Author

Topic: Review of the Spondoolies-Tech SP10 „Dawson“ Bitcoin miner (1.4 TH/s) (Read 13641 times)

jr. member
Activity: 79
Merit: 1
Initial post:

The SP10 offers the following modes of operation (first three columns taken from the webinterface, last column added by me):

Code:
1.00Th / ~720W  / ~quiet / 1,39 GH/s/W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W
1.43Th / ~1350W / normal / 1,06 GH/s/W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W

As you can see, the slowest (or "underclocked") mode is most efficient, whereas the two fastest modes offer about 20% less performance per Watt.

Update 24.04.14:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Update 26.04.14:

More pictures: http://imgur.com/a/afpMF#0

how did you do this? i keep on getting 1150 at the wall.
1.00Th / ~720W  / ~quiet  is not possable.

720-900 watts is not possable. this 1150 watts makes my power bill $1,200 a month. i need to lower it to 720 - 900 watts. how do i lower it?  what is the correct settings to lower the power.  nothing is working.


i got 745 - 800 watts at wall  with the following settings

i have 800 watts at wall with the following settings hash is 1051 ghs.

fan speed  60
start voltage top 0.59
start volts bottom 0.59
max voltage 0.61
max power 800
dc2dc current limit 62

ghs is at 1051 avg and 933  5sec after 19 min uptime.  done on a 110v wall outlet.

jr. member
Activity: 79
Merit: 1
Initial post:

The SP10 offers the following modes of operation (first three columns taken from the webinterface, last column added by me):

Code:
1.00Th / ~720W  / ~quiet / 1,39 GH/s/W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W
1.43Th / ~1350W / normal / 1,06 GH/s/W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W

As you can see, the slowest (or "underclocked") mode is most efficient, whereas the two fastest modes offer about 20% less performance per Watt.

Update 24.04.14:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Update 26.04.14:

More pictures: http://imgur.com/a/afpMF#0

how did you do this? i keep on getting 1150 at the wall.
1.00Th / ~720W  / ~quiet  is not possable.

720-900 watts is not possable. this 1150 watts makes my power bill $1,200 a month. i need to lower it to 720 - 900 watts. how do i lower it?  what is the correct settings to lower the power.  nothing is working. 
jr. member
Activity: 79
Merit: 1
Initial post:

The SP10 offers the following modes of operation (first three columns taken from the webinterface, last column added by me):

Code:
1.00Th / ~720W  / ~quiet / 1,39 GH/s/W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W
1.43Th / ~1350W / normal / 1,06 GH/s/W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W

As you can see, the slowest (or "underclocked") mode is most efficient, whereas the two fastest modes offer about 20% less performance per Watt.

Update 24.04.14:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Update 26.04.14:

More pictures: http://imgur.com/a/afpMF#0


what is the correct settings to get 720-900 watts at wall?

fan speed 50
start volts top 0.635
start volts bottom 0.635
max voltage 0.642
max power 800

dc2dc current limit 50

please tell me the correct settings on each one of them so i can get 1ths at 720-900 watts at wall this 1250 watts at wall is killing my power bill with 6 of them running.
 i am still pulling 1150 at the wall and it will never change i have the max watts at 800 and it still pulls 1150 at my wall. 
 
member
Activity: 119
Merit: 100
noobie
BTW: Anybody interested in my SP10s? Need to get some space....

go may go to the marketplace : https://bitcointalk.org/index.php?board=51.0
legendary
Activity: 2338
Merit: 1124
BTW: Anybody interested in my SP10s? Need to get some space....
donator
Activity: 543
Merit: 500
In Germany you have to pay 19% VAT.
member
Activity: 119
Merit: 100
noobie
hi, sorry for bringing this up... i could find somewhere to ask.. did anyone delivered it to europe? how was the custom fees??
Custom fees are 0%
Tax may and will apply and the % depends on your country.
hero member
Activity: 784
Merit: 1000
Pools Of Honor
hi, sorry for bringing this up... i could find somewhere to ask.. did anyone delivered it to europe? how was the custom fees??
donator
Activity: 543
Merit: 500
Just a small update. The unit is still running great, only once I received an idle notification from BTCGuild. Somehow cgminer crashed, but I could just click on "Start" again, didn't even need to reboot the unit.

