Hope you got smart and put both bad boards in one for the return
Um ONLY if Bitmain has said that you can do that! Breaking the seals over the screws to swap boards w/o authorization is a great way to void the warranty.... |
|
|
|
Great - Thanks! BTW - How do I see what the current release and/or firmware I have? The system/overview info for both R4s appear to be identical....  Thanks again (sorry for the newbie questions) System > Overview, look at the date after File System Version. That is actually the firmware version. For Bitmains downloads the date is given in YR-MO-DAY-24hr time format eg. 201612020035 is 2016 Dec 2 00:35 Eastern Time |
|
|
|
yup really looks like reliability is finally fixed. Instead of trying to push to limits and then dying like flies
Ah the time frames of Youth... A few miners with a few days of run time on them does not equal better reliability than a s9. Most folks have zero trouble with the s9's at the start (and minimal issues during its lifetime). Out of the 16 s9's starting from batch-1 that I have, only 2 bad boards to date. At least give them a week, preferably > a month and THEN make that call... Now better than the latest R4's -- based on my 1 bad out of 3, so far I'll give ya that one. |
|
|
|
Thanks for giving us a useful website for the miners but I'm currently looking for a place where I can get the pandaminer or similar for cheap. I don't want to buy from their website directly because it's 2.4Btc but I want something cheaper like 1.5BTC miner with the qualities of the 8 core pandaminer And I'd like to fly and spit diamonds. Don't see that happening either. Also the Panda is for Altcoins not BTC so asking in wrong area. |
|
|
|
Yep, right spot. Yours is dated Nov 25, 2016 Just as with my 1st batch-6 (with bad board after 22hrs so is 1/2 speed, Dec 2 firmware), this latest batch-6 has eerily smooth hashrate compared to the s9's. ref:  I wonder if it has to do with this bit of code the log spits out and the miner is per-chip tweaking to give advertised spec? miner total rate=8898GH/s fixed rate=8700GH/s The R4 Batch-4 in my main farm at work is just as rugged looking as those s9's are. Hmm... |
|
|
|
Got my other R4 batch-6 today: so far so good, holding a very even 8.7-8.9THs and all temps under 100C Interesting thing: My 1st batch-6 (which had a board crap out 22hrs after running it) which I got last Tuesday has firmware dated Dec. 2 2016 - same as what Bitmain has up on their site. This new one firmware date is Dec 9 2016... Hmm, beta testing on us customers again? |
|
|
|
Glad to see it worked  re: this isnt an easy fix by any means soldering to r17 is a royal pain , c76 isnt to bad but id imagine u could just solder to gnd on the pcie plug instead of the side of c76. that would make it a lil easier Yessss it might/probably would work but -- when it comes to voltage regulators it is always best to keep the various returns as close to the regulator chip as possible. Really, that applies to all signal paths in any circuit... Minimizes phantom parasitic components. |
|
|
|
|
Tried it with my dead board -- by itself even after several hard/soft boots the bad card never reports more than 8 chips and refused to even try hashing. Reconnected the good board and back to 4.4THs on the good board and only 450GHs on the bad board.
Looking at the log in several places it reports the asic number is stored in the PIC and then jumps over several tests both for the good board and the bad one. Wondering if resetting or (god forbid) reflashing the miner might reset that info stored in the PIC...
Really hoping that the other R4 B6 arriving on Monday does better!
|
|
|
|
As far as I can see Bitmain doesn't have any significant IP (patents etc), so I would hope as bitcoin continues to be adopted that eventually more chip foundries and manufacturers will get in the game and drive up competition, which will mean better products at better costs. More chip manufacturers we may see. More Foundries to produce their chips -- not going to happen. Aside from fabs ran by Intel and IBM there is only TSMC, Samsung, and Global Foundries that are capable of producing 16/14nm (and eventually smaller node) chips. Samsung is largely producing for their own products with so far BW chips being the only exception. That leaves only TSMC and GloFo to run boutique chips like mining ASICS so end of story there. |
|
|
|
A bad chip is a bad chip. The way the boards are laid out it only takes one bad chip to take a board out. There isnt much you can do except warranty it.
