70F outside is far from a crisis for the miners, it's just a harbinger of what is to come. Perhaps you like your house at 85 to 95F; I, and especially my wife, do not. We are Minnesotans. We are used to temps of 10F outside, not 100F. I keep my miners in the garage. The garage doesn't have any windows. I prefer not to advertise the fact that I have lots of nice electronics in there. With the door closed and it's upwards of 70F outside, my miners (Ten S1s and six Cubes) bring the garage to 85F in under a half hour and the miners' temp sensors quickly head north of 75C. The cubes drop 10GH apiece at the high clock rate. They don't have temp gauges that I can see via the web interface. When the Cubes go over whatever is their maximum temp, they drop down to the low clock rate. The error rate of the S1s increases from under 100 per hour to several ten thousand per hour. The number of invalid/discarded/rejected shares jumps from tens to tens of thousands. Then the chips on all the miners randomly start getting flagged with X's a few at a time. So no, the miners are not happy at 50C or higher ambient temps. When it hits 80F and 90F outside, forget it, the garage turns into an oven. Secondly, I doubt your miners are doing "just fine." I'd be interested what temp the S1s are reporting on their miner status web page. If the room is already 50+C, I'd bet they are running at 70 to 80+C. At those temps you can look forward to cumulatively degrading hashing rates and/or chips failing. Me, being an electrical design engineer, I want my miners to perform as best they can for as long as they can. Right now, when they are operating below 50C, they get hashing rates around 220GH consistently, with no permanent X's on any chips. Lastly, again, being an electronics design engineer, I tend to be a little anal about keeping my electronics running in the "low to mid" temp range stated on a devices data sheet, if I can, and stay far away from the "absolute maximum" temp range. So, all those who like to bake their miners and shorten their useful life, you can line up with AbiTxGroup. I'll keep it cool thank you. It's also about airflow, not just temps. In many ways, 20k CFM of -10 degree air = 10k CFM of -20 degree air. Cut a hole and mount an exhaust fan high up in your garage. Cut another one and mount an intake fan low on the far side (north/shaded side is best for intake). With the right capacity fans, you can make it so that your garage never rises more than a couple degrees above the exterior temps. This will be much cheaper and less trouble-prone than your proposed water-cooling design. Speaking of which, the "thermal mass" of a 340Gal tank isn't going to do jack. (Although of course it's a good idea to have a reservoir in case something goes wrong with your plumbing.) If you've got cool water coming in from your tap, running through your system, and then running down the drain, the system is going to use so many GPM of cool water regardless of whether it's initially stored in a tank. The stagnant air in your garage hits 85F in half an hour, with exterior temps of 70F. Your miners are putting X BTU's into the garage at a constant rate. The rate of heat exchange is related to the temperature differential. At 15F differential, there is enough heat exchange going on between the garage and the exterior to maintain equilibrium. This is how all cooling systems work. All you're doing is moving heat from one place to another via a medium. The "thermal mass" of the medium of heat exchange doesn't matter. What matters is the overall heat exchange rate. The amount of water stored in my car's cooling system doesn't make a difference. Once it hits the minimum level required to fill the radiator, engine block, and all connecting lines, having an extra gallon or two in there won't make a difference. (It will take a bit longer to reach equilibrium temps from a cold start, but other than that, no difference.) What WILL make a difference is if my water pump goes out and the flow of water through the engine stops. In that case, it won't matter how many gallons of coolant I have in my reservoir; the engine will overheat. Same with the radiator being blocked. If air is not flowing over the radiator at a sufficient rate, the engine will overheat. (Dependent upon the air temps, of course -- I know you northerners sometimes block your radiators with cardboard in the winter when the temps hit the negatives. In this case, there is sufficient heat exchange even without any airflow, due to extreme temperature differential.) I apologize for the continued derailment of the thread. I'll keep the reply shorter this time. You have a substantial misunderstanding