Swedish ASIC miner company kncminer.com

[davewr2013]

Does anyone know where I can buy more thermal blocks of the same size as what is in the November Jupiter that cover the new VRM's?  Mine didn't come with any and it performs like an October Jup.

Newark, Digikey Mouser.

For example:
http://canada.newark.com/webapp/wcs/stores/servlet/Search?st=thermal+pad&catalogId=15003&categoryId=800000006679&langId=1&storeId=10196

They can ship overnight.

Otherwise a local computer dealer may have the thermal grease or the pads.

[Paladin69]

Does anyone know where I can buy more thermal blocks of the same size as what is in the November Jupiter that cover the new VRM's?  Mine didn't come with any and it performs like an October Jup.

Newark, Digikey Mouser.

For example:
http://canada.newark.com/webapp/wcs/stores/servlet/Search?st=thermal+pad&catalogId=15003&categoryId=800000006679&langId=1&storeId=10196

They can ship overnight.

Otherwise a local computer dealer may have the thermal grease or the pads.

Thanks but they are so many options!  My head asplode.  We know it is 1cm thick I guess...which is 10mm, the largest they carry is 2mm it looks like...

[vesperwillow]

Anyone thought about building a custom case. I was thinking about a better (non-rack) design which is visually appealing to turn the rig into a spouse friendly space heater, internally mount the PSU (and rats nest of cables) and at the same time improve cooling.

ASIC boards are ~5" wide.  Optimal airflow would be units side by side in a line. Now put the PSU (on its side) at the end.  This would make the case ~24" wide.

Standard PSU is about 7.5" deep plus you need some room for routing the cables so say 12". 

So imagine a wooden box 24" x 12" x 7".  The entire front and back would be a grill for improved airflow.  The fans includes are pretty crappy so I think by upgrading to some higher airflow (but still low volume) fans one could get rid of the "intake" fans.  If needed it should be possible to mount a fan on each side of the heatsink in a push pull config.

Excuse my horrible ASIC art but here is a top view.  Front and back are mesh screens.  Arrows indicate airflow.  PSU intake could be on the side of the case or if case is made larger draw flipped 180 to draw from the inside.

Code:

|---------------------------------|
|   |      |      |      |        |
|   V      V      V      V        |
|[ASIC1][ASIC2][ASIC3][ASIC4][PSU]|  <--
|   |      |      |      |     |  |
|   V      V      V      V     V  |
|---------------------------------|

Still a rough idea at this point but wondered if anyone had considered something like this?

I literally came up with this same idea a month back when looking into the KNC design. Makes it more compact/easier to put in places (especially under a couch haha, blow warm air on your feet). I say go for it. If I can ever get more modules, I might. My mercury wouldn't benefit lol

[Puppet]

Anyone thought about building a custom case. I was thinking about a better (non-rack) design which is visually appealing to turn the rig into a spouse friendly space heater, internally mount the PSU (and rats nest of cables) and at the same time improve cooling.

ASIC boards are ~5" wide.  Optimal airflow would be units side by side in a line. Now put the PSU (on its side) at the end.  This would make the case ~24" wide.
..
Still a rough idea at this point but wondered if anyone had considered something like this?

If anyone wants to work out a 3D design that uses elements smaller than 20x20x20 cm (and perhaps square wooden  or aluminium booms), Id be happy to try 3D printing it.

[davebodger]

Does anyone know where I can buy more thermal blocks of the same size as what is in the November Jupiter that cover the new VRM's?  Mine didn't come with any and it performs like an October Jup.

Has it been confirmed that those are actually termal blocks to cool the VRMs?
They look far too thick to be effective, and I have never seen any that thick before when looking to buy pads for GPUs and CPUs myself.
If they were 1mm pads on either side of a bar of aluminium or copper I could understand it, but at that thickness any soft/foam material is more likely to work as an insulator.

I have checked my November Jupiter and it does not have them.
Hashing happily at 657 GH/s presently, pulling 855 Watts from the wall through an EVGA 1000P2 PSU.

My guess, without further info, is that they are packing material used to either aid assembly or to protect the VRMs during assembly operations.
That would mean that somebody forgot to pull them out before shipping. 

Good luck in your search.

[RickJamesBTC]

I switched to ghash.io, and immediately my HW errors dropped below 1.5% on all machines and have stayed there. Bye bye eligius I suppose!
Anyone have a good starting point for the difficulty setting with ghash? I'm at 256 right now.

[Biffa]

Does anyone know where I can buy more thermal blocks of the same size as what is in the November Jupiter that cover the new VRM's?  Mine didn't come with any and it performs like an October Jup.

Has it been confirmed that those are actually termal blocks to cool the VRMs?
They look far too thick to be effective, and I have never seen any that thick before when looking to buy pads for GPUs and CPUs myself.
If they were 1mm pads on either side of a bar of aluminium or copper I could understand it, but at that thickness any soft/foam material is more likely to work as an insulator.

I have checked my November Jupiter and it does not have them.
Hashing happily at 657 GH/s presently, pulling 855 Watts from the wall through an EVGA 1000P2 PSU.

My guess, without further info, is that they are packing material used to either aid assembly or to protect the VRMs during assembly operations.
That would mean that somebody forgot to pull them out before shipping. 

Good luck in your search.

I pulled one off, its definitately thermal pad material. I've also measured the temp of the vrm under the thermal pad and the same VRM's with the thermal pad removed and without the pad they are definitely hotter. They are also on every board in the same place in my Nov Jupiter. Maybe some machines needed them and some didn't?

[Biffa]

I switched to ghash.io, and immediately my HW errors dropped below 1.5% on all machines and have stayed there. Bye bye eligius I suppose!
Anyone have a good starting point for the difficulty setting with ghash? I'm at 256 right now.

512 seems to be optimal for 550GH/s

There is a thread here https://asktom.cf/index.php?topic=274023.0

In which ckolivas suggests

TL:DR
Set your diff to your worker's hashrate in GH / 1.4

555/1.4 = 396 Difficulty, but not sure if you have to round it up.

[Entropy-uc]

Thanks for all your efforts on this D&T.  This worked perfectly on my dead machine as well.

[joeventura]

Anyone know how to get cached copies/screen grabbed images of the KNC site from back in June/July/August/September. Also any modifications to their order TOS.

Archive.org  ??


https://web.archive.org/web/20130822044311/https://www.kncminer.com/categories/miners

oh and welcome to page 1300

[RoadStress]

Anyone thought about building a custom case. I was thinking about a better (non-rack) design which is visually appealing to turn the rig into a spouse friendly space heater, internally mount the PSU (and rats nest of cables) and at the same time improve cooling.

ASIC boards are ~5" wide.  Optimal airflow would be units side by side in a line. Now put the PSU (on its side) at the end.  This would make the case ~24" wide.
..
Still a rough idea at this point but wondered if anyone had considered something like this?

If anyone wants to work out a 3D design that uses elements smaller than 20x20x20 cm (and perhaps square wooden  or aluminium booms), Id be happy to try 3D printing it.

Sign me up if that happens.

[FUKT]

Anyone know how to get cached copies/screen grabbed images of the KNC site from back in June/July/August/September. Also any modifications to their order TOS.

They have quite a few archived versions on the wayback machine at http://web.archive.org/web/*/http://www.kncminer.com .

The problem is the cached copies do not include the flash banner KNC used on their website. I'm specifically looking for the October 15th order banner and the September banner or such.

[Biffa]

Anyone know how to get cached copies/screen grabbed images of the KNC site from back in June/July/August/September. Also any modifications to their order TOS.

They have quite a few archived versions on the wayback machine at http://web.archive.org/web/*/http://www.kncminer.com .

The problem is the cached copies do not include the flash banner KNC used on their website. I'm specifically looking for the October 15th order banner and the September banner or such.

Try using the wayback machine to search for https://www.kncminer.com/pictures/carousel/5390.png

[The Avenger]

Anyone know how to get cached copies/screen grabbed images of the KNC site from back in June/July/August/September. Also any modifications to their order TOS.

They have quite a few archived versions on the wayback machine at http://web.archive.org/web/*/http://www.kncminer.com .

The problem is the cached copies do not include the flash banner KNC used on their website. I'm specifically looking for the October 15th order banner and the September banner or such.

Do you need images? The newsletters contain all their promises. "Not later than 15th October" promises were made in these two newsletters:
https://www.kncminer.com/news/news-33
https://www.kncminer.com/news/news-37

[edgar]

ahhh... SEPTEMBER was the 'promise'

...but the KnC fluffers will tell you a promise isnt a promise unless they say 'we solemnly promise'.

lets see how that holds up when they get sent a box of wank... and NO support at all (but maybe being a fluffer DOES have benefits)

[FUKT]

Anyone know how to get cached copies/screen grabbed images of the KNC site from back in June/July/August/September. Also any modifications to their order TOS.

They have quite a few archived versions on the wayback machine at http://web.archive.org/web/*/http://www.kncminer.com .

The problem is the cached copies do not include the flash banner KNC used on their website. I'm specifically looking for the October 15th order banner and the September banner or such.

Do you need images? The newsletters contain all their promises. "Not later than 15th October" promises were made in these two newsletters:
https://www.kncminer.com/news/news-33
https://www.kncminer.com/news/news-37

Thanks, thats what I needed.

[Humax]

What's with this another pre-order crap?!
We have all (first batch buyers) helped KnC startup with their business and it is now up and running like a train.
Now they treat their customers like shit and put up another pre-order BS!

And the big question is when are they going to deliver the Neptune?
They say Q1/Q2. Looking back to their original promise "before September" and then "all payed order before 15th of October", which obviously they have never made it.
Maybe when lucky the first unit will be shipped at end of June 2014?

I just hate people who can't keep their words!
Anyway, couldn't be worse than BFL.... 

[Phoenix1969]

What's with this another pre-order crap?!
We have all (first batch buyers) helped KnC startup with their business and it is now up and running like a train.
Now they treat their customers like shit and put up another pre-order BS!

And the big question is when are they going to deliver the Neptune?
They say Q1/Q2. Looking back to their original promise "before September" and then "all payed order before 15th of October", which obviously they have never made it.
Maybe when lucky the first unit will be shipped at end of June 2014?

I just hate people who can't keep their words!
Anyway, couldn't be worse than BFL....  

And as a reward, we all got a 3000 dollar discount on Neptunes...!  (the first 1200 were for us!)
I saved 9 grand on that deal.
I consider that alone a hefty reward!
Yes, each delivery batch was a few days late, but like close to 50% over spec on performance from our purchase date.
Not a bad trade-off, eh?
Imagine the ASIC race like the Kentucky Derby...
Your horse may have been slower out the starting gate than predicted... but...
He was faster than the rest, and won the race running twice as fast as predicted...... by a mile...
Do we shoot him anyway?    Would you bet on him again?
I'm betting the trainer of that particular horse works on the gate launching, and improves.  

[KGBSlim]

Between me and my friend we ordered 3 Jupiters for November, recieved 2 in November and 1 on Dec 5th. All 100GHs+ over what we thought we had ordered.. Wondering how many GHs they will 'over-deliver' on the Neptunes?

[RHA]

I switched to ghash.io, and immediately my HW errors dropped below 1.5% on all machines and have stayed there. Bye bye eligius I suppose!
Anyone have a good starting point for the difficulty setting with ghash? I'm at 256 right now.

I confirm that. Switching to other pools changes hardware error rates shown by cgminer from 14% to 1.6%.

I've performed a long test with Eligius, writing down all cores numbers turned off and on. The turning off starts not immediately but after a dozen of minutes.
The cores turned off are purely random. Then cgminer turns them on (one by one). They are still randomly turned off.
(Experimental "tuning" version of cgminer 3.8.2 is no better.)

It looks like something (a config string or data sent from Eligius) causes something wrong (like buffer overrun) with November miners,
making cgminer working erroneously, especially spoiling random nonces (what in effect degrades effective hashing rate and causes cgminer to turn off good cores gradually).

(It doesn't always happen - maybe it depends on which server at Eligius it connects, I guess.)