even using the old firmware I still get a lot of corruption.
I think maybe the SD card is broken. Whre can I find a image file for SP30?
thanks.
11/04 20:53:20:Data corruption 1 on read 2fe:e1 (e1817f00:20) got 3fe:e0 (e081ff3f:ffffffed) - ql=52
11/04 20:53:20:Data corruption 2 on read 3fe:e0 (e081ff00:20) got 3fe:e1 (e181ff3f:ffffffed) - ql=52
11/04 20:53:20:Data corruption 3 on read 3fe:e1 (e181ff00:20) got 4fe:e0 (e0827f3f:ffcfffed) - ql=52
11/04 20:53:20:Data corruption 4 on read 4fe:e0 (e0827f00:20) got 4fe:e1 (e1827f3f:fffbffed) - ql=52
11/04 20:53:20:Data corruption 5 on read 4fe:e1 (e1827f00:20) got 5fe:e1 (e182ff3d:ffffffed) - ql=52
11/04 20:53:20:Data corruption 6 on read 5fe:e0 (e082ff00:20) got 6fe:e0 (e0837f3f:ffffffed) - ql=52
11/04 20:53:20:Data corruption 7 on read 5fe:e1 (e182ff00:20) got 6fe:e1 (e1837f3f:ffffbfed) - ql=52
11/04 20:53:20:Data corruption 8 on read 6fe:e0 (e0837f00:20) got 7fe:e0 (e083ff3f:ffffffed) - ql=52
11/04 20:53:20:Data corruption 9 on read 6fe:e1 (e1837f00:20) got 7fe:e1 (e183ff3f:ffffffed) - ql=52
11/04 20:53:20:Data corruption 10 on read 7fe:e0 (e083ff00:20) got 8fe:e0 (e0847f3f:ffffffed) - ql=52
i get a lot of similar logging on an SP31 unit that has 3 consistently bad chips, and 1-2 bad chips that are a bit inconsistent in when they show up |
|
|
|
Bitmain has zero pressure to do anything besides maximize their return on their current S5 miner. They have zero serious competition at the moment, and seem to be able to produce and sell S5's "from stock". While they don't infinite pricing power, they can even tinker with prices enough to maximize their total income from the S5 product. They are only going to make N (pick your value for N) S5's, hence they want to maximize N*Price, don't they? If in a few months if they still have some S5's, they can have a "fire sale, an S5 for $250 quantity one".
While the Avalon box is nice, it's price is still well above Bitmain's price for the S5.
at this stage production cost is a huge factor. IIRC the SP20 cost spondoolies >$400 to build, and even as a very stripped-down design, the S5 probably still costs bitmain $300/unit or so to produce. As long as bitcoin price stays between $220-280/BTC, difficulty will stagnate as any new gear simply offset 1w/GH stuff going offline. But if BTC goes to something like $400, these units will start selling like crazy and difficulty could quickly spiral upwards. Its likely the reason why larger farms are selling coins rather than holding them |
|
|
|
one of my hashboard has failed once again, sent for RMA. this is the status for the other hashboard. is the HW error too high?  your temperature is pretty high - try reducing the frequency to 325MHz and see if that helps |
|
|
|
I also can share my experience with RMA
I bought some units SP20 in last days february.
In the second week I was trying to get the bestperformance from my SP20 and I downgraded to 2.5.12, with utility in web manager.
After this, my SP20 is not more SP20. It shows like a SP10 in webmanager and not mining.
Tried recovery sdcard, flashing ok, got chritsmas lights but really dont work.
Tried boot up SDcard and it works ok.
Tried with controlboard from other sp20 and works ok.
support take one hour to test and check, and it conclusion is "unable to boot from internal flash, boots successfully from SD, boot problem not recoverable from remote".
so I asked a warranty replace for controlboard, because it is not the original condition work.
The answer from Barbara is:
"... our policy not to replace machines that work with SD card reboot... "
My reply is that machine must work like original, and with more reason, it is damaged because a fail from upgrade firmware system (i get the upgrade with a nonvalid firmware for SP20).
sp20 can work for some days, but if I reboot the machine I am not sure if this is going to work or the SD will be corrupted ... just a incognite and I have a misvalued and problematic miner.
spondolies support recognise "...work with an SD it's not as reliable as the internal ..."
My goal is get a new controlboard, not the complete sp20, but they refuse to do it.
Imagine you can not boot your computer from harddisk and only with pendrive ... imagine you can only start your car if the door is open ... imagine ... you can imagine lots of situations but when a product is not working as original and it represent a value lose I think it must be fixed.