of the system I am designing and using. I am currently using this method of water cooling. It works as designed. NO airflow AT ALL is required. NO garage modifications are required, short of adding some plumbing. The heat is transferred to the water by waterblocks. The heated water is then returned to the tank. When the water in the tank is too hot to keep the miners at optimal temperature, the hot water is DISPOSED down a drain. The heat leaves with the water. The purpose of the thermal mass water tank is to recirculate the water until it reaches the temperature where disposal is warranted. The disposed water is replaced by fresh cold water from the water main at a much slower rate due to recirculation of the tank of water. This is not a closed system like a cars water cooling system which has to dispose of the heat to the air. When I was running the miners with just fans, I just opened the bottom of the garage door to exchange air. That worked fine, but I know worse is coming. That was when I started designing this water cooling system. The garage only heated up so severely when the garage was closed up AND the miners are running with fans only. That is not the way I run the miners. It was an experiment to see how fast the heat accumulated. The 70F degree air outside right now is NOT the concern. It is the 85+F temps that are coming that are the problem. When starting with such temps no amount of air exchange between indoors and outdoors will cool the miners below the ambient temperature. In fact, considering the 570+ watts of heat dissipation by the miners, the chips have little chance to get below 70C (158F). As I noted above, I have observed significant performance degradation, malfunctions and even chip failures begin when the S1 web interface is reporting temps of 60C (140F) or higher. End of thread derailment. Everyone here knows how water cooling systems work. Please get a blog for this stuff - or visit the hardware topics... |
|
|
|
I, too, have several addresses that I moved my miners from well over 10-days ago from that have been saying "a manual payout no sooner than a few seconds from now." for going onto a week.
I, too, happen to be one of the douche-bags that's more than a little curious about the NMC backlog...
I'm also growing more than a little frustrated that everytime my miners hit the payout queue, my payouts are stuck behind 12, 20, 40+ block delays...
My last 4 payouts have been right on schedule. Today's was delayed by a few hours, but that's it. M I don't doubt that, and I'm sure that everyone's mileage varies; however, almost everyone of my payouts over the last month has been delayed. My last one was only about 6 blocks delayed, but I'm currently in the queue again now for a small payout and I'm *finally* only 12-blocks back--this one started out yesterday sometime at around 26-blocks back. The record I recall seeing on one of my addresses was a 54-block delay. All said, I'm actually quite patient and I don't *need* the BTC; however, on top of everything else that's been going on, I just see hiccups everywhere. This causes me concern. Luke-JR's most recent post a few minutes ago does give me some solace that WK didn't take a stiletto to the grill, but the wrinkles in the pool are proving a little less than comforting, atm. |
|
|
|
Stats server is still under DDoS attack, causing some screwy stats in some random cases because of database lag from the attack.
Pool server, as usual, is not actually affected and your earnings for your accepted shares are not affected.
If the DDoS is ongoing, then you are doing an awesome job! My stats are back to normal for almost 24hrs now without incident and the payout queue has caught up nicely  It is and has been ongoing for a while... I think 3 weeks now.  I've put some scripts in place to protect the web server a bit more and it is helping. This weekend I will be working on a major overhaul to Eligius's back end which should help with any remaining pool side stability problems caused by these DDoS attacks. If I have time I will try to do the same to the stats, and try to finish the remaining code needed to re-enabled namecoin payouts and do namecoin backpay. Unfortunately I've been extremely busy all around with non-Eligius items, and the time I have been putting into Eligius has been DDoS mitigation So, better times to come. Some important notes:High hash rate miners (I'll start with 10Th/sec+): Please send me a private message, email (forum name at gmail), submit a support ticket, or message me on freenode IRC with a signed message from your mining address requesting private server access if you desire it. I plan on having a private node setup (beyond the current automated stuff) this weekend for top miners to use exclusively. This node will have higher work difficulties and other settings best for high hash rate miners. Will be a trial run, and we'll take it from there. I assume all will go well though. This server will function with the rest of the pool normally and you won't notice any difference besides potentially improved stability. Stats and everything else remain the same. I will say that you must be mining on Eligius at 10Th+ for at least 24 hours to qualify. Next, this weekend the "KnC port" (12234) will be redirected to the general stratum mining port (3334). Anyone still using port 12234 will still be able to for now, but please switch back to 3334. The firmware issues seem to have been resolved by KnC, so, just upgrade your miners and you should be fine on the main pool. This will prevent the KnC port failsafes (blocks that pay the cold wallet 1ChANGe address) and help the payout queue stay shorter longer (less manual payouts). Thanks! It's going to be a pretty nice weekend, and I'll be inside working on the pool, so... *queue death stares from g/f*  -wk Save for one silly post about the engraving the offline wallet, anyone else take notice that WK hasn't posted in 10-days? Either the death stares from his gf were no joke, or perhaps the DDoS has finally caused WK to throw in the towel here?? I'm a longtimer here on Eligius, and I've been as vocal as anyone at supporting the pool during thick and thin, all the way back to the days when it sports a fraction of a % of the total network hashrate; however, I do admit that the signs of burnout in his latest posts were *extremely* evident, despite everyone (myself included) throwing more and more donation to try to make the work on the pool pay... Now, I will admit that I am seriously concerned at the lack of input from WK... this thread has been pretty much hijacked by a bunch of dickweed bickering over who's rigs are hot, how much water is in California, and whether or not anyone other than Eskimos live in Minnesota... I, personally, would LOVE to hear how that last weekend of work--two weekends ago--went? I, too, have several addresses that I moved my miners from well over 10-days ago from that have been saying "a manual payout no sooner than a few seconds from now." for going onto a week. I, too, happen to be one of the douche-bags that's more than a little curious about the NMC backlog... I'm also growing more than a little frustrated that everytime my miners hit the payout queue, my payouts are stuck behind 12, 20, 40+ block delays... I'm not trying to stir up trouble, and I'm sure this post will probably get me flamed by all the *pros* on here--but can we PA-LEASE get some focus back on this thread?? Where's WK? Is he ALIVE? Hello...  |
|
|
|
So where can we see some info on what you have in the works MrTeal?
I got BTC burning away ready to get in on a new 28nm miner.
+1 Details...? |
|
|
|
even the best psu will fail if pushed to 80% of its rating for 24
so this would indicate your unit has an absolute max of 440w usable.
Running 750 watt Corsairs here at 700 watts with no problems, so I think a good supply makes a difference. Then again I ran it at 800+ for a few days, so a really good power supply makes a difference. C I've been running 2 off of this for couple of months. @750w, sounds like maybe the same Corsairs. No other draw - zero problems. ymmv http://www.newegg.com/Product/Product.aspx?Item=N82E16817139051-r |
|
|
|
Hi,
where can I find the firmwares & flash-tool - any idea? I look around everywhere... 3 boards make some problems (SICK) and I want to test around if another firmware can fix it.
Thanks!
- Hope everything is OK at you Lucko... was it snowing strong again?
MrTeal maintains this stuff on the Chili thread: https://bitcointalk.org/index.php?topic=304250.msg3257013#msg3257013I have 2 of Lucko's boards. One of them runs better with the 1.1v limiting fw, while the other works best with the "standard" fw. |
|
|
|
|
The only symptom that sounds different for me, they never show SICK in bfg -- they just disappear after a few minutes.
I only have 2 cards. They both do the same exact thing as this. They start normal and ChiliFlash shows everything within normal parameters. Start mining and they ramp up to about 22-26ghs for anywhere from 30-60 sec. If I don't do the warming, they start throwing more errors, hash rate falls until I get a flood of the temp errors and then they'll disappear from the bfg console.
If I kill bfg and go back to ChiliFlash, the device still opens, but it shows only 2 engines still enabled. Power cycling the Chilis is required to bring them back.