Iv'e just disconnected the faulty board and I'm mining with the remaining board. Bitmain want the entire R4 back not just the hash board - I'm waiting until after CNY to ship it off, no need to make a bad situation worse. I just joined the sick R4 Club... ref https://bitcointalk.org/index.php?topic=1597639.msg17559940#msg17559940For now at least the bad board is still somewhat working with 55 chips vs the normal 63 and giving 1-1.4THs vs the usual 4.4THs . Lovely timing....  |
|
|
|
and losing about 80 to fedex because they wrote aluminum not computer on the invoices. Oddly enough same happened to me with the 1st R4 I got from Bitmain wheny they were briefly only DHL or FedEx: On shipper it lists 'Aluminum' and FedEx dinged me for $35 because of it. My 2nd R4 ordred last week had UPS back as option. Used it, got miner yesterday with no extra charges... Yay FedEx... |
|
|
|
Ah what wonderful timing.... Of course CNY is very soon starting and.... Yesterday got my 1st R4 batch-6 that I ordered last week. Was running perfectly with amazingly smooth hashrate. Emphasis on was... After 22hrs the top board (chain-8) now at best reports 55 chips vs the normal 63 and does not report temps. The 2nd batch-6 I ordered shows as being picked and will prolly ship tomorrow. Hope it fares better... Still the same amazingly smooth hashrate though, banging around 5-5.7THs of course mostly from the good board (4.4THs) vs bad board of 1-1.4THs. The top 2 with their variance are what I normally see from ALL of my s9's. The start-stops seen are from my testing to verify bad board. Even my R4 batch-4 shows the same spikey throughput as the s9's but the R4 batch-6... scarily smooth. Lovely time to start the RMA process eh?  To save space, links to screen shots of tests I ran. Tried soft-boot, hard-boot, and swapped data cables between top board (chain-8) and bottom board (chain-7). Yes I know the logs show the chains as 6 and 7 but whatever. The GUI calls them 7 and 8 and so will I. http://imgur.com/fd2ukLxhttp://imgur.com/OFTZqpthttp://imgur.com/mI0a7JJhttp://imgur.com/jlF0MnZ This will of course be part of the info I send Bitmain as will be the interesting Kernel Logs resulting from the tests. All of the logs below are massively snipped to stay under forum size limits. However the full logs show very extensive logging going on - right down to the number of good cores per-chip. along with which core number(s) is bad in each chip(s). Very nice to see that! After soft boot Problem 22hrs after getting it on 1-18-2017: Chain-8 (top board) reporting very low temp and only hashing ~1.4THs vs expected 4.4THs
Kernal log after soft-boot, Chain-8 (top board) reports dead with no chips. Logs are snipped to keep under Forum msg size limits.
<massive snip>
total valid nonce number:16108
total send work number:57456
require valid nonce number:57456
repeated_nonce_num:0
err_nonce_num:8726
last_nonce_num:607
chain[6]: All chip cores are opened OK!
Test Patten on chain[6]: OK!
chain[7]: some chip cores are not opened FAILED!
Test Patten on chain[7]: FAILED!
restartNum = 2 , auto-reinit enabled...
Chain[6] Chip[3] TempTypeID=1a middle offset=-78
Chain[6] Chip[3] local Temp=48
Chain[6] Chip[3] middle Temp=95
Chain[6] Chip[4] TempTypeID=1a middle offset=-67
Chain[6] Chip[4] local Temp=60
Chain[6] Chip[4] middle Temp=102
FAN PWM: 65
Note the absence of 1 board...
-------------------------------------------------------------------------------- After hard-boot Following the results of a soft-boot (chain-8 not hashing at all) did a hard-boot.
Kernal log after hard-boot, at least this has chain-8 back to partially working (@ 1.4THs vs expected >4.4THs):
<massive snip>
chain[6]: All chip cores are opened OK!
Test Patten on chain[6]: OK!
chain[7]: some chip cores are not opened FAILED!
Test Patten on chain[7]: FAILED!
restartNum = 2 , auto-reinit enabled...
Chain[6] Chip[3] TempTypeID=1a middle offset=-79
Chain[6] Chip[3] local Temp=46
Chain[6] Chip[3] middle Temp=83
Chain[6] Chip[4] TempTypeID=1a middle offset=-66
Chain[6] Chip[4] local Temp=59
Chain[6] Chip[4] middle Temp=99
FAN PWM: 64
Chain-8 is partly back running 1.4THs.
-------------------------------------------------------------------------------- Data cable swapped between top and bottom boards, hard-boot, problem follows top board Kernal log after swapping data cables between boards and hard-boot. As expected reporting followed the bad (top) board. Bad board again running only ~1THs.
<massive snip>
chain[6]: some chip cores are not opened FAILED!
Test Patten on chain[6]: FAILED!
chain[7]: All chip cores are opened OK!
Test Patten on chain[7]: OK!
restartNum = 2 , auto-reinit enabled...
Chain[7] Chip[3] TempTypeID=1a middle offset=-79
Chain[7] Chip[3] local Temp=45
Chain[7] Chip[3] middle Temp=96
Chain[7] Chip[4] TempTypeID=1a middle offset=-65
Chain[7] Chip[4] local Temp=61
Chain[7] Chip[4] middle Temp=104
FAN PWM: 65
---------------------------------
As expected error followed the top board.
Put data cables back as normal, now orginal error back with chain-8 hashing at ~1.4THs, no temps reported for it.
|
|
|
|
The "new" chips are 0.001J/GHs at least less efficient than the previous S9 chips.
Same chip running at a higher operating voltage, thus the reduced efficiency. Up side, it SHOULD allow better overclocking at more reliability *IF* you can keep the bloody things cool enough, as they're not pushing the "bottom end" of where the chip will run at all (which seems to be a lot of the issue with the S9). Agreed. The power specs are rather on-par with the s9 batch-1 and 3 I have. Both use a high Vcore and are Hungry. At full rated clock of 650MHz the batch-1 tops out at just over 1400w after settling down from warmup (briefly >1500W). To keep temps down have that one clocked at 600 and it still averages over 13.5THs  |
|
|
|
And I too picked up another R4 batch 6 today since BTC is over $900 again Need it to replace the 2 s7's I've been running at home for Winter heating. Just too power hungry at 15-cents/kWh even if they do reduce the gas bill.. At least the 2x early batch s9's in the basement still fully pay for their feed and more. So far very happy with the batch 4 I picked up earlier which is running silently sitting in my office by my desk. |
|
|
|
Anyone have firmware that removes the auto frequency and allows any kind of tuning to freq or voltage?
Does not exist. As has oft been said before: Auto-tune exists so Bitmain can use all the rather variable performing dies (chips) they can get from each wafer produced by TSMC. By being able to have each hash board run at its 'optimum" (in my book - run reliably at all) Vcore/freq/Temp settings Bitmain is able to mix and match different performing boards to produce advertised specs. Once Auto-tune finds those volts/freq/temp settings per-board that it thinks are best the information is stored in the PIC of each hashboard for faster reboots (methinks). |
|
|
|
Are their any posts that can explain the issues with antman? Why are they a scam site?
1st glaring reason: Their name! ANY website that uses a name that similar to legit sites is certain to be either an outright scam and/or could well be a phishing site to spread malware! Welcome to the internet.... |
|
|
|
Why would there be S10, there is no advancment in manufact. tech. Most likely next will be on 7nm at mid/late 2018
Given that ALL companies making chips at the 14/16nm node are having issues with them -- including the likes of IBM, Intel, Samsung, AMD, and others -- 16/14nm mining ASICS are are going to be here for a long while before ASIC and other boutique application chip mfgs seriously considering financing chips at lower nodes. |
|
|
|
Can you provide the BM v.1 firmware?
As has been often repeated here: ALL miners from batch-16 and up requires BM v2 auto-tune firmware. The hash boards will not work with BM v1. |
|
|
|
just edited the list to move dps-2000bb from working , to working but unreliable
, i have gone threw aot of dps-2000bb in my dc on 277v power i only get about 1-2 months out of them so i dont suggest using them with 277v
i have ibm 2880's that are 2 years old still going strong on 277
.... So -- why -- don't you just pick up a few buck/boost transformers with 32v secondaries and knock the 277v line down to more reasonable 245vac??? They are not expensive. A DPS 2kw supply would need one rated to handle 266VA, call it 300VA with margin. |
|
|
|
Brand new account + a client who trusts you to find 200 S9's != warm fuzzy
NOT Sounds along the line of why I chose my Forum name so many years ago. ref the leaked AMT/A1 Skype chats link in sig. |
|
|
|
|
Show Posts
