Bitmain recently released the APW3++ 1600 watt power supply for $145 (including shipping). This power supply can now power an Antminer S9, S7 or L3+ on a standard USA 15 amp 110v power outlet. We wanted to see how this new power supply compared to the EVGA 1600 G2 power supply. We tested both power supplies using the stock L3+ miner frequency level of 384M and then with the overclock speed of 462M. Our results can be found below.
Stock 384M @ 504 mh/s
AW3++ (850w)
EVGA 1600 (840w)
Overclocked 462M @ 610+ mh/s
AW3++ (1025w)
EVGA 1600 (1015w)
Our data revealed that increasing the frequency to 462M increased the miners watt usage but also increased hashing rate to 610-650mh/s (stock is 504 mh/s) The EVGA 1600 was slightly more efficient (10 watts less) when compared to the APW3++. The EVGA 1600 G2 retails for $370 and the APW3++ costs around $145 depending on your location. Since the EVGA will set you back $225 more than the APW3++ and only uses 10 watts less, it would take approximately 8,333 days to begin to see any savings (using our $0.12 price per kilowatt cost for power). It is highly unlikely the power supply would last the 22 years to break even with the APW3++.
From our results, the APW3++ looks to be a great power supply at a reasonable price. We were very happy with the results from both power supplies specifically after increasing the frequency of the L3+, as the overclocking led to a 20% increase in total hashing power.
Here is a video showing our results
20% increase is quite a big change, nice job.
Where did you get the Antminer L3+, directly from bitmain?
Yes well worth the extra power use. I ordered from Bitmain back in April and just received the miner earlier this week. Thank you for watching!
This post received a 2.6% upvote from @randowhale thanks to @sticker! For more information, click here!
Shizer, Are you shortening the life of your L3+ by pumping up the volume? Will you get 20% less now but the machine will last 5 more years if you skip the overclocking? What wear and tear will you cause... comprend pa?
@fantastico If you check out older model miners such as the Antminer s1 or even the s7, miners make the largest profits in the first few months. As the difficulty increases and new more advanced mining equiptment is created then the units we use today are less profitable, i would be very surprised if this miner is profitable in 3 years from now. This is why I believe over clocking is the best option in terms of profit. The difference between over clocking and not overclocking is Bitmain has found the most efficient speed to mail the miner where the hashing power compared when compared to power use is at its lowest rate. Overclocking just leads to additional power usage.
Ok, then how about the power question. L3+ was designed to run off of 800W. So you have your antminer running at the higher frequency. Will that harm the machine?
Curious to know how are the machines doing so far? I'm seriously thinking of overclocking mine after reading & seeing your vids. I makes perfect sense but curious to know how are your miners doing now. Thanks
I've been running 6 L3+'s for almost 2 months now and using the overclock specs Killenmes states above and they are spot on. I've got an average of 600 mhs per machine ;
and zero board faults or problems. Now trying to find the same good information on the Bitmain D3 Dash Miner. The above source for info is credible.
What kind of HW errors you getting per day? Do they negatively affect your effective hashrate? For example it is doing 600 MH but with the 1000 HW it is making a certain percentage of that reported MH useless? have mine now set at 500 frequency doing 650 MH but my average seems to be around 610 620 in the RT is 650; not sure what the actual difference is. JR large amounts of HW have a long-term negative affect on the minor? I have had mine over clocked from 462 to 500 frequency for the past 2 to 3 weeks now and the only issue I see is the high HW. I have my L3 plus set up in my garage where it gets down to 32°F at night and up to 50° during the day. My chip temps are all below 60 and PCB temps are under 50, I’ll try to find a way to share a screenshot. I’m running my machine or nice hash and I should really be doing experiments but it would require me sit in there for a while with the bitcoin price fluctuations. At first my experiment was to change it every 12 hours and look at how much USD was earned but the fault there was that USD earned was based on the current BTC price - same issue using amount of BTC earned per hour or six hours or day. And I don’t know how much of a price fluctuation there is for being paid at any given time for your work. For example at noon someone could be paying $1 per MH (well 0.001 btc/mh/hr) and then at 1 AM someone could be paying you two dollars per mega hash – for this example of the numbers were completely pulled out of thin air to illustrate the point. Does anyone know if your paper mega hash is constant or constant enough to do a test or is there way to find out how much you are currently being paid per mega hash by users. I knew you could go to the marketplace and see what the current offers are but that is it necessarily what you’re being paid. I wish I had 2 as I would keep pushing it up because if someone mentioned earlier most of the money earned is in the first few months and these went from making $75-$100 a day down to $25 per day. I know they just released another batch for release in December so better enjoy that $25 a day while you can
Killn - do you have overclock info for D3 yet ?