Pages:
Author

Topic: [ANN] Spondoolies-Tech - carrier grade, data center ready mining rigs - page 50. (Read 1260223 times)

legendary
Activity: 1526
Merit: 1013
Make Bitcoin glow with ENIAC
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....

But with a more standard controller board. Those SP3X ones were okay up until the point they prematurely failed.

Which was rare.

They are not rare, at all.


I think we had a survey of failed SP-tech kit that showed failures were extremely rare. They occur, sure but its not like you can just say the above dogie without a caveat.
That wasn't a real survey or any measure of failed equipment, that was Guy paying for a list of frilly questions like:
  • Is the most trusted?
  • Has the best firmware?
  • Has the best support?
  • Cares most about its customers? [lol]
  • Is the most professional and pays its debts?
  • Would you buy from next
  • Gives value for money
  • Hardware performs to your satisfaction

Other reasons why don't get as many failed reports as you otherwise would:
  • When you have a problem you email Zvi, who emails internally to order a replacement. There is just no fixing those things once they go bad so no point posting.
  • 20-40x less individual customers than Bitmain (which you're comparing them to) so even with the same failure rate you get 20-40x fewer reports.


No not that one, Phil ran another one about reliability, it wasn't fancy he just asked people about reliability, but I just dug it out and it was about SP20's not SP3x but the people who responded showed a 95.8% reliability rate. And no-one paid anything to anyone. (in the one you allude to Guy actually donated to charity)

You've become very bitter dogie, you've fallen a long way, sly digs and innuendo and half truths seem to be your forte lately which is a shame really.

Plus you missed out the bit where I said I hadn't had any controller failures in my experience which isn't huge, but I have had over 20 SP3x machines some of which are still going strong from May 2014 and never had a controller failure.

Maybe if you have real statistics about the failure rate of SP3x controllers you could give them to us here. Waffling on about numbers you know nothing about doesn't help. And I wasn't comparing them to anything, I just pulled you up on your opinion regarding the controller on the SP3x range of equipment being problematic, which I don't believe is the case.

But regardless of this, if they brought out a SP3x rackmount form factor up to 2U in size but with 10-20TH at a competitive cost. I'd definitely buy them no questions asked.

So you admit that you were wrong, but you keep attacking dogie?

Why is dogie held to a completely different standard than everyone else?

Besides, the philipma surveys were a joke. The fee debacle meant that mostly die-hard spondoolies fans ever bothered to take part. And what was it in the end? 60 people? Yeah, that's statistically significant. Pfff.....

Nah, give me dogie-style reviews over twisted science any day of the week!
legendary
Activity: 3234
Merit: 1220
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....

But with a more standard controller board. Those SP3X ones were okay up until the point they prematurely failed.

Which was rare.

They are not rare, at all.


I think we had a survey of failed SP-tech kit that showed failures were extremely rare. They occur, sure but its not like you can just say the above dogie without a caveat.
That wasn't a real survey or any measure of failed equipment, that was Guy paying for a list of frilly questions like:
  • Is the most trusted?
  • Has the best firmware?
  • Has the best support?
  • Cares most about its customers? [lol]
  • Is the most professional and pays its debts?
  • Would you buy from next
  • Gives value for money
  • Hardware performs to your satisfaction

Other reasons why don't get as many failed reports as you otherwise would:
  • When you have a problem you email Zvi, who emails internally to order a replacement. There is just no fixing those things once they go bad so no point posting.
  • 20-40x less individual customers than Bitmain (which you're comparing them to) so even with the same failure rate you get 20-40x fewer reports.


No not that one, Phil ran another one about reliability, it wasn't fancy he just asked people about reliability, but I just dug it out and it was about SP20's not SP3x but the people who responded showed a 95.8% reliability rate. And no-one paid anything to anyone. (in the one you allude to Guy actually donated to charity)

You've become very bitter dogie, you've fallen a long way, sly digs and innuendo and half truths seem to be your forte lately which is a shame really.

Plus you missed out the bit where I said I hadn't had any controller failures in my experience which isn't huge, but I have had over 20 SP3x machines some of which are still going strong from May 2014 and never had a controller failure.

