Swedish ASIC miner company kncminer.com - page 1685.

erk

hero member

Activity: 826

Merit: 500

Quote from: Transam808 on September 10, 2013, 07:21:52 PM

Quote from: btc_uzr on September 10, 2013, 07:03:07 PM

http://www.nytimes.com/2013/08/17/us/as-worries-over-the-power-grid-rise-a-drill-will-simulate-a-knockout-blow.html?pagewanted=all&_r=2&
whoa, maybe better get some solar panels..image they blow something up by accident during the drill Cheesy

One thing that sucks about solar panels. Is if its a grid tide system like mine and most are. They do not work when the power go out. Its a safety future so you don't feed electricity back in to the lines and end up frying the guy working on them on the other end. But you can always do a battery backup. Witch in that case they would still work. But cost and maintenance of batteries Can be costly, So most people don't bother. Easier to just have a generator for a black out.

You can buy hybrid inverters that will grid feed and charge batteries so you can continue to mine if the grid goes out. They obviously wont feed the grid from the batteries for safety.

Transam808

newbie

Activity: 47

Merit: 0

Quote from: btc_uzr on September 10, 2013, 07:03:07 PM

http://www.nytimes.com/2013/08/17/us/as-worries-over-the-power-grid-rise-a-drill-will-simulate-a-knockout-blow.html?pagewanted=all&_r=2&
whoa, maybe better get some solar panels..image they blow something up by accident during the drill Cheesy

One thing that sucks about solar panels. Is if its a grid tide system like mine and most are. They do not work when the power go out. Its a safety future so you don't feed electricity back in to the lines and end up frying the guy working on them on the other end. But you can always do a battery backup. Witch in that case they would still work. But cost and maintenance of batteries Can be costly, So most people don't bother. Easier to just have a generator for a black out.

btc_uzr

sr. member

Activity: 476

Merit: 250

let's have some fun

http://www.nytimes.com/2013/08/17/us/as-worries-over-the-power-grid-rise-a-drill-will-simulate-a-knockout-blow.html?pagewanted=all&_r=2&
whoa, maybe better get some solar panels..image they blow something up by accident during the drill Cheesy

DeathAndTaxes

donator

Activity: 1218

Merit: 1079

Gerald Davis

Quote from: canth on September 10, 2013, 05:27:36 PM

I happen to have one of these PSUs kicking around. I've looked at the specs and it seems like it would fit the bill, but I'd appreciate a 2nd set of eyes. Any glaring reason that this unit wouldn't be a decent PSU for a Jupiter?

http://www.newegg.com/Product/Product.aspx?Item=N82E16817103088

Should be good. Has 4 pairs of 2 PCIE connectors so you can use 1 from each pair without any worries of overloading the wiring. Has a single 83A 12V rail which should be more than enough. 80-Gold efficiency and more headroom then KNC recommends. Can never know for sure until you test it but I can't see any reason it wouldn't work.

DeathAndTaxes

donator

Activity: 1218

Merit: 1079

Gerald Davis

Quote from: Wesly on September 10, 2013, 05:41:46 PM

There has been quite a bit of discussion on which PSU to purchase at the KNC forum. The bigger 1000+ Watts PSU will not only offer better efficiency at around 50% load, but the lighter load (relative to max load) will also be easier on the PSU, extending its life and reliability. The extra wattage headroom might also be useful in case the KNC chip turns out to be highly-overclockable. On the same PSU KNC forum thread, one of the KNC staff has already l leaked Jupiter will run at an expected speed of 500 GH/s!

True and maybe some people can just wait for more info but there is a point of diminishing returns. Lets say the Jupiter is 500 MH/s. >850W @ 12VDC would mean >944W @ 120VAC. That is >1.8 J/GH. Already close to double what bitfury does @ 55nm and more than double all other 28nm offerings. I certainly hope it doesn't require even MORE power than that, I am hoping it will be significantly less power even when overclocked. Then again it isn't like more capacity hurts I was just pointing out the "much lower efficiency" at high load is pretty much a misnomer at this point.

The Avenger

hero member

Activity: 798

Merit: 1000

Quote from: btc_uzr on September 10, 2013, 05:44:55 PM

Quote from: rolling on September 10, 2013, 05:38:49 PM

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Yes, I think you will need a 5v connector as well. I believe all ATX power supplies have them. When KNC announced the case, they mentioned the 5v connector as well as the 12v.

https://www.kncminer.com/news/news-28

Have read this too, but I guess the 12V are the asic power supply.
Molex is 12V and 5V, so you have the choice => http://en.wikipedia.org/wiki/Molex#Molex_connector

Does your psu have such a small 5V connector ? Did they mention we need anything in addition to the psu (except for the 'paper-clip') ?
Both no, isn't it ?

