Topic: Hacking The KNC Firmware: Overclocking - page 11. (Read 144362 times)

the plate/bar is not as large as the one currently installed on the Neptune, nor will the foam be intruding on airflow over the VRMs.  It is the same design as was used to hold the radiators and ASICs together on the Jupiter.

This how the Arctics were mounted to the ASICs on the Jupiter ...



and using only the two holes, front and back, the end result opens the board (VRMs) for better air flow ...



off to sleep ... hopefully tomorrow is a better day for "receiving" 

i don't think anyone has done that...but why would you still be using the case? you get so much better temps outside the case

I am sorry but cannot understand the usage of your plates.

Can you please explain a little more?

Elenelen,
  I'm sure you know all about the "Super Jup" fiasco.  It looks (from the pics you posted and the info. KnC has show on their site) that the boards are the same dimensions as the 28nm boards.