Maybe if you have real statistics about the failure rate of SP3x controllers you could give them to us here. Waffling on about numbers you know nothing about doesn't help. And I wasn't comparing them to anything, I just pulled you up on your opinion regarding the controller on the SP3x range of equipment being problematic, which I don't believe is the case.

But regardless of this, if they brought out a SP3x rackmount form factor up to 2U in size but with 10-20TH at a competitive cost. I'd definitely buy them no questions asked.
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....

But with a more standard controller board. Those SP3X ones were okay up until the point they prematurely failed.

Which was rare.

They are not rare, at all.


I think we had a survey of failed SP-tech kit that showed failures were extremely rare. They occur, sure but its not like you can just say the above dogie without a caveat.
That wasn't a real survey or any measure of failed equipment, that was Guy paying for a list of frilly questions like:
  • Is the most trusted?
  • Has the best firmware?
  • Has the best support?
  • Cares most about its customers? [lol]
  • Is the most professional and pays its debts?
  • Would you buy from next
  • Gives value for money
  • Hardware performs to your satisfaction

Other reasons why don't get as many failed reports as you otherwise would:
  • When you have a problem you email Zvi, who emails internally to order a replacement. There is just no fixing those things once they go bad so no point posting.
  • 20-40x less individual customers than Bitmain (which you're comparing them to) so even with the same failure rate you get 20-40x fewer reports.
legendary
Activity: 3234
Merit: 1220
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....

But with a more standard controller board. Those SP3X ones were okay up until the point they prematurely failed.

Which was rare.

None of mine have failed and I have been an early adopter, I've even bought second hand ones that are still going strong.

I think we had a survey of failed SP-tech kit that showed failures were extremely rare. They occur, sure but its not like you can just say the above dogie without a caveat.
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....

But with a more standard controller board. Those SP3X ones were okay up until the point they prematurely failed.
hero member
Activity: 994
Merit: 504
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....
i also hope they will develop a small version of sp50
hero member
Activity: 742
Merit: 500
I'd buy something similar to sp3X!

Throw 2 boards and 2 psu in one case and give us that thing!

3.4kw and 22th....
copper member
Activity: 2898
Merit: 1465
Clueless!
With the S7 shipping delays and issues (based on forum posts) I would love something similar to the SP20.
100% of my SP20s are still going strong with zero care from me, while 2 S5s and 3 S4s have had board failures and lock ups.
I did have 2 boards fail on a SP20, however I did buy them used and at a huge discount so even then I still already hit ROI.
Please Spondoolies-Tech, release something smaller than the 110TH titan.

The problem is if they are gonna fill up their new venture data hall..it makes no sense to sell or make small. They already took the
step to sell 'bulk orders' of SP50's but that imho will only be to recoup some costs and such. They did not even have to go that far.
They could have done like KNC and made it all private.

It unfortunately imho is not worth their time nor their expertise to do anything more then they already are with the sp50 and myself
I bet the sp50's that are sold on 'bulk orders' are down the road after the data hall has been up a bit and limited.

It just seems to be the way ASIC makers of any type are now. Why sell equipment that competes against your own data hall and add
to btc difficulty.

edit: Just to be clear I hope I'm Dead wrong on the above and they do make smaller units. I just continue to think it is beyond 'unlikely'.
hero member
Activity: 979
Merit: 510
With the S7 shipping delays and issues (based on forum posts) I would love something similar to the SP20.
100% of my SP20s are still going strong with zero care from me, while 2 S5s and 3 S4s have had board failures and lock ups.
I did have 2 boards fail on a SP20, however I did buy them used and at a huge discount so even then I still already hit ROI.
Please Spondoolies-Tech, release something smaller than the 110TH titan.
legendary
Activity: 1120
Merit: 1001
my guess is the SP50 does not exist ... "yet"
they just put it on their page to stall from s7 buyers
legendary
Activity: 2128
Merit: 1005
ASIC Wannabe
man this all got out of hand quick. yeah, i agree that in a large installation a battery backup is not really necessary unless you have some reason to believe that your power may be intermittant/inadequate and having even a 20-60 second backup battery to prevent needing a full restart (can take a minute or two) or otherwise affecting the chips via thermal changes (unlikely, but still has some effect on the overall product life)

Anyways, seems like a whole lot of speculating and waiting for a reveal
hero member
Activity: 700
Merit: 501
https://bitcointalk.org/index.php?topic=905210.msg
Typical ego driven puffering... This entire conversation started because of a misidentified piece of hardware (no it wasn't a UPS) and the point was there are some circumstances, albeit limited where it may make sense.. I am not going to presume I know all of the details of the environments you've deployed hardware in, big or small, and you don't know the specifics of every miner who has ever deployed hardware..

