Is there not a better or more efficient and large-scale way to supply power to the miners than by purchasing individual PSU's for each? IF you are talking about a few hundred KW of power on up, then yes using a high DC bus fed from the mains followed by local bus DC-to-12vDC is worth considering. HP, IBM and others all have DC fed 12v PSU's available just for that purpose. The power eff advantage comes from losing all of the the normal input brute-force mains-to-DC and PFC power conversion stages found in each of the normally AC-fed local 12v PSU's. Instead a single several hundred KVA 3-phase transformer with full-wave rectified delta-wye secondary provides what is essentially pure DC with less than 0.7% ripple and no further filtering needed. Typically that high DC bus is around 300vdc or more which is also what the input stage to a normal 208vac PSU feeds to the power conversion stage that make the 12v. |
|
|
|
That actually makes me think, you wouldn't get temperature data since it uses different data lines but I bet you could drive an S5 board with a USB/UART adapter using vh's new code. That's basically what the stick is, is a 2-chip version of an S5 board.
Well I *do* have one more s5 if you need one to play with... |
|
|
|
Well that is slightly easier than altering the /etc/rc.local file every time and rebooting.
Um, ja... Back when I ran my 10GHs BFL cubes I tried using cgminer and for the life of me could not get the hang of it. BFGminer was plug-and-play with a CLI. I do believe there are some 3rd-party GUI programs for it as well. |
|
|
|
Nice that as I recall you have a technical college in yer area. Ah student labor  |
|
|
|
|
On the Compac at least yes. It just stops then restarts the miner @ whatever freq in MHz you typed in when prompted. I never tried on other BM based miners but since the s1-s3 also were using OpenWRT should be the same. Hell, it might even support the s5 since those are where the Compac chips came from...
|
|
|
|
Somewhere on github there exists a cgminer fork with Compac support built right into the Icarus driver. That was sorta VH's "skills test" before working on developing drivers for new stuff, and I think it runs a bit smoother than Novak's original setup. I don't know if he plans to integrage the new 2-chip Compac code into that or not, or just scrap it - we're going to test how well the 2-chip code works by default on an original single-chip stick one of these days.
Someone did put together a Minera image with cgminer-gekko built right in but the link is buried about 90 pages back and might be broken anyway since that was a year ago.
Ya also should know that BFGminer supports the Compac as well, no outside drivers needed (eg Zaadig)  https://bitcointalk.org/index.php?topic=877081.0 From what I can tell BFGminer is largely based on OpenWRT. I use the latest version to run the Compac I got from phillipma, in setup told it to search for miners and bingo found the Compac, runs it fine, can even change freq at will. btw: it also supports directly running s1, 2, &3's on Pi's... |
|
|
|
Reckon those have a switch to select CC or CV? Looks like CV mode would be fine.
Agreed. If it has a selectable CV mode to power banks of luminaires or panels each with its own CC circuits. Just be aware that most LED supplies are constant-current devices to drive things like these https://images.jetcdn.net/md5/a0f47f3aa6d7435aa4fe26be51b01019.500 (in that particular model min threshold voltage is 46v though) |
|
|
|
Folks, that supply is for LED lighting and is specifically made for power LED strings in lighting panels! It is not a general purpose power supply. Those supplies are made to deliver stable fixed or adjustable (when using dimming funct) current - not stable voltage! Typically that means that there is a set max voltage to light the LED's (typical conduction threshold drop is 1.1 to 1.6v per-LED depending on the type) and then voltage drops as needed to maintain desired current through the string. (LED brightness is set by the amount current flow through them and that is only indirectly controlled by voltage). Now yes, if the current is dialed to max then the voltage adjust might do fine and keep ya at 12v but - point is these supplies are not designed to hold voltage steady. If you exceed the dialed in current the voltage out will drop as low as needed to produce the set current and odds are that voltage will be far below what a miner needs leading to possible Vcore regulator damage. Just putting this warning out there.... |
|
|
|
<snip>
my powered hubs have a brick, which output is 12V through round connector (OK, barrel jack). Then, you connect the barrel jack to powerhub itself, which then provides 5V to USB. therefore, one can just take the brick and shove the barrel jack into pod, I assume. i don't see why not.
Of course, if you want to buy larger PSU, go for it. It happens that I have a multitude of such powered hubs and thought of re-purposing them.
Then yer golden. Edited my post to reflect not using a standard USB hub supply vs a custom or BC1.2 high power hub & supply. |
|
|
|
<snip>
I have, perhaps, a naive question. It might have been mentioned somewhere already, my apologies.
Say, 10-chip BM1385 pod needing 60W. What would be the proper/best way to power it?
I get the sticks, you connect powerhub's little brick to powerhub via round connector, then power the stick via USB port.
the hub would have the same type of connector (round), PCIe or something else altogether?
EDIT: OK so round powered USB connector=barrel jack, so i assume that as long as it provides 60W, it should be OK to just re-purpose the powered hub power supply.