As of now:

System Uptime
74 day 23 hr 3 min 58 sec
CGMiner Uptime
21 day 1 hr 24 min 20 sec
donator
Activity: 543
Merit: 500
Well, once I had days of downtime because I thought it was a good idea to upate the AMD driver on Linux.

Now I know what "Never touch a running system" means Cheesy
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
Code:
System Uptime
41 day 22 hr 34 min 31 sec
CGMiner Uptime
37 day 22 hr 29 min 11 sec
This is really the most stable mining system I've used. Wish the BitFurys or GPU rigs were that stable Wink
Mine never reaches that high because I'm a good boy and update the firmware Wink
donator
Activity: 543
Merit: 500
Code:
System Uptime
41 day 22 hr 34 min 31 sec
CGMiner Uptime
37 day 22 hr 29 min 11 sec
This is really the most stable mining system I've used. Wish the BitFurys or GPU rigs were that stable Wink
member
Activity: 119
Merit: 100
noobie
don't forget that it's not USA everywhere..... damage your earing for few buck is not worth it... in add of the compagny is in Israël, so I believe that if you want to sue them, their answer will be GTFO NOOB.
hero member
Activity: 490
Merit: 500
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

I don't understand how people can sleep in the same room with their rigs??? I go sleep to other room or otherwise i couldn't sleep when hearing the noise.. That just doesn't sound a smart thing to do.

I often sleep in the same room if I'm too lazy to move, you get used to the noise pretty quick. It's also a constant noise so it's less disruptive than traffic for example.

Agree. And the noise can't be worse than my tinnitus.

Well if you sit in a room with noise over 70dB it will eventually damage your hearing so beware!

Expect law suits in the future for Bitcoin manufacturers if they don't make the buyer aware and sell it under some kind of disclaimer. Tongue
http://www.thompsons.law.co.uk/workplace-illnesses-and-diseases/industrial-deafness-personal-injury-compensation.htm
donator
Activity: 543
Merit: 500
Having gone through this whole BFL and Avalon (and partially BitFury) mess I takes nothing as granted Smiley
legendary
Activity: 3234
Merit: 1220
Web UI screenshots can be found here: http://imgur.com/a/H3dee - Nice, clean and functional.

The SP10 is now running for over two weeks without the slightest problem. It's just delivering rockstable 1,4 TH/s.

If I compare that to my BitFury units, which cry for attention all the time....

I think thats getting pretty much the norm now. My 1T Dragon miner has been up for just about 6 weeks (5w29d) without a problem at a stead 1020Gh/s Avg.

donator
Activity: 543
Merit: 500
Web UI screenshots can be found here: http://imgur.com/a/H3dee - Nice, clean and functional.

The SP10 is now running for over two weeks without the slightest problem. It's just delivering rockstable 1,4 TH/s.

If I compare that to my BitFury units, which cry for attention all the time....
hero member
Activity: 532
Merit: 500
seems like a high end high quality server product. bottom line.1u and a high noise level are just part of this. hosted mining is the way of the future. power also is a issue not sustainable by hobby miners homes generally speaking. spondoolies is just building the next logical step and moving the mining industry forward. away.from the hobbylike setups.
legendary
Activity: 2338
Merit: 1124
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

I don't understand how people can sleep in the same room with their rigs??? I go sleep to other room or otherwise i couldn't sleep when hearing the noise.. That just doesn't sound a smart thing to do.

I often sleep in the same room if I'm too lazy to move, you get used to the noise pretty quick. It's also a constant noise so it's less disruptive than traffic for example.