I think that's it. The molex to the right of the beaglebone is probably providing the juice (through the +5v red wire that you point out). So that's all good, will just be using regular PSU connectors.

btc_uzr

sr. member

Activity: 476

Merit: 250

let's have some fun

Quote from: rolling on September 10, 2013, 05:38:49 PM

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Yes, I think you will need a 5v connector as well. I believe all ATX power supplies have them. When KNC announced the case, they mentioned the 5v connector as well as the 12v.

https://www.kncminer.com/news/news-28

Have read this too, but I guess the 12V are the asic power supply.
Molex is 12V and 5V, so you have the choice => http://en.wikipedia.org/wiki/Molex#Molex_connector

Does your psu have such a small 5V connector ? Did they mention we need anything in addition to the psu (except for the 'paper-clip') ?
Both no, isn't it ?

Wesly

full member

Activity: 185

Merit: 100

Quote from: DeathAndTaxes on September 10, 2013, 03:49:13 PM

Quote from: chrono030 on September 10, 2013, 02:36:53 PM

Quote from: DimensionsOfHell on September 10, 2013, 02:15:46 PM

Alright guys, some of you have a lot more knowledge then me in the PSU area, so if you could help me out here, I'd appreciate it.
I have a Saturn on order, which I plan to upgrade to Jupiter once they start offering the modules separately. The PSU that I have picked out seems like a perfect fit to me, and at a fairly good price. If you guys could look it over and let me know, that would be great. And if it is a good one, it could turn out to be a good deal for everyone.

http://www.newegg.com/Product/Product.aspx?Item=N82E16817139011

Should be the minimum acceptable wattage for a power supply. I'd recommend going with something around 1kw... They are much more efficient at 50 % load versus 80 - 90 % load.

That hasn't been true for the last decade but it is a myth which dies hard.

The minimum efficiency for an 80-Gold PSU is 90% at 50% load and 87% at 100% load. The curve is very flat.

For example here is the test for the PSU linked to:
http://www.plugloadsolutions.com/psu_reports/CORSAIR_CMPSU-850HX_ECOS%201464_850W_Report.pdf

50% load = 90.38% efficiency
75% load = ~90.0% efficiency (estimated due to their horribly low res chart)
100% load = 87.06% efficiency

So using a larger PSU would save you maybe 1% in power, 3% under the extreme example of 100% load vs 50% load. At 1% power on 600W 24/7 would save you 53 kWh per year. At $0.10 per kWh we are talking $5. Spending $50 more on a PSU makes the break even 10+ years. At once time PSU were pretty shitty with narrow "sweet spot" (40% to 60% of peak power) and horribly efficiency outside of that (<70%). Those days have been gone for a long time with most quality PSU having an almost flat efficiency "curve" from 20% to 100%. It is nice companies keep making more and more efficient PSU (Platinum and Titanium are also possible) but once you get to ~90% it starts being diminishing returns. Still Google and facebook use 12V only PSU and the get about 95%+ efficiency. Really there is no reason for PC PSU to be as complex as they are. Maybe some future standard will improve that.

There has been quite a bit of discussion on which PSU to purchase at the KNC forum. The bigger 1000+ Watts PSU will not only offer better efficiency at around 50% load, but the lighter load (relative to max load) will also be easier on the PSU, extending its life and reliability. The extra wattage headroom might also be useful in case the KNC chip turns out to be highly-overclockable. On the same PSU KNC forum thread, one of the KNC staff has already leaked that Jupiter will run at an expected speed of 500 GH/s!

http://forum.kncminer.com/forum/main-category/hardware/107-psu-recommendation-jupiter-the-largest-planet/page3

rolling

hero member

Activity: 824

Merit: 712

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Yes, I think you will need a 5v connector as well. I believe all ATX power supplies have them. When KNC announced the case, they mentioned the 5v connector as well as the 12v.

https://www.kncminer.com/news/news-28

DimensionsOfHell

full member

Activity: 346

Merit: 100

Quote from: DeathAndTaxes on September 10, 2013, 03:49:13 PM

Quote from: chrono030 on September 10, 2013, 02:36:53 PM

Quote from: DimensionsOfHell on September 10, 2013, 02:15:46 PM

Alright guys, some of you have a lot more knowledge then me in the PSU area, so if you could help me out here, I'd appreciate it.
I have a Saturn on order, which I plan to upgrade to Jupiter once they start offering the modules separately. The PSU that I have picked out seems like a perfect fit to me, and at a fairly good price. If you guys could look it over and let me know, that would be great. And if it is a good one, it could turn out to be a good deal for everyone.

http://www.newegg.com/Product/Product.aspx?Item=N82E16817139011

Should be the minimum acceptable wattage for a power supply. I'd recommend going with something around 1kw... They are much more efficient at 50 % load versus 80 - 90 % load.

That hasn't been true for the last decade but it is a myth which dies hard.

The minimum efficiency for an 80-Gold PSU is 90% at 50% load and 87% at 100% load. The curve is very flat.

For example here is the test for the PSU linked to:
http://www.plugloadsolutions.com/psu_reports/CORSAIR_CMPSU-850HX_ECOS%201464_850W_Report.pdf

50% load = 90.38% efficiency
75% load = ~90.0% efficiency (estimated due to their horribly low res chart)
100% load = 87.06% efficiency

So using a larger PSU would save you maybe 1% in power, 3% under the extreme example of 100% load vs 50% load. At 1% power on 600W 24/7 would save you 53 kWh per year. At $0.10 per kWh we are talking $5. Spending $50 more on a PSU makes the break even 10+ years. At once time PSU were pretty shitty with narrow "sweet spot" (40% to 60% of peak power) and horribly efficiency outside of that (<70%). Those days have been gone for a long time with most quality PSU having an almost flat efficiency "curve" from 20% to 100%. It is nice companies keep making more and more efficient PSU (Platinum and Titanium are also possible) but once you get to ~90% it starts being diminishing returns. Still Google and facebook use 12V only PSU and the get about 95%+ efficiency. Really there is no reason for PC PSU to be as complex as they are. Maybe some future standard will improve that.

Thank you very much for that. That was similar to my line of thinking. I couldn't justify the $50 difference in price, when the efficiency of the unit won't be that huge.

My mind is now set to that unit. :-)

canth

legendary

Activity: 1442

Merit: 1001

I happen to have one of these PSUs kicking around. I've looked at the specs and it seems like it would fit the bill, but I'd appreciate a 2nd set of eyes. Any glaring reason that this unit wouldn't be a decent PSU for a Jupiter?

http://www.newegg.com/Product/Product.aspx?Item=N82E16817103088

btc_uzr

sr. member

Activity: 476

Merit: 250

let's have some fun

Quote from: Henchman24 on September 10, 2013, 05:14:17 PM

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Possibly the molex to the right provides power?

I assume so as well, since we have a psu for the asics already in place.
The small 5V connector might be used in cases (not in ours) where you run the SoC board and nothing else....like a Rpi.

vulgartrendkill

hero member

Activity: 518

Merit: 500

Quote from: Henchman24 on September 10, 2013, 05:14:17 PM

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Possibly the molex to the right provides power for it?

I believe that is the case

Henchman24

sr. member

Activity: 378

Merit: 250

Quote from: The Avenger on September 10, 2013, 04:37:58 PM

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

Possibly the molex to the right provides power for it?

The Avenger

hero member

Activity: 798

Merit: 1000

Quote from: mbbc on September 10, 2013, 03:42:58 PM

Quote from: dropt on September 10, 2013, 12:35:35 PM

Quote from: chrono030 on September 10, 2013, 12:24:36 PM

I assume that it is to power the on board Linux OrSOC device?

It looks like a Beaglebone. Further, the cut-out to the left looks as though it might be for a rPi.

You could be right, here's a comparison:

Also note the round edges.

Well spotted! It's a dead ringer. This'll be what the 5v requirement was about.

Which makes me wonder - are we going to also need a separate 5v power supply adaptor? Or is there such an adaptor that can convert one of the connectors from a PSU to that fitting?

crumbs

full member

Activity: 210

Merit: 100

Quote from: dropt on September 10, 2013, 04:10:52 PM

F'ing SWEET. Where do I collect?

& mbbc: New PR guy will taek caer of stuffs 4U.
vesperwillow: Your new kitty is busy, shooping graphical designings 4 Ken's new website. B patient pl0x.

-- Advisory Cats, Active Mining & Catteries.

Edit: Wrong thread Embarrassed