Almost. AFAIK the USB does not supply power. It's just coms. Power input is only through the barrel (a 2-pin power connector) or PCIe jacks and power input is 12V nominal, not the 5v as a ( edit: typical standard) USB hub supply usually provides. Example of a PCIe to barrel adapter... https://www.google.com/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&cad=rja&uact=8&ved=0ahUKEwj_k4mH5_HQAhWLgFQKHbtpDUEQjRwIBw&url=https%3A%2F%2Fbitcoinware.net%2Fproducts%2Fgridseed-asic-accessories-highest-quality-pci-e-to-barrel-plug-power-adapter&psig=AFQjCNHL5XxyX0y60Xz6l2YdsEG_VnWtEw&ust=1481740337415027 |
|
|
|
I have to admit I think the R4 is one of the most interesting miners available. I am curious if R4 owners have seen the issues that the S9 has. I wanted to jump on the batch 2, but without some more data on reliability I decided to wait. I would love to get your feedback on the R4. I am very tempted if another batch is released to pickup one or two to give them a try. I think the idea of mounting them on the side makes sense. Has anyone rigged up a mount yet for side mounting these? Thanks for sharing your R4 with us!  My 1 concern with mounting them on their side is the fan bearings. Odds are they are simple sleeve bearings made for the axle being horizontal meaning minimal shaft support if tipped on its side. Pretty safe bet what support there is is not a ball bearing which would make things more forgiving... Anyone have a closeup pic of the bearing area? |
|
|
|
|
PM sent with order/payment info for 2x of the 2Pac's
|
|
|
|
|
I've always used UPS and until the last 3 orders, for the past 3 years all went from Shenzhen to Anchorage and then DTW. The last few s9's I've gotten, the first of them took the scenic route: went from Shenzhen > Soul Korea > Seattle > DTW, the last 2 orders were Shenzhen >Narita > Osaka (via truck I'd think?) > Anchorage > DTW.
Most odd.
|
|
|
|
|
As an FYI reminder, autotuning was introduced with the batch 16 and 17 production runs.
|
|
|
|
|
Running 10 s9's (along with 15 s7's) here, from batch 1 to batch 23, only 2 board failures to date: 1 from my batch 1 and 1 board from a batch 17. Both sent to BitmainWarranty in CO - yes even the one from batch 17 which was still well under the 90-day. Just faster, no customs issues, and cheaper shipping inc being fully insured for full $450 value.
All run in a clean industrial environment, most in 2nd floor parts storage area - gets up to 90F in the summer.
PSU's are mostly Bitmains 1600w which also run the bulk of my s7's along with a few IBM 2kw PSU's using Sidehackhacks breakout.
I agree with your ID layout. After all, the boards slide in/out of the exhaust end of the case so makes sense to use that as the reference end. Looking at it and moving left-to-right we have boards 1-2-3.
Controller socket numbering OTH... They are all over the map from batch to batch. Think my B1 has 5 sockets on it and uses every-other-one....
Definitely agree on ESD: ALWAYS either ground yourself before touching/handling a hashboard or better yet ware an ESD grounding wrist strap.
If you do do repairs on the boards or at least some trouble shooting, it would be nice to know what the failures are. So far Bitmain and BitmainWarranty just give excuses why the repairs are not fully documented and why they can't tell us what failed. Given the low cost of the (2) repairs I've had done so far I'm thinking it may be PIC data getting munged and just needing to be reprogramed or if an actual physical part, something in the Vcore reg or perhaps a node-bypass cap.
|
|
|
|
I got the R4 today as well and it is huge indeed Working nicely and quiet. I bought Enermax Revolution 87+ 1000W PSU from mammon for this and it is a great choice (I also have previous good experiences with this PSU). You just need a PCI-E power splitter or one more modular PCI-E cable and you're good to go with this PSU. But I still asked again Bitmain for the possibilty to manipulate the settings myself. As I see it in the kernel log, it is the PIC firmware version in the hash board, which is programmed for the auto-tune setting and once the controller notices it, it disables user freq settings possibility. <snip> Welllll.... it all rather depends on what changes there are to the hashboard circuitry vs the first batch. This applies to the s9 boards as well. Can you take a pic of the front & back of the Vcore regulator area of an R4 board? Mainly interested in seeing if someone else can post a pic of a batch 1 so some differences can/might be seen. If there are any then probably game over right there. That follows with someone hooking up a PIC programmer and pulling the code from the PIC and then knowing how to read and hack it. Programmer I got from doing the Sidehack UV/UC mods to my s7's. Reading & hacking the firmware is definitely outside of my province. |
|
|
|
I know you can use the insteon or X10 switching plugs.
But I'd rather switch the 5vdc line for the PSU.
I travel a lot and am looking for something to do this remotely, from anywhere in the world effectively.
Wifi or wired based controller with a web or phone app?
I'm sure guys are doing it. Just trying to trackd own my best choice, and find something I can buy and have delivered quickly.
Lot's of ebay options, but the lack of info and wait time to get it. Is to much.
So let's see what you guys have going? Rasp pi maybe and usb relays?
Do a search for RasPi or BeagleBone or Audrio and home automation. You'll find tons of stuff including the relay boards. I don't use it myself and rarely need to power cycle my miners but - would have come handy at least a couple times... |
|
|
|
Also pitch is good not a nasty whistle like the s-5 s-7 s-9 That is a direct result of using the longitudinal blower vs fans. The blower produces a linear stream of airflow vs the mass of highly turbulent air a fan makes. Smooth airflow hitting the heat sinks means minimal noise from turbulence and also more to the point of 'pitch', with no rotating blades moving in a circle very close to the heatsink fins there is no possibility of the siren effect. Should be just a soft 'whoosh'. |
|
|
|
Anyone that purchased multiple units seeing a large range of chip temps by unit?
One of my units is showing 105/79 and 93/103. The other that is sitting right next to it, with same airflow and ambient temps showing 70/70 and 55/75. Doesn't make sense. Do you think the lower temperature unit is reading the temps incorrectly by chance? The other 2 R4 batch 1 unit chip temps I have in the room are all about 88c~
Ja temps on the auto tune miners will vary a lot from miner to miner. Part of what auto-tune seems to do is look at hashrate vs temp and use that data as part of the Majik Sauce of speed, temp, and Vcore used to produce as-advertised total hashrate. That process lets them be less picky about what params are best for any particular group of chips from the Foundry and that in turn gives them a lot more usable chips. -- If any individual hashboard(s) under-performs they make sure a better one is in the same miner to give advertised speed. |
|
|
|
|
Show Posts