Agree. And the noise can't be worse than my tinnitus.
donator
Activity: 1414
Merit: 1051
Spondoolies, Beam & DAGlabs
donator
Activity: 543
Merit: 500
legendary
Activity: 1904
Merit: 1007
2$/G Is a little for me, i think 1.2-1.3$/G is a reasonable price for everyone

You can get under 1$/GH in my group buy, but for pre-ordering. Nobody has 1.2-1.3$/GH for in stock items.
newbie
Activity: 56
Merit: 0
2$/G Is a little for me, i think 1.2-1.3$/G is a reasonable price for everyone
legendary
Activity: 1904
Merit: 1007
Will give the Turbo mode a chance too.

Code:
SP10 1,497.78 GH/s

Looking good!
donator
Activity: 543
Merit: 500
This is the future but I am not sure if mineral oil will be the mass deployment coolant.  It is still pretty messy to service.   I am working on a project with another coolant that is a little less messy.   Regardless this is cool and green tech.  I love it.
It should be possible to deploy immersion cooling at large scale, as shown here: https://bitcointalksearch.org/topic/visit-of-asicminers-immersion-cooling-mining-facility-346134
legendary
Activity: 1330
Merit: 1026
Mining since 2010 & Hosting since 2012
Liquid submersion cooling is the next step:

http://www.grcooling.com/

This is the future but I am not sure if mineral oil will be the mass deployment coolant.  It is still pretty messy to service.   I am working on a project with another coolant that is a little less messy.   Regardless this is cool and green tech.  I love it.
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

I don't understand how people can sleep in the same room with their rigs??? I go sleep to other room or otherwise i couldn't sleep when hearing the noise.. That just doesn't sound a smart thing to do.

I often sleep in the same room if I'm too lazy to move, you get used to the noise pretty quick. It's also a constant noise so it's less disruptive than traffic for example.
sr. member
Activity: 452
Merit: 250
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

I don't understand how people can sleep in the same room with their rigs??? I go sleep to other room or otherwise i couldn't sleep when hearing the noise.. That just doesn't sound a smart thing to do.

He mentioned living room, not Bedroom
legendary
Activity: 938
Merit: 1000
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

I don't understand how people can sleep in the same room with their rigs??? I go sleep to other room or otherwise i couldn't sleep when hearing the noise.. That just doesn't sound a smart thing to do.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

send it to me I have a  place for it.


seriously I am in Howell NJ 07731.

pm me with a price and I will take it off your hands.
legendary
Activity: 1904
Merit: 1007
Could we get a video of the fan noise.  Start video, turn unit own, struggle to hear you talk over noise  Wink
That should be possible Wink

--

I now have some more numbers to present:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Thank you for the review too. I was using the Normal mode until now. Will give the Turbo mode a chance too.
legendary
Activity: 1666
Merit: 1000
Could we get a video of the fan noise.  Start video, turn unit own, struggle to hear you talk over noise  Wink
That should be possible Wink

--

I now have some more numbers to present:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Thanks again for the information!!
donator
Activity: 543
Merit: 500
Could we get a video of the fan noise.  Start video, turn unit own, struggle to hear you talk over noise  Wink
That should be possible Wink

--

I now have some more numbers to present:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.
newbie
Activity: 22
Merit: 0
Liquid submersion cooling is the next step:

http://www.grcooling.com/
legendary
Activity: 1666
Merit: 1000
Thanks for the review -- looking forward to more.

