The 24/7 model has grey-colored fan blades, where the other one has white.
I believe the coolers that came in my Saturn have the white-style fans.
Oh no, color clash! What will I do?
Lol.
Oh no, the ones used by knc are not designed for 24/7 use!
What will we do?
Apart from have significant facts obscured by retards shills fools god knows who, for god knows what reasons? Could this explain the never-ending fluctuations on these rigs that defy logic regarding cooling(or heating)
Not lol, at all!
It makes it a little risky to reply when you set up that likely one of those "bad" types is going to respond.
I'm going to take a shot at it though. I work in a closely related branch of Math -- NP style problems. We sometimes see this behavior in the activity of a "solver" (algorithm) which is tackling a large problem -- i.e. trillions of operations to do an approximation...
If I were going to guess -- without doing much analysis -- the answer would be fairly simple...
Every attempt to "solve" the problem which is centered about a given difficulty has roughly equal probability.
However there is a certain amount of variation in the problems we are sent to solve.
If you watch CGminer you will see that most of the problems center around a given difficulty... Say you set Bitminter to use a minimum difficulty of 256 -- what do you see? Try 512 then 1024 -- what do you see as you watch the solution line roll by? (I actually did this testing because I too was curious.)
At difficulty 256, all of the problem will be centered around some value of say 500 -- just because... (it could be 450 or 750 or whatever -- I have not done more than a casual observation that it appears to be so.)
Then if the pool assigns a difficult of 128 say -- then you will see most difficulties come in close to say 384.
Assume difficulty 512: In all cases there is variance up and down from the new average -- which must be greater than 512 but is almost certainly less than 1024 -- since most of the "problems" are less than 1024 in difficulty.... sometimes it's like a random walk where for a while the difficulties tend to cluster around 1000 say -- but then the average drifts back down -- often slowly -- sometimes quickly.
What you see is a drifting up and down of the hash rate -- but -- it has more to do with the pool sending you a variance in the difficulty.
I suspect the chips are behaving in a way that the hash rate is constant -- but the presented problem difficulty is varying in the way I described... so you get a varying throughput based on the difficulty.
One of the specialists that hang out in this group and works or dabbles in probability theory could explain this better -- but might insist on using the dreaded
squiggles that you see in this type of math.
To summarize -- the difficulty presented to your miner will be varying about a central value in a somewhat random fashion. The variance may present itself in such a way as to mimic a pattern that you believe is meaningful -- and it could well be. Proving it might be tough -- the meaningful bit I mean.
As for the random walk part -- go back to the old standby of flipping coins... Heads == +1 tails == -1 ...
Once you drift above or below the zero line you can stay positive for quite a while sometimes (or negative).
Combine the random walks, and the natural variation and what you are looking at could be described quite easily in some mathematical notations.
CGMiner: The 5 Second response wanders up and down slowly. The long term response on CGMiner stays rock steady after a day or two with only long term drift... The fast response display just above the scrolling part varies fairly rapidly.
It seems to me to be related to the difficulty being presented -- not to a varying response of the Miner Chips.
That's is pretty much my best effort in plain English...
If you are having trouble seeing this -- try the tests I mentioned above of setting new minimum difficulties in your worker setup if you are on Bitminter. You may lose a few shares while you play -- but it should not be significant.
Cheers.