Not to confuse the thread, as so many have the 1.1v fw working for them, but for some reason neither of my boards worked better with it -- they both actually sermed worse, as o barely had time to pick up the hairdryer before they'd die.... I decided to reflash the "standard" Chili14e fw before giving up on them and to my surprise, they were stable enough to be able to use the warming technique. Once I got this worked out, based on all the previous cooling/mounting/shielding research of others on this thread, now they run nice and steady about 32-34ghs -- up for well over 48 hrs now.
I have 2 old Arctic USB fans rigged on a hub pointing down onto the pcbs for long term cooling -- hairdryer on cool to cool it down during the startup dance...
|
|
|
|
+1 the hairdryer, af_newbie. Get the hairdryer ready, then start it hashing. As soon as it's hashing, start blasting the top of the fets with it. Give it a minute -- the hash rate should stabilize for about 30 sec, then it'll start to fall -- switch over to cool air and start cooling the back of the pcb -- when you get good cooling on the back, hash rate will start climbing along with the reported asic temp -- when airflow keeps this around 68-70 degrees, it should run. I've been average 32-33ghs for over 48 hours now after getting this one going like this. One twist - I'm running the "regular" Chili14e fw - the 1.1v limited version would crash this board after about 30 sec of hashing, no matter what... Good luck. |
|
|
|
Update on my one running board. It ran around 24-28ghs for about an hour the first day - right up until it just disappeared from the peon. I reset a time or two and couldn't get it to run for a more than a minute at that point. I redid all the cooling and moved it to a windows box -reflashed the fw to the 1.1v limiting version. I was having tons of stability issues after that. I couldn't do anything to keep it hashing for more than 30 seconds. Hairdryer, nothing. {I also was getting some weird comm port issues, but I think that was the windows machine puking} I moved it to a different host again (same psu) and decided to reflash back to the "normal" Chili14e fw. After that, things seemed to work much better for me. I do have to do the hairdryer trick to get it running. And it's hanging upside down with a cardboard shroud over the caps & fets - with the backside open to allow a small usb arctic fan on blast on the back of the pcb - it's a headache, but it's been hashing between 32-34ghs for about 18 hours like this reporting about 6% errors. Seems the temp in bfg stabilizes at 70c when you get the airflow just right - then it runs. Still waiting on replacement capacitor to fix the other one. https://s3.amazonaws.com/img_host/20140131_234257.jpghttps://s3.amazonaws.com/img_host/20140131_234307.jpghttps://s3.amazonaws.com/img_host/20140131_234339.jpg |
|
|
|
Got this first board together with the Evo212 and it's hashing between 30-32GHS. Heatsinks for the backside delayed so running without - but it is upside down in 14.4C ambient.  Figured I'd try my luck on minepeon, not optimistically -- but it fired right up and ramped up to here. Temp in peon showing 51C. Don't want to jinx it for now, so just gonna watch it for a while. I'll reseat the heatsinks move it to another machine I can tweak and will push it later, but seems a decent set of chips so far... Glad it's alive! Nice board. Thanks Lucko! |
|
|
|
Damn! That was fast. Thank you very much MrTeal! |
|
|
|
I have found a local shop that has almost everything I needed to get my boards repaired from shipping. Getting them interested in bitcoin now, too...  We have one of my boards fixed up and ready to go. Putting cooling on and powering it up tonight - everything looks ok this far. One of the capacitors for the other one was damaged beyond repair though. The contacts were bent badly when it came off the board. It wouldn't make good contact that way so had to try to straighten it a little. Snapped it the rest of the way trying... My shop doesn't have these exact caps at hand, but they gave me a part # for it to order a few. Lucko, MrTeal or ChipGeek - would one of you be able to recommend a proper source/part # for getting these? EDIT: I'll take any recommendation from this crowd - I trust a lot of you more than myself, but I figured I'd ask the gurus first... |
|
|
|