Don't be so presumptuous..
Well, I do know a lot about hardware and I can recognize real technical problems from the cargo-cult and voodoo-chicken-blood inspired pseudo-solutions.
You would however have to check first that it is acceptable to have the Controller Board powered and the Hash Boards not? For instance Bitmains latest S5+ & S7 both require that the Hash Boards are powered before the Controller Board is fired up. It's all to easy to pontificate using generalities when we are not in possession of the detail. Smiley

Rich
I wouldn't be surprised if the engineering staff at that Chinese company doesn't have a real engineering degree but some sort of "work experience" degree after serving in one of those remote technical help call centers.

Edit: https://en.wikipedia.org/wiki/Latch-up is something discussed in every electronic engineering course and any competent student would know how to deal with and avoid this problem.
End edit.

It is very easy for me to pontificate here because I have all the experience required:

1) I helped many very non-technical artists help set up and maintain their "render farm" which is the finishing stage of ray tracing in computer animation.
2) I worked for many years in hardware design and I know how to recognize common hardware faults and properly fix them.

The "home mining" market is based on pandering to the lowest common denominators of the users, the ones who even have to be explained where on the wall they have to plug in their device.

I had a good fortune on having worked with non-technical people in other markets (animation artists and medical doctors) where such pandering is neither expected nor provided. I really do know what kind of technical work can be expected from people without technical education but with common sense and self-respect.



I may have misunderstood, but I would like to address a couple of your points for clarification.
Are you suggesting a degree provides a better education and skill sets than real OTJ experience?
I've hired, fired, and worked with all types over the past 25 years and I will take the person with 25 years of OTJ experience over the kid who just graduated and thinks he is hot shit any day of the week. Obviously both are preferred, but someone with solid field experience can be filled in quickly on any areas where they may need to improve.
In that same vein I push my kiddos to college and one starts next year because that is definitely the better building block if we talk about what the best thing you should do once you graduate high school, but please never forget how many monumental businesses are founded and successful by people who are college dropouts or never went at all.
My world runs on experience, documentation, and customer service. The kid with his degree usually has never been in the same positions with customers as our high level engineering and support personnel. 
Further, the base your knowledge is built from can be fantastic depending on the school but it surprised the hell out of me when some 27 year old guy decided to measure the resistance of a cable he removed it from where it was installed so he could get the ends close enough together for his meter leads to reach. (By the way, for anyone who doesn't get it, you simply use another piece of wire to extend your meter lead, or one of several other solutions.)
I'm not saying it was his school's fault, he is a simple minded, lazy guy. However, real world experience would have busted that bubble and taught him how to think about things well before that.

I had someone I very much respect, and who is doing very well for himself tell me that at the pace technology changes that what he learned in college was how to teach himself how to do things, and not so much about the specifics. We have some of the most brilliant minds in our country who will tell you the same thing. The best schools provide you with the skills you need to be prepared and take advantage of the situations you find yourself in after school. They do not know the future and do not teach based on guess work. My experience at school mimicked his to a degree, but I did not give myself the opportunity to go to a school of the caliber where he attended.

Specifically pertaining to BITMAIN since I believe is where you directed your comment, it is obvious they have skills to make a good product and I do not think you can say with any certainty if their engineering team have degrees, have 500+ years of experience between them, or both.
I am not defending bitmain as much as I am questioning the logic you chose to use in making your case. It doesn't stack up.
Now if you want to discuss the levels of customer service, communications, and other items I'd be happy to make my own opinion more known than it is already, but I do not see how you can make such a generalization without personal relationships in place with the people you mentioned.
There are great minds at many levels, and being a great engineer, repair technician, manager, or almost any position you want to name who deals with an end user in any way requires knowing how to treat people with respect and manipulate whatever you are working on to be the best experience for the customer through their entire experience. You only get there if you have the real world experience, in some cases climbing the ladder through a call center.
I am not advocating anyone skipping college at all, but I do not think it is fair to generalize a group of people based on what has been stated here.