So I ask people and spondoolies guys in this forum your opinion about this.
Do you think this is the correct policy from spondoolies ?
how often do you reboot? If it runs with a BootSD it should be stable indefinitely (unless a power disruption or reboot mucks the sd card and you have to re-image it again) most of my hardware runs weeks or months once its plugged in and configured. The control board probably costs SP-T something like $150 so its easy to see why they are hesitant to ship you one when a $5 sd card does the trick |
|
|
|
Just to point one thing, maybe it was already written.
Pencilmoded Antminer S1 if connected to power meter or in the timer switch in 80% of the time it will show lots of xxxx on ASIC chips.
If connected directly to power source all ASICS will be fully working.
Tested with few S1 as well as few power supplies (high quality server power supplies).
Didn't try without pencil mod, maybe it works fine...
put a surge bar between the timer and the PSU - it should offer a small amount of capacitance to prevent ripple (if the timer is causing that sort of issue) |
|
|
|
I think the gap between the BFL and the Antminer is probably filled with some of the BitFury hardware?
It was, and those little H-boards were well ahead of the curve. Some people are only just now replacing them from their farms. IIRC the bitfury gear was actually more in the 0.9-1.3w/GH range depending on overvolting. I still have my 6-card rig from the september batch hashing away, but i think its drawing >1.1w/GH and is time to be unplugged any day now if the price doesnt break $250 upwards |
|
|
|
After making the changes to my S3+ the new frequencies show up in the gui. I selected 262.5 and input voltage as 0750. When I restarted the box it came back still hashing at 450gh, but the error rate jumped from basically 0 to 0.3%. I set the frequency then to 250 and left voltage as 0750, restarted, and it came back hashing at 500gh with 0 error again. Why wouldn't the 262.5 frequency work?
0.3% error isnt much, but poor performance doesnt always show as errors - you could need slightly higher voltage look at the discard rate - if its more than 100% of the accepted sharees, you might have a problem and need to tune lower |
|
|
|
If the miners are mining to a local pool, and the pool is connecting to the bitcoin network over satellite, then the lag will be minimal. Trying to actually mine over satellite would be awful.
+1. assuming that each individual location acts as a local pool (which would be the best way to operate, and likely the case), then the ping time for a satellite connection should be irrelevant so long as the datacenter is still receiving new work. Bitmain mentioned it is a main datacenter, which i assume means >1PH of equipment. with a local pool/work distribution system, there should be less than 3% orphans from lag since theres still such computing power |
|
|
|
My entire point was : Guys with my "RMA experience is" ... make sure you have a unit not running in the same condition as me before sending it for an RMA.
My entire unhappiness is about the two options on the table.
1. 200$ for getting your unit back as is.
2. Pay 500$ for 2 new blades plus 200$.
so you shipped the entire unit for the RMA? IMO/IME, basic home diagnostics are usually required related to bitcoin miners. If it was a single bad board, shipping just that board for repair/replacement without having to ship the rest of the unit would allow you to continue mining with 1/2 the SP20 and drastically reduce shipping weight/cost I did that way back when a Bitmain S1 stopped hashing on one board. pulled just the board off the frame, mailed it to them, got back a brand new hashboard that i pasted and re-installed. Bitcoin isnt mainstream enough that the RMA/DOA process is smooth - being able to do modular components and repairs is crucial. (I look forwards to the SP100 having boards that can be removed with no more difficulty than in the SP20 or antminer S2 - it would make RMA simpler) ----> on a related note, i have a few SP3x units in various states of 'angst', and when looking under the hood its pretty clear that the SP3x series is not easily repaired - there are a LOT of different components and screws that lock the hashing boards in place, and thus make RMA of a single bad board nearly impossible (same with trying to frankenstein a pair of half-working units into a singular 100% unit) |
|
|
|
seriously - the paper clip technique is not a big deal - theres literally milliwatts of power at the most that travel through it - it wont short out to anything (its already GRN-GND), it takes 5seconds and <$0.01 to do (paperclip, thin wire, etc will work)
True. I wouldn't run it without the resistor to a low voltage pin. KnC had a problem early on with Jupiters blowing up a capacitor. My theory is that without the actual motherboard the low voltage wasn't there for a start up current ramp and as a result some power supplies hit the miner with voltage much higher than 12vdc blowing up the cap. Look for this on ebay "JUMP START EPS/ATX 24 PIN POWER SUPPLY WITHOUT A MOTHERBOARD MADE IN USA" i think any good quality PSU will avoid problems like this, and even if not 12V is often +/-1% - anything under 12.3V should be safe |
|
|
|
|
seriously - the paper clip technique is not a big deal - theres literally milliwatts of power at the most that travel through it - it wont short out to anything (its already GRN-GND), it takes 5seconds and <$0.01 to do (paperclip, thin wire, etc will work)
|
|
|
|
|
well, in that case i look forward to avalon making a larger, affordable unit (something rackmountable and <4U in size) that is similarly quiet.
S3 noise is the ideal for me - tucked in the back corner of a stockroom it is audible but not overbearing
|
|
|
|
Nice product, but very overpriced.
Shipping for one piece (ehash) 0,6BTC? It is aprox. 144USD (if BTC 240) !! And almost no discount for shipping if more pieces.
(Bitmain S5 - 55USD only and if more pieces under 30USD)
For that price does not make any sense to buy.
The premium price is for the quiet it provides.This is a great product if you have an office space with free power . Put an s-5 in your office and see what your boss thinks. Many places have a need for a space heater even in the summer. Due to Air Conditioning that is controlled by the landlord not the tenant. My friend rents 20% of the top floor of a bank. He pays 20% of the power bill as he is a sublet to the main tenant . He is underheated in the winter and over a/c in the summer. Right now I have 2 s-3's in his office they have been running since July of 2014. I will replace with these s-3's with my avalon 4.1's in late May. So this is a perfect piece of equipment for his needs. You are correct that is costs a lot it is not for a big farmnot a worthwhile premium. you can make the antminer S5 quieter with a $5-20 fan mod and a little underclocking. The S3 is the alternative, and very inexpensive. I want 10TH of equipment today, and honestly neither avalon of bitmain are appealing options Which fans are you talking about? Even modded the S5 is no where near the sound level of the Avalon4. I had a good experience running the fan at 7V (comparable to an SP20 at 20-30%) and as low as 5V (Sp20 at 10-15%) when very cold air was used at the inlet. thats a 5$ mod at the most. otherwise there a few options for >100cfm quiet fans i havent personally heard the avalon, but noise is secondary to well-priced hashrate right now |
|
|
|
my guess:
1) is the BBB getting power through gpio? try powering it with a 2A usb cable to see if it runs
2) bad sensor. it happens
3) it happens, poor SMD placement
4) try usb cable for power, try a different network cable (or check connection at router)
5) power delivery - maybe use 2x PCIe connectors from a better PSU
6) same as ^
ive had a raspberri pi stop accepting power through GPIO, and only work if usb-powered. its possible you have a similar issue, or a bad OS image on the BBB
1. Antminer S5 BeagleBone Black does not have a USB connector. BeagleBone Black runs, only yellow led not work. 2. 3. 4. Antminer S5 BeagleBone Black does not have a USB connector. The second network cable to the same. 5. I tried a different PSU. No help. 6. same as ^ well, i would think the best you can get for 1-3 is a coupon since the miner otherwise mines (?) 4) you could switch it with a working BBB - sounds like its either not getting power or has a more serious issue. 5/6) might be bad boards then - RMA? |
|
|
|
|
under 'asic stats' on the spondoolies you can see the current frequency for each individual chip, based on temperature and other optimizations
|
|
|
|
Nice product, but very overpriced.
Shipping for one piece (ehash) 0,6BTC? It is aprox. 144USD (if BTC 240) !! And almost no discount for shipping if more pieces.
(Bitmain S5 - 55USD only and if more pieces under 30USD)
For that price does not make any sense to buy.
The premium price is for the quiet it provides.This is a great product if you have an office space with free power . Put an s-5 in your office and see what your boss thinks. Many places have a need for a space heater even in the summer. Due to Air Conditioning that is controlled by the landlord not the tenant. My friend rents 20% of the top floor of a bank. He pays 20% of the power bill as he is a sublet to the main tenant . He is underheated in the winter and over a/c in the summer. Right now I have 2 s-3's in his office they have been running since July of 2014. I will replace with these s-3's with my avalon 4.1's in late May. So this is a perfect piece of equipment for his needs. You are correct that is costs a lot it is not for a big farmnot a worthwhile premium. you can make the antminer S5 quieter with a $5-20 fan mod and a little underclocking. The S3 is the alternative, and very inexpensive. I want 10TH of equipment today, and honestly neither avalon of bitmain are appealing options |
|
|
|
I bought a 20 Antminer S5.
