Tinua
|
|
May 03, 2012, 11:34:08 PM |
|
Beside..the temperatur from my singles are only betwen 51°C (124°F) and 63 °C (145°F) with roomtemperatur from ~20 °C (69°F)!
|
|
|
|
Epoch
Legendary
Offline
Activity: 922
Merit: 1003
|
|
May 03, 2012, 11:36:12 PM |
|
I read allways that the BFL-Singles need ~80Watt. My PC with 1 mining GPU HD5770 and 15 BFL-Singles v3 suck ~920 Watt from the wall. My PC and GPU need 180 Watt; the 15 BFL's need 740Watt. This give only ~50Watt/Single ; this with the orginal China-Powersupply. Or is something wrong with my measurement? I put a Kill-A-Watt meter on a cluster of 10 Rev3 Singles, using the stock BFL PSU and stock fans. Here are my results: 10 Singles at idle: 181W 10 Singles + 1 host PC at idle: 206W 10 Singles + 1 host PC mining: 820W Since my PC is using 25W (Asus Eee Box EB1012, dual-core Atom processor), my 10 Singles use exactly 820W-25W or 795W (this is at the wall using the stock PSU). Thus each Single uses 79.5W at full load. These are Rev3 Singles, with the stock heatpipe cooler and stock fan. I am running these off of 120V/60Hz. Power draw shouldn't be much different for 230V/50Hz.
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
May 03, 2012, 11:38:37 PM |
|
Tinua you were asked HOW you measured the load 3 times but have never answered?
So here goes #4.
HOW did you measure a load of 50W? As in describe the steps, tools, and measurements that lead you to believe the single is using 50W.
|
|
|
|
Tinua
|
|
May 03, 2012, 11:47:59 PM Last edit: May 04, 2012, 12:40:16 AM by Tinua |
|
Tinua you were asked HOW you measured the load 3 times but have never answered?
So here goes #4.
HOW did you measure a load of 50W? As in describe the steps, tools, and measurements that lead you to believe the single is using 50W.
I Put the kill a watt meter on 1 host PC with an HD5770 and start to mine. Then i connect 15 Singles and start to mine. 1 host PC + GPU mining: 180W 1 host PC + GPU + 15 Singles idle: 412W 1 host PC + GPU + 15 Singles mining: 920W
|
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
May 04, 2012, 12:49:56 AM |
|
Tinua you were asked HOW you measured the load 3 times but have never answered?
So here goes #4.
HOW did you measure a load of 50W? As in describe the steps, tools, and measurements that lead you to believe the single is using 50W.
I Put the kill a watt meter on 1 host PC with an HD5770 and start to mine. Then i connect 15 Singles and start to mine. 1 host PC + GPU mining: 180W 1 host PC + GPU + 15 Singles idle: 412W 1 host PC + GPU + 15 Singles mining: 920W OK so you are using a kill-a-watt. That's good, they are generally accurate. This interested me: They changed the power phase in REV3, it looks like a much more efficient one.
It is possible that the changed power subsystem, coupled with the high voltage input has increased your efficiency by that much. However, I wouldn't expect that either one or the other is solely responsible for the boost.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
May 04, 2012, 05:24:17 AM |
|
Tinua you were asked HOW you measured the load 3 times but have never answered?
So here goes #4.
HOW did you measure a load of 50W? As in describe the steps, tools, and measurements that lead you to believe the single is using 50W.
Got my BFL (single top fan with the new heat sink) from someone else today. My Watt meter was measuring 50W (extra) when I checked WTF was going on ... After watching it for a while ... it jumped up to 75W. Answer is simple: the POS is throttling. Even 75W wasn't full 830Mh/s
|
|
|
|
Tinua
|
|
May 04, 2012, 07:09:32 AM |
|
Tinua you were asked HOW you measured the load 3 times but have never answered?
So here goes #4.
HOW did you measure a load of 50W? As in describe the steps, tools, and measurements that lead you to believe the single is using 50W.
I Put the kill a watt meter on 1 host PC with an HD5770 and start to mine. Then i connect 15 Singles and start to mine. 1 host PC + GPU mining: 180W 1 host PC + GPU + 15 Singles idle: 412W 1 host PC + GPU + 15 Singles mining: 920W OK so you are using a kill-a-watt. That's good, they are generally accurate. This interested me: I just see, that " kill a watt" is an brand. So i use not this!! I use one from "Brennstuhl". I will buy an other one today with different brand for have an second measurment. I just check the rig after 7h mining: still 930Watt. Single temperatur betwen 52°C and 63°C!
|
|
|
|
fizzisist
|
|
May 04, 2012, 07:23:25 AM |
|
I just see, that " kill a watt" is an brand. So i use not this!! I use one from "Brennstuhl". I will buy an other one today with different brand for have an second measurment.
I just check the rig after 7h mining: still 930Watt. Single temperatur betwen 52°C and 63°C!
Better would be to buy an ammeter and measure the DC current at the board level. Apparently these kill-a-watt devices can get very confused with inductive loads (like a switching power supply).
|
|
|
|
Tinua
|
|
May 04, 2012, 10:56:07 AM |
|
I just see, that " kill a watt" is an brand. So i use not this!! I use one from "Brennstuhl". I will buy an other one today with different brand for have an second measurment.
I just check the rig after 7h mining: still 930Watt. Single temperatur betwen 52°C and 63°C!
Better would be to buy an ammeter and measure the DC current at the board level. Apparently these kill-a-watt devices can get very confused with inductive loads (like a switching power supply). You are absolutly right!!! Looks like "this midle price" kill a watt devices have BIG- problem with iductive loads!! I connect my new "cheap" devices and it shows: 1 host PC + GPU mining: 180W (old 180W) 1 host PC + GPU + 15 Singles idle: 418W (old 412W) 1 host PC + GPU + 15 Singles mining: 1370W (old 920W) This give 1190W for 15 Singles = 79.33W/SingleMystery solved!Means: more expensive german-tools are not allways better! Sorry for the revolt!!!
|
|
|
|
bulanula
|
|
May 04, 2012, 12:14:04 PM |
|
I just see, that " kill a watt" is an brand. So i use not this!! I use one from "Brennstuhl". I will buy an other one today with different brand for have an second measurment.
I just check the rig after 7h mining: still 930Watt. Single temperatur betwen 52°C and 63°C!
Better would be to buy an ammeter and measure the DC current at the board level. Apparently these kill-a-watt devices can get very confused with inductive loads (like a switching power supply). You are absolutly right!!! Looks like "this midle price" kill a watt devices have BIG- problem with iductive loads!! I connect my new "cheap" devices and it shows: 1 host PC + GPU mining: 180W (old 180W) 1 host PC + GPU + 15 Singles idle: 418W (old 412W) 1 host PC + GPU + 15 Singles mining: 1370W (old 920W) This give 1190W for 15 Singles = 79.33W/SingleMystery solved!Means: more expensive german-tools are not allways better! Sorry for the revolt!!!Can you post a picture of this device ? If my power meter is wrong then I am I believe I have 3 of them like this : Which one are you using to get the bad results ? Thanks !
|
|
|
|
Tinua
|
|
May 04, 2012, 12:46:04 PM |
|
I just see, that " kill a watt" is an brand. So i use not this!! I use one from "Brennstuhl". I will buy an other one today with different brand for have an second measurment.
I just check the rig after 7h mining: still 930Watt. Single temperatur betwen 52°C and 63°C!
Better would be to buy an ammeter and measure the DC current at the board level. Apparently these kill-a-watt devices can get very confused with inductive loads (like a switching power supply). You are absolutly right!!! Looks like "this midle price" kill a watt devices have BIG- problem with iductive loads!! I connect my new "cheap" devices and it shows: 1 host PC + GPU mining: 180W (old 180W) 1 host PC + GPU + 15 Singles idle: 418W (old 412W) 1 host PC + GPU + 15 Singles mining: 1370W (old 920W) This give 1190W for 15 Singles = 79.33W/SingleMystery solved!Means: more expensive german-tools are not allways better! Sorry for the revolt!!!Can you post a picture of this device ? If my power meter is wrong then I am Which one are you using to get the bad results ? Thanks ! The one with bad results is: Brennenstuhl - Typ: PM230 The one with accurat result (i hope) is: GA - Typ: EMT707CTL I read in internet, this one have bad measurment with low load (<20Watt), bud ok with mutch load.
|
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
May 04, 2012, 01:47:57 PM |
|
Yeah, that's usually the case.
|
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
May 04, 2012, 02:04:34 PM |
|
I have my wires broken out, I can measure a Rev3 unit with a clamp meter tonight for actual board draw.
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
May 04, 2012, 02:05:22 PM |
|
I have my wires broken out, I can measure a Rev3 unit with a clamp meter tonight for actual board draw.
Can you try it on 240v too?
|
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
May 04, 2012, 02:08:07 PM |
|
Yeah, I will try on both 120v and 240v and see if there's a difference. Though wouldn't the PSU mask that difference? I guess I can measure at the wall, too... although I will have to do some rewiring to measure 120v at the wall.
I was just going to measure the DC current between the PSU and BFL
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
May 04, 2012, 02:09:04 PM |
|
Yeah, I will try on both 120v and 240v and see if there's a difference. Though wouldn't the PSU mask that difference? I guess I can measure at the all, too... although I will have to do some rewiring to measure 120v at the wall.
Ah I forgot that you were measuring DC. Yeah, it would only make a difference at the PSU on the AC side.
|
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
May 04, 2012, 02:14:01 PM |
|
Incidentally, I can also authoritatively say that 20ga wire will melt if you try to run a BFL on it.
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
abeaulieu
|
|
May 04, 2012, 02:58:36 PM |
|
Incidentally, I can also authoritatively say that 20ga wire will melt if you try to run a BFL on it.
Possibly the best post in this thread.
|
|
|
|
Cablez
Legendary
Offline
Activity: 1400
Merit: 1000
I owe my soul to the Bitcoin code...
|
|
May 04, 2012, 05:45:10 PM |
|
Incidentally, I can also authoritatively say that 20ga wire will melt if you try to run a BFL on it.
How long was the run? <3'?
|
Tired of substandard power distribution in your ASIC setup??? Chris' Custom Cablez will get you sorted out right! No job too hard so PM me for a quote Check my products or ask a question here: https://bitcointalk.org/index.php?topic=74397.0
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
May 04, 2012, 06:12:52 PM |
|
6" or 8" or so?
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
|