Hi Lucko - first, I want to say I am NOT upset, in any way. These things happen and my feedback is to hopefully help other's still waiting to get their boards without damage. I am happy I was able to be part of this gb - you have been excellent to work with and I would join you in more projects in future, without doubt. Thanks to MrTeal, too, for licensing the design to our group. I'm sure the outcome hasn't been what you hoped, but you've been a big help getting things straight. Back to boards and shipping. I believe that if there was something rigid or semi-rigid placed in between the boards before they are wrapped that this wouldn't happen. A small piece of cardboard, or even a small foam packing sheet would do. I believe that your packing is a good plan, but it looks like it allowed the boards to shift laterally on top of each other during loading dock football shipping, so the end of one board pushed the caps laterally and off the solder. I think that's the core problem. It looks like the caps were pushed away from the boards at the top: EDIT: In reviewing my own pictures, it becomes obvious that the power block is what pushed these same caps off when the boards are rotated to each other...https://s3.amazonaws.com/www.annagreentechnology.com/lucko-chili/20140127_133615.jpgThat said, the solder doesn't look very good to begin with either, as you can see from this pic. Neither the cap nor board in this pic shows signs of decent soldering, and there's absolutely nothing at all to clean off to redo it... I don't think this was the main issue, but I don't think it helped: https://s3.amazonaws.com/www.annagreentechnology.com/lucko-chili/20140127_133622.jpgThen there's the box - like GrapeApe indicated of his, no signs of "abuse". The boards were definitely a tight fit in it and it has clearly traveled a long way (my have even been examined by a friendly government safety representative at some point) - but it doesn't look like it was "abused": https://s3.amazonaws.com/www.annagreentechnology.com/lucko-chili/20140127_133357.jpghttps://s3.amazonaws.com/www.annagreentechnology.com/lucko-chili/20140127_133406.jpgSummary:I hope this helps figure out how this has happened on 2 shipments. I'm not upset, although my boards are not hashing yet. I'll be attempting to fix these myself later today when I have time to warm up the soldering iron. I have the tools and I can do this work, but I am not an electrical engineer, so I can only hope it goes smooth and there is no other unseen damage. If anyone has any suggestions or guidance on things I should look out for, or be careful of, or general tips on cleaning/re-soldering/testing things - I welcome any help. Seems simple enough so I'm confident it won't be an issue. Not as lucky as GrapeApe in having ChipGeek right up the road - I'm in the Charlottesville Virginia area, so if there's anyone around me that has some better thoughts, I'm all ears... Thanks - sorry for the long-read. |
|
|
|
|
I received my boards today, as well. Unfortunately, they both have same damage as GrapeApes... :-(
Posting from my phone so not sure how to include pics in forum -- will post when home later.
Package does not appear damaged or mis-handled...
EDIT: Not going to give anyone a hard time here, but this is 100% due to bad packing. Boards shipped face-to-face wrapped together in single piece of bubble wrap in a box that they *barely* fit in... No offense Lucko, but these were never going to go all the way around the world undamaged. I can fix these, but definitely suggest better packing for all future shipping. No doubt in my mind.
|
|
|
|
I, for one, think threats of suing someone are chickenshit... How many Chilis are out there hashing? How would this gb be doing without his help troubleshooting? Design your own board and get to hashing. Leave MrTeal alone. I don't know him. Don't know you. Don't know anyone on here, but what I've seen on this forum, bitcoin is bringing out the worst in quite a few of you.  EDIT: not to mention, these people bitching are probably the ones not willing to accept the 30-32GHS firmware that both MrTeal and Luckos Chilis were offered to do... probably whining you can't overclock the shit out of it... piss off. |
|
|
|
any news about patch kncminer ?
He's coded a special ver of cgminer AND provides a special port for you precious little device... {Ignore you now} |
|
|
|
Sorry to be a pain the ass.
I think it's unfair to yourself to think this is your fault, GrapeApe... [edit] Not that it's Lucko's either! |
|
|
|
Ugh... Sorry to see that GrapeApe...  I too am expecting 2 boards - delivery was estimated for today, but now delayed until Monday for unknown reason. Praying my boards won't be delivered in the same condition... |
|
|
|
Looks like I'm on the 24th! Thank you, Lucko! |
|
|
|
|