Only 14 of them are perfectly fine, 6 is not.
1 miner . Yellow LED an BeagleBone Black does not work
2. miner One board shows temp 0
3.miner. Very confusing error. Mined, but will not light up green blinking LED..
Someone in factory Scraped off the resistor...
4. Miner. Ohne BeagleBone Black working only 1 hour, and and after it is dead.
No network, no network leds .
5. One of the miners shows on one blade only 6 chip ( 6 x o )
After only a couple of minutes shows only 6 x
6. Only one blade is working. I get the ----- or xxxxx . The first few minutes are ooooooo , but only half the speed.
Which of these errors go under warranty?
my guess: 1) is the BBB getting power through gpio? try powering it with a 2A usb cable to see if it runs 2) bad sensor. it happens 3) it happens, poor SMD placement 4) try usb cable for power, try a different network cable (or check connection at router) 5) power delivery - maybe use 2x PCIe connectors from a better PSU 6) same as ^ ive had a raspberri pi stop accepting power through GPIO, and only work if usb-powered. its possible you have a similar issue, or a bad OS image on the BBB |
|
|
|
Noctua NF-F12 IndustrialPPC 3000RPM PWM 120mm High Performance Fan x 2 ( push-pool )
Conceived for industrial heavy duty applications that require enhanced cooling performance and advanced ingress protection, the NF-F12 industrialPPC (Protected Performance Cooling) is a ruggedised high-speed version of the award-winning retail model. Thanks to the outstanding aerodynamic efficiency of the NF-F12 design and the use of a novel three-phase motor, the industrialPPC version provides superior airflow and pressure capacity while keeping noise levels and power consumption moderate as against comparable high-speed fans. While its fibre-glass reinforced polyamide construction and certified water and dust protection make the NF-F12 industrialPPC suitable for operation in challenging environments, the renowned SSO2 bearing technology guarantees an MTBF of over 150,000hrs. Topped off with Noctua's trusted reliability and six year manufacturer's warranty, the NF-F12 industrialPPC is an ideal choice for highly demanding applications that require superior flow rates and ultimate dependability
JUST PERFECT!
Read the thread man. I must have said the only real world workable dual fan solution (if you don't have any S3 fans) is to buy two of these fans in a push / pull. Must have mentioned it over 10 times. For one fan setup you need at LEAST a 38mm push with at least 185-200 CFM, max 4000 RPM's. For a dual fan setup you would need two 25+ mm in a push pull with at LEAST 110-125 CFM, max 3000 RPM's. If the ambient temps increase considerably on the dual fan setup some kind of air conditioning will or may be needed. Most other fans cannot take the pressure or keep up with the demand then the Noctua's, hence the 6 year warranty. They are also IP52 rated, which can take dripping water and some spray and dust won't even bother it. The motor is completely sealed of course. Have fun!!! \ IMO, a way simpler and cheaper option is to run the stock fan at lower power. something like this for driving the fan at 5V or 7V:  I found that at 7V the noise is about 60%, but airflow is still about 75-80% and its suitable for slightly above stock speeds at room temp at 5V the noise is about 30% of stock, but airflow is about 55-65%, and can offer about 1.05TH at room temperature with home-friendly noise levels |
|
|
|
Hey anyone. I have an SP30. I have tried every version of firmware and contacted support. i have Loop0, Loop1 and Loop2 that all do not hash. Below are the asic stats section. i have tried enabling and reenabling to no avail. anyone know what is going on. I purchased this from minersource 1 month ago new. Spondoolies no longer supports rma for these. this sucks. i have opened the unit up. all cables look fine, nothing looks or smells burnt. again this sucks.
LOOP[0] OFF TO:0 (test serial failed)
0: disabled (serial problem)
1: disabled (serial problem)
2: disabled (serial problem)
LOOP[1] OFF TO:0 (test serial failed)
3: disabled (serial problem)
4: disabled (serial problem)
5: disabled (serial problem)
LOOP[2] OFF TO:0 (test serial failed)
6: disabled (serial problem)
7: disabled (serial problem)
8: disabled (serial problem)
has it been like this sine day1, or just more recently? my guess is that the ribbon cable inside may be loose or poorly assembled. (I had an Sp20 where one loop was disabled because a single wire in the ribbon cable wasnt properly 'bitten' by the 'teeth' in the connector (its basically an IDE cable). 25min was more then enough to pop open the connector, see the offset wire, reseat it properly, and close everything up again for use |
|
|
|
|
Show Posts