Could we get a video of the fan noise.  Start video, turn unit own, struggle to hear you talk over noise  Wink

Dogie wins the price for keeping it in the living room -- diehard miner!  Grin
donator
Activity: 543
Merit: 500
Hey guys, it would be nice if you could keep this all time favourite discussion "mining vs. buying BTC" out of this thread Wink
I would like to have this thread dedicated to my SP10 review and questions regarding this unit/review.
(I'm not saying either option is better. Everyone has to do due diligence. Then, buy mining gear or buy BTC or buy nothing.)
donator
Activity: 543
Merit: 500
This is not going to reduce the loudness by a single dB, sadly.
donator
Activity: 543
Merit: 500
member
Activity: 114
Merit: 10
No more Crypto in this world
Bet i could find a way to quiet it down. I have a fuw idea's
hero member
Activity: 686
Merit: 500
FUN > ROI
How would that look like? Currently, cool air is sucked in at the front and blown out of the case at the back by a battery of fans (see http://imgur.com/a/kASCu#2 ). So the air is moving in a "straight line", I really don't believe it does get any more efficient than that. Also, the PSU adds some noise. Having such a small form factor  (it's really small considering it delivers up to 1.4 TH/s) just limits what you can do in terms of cooling.
The air can only move in a 'straight' (a CFD simulation would be fun to watch) line because the case is pretty much closed except at the two ends Smiley  But in terms of what it would look like - pretty much exactly what it sounds like, replace the axial fans with a cross flow fan.  Normally I'd be concerned about space, but given this shot (finally - thanks, Dogie!), I'd say the space for one is definitely there: https://farm8.staticflickr.com/7243/13540240174_de9005ae25_z.jpg .  The main concern I'd have is whether the displacement (velocity and volume) made up for by less dead space is enough to make up for the loss of displacement from generally being less efficient.  Note that this is just wrt noise - obviously these devices are not meant to be put in bedrooms in the first place and sacrificing silence for performance is the norm Smiley
donator
Activity: 543
Merit: 500
Seriously, though, I'd still see if a cross flow fan setup would work.
How would that look like? Currently, cool air is sucked in at the front and blown out of the case at the back by a battery of fans (see http://imgur.com/a/kASCu#2 ). So the air is moving in a "straight line", I really don't believe it does get any more efficient than that. Also, the PSU adds some noise. Having such a small form factor  (it's really small considering it delivers up to 1.4 TH/s) just limits what you can do in terms of cooling.
hero member
Activity: 686
Merit: 500
FUN > ROI
Seriously, though, I'd still see if a cross flow fan setup would work.
member
Activity: 70
Merit: 10
Burn the wings off those bastards.
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.

It is sooooo loud it affects his typing!  Cheesy
hero member
Activity: 686
Merit: 500
FUN > ROI
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.
WHAT? WHAT DID YOU SAY? SPEAK UP!
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
I STILL HAVE A UNIT IN MY LIVING ROOM, YOU GET USED TO THE NOISE LEVELS AFTER A WHILE.
member
Activity: 64
Merit: 10
If you ever visited a real data center, you know these fans. The dB don't matter there. And that's wat this Unit is designed for: 1.25 U rack mountable for a datacenter rack.

Thanks for your review. Makes the SP30 Group buy decision even harder Smiley
hero member
Activity: 700
Merit: 504
Run a Bitcoin node.
So I take it is loud. LOL.
After reading through the comments other people, who received their units earlier, have made, I was expecting something, but my expectations were greatly exceeded Wink

why simple put better quality fan
Well, this is a 1U device, and power consumption is above 1 kW. Looking at the webinterface, there is a temperature reading:

Temp Front / Back
28 °C / 82 °C

I have not opened the unit yet (will do that, but there are already pictures of an open SP10), but I guess the thermal management is well-engineered. So I don't see any way to make this unit more quiet, at least not using air cooling. You have to dissipate the heat somehow. The problem isn't that the fans are of bad quality, but that they are spinning at (estimated) 4000-6000 rpm. Even if the fans had loud/bad bearing, air movement will always be magnitudes louder at this level. Even if Spondoolies-Tech manage to build in a better (read: more quiet) fan management, this unit will never be "quiet".
Exactly.  Small fans are required for the 1.25U form factor.  Basically this is a server-grade machine, and sounds like it.  It's meant for racks, server rooms, and data centres where noise is expected and not an issue.  Regular servers are loud too.