If you have experienced working with so many levels of people then you must understand common sense and self respect exist at every floor where the elevator stops. I agree some locations are better setup with personnel who can handle things better than another facility. What I do not understand is you seem to state common sense exists more in the Dr. Office than in the welder's shop? I again have to disagree. I find it depends on too many variables to come to such a conclusion. Common sense is not something which is taught in college, and certainly being educated more than someone else does not mean you have more common sense.

I work with one particularly bright engineer who has an amazing work ethic, can troubleshoot an issue quickly, and he is the person you want with you on a project, but, he cannot deal with people well in a customer to employee setting. His social skills are atrocious. He talks to everyone the same way and doesn't have the common sense to handle many parts of what is needed to put equipment on someone's floor and get fully paid for doing so.
I am saying common sense is not a variable to be considered when everyone's money spends.

On another note, maybe this is where you were headed with that comment... There are certainly customers who do not make it worthwhile to continue doing business with. I can handle almost anything from an end user, but there are very few, straight lines people should not repeatedly cross, and many of the people who do so probably do not realize what they are doing, and the ones who do realize it, are going to keep doing it to every vendor they have because it is their personality and it has worked for them in the past.
I think this still says the same thing, and that is common sense has nothing to do with it. The person may be an idiot or brilliant but in the end if I cut them off they lost the best they would have ever worked with. In the long run neither had the common sense to stop what they were doing to better position themselves in the future.

Interesting variables to use in defining a person, but it is one person and to make a determination at that level you need many more facts.

I would say if when discussing the company as a whole  we may actually agree, but when generalizing that same entire company of individuals we simply do not have enough information. We know the real decisions are made by the owner, CEO, the board, etc and overall we see consistency in specific situations, but that has nothing to do with a decision the engineer made and everything to do with the person responsible for the end user experience.

I hope you aren't the guy who always goes in badmouthing the last person who was there.
Personally I do not badmouth my competitors, partners, employees, etc to the end user. Setup a positive experience and act on it. When you make a mistake, fix it, when you can make an improvement, do so. Always be respectful, explain what is happening at a level the person understands, and provide instructions you know anyone could execute. Common sense has nothing to do with these things, but experience does.
Hell, anything can be improved.   
legendary
Activity: 2128
Merit: 1073
I'll just quote this in the entirety for the posterity. I was actually expecting the visits and comments from the competing hustlers.

Anyway: everyone knows that the proof of the cook is in the pudding. Proof of the EE is in that his design doesn't suffer from the problems that are taught at the undergraduate levels. I don't have an easy way to dig through my textbooks, so I'll just cite what I remember off the top of my head:

Intel 8080 was from 1974 and required careful sequencing of power up to work properly and avoid damage.

Intel 8085 was from 1975 and was free of power sequencing requirements.

Zilog Z80 was from 1976 and was not only free from power sequencing requirements but had a patented way of powering up and resetting to enable special debug mode.

(From Wikipedia) In 1977 engineers at Hughes Aircraft invented general circuit for protecting CMOS IC's from latch-up.

I'm not in the mood for devising of the wall jokes, but this situation really calls for some satire.

How about:

Spondoolies: the honey badger of SHA-256 mining: it comes up correctly no matter in which order you plug in its subsystems.

Maybe someone has or can cheaply find an undergraduate digital logic design textbook from 1980 (or later) and send one for the benefit of the designers at BitMainTech?

Edit: Also, for somebody who's a native English speaker or a decent poet: can someone come up with a version of "Timex: takes on lickin' and keeps on tickin'". Something like "Spondoolies: takes on plugin' and keeps on hashin'".
 
I wouldn't be surprised if the engineering staff at that Chinese company doesn't have a real engineering degree but some sort of "work experience" degree after serving in one of those remote technical help call centers.
That was the dumbest thing I've heard in a while. Many of the sales staff have post graduate electronics degrees.


It is very easy for me to pontificate here because I have all the experience required:
1) I helped many very non-technical artists help set up and maintain their "render farm" which is the finishing stage of ray tracing in computer animation.
2) I worked for many years in hardware design and I know how to recognize common hardware faults and properly fix them.
tldr something something computers in the 1990's so clearly know more than ASIC, electrical and software engineers combined. Could throw together an S7/SP50 killer in a weekend /s.
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
I wouldn't be surprised if the engineering staff at that Chinese company doesn't have a real engineering degree but some sort of "work experience" degree after serving in one of those remote technical help call centers.
That was the most ignorant thing I've heard in a while, "they're in a foreign country so they must be dumb and don't know anything". Many of the sales staff have post graduate electronic design degrees.


It is very easy for me to pontificate here because I have all the experience required:
1) I helped many very non-technical artists help set up and maintain their "render farm" which is the finishing stage of ray tracing in computer animation.
2) I worked for many years in hardware design and I know how to recognize common hardware faults and properly fix them.
tldr something something computers in the 1990's so clearly know more than ASIC, electrical and software engineers combined. Could throw together an S7/SP50 killer in a weekend /s.
legendary
Activity: 2128
Merit: 1073
Typical ego driven puffering... This entire conversation started because of a misidentified piece of hardware (no it wasn't a UPS) and the point was there are some circumstances, albeit limited where it may make sense.. I am not going to presume I know all of the details of the environments you've deployed hardware in, big or small, and you don't know the specifics of every miner who has ever deployed hardware..

Don't be so presumptuous..
Well, I do know a lot about hardware and I can recognize real technical problems from the cargo-cult and voodoo-chicken-blood inspired pseudo-solutions.
You would however have to check first that it is acceptable to have the Controller Board powered and the Hash Boards not? For instance Bitmains latest S5+ & S7 both require that the Hash Boards are powered before the Controller Board is fired up. It's all to easy to pontificate using generalities when we are not in possession of the detail. Smiley

Rich
I wouldn't be surprised if the engineering staff at that Chinese company doesn't have a real engineering degree but some sort of "work experience" degree after serving in one of those remote technical help call centers.

Edit: https://en.wikipedia.org/wiki/Latch-up is something discussed in every electronic engineering course and any competent student would know how to deal with and avoid this problem.
End edit.

It is very easy for me to pontificate here because I have all the experience required:

1) I helped many very non-technical artists help set up and maintain their "render farm" which is the finishing stage of ray tracing in computer animation.
2) I worked for many years in hardware design and I know how to recognize common hardware faults and properly fix them.

The "home mining" market is based on pandering to the lowest common denominators of the users, the ones who even have to be explained where on the wall they have to plug in their device.

I had a good fortune on having worked with non-technical people in other markets (animation artists and medical doctors) where such pandering is neither expected nor provided. I really do know what kind of technical work can be expected from people without technical education but with common sense and self-respect.

hero member
Activity: 588
Merit: 500
Even if by strange happenstance one was forced to use mining controller with a hardware fault that causes SD card corruption the sensible managerial solution is to keep only the controller on the centralized low-power UPS not all the power-hungry mining chips.

You would however have to check first that it is acceptable to have the Controller Board powered and the Hash Boards not? For instance Bitmains latest S5+ & S7 both require that the Hash Boards are powered before the Controller Board is fired up. It's all to easy to pontificate using generalities when we are not in possession of the detail. Smiley

Rich
legendary
Activity: 1428
Merit: 1000
https://www.bitworks.io
that said, having a UPS (or even a capacitor-based backup) would be useful in areas where theres a risk of power fluctuations or 'blips' that could otherwise disrupt operations or cause reboots in the miners. The number of times ive corrupted SD cards in an RPi due to a 'flicker' (even in an office building) is insane, and some devices cant handle fluctuations very well
This is just a symptom of pathetic system administrator ineptness to allow the file system corruption upon a power fluctuation.

Any "farm" scale computer operation has to be run with read-only file systems, configurations fetched via bootp (67&68/udp), error logging via syslog (512/udp) and monitoring via SNMP (161&162/udp).

I presume that Spondoolies initially hoped to be able to still make profit while pandering to the individual miners that are mostly ignorant of the proper network and system administration practices. And their move away from serving that market is to realize the savings from serving the markets where one can demand certain standard of computer literacy.

Even if by strange happenstance one was forced to use mining controller with a hardware fault that causes SD card corruption the sensible managerial solution is to keep only the controller on the centralized low-power UPS not all the power-hungry mining chips.

This is why all the companies serving the "retail" markets have those calling centers in India with personnel trained and prepared to deal with the complete ignorant.
 

Typical ego driven puffering... This entire conversation started because of a misidentified piece of hardware (no it wasn't a UPS) and the point was there are some circumstances, albeit limited where it may make sense.. I am not going to presume I know all of the details of the environments you've deployed hardware in, big or small, and you don't know the specifics of every miner who has ever deployed hardware..

Don't be so presumptuous..
legendary
Activity: 2128
Merit: 1073
that said, having a UPS (or even a capacitor-based backup) would be useful in areas where theres a risk of power fluctuations or 'blips' that could otherwise disrupt operations or cause reboots in the miners. The number of times ive corrupted SD cards in an RPi due to a 'flicker' (even in an office building) is insane, and some devices cant handle fluctuations very well
This is just a symptom of pathetic system administrator ineptness to allow the file system corruption upon a power fluctuation.

Any "farm" scale computer operation has to be run with read-only file systems, configurations fetched via bootp (67&68/udp), error logging via syslog (512/udp) and monitoring via SNMP (161&162/udp).

I presume that Spondoolies initially hoped to be able to still make profit while pandering to the individual miners that are mostly ignorant of the proper network and system administration practices. And their move away from serving that market is to realize the savings from serving the markets where one can demand certain standard of computer literacy.

Even if by strange happenstance one was forced to use mining controller with a hardware fault that causes SD card corruption the sensible managerial solution is to keep only the controller on the centralized low-power UPS not all the power-hungry mining chips.

This is why all the companies serving the "retail" markets have those calling centers in India with personnel trained and prepared to deal with the complete ignorant.
 
legendary
Activity: 2128
Merit: 1005
ASIC Wannabe
https://www.youtube.com/watch?v=DgHuyItPUfg
"BTCS Preparing for Spondoolies SP50 Deployment"

"Ready to accept SP50's, I assume they should be off the assembly line."

Well that answers some of our questions: SP50s are actively being produced, the pictures weren't renderings, queue up now for hash rate.
I just see a bunch of empty racks and transformers. Who knows if they are anywhere near close to production.

you dont drop several hundred thousand (maybe a million+?) on racks, cooling, and transformers unless you have a near-term plan to use them (and BTCS=SP-Tech, so be sure they are gonna be the first)

http://www.emersonnetworkpower.com/en-CA/Products/ACPower/LargeFacilityUPS/Pages/LiebertSeries610OnLineUPS1001000kVA.aspx
thats what they have in the second hall, 225kVa (208V i presume) UPS systems. Point me to any other mining facility (besides  legit datacenters/colocations) that use these to prevent electrical downtime

okay after a bit of extra searching it seems emerson offers both 225KVA UPS and 225KVA PDU. I was a bit hasty in my googling, but i imagine this is the *correct* PDU:
http://www.emersonnetworkpower.com/documentation/en-us/products/acpower/powerdistribution/documents/sl-20045.pdf
It offers the benefits of acting as a transformer, fuse panel, and PDU all within a single unit. fewer cables and devices, better efficiency, and no battery.

that said, having a UPS (or even a capacitor-based backup) would be useful in areas where theres a risk of power fluctuations or 'blips' that could otherwise disrupt operations or cause reboots in the miners. The number of times ive corrupted SD cards in an RPi due to a 'flicker' (even in an office building) is insane, and some devices cant handle fluctuations very well
legendary
Activity: 1498
Merit: 1030
It simply does not make sense from a business point of view. Batteries are expensive and I want to see those people replacing them every 3 years.

 If someone is having to replace a UPS battery every 3 years, they're buying some truely BAD quality batteries for their UPS or their power reliability is VERY VERY bad or they have BOTH poor quality batteries AND poor power reliability.


 Another thing to keep in mind though - probability that if a MAJOR mining facility has lost power, it's ISP has also lost power (but the ISP probably has UPS backup and likely is still up and running for a while).


 If you have high reliability on your power, then it probably makes no sense to put your mining gear on a UPS - but high reliability power is NOT a given in most of the world.
Pages:
Jump to: