KnCMiner Jupiter Miner First Impressions

[Sitarow]

Kano make a good firmware for Jupiter please

I think that firmware v0.9.4 drops the voltage to the chips further. It pulls 540 watts at 120v now.

[jelin1984]

Yes but not the same hashing

How was the watt for 0.91 firmware

[Sitarow]

Yes but not the same hashing

How was the watt for 0.91 firmware

I have never used it but from what I was told similar to v0.9.2

[DPoS]

Yes but not the same hashing

How was the watt for 0.91 firmware

I switched back to .91 on my crappy Jupiter and it acted the same so I switched it back to .94 like the other one  (why one hashes 50Gh less than the other, who knows..)

paid, real hash will be a problem until they fix the flushwork

[Sitarow]

Okay Update.

This is the results after 1hr on firmware v0.9.4 with the case cover off and the large fan placed on top pushing air down.

Yes I needed to restart the miner because of the api configuration settings. (Thanks Kano for the help!)

Results after 1hr on firmware v0.9.4 with adequate cooling.





Note: You can see from the results that this performance is far better then the results we have seen with the case cover on and relying on the stock fans to move cool air into the unit and the hot air out.

Wanted to add that the power at the plug is at 570 watts.

[mininganon]

My Jupiter on .94 fw after 4hr is running at 495watts and hashing at 435ghs and dropping?? It was at 480ghs@535 when it started up and settled. Also had cg crashing when the room went over 75F, mind you I have the top off with a ac fan blowing downward and another doing cross sweeps. This new firmware is far from a finished...

At 495watts
44.0 ℃
43.5 ℃
43.0 ℃
33.0 ℃

[DPoS]

thanks for the update Sitarow...  proving our assumptions perfectly


I do agree with someone on KNCforums who said that for jupiters that have a lot of HW errors, it is better to use .93
That is true, my crappy jupiter is showing hashing on BTCguild at 446 when it was barely around 410 on .94  (same cooling of fans with top off, but not as good as Sitarow yet)

My second jupiter probably is fine on .94 but just needs even more cooling.. avg 463 on BTCGuild



PS- mininganon, maybe switch to .93   it seems miners that nosedive run more stable on it.   You have it cooled well

[Sitarow]

thanks for the update Sitarow...  proving our assumptions perfectly


I do agree with someone on KNCforums who said that for jupiters that have a lot of HW errors, it is better to use .93
That is true, my crappy jupiter is showing hashing on BTCguild at 446 when it was barely around 410 on .94  (same cooling of fans with top off, but not as good as Sitarow yet)

My second jupiter probably is fine on .94 but just needs even more cooling.. avg 463 on BTCGuild



PS- mininganon, maybe switch to .93   it seems miners that nosedive run more stable on it.   You have it cooled well

Thank you for providing your experience. It helps with getting others to tweak their systems for their environment.

Here are the results after 2 hours on v0.9.4.

[DPoS]

glad .94 is working for you.  I think some miners may work better on different firmware right now

My subpar Jupiter that I put on .93 is actually climbing up now.  ~472
So I put my other Jupiter on .93 and it is now at ~500

I'll check back in a couple hours to see how it stays

[Sitarow]

glad .94 is working for you.  I think some miners may work better on different firmware right now

My subpar Jupiter that I put on .93 is actually climbing up now.  ~472
So I put my other Jupiter on .93 and it is now at ~500

I'll check back in a couple hours to see how it stays

I forgot to include the 50btc.com results

Here it is. Hardware errors are down to 7.4% and power is at 580 watts at 120v

[DPoS]

looking good there at 540+  and that graph is more detailed than BTCGuild


update on mine.. unbelievably my crappy jupiter is now 499 on BTCGuild and my other is still holding above 500 at 507

I can't believe .93 is working that much better than .94 on my miner...they shipped with .91 of course  


PS-was nice to cross the Th barrier:  Totals (All Active Workers) 1,009.40 GH/s

[mininganon]

Tryed .93 again now that I have the top off and mass fans and it hashes for about 90 seconds or so then cg crashes, repeat, repeat, etc...

Putting .91 back in action.  .94 for me is not profitable for me, stable or not taking an 80ghs hit after 3 hours of operation isn't acceptable. My main problem with .94 and .91 is this slow death of ghs the longer it runs.

EDIT: Hey is there a way for me to make my Jupiter auto reboot each hour, like a script or some cron job entry I can make? This could fix my .94 issue.

[DPoS]

Tryed .93 again now that I have the top off and mass fans and it hashes for about 90 seconds or so then cg crashes, repeat, repeat, etc...

Putting .91 back in action.  .94 for me is not profitable for me, stable or not taking an 80ghs hit after 3 hours of operation isn't acceptable. My main problem with .94 and .91 is this slow death of ghs the longer it runs.

I will count my blessing with .93 then...  both miners hashing over 500 now which is blowing me away since I was in the low 400's running .91 or .94

I agree, do what's getting you the most coin at this point

[DPoS]

EDIT: Hey is there a way for me to make my Jupiter auto reboot each hour, like a script or some cron job entry I can make? This could fix my .94 issue.

there's some creative ways, but probably a simple way with cgminer that one of the devs would know


update on my .93 miners:

the previously subpar miner is running 538 at BTCguild now  (crazy)
the other one is at 510

I couldn't be happier switching to .93 

[Sitarow]

5 hours into mining with open case and fan on firmware v0.9.4

[-Redacted-]

I must have gotten a pretty good one right from the factory then, because mine is exactly as it came from KnC (actually it has one dead front fan) and runs like this, with the case closed, on a corsair hx850.  I split the output to 2 separate BTCGuild pools.



Both pools/workers on BTCG have virtually identical charts:

[sickpig]

I must have gotten a pretty good one right from the factory then ....

definitely

what's your env temp?

edit:

here are the number for my 2 jups

Code:

	Jup1	Jup2
E	55599	54986
A	6156012	6198900
R	42206	31873
HW	365264	314332
		
			Jup1	Jup2
Accepted hrate: 	475.55	484.2
Full hashrate:		478.81	486.69
HW error %:		5.57%	4.80%
HW error hashrate:	28.22	24.55

[amencon]

I must have gotten a pretty good one right from the factory then, because mine is exactly as it came from KnC (actually it has one dead front fan) and runs like this, with the case closed, on a corsair hx850.  I split the output to 2 separate BTCGuild pools.



Both pools/workers on BTCG have virtually identical charts:

Nice speeds for sure.  I've tried all FWs and can't seem to get much over 470-480 steadily.  Oh well, first world problems right?

Also thought I had a dead front fan but for whatever reason it was jammed even though I couldn't see any obstruction.  Little muscle and it popped free and seems to be working great now.

[samurai1200]

Been running almost nonstop since Monday afternoon (gmt-8) on the firmware mine shipped with, 0.91.



Hesitant to update since people are so hit-and-miss with them.

I felt the caps that were blowing on other peoples' units with the back of my finger, after running the rig for a few hours, and none of them were getting hot -- indicates not much voltage/current ripple. Maybe faulty voltage regulators are putting unexpected stress on those caps?

Also, I did just notice that -Redacted- is running at diff 128. Would that have any effect on hashing speed? I would assume not, but i don't understand the concept of 'difficulty' very well.

[Xialla]

24h without any cgminer restart (.94)