If you want to run this at home, buy a quiet rack cabinet, e.g. http://www.acoustiproducts.com/en/ucoustic_active.asp
You can even get them in wood, they look like a nice piece of office furniture: http://www.acoustiproducts.com/en/ucoustic_wood.asp

donator
Activity: 543
Merit: 500
So I take it is loud. LOL.
After reading through the comments other people, who received their units earlier, have made, I was expecting something, but my expectations were greatly exceeded Wink

why simple put better quality fan
Well, this is a 1U device, and power consumption is above 1 kW. Looking at the webinterface, there is a temperature reading:

Temp Front / Back
28 °C / 82 °C

I have not opened the unit yet (will do that, but there are already pictures of an open SP10), but I guess the thermal management is well-engineered. So I don't see any way to make this unit more quiet, at least not using air cooling. You have to dissipate the heat somehow. The problem isn't that the fans are of bad quality, but that they are spinning at (estimated) 4000-6000 rpm. Even if the fans had loud/bad bearing, air movement will always be magnitudes louder at this level. Even if Spondoolies-Tech manage to build in a better (read: more quiet) fan management, this unit will never be "quiet".
legendary
Activity: 2408
Merit: 1004
why simple put better quality fan
hero member
Activity: 924
Merit: 1000
So I take it is loud. LOL.

Waiting for a more detailed review of the unit.

Do you think there is anyway beside firmware to baffle the sound?
donator
Activity: 543
Merit: 500
Initial post:

The SP10 offers the following modes of operation (first three columns taken from the webinterface, last column added by me):

Code:
1.00Th / ~720W  / ~quiet / 1,39 GH/s/W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W
1.43Th / ~1350W / normal / 1,06 GH/s/W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W

As you can see, the slowest (or "underclocked") mode is most efficient, whereas the two fastest modes offer about 20% less performance per Watt.

Update 24.04.14:

Code:
Mode (Web UI)                                   | Performance | In/Out Temp   | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle                                            |    0 Gh/s   | N/A           |        35 W
1.00Th / ~720W  / ~quiet     / 1,39 GH/s/W      |  1139 Gh/s  | 29 °C / 67 °C |  780- 820 W
1.35Th / ~1100W / ~quiet     / 1,23 GH/s/W      |  1341 Gh/s  | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal     / 1,06 GH/s/W      |  1412 Gh/s  | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W      |  1456 Gh/s  | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------

The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature  (3°C more than "normal") and a 5% reduction in performance compared to the standard mode.

So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption.

Update 26.04.14:

More pictures: http://imgur.com/a/afpMF#0
donator
Activity: 543
Merit: 500
I've been selected as a reviewer by Spondoolies-Tech for the SP10 miner.

Here is a link to their official thread: https://bitcointalksearch.org/topic/ann-spondoolies-tech-carrier-grade-data-center-ready-mining-rigs-521520
And the website: http://www.spondoolies-tech.com/

I received the unit today and as I was not prepared for that I will not be able to give a full review right away, so I will start with some impressions first Smiley

(Link to full gallery: http://imgur.com/a/kASCu#0 - Pictures are from my unit)



Note the text on the fans - "Xtreme" - that's exactly what they are. Make no mistake, this unit is loud. It's louder than my hairdryer. It offers multiple performance modes which also has an effect on the loudness. "Quiet" mode is about as loud as my hairdryer Cheesy

I guess there is no doubt that Spondoolies-Tech is a legit a company, as they already have delivered regular orders. (I guess my sample unit is from April batch.)

I've been running a 3 TH/s mining operation using BitFury hardware, which was very cumbersome to setup. The SP10 is incredible easy to setup - plugin power and LAN, have a look at your DHCP server or run a network scanner (I used http://www.softperfect.com/products/networkscanner/ ), navigate to that address using a webbrowser - done.

The hardware by Spondoolies-Tech is not for hobby mining, that's for sure. Their hardware is a professional as it can get [which also manifests in things like the loudness Wink ]


To be continued... (I will also happily answer any questions you may have.)
Jump to: