If I'm not mistaken, that VRM PDF you linked is not the right one. That 50A unit was used in the Neptune only. KNC cheaped out even more by using a 40A unit in the Titan.

I think you're a little behind on the news LOL! It's (I thought so anyways) well known that both Neptune and Titan cubes pull approximately 360w to 400w from that poor PCI-e connector. A 6 pin PCI-e connector is rated at 75w and an 8 pin is rated at 150w. I'm not sure why, since all it does is add 2 ground wires while only leaving 3 12v wires. Where did you get the 275w figure from a 6 pin from?

I'm out running errands right now, I'll get you the measurements in a few hours. 

They absolutely are LOL. I calculated it myself by measuring the total DC power draw from the 3 server PSUs VS the AC power draw measured at the wall. From there it was easy to calculate; 94.13% AC/DC efficiency. Take a look at this: http://www.plugloadsolutions.com/psu_reports/POWER%20ONE_PFE1100-12-054NA_1100W_SO-152_80_Report.pdf

I'm using 3 of those to power 5 cubes. I could power them all with 2, but the PSUs fan gets way too loud near 100% load. They're completely silent at ~50% load, and a bit more efficient to boot. You must be using only 4 cubes if you're that far off from me?

Too bad there's no way to correct the extremely wrong DCDC values.  And I don't use the protection, don't need it in my case. Cold Canadian weather FTW!

What do you mean not everyone will be able to adjust? I personally think adding an option to change the efficiency rating is a very nice added feature, not clutter. I'd like the option to always be avilable, not only a 1 time setup. If you don't want it in the main "Advanced" page, you could put it in the "System" page?

I was thinking I'd get you an actual reading at the cube by measuring the amperage and volts to get the best accuracy possible for a cube.

Yep. I put the jumper directly on the PSU to avoid having that thick ass wire around for nothing. Easy to do if your wires are coloured. Just look for which pin connects to the green wire, and short it to any pin that connects to a black wire.

So, time for me to post an RMA experience with KNC. They were absolutely fantastic. I seriously could not have had better service. I almost feel bad for ripping them off in the process! (almost)

I've had my Titan up and running since October 6th of 2014. All dies were working initially, but a die crapped out 2 days later. A week later, a second die down. 2 weeks after that a third. 2 dies on one cube, and 1 die on an other. KNC only agreed to replace the cube with 2 dead dies, which pissed me off to the extreme after having paid $10000+ for the miner. I set up an RMA for the cube, but never followed through out of fear of the probably horrible service I would receive from KNC. Fast forward 10 months, and I now had a 5 cube Titan. I bought a cube with 2 working dies on this forum, but it turns out only 1 die worked when I received it. 

Fast forward a bit more to the end of September when I realize that I have to act now, or my warranty will be expired in early October. Get back in touch with Kurt @ KNC via email, explain (lie) that my 1 dead die cube is now down to 2 dead dies, and my 2 dead die cube is now down to 3 dead dies (really this is the cube I bought on this forum). Kurt promptly issues an RMA number and has me send both cubes back to KNC. $200 and 4 days later, my cubes arrive in Sweden on a Wednesday. Kurt informs me that they can't repair the cubes and that they've been replaced with new ones. They're shipped back out to me 2 days later on Friday, and get here early the following week. They've been hashing perfectly for the past 3 days, and I'm a happy camper.
 

Gentarkin, I have a request. Power readings in the advanced page are way off. It isn't even close! If I get you a proper reading of how many watts a cube consumes at 325MHx and -0.0366v would you be able to adjust them? Also, would be great to have the ability to set our own Wall efficiency. I'm running 94% efficiency PSUs, so 85% is no where near accurate.

You've got a strange way of looking at business. Being overloaded with work is a good thing. The opposite is what would be rough.

I think you need to do some more (up-to-date) reading before you spend any money.

What speed were you seeing at 675?

There's always a risk when it comes to overclocking. I personally have never destroyed HW by only upping the clock speed and not increasing voltage, but it's possible.

675 wasn't really an improvement over 650 for me, same thing for 700MHz. I'm running 650MHz for now.

Still have my package if you're interested.

Interesting. Post your results after you let it run with the fans on automatic, like mine are set to.


For now, a DPS2000B.

Someone shit in your Corn Flakes this morning? 

Well, that about confirms it. Now we just need someone with a default clock speed of 575 M to chime in.

Yeah, seems like I definitely got a good one. Ambient is about 20c right now, so that helps a bunch. Running 675MHz now, averaging 5260. Not much gain over 650MHz, looks like I've hit the point of diminishing returns.

Thanks for confirming my theory.

Cool, you must have one of the first public ones. I've got  feeling you won't have stickers.

Have I mentioned how much I like the S7 so far?






Some quick stats:

625MHz = 1220w DC (From PSU) 1300w AC (From WALL) 5015 gh/s 12.05v at hash board
650MHz = 1267w DC (From PSU) 1352w AC (From WALL) 5220 gh/s 12.05v at hash board

Next stop, 675MHz.  

I just noticed that my hash boards are labeled 600M, which I'll go ahead and assume means they've been tested to work properly at 600 MHz.



Can people who have received units with stock clocks of 575MHz confirm my theory? Also, for the few of you who received a unit before bitmin's time off, could you verify if you have a sticker at all? I'd guess there isn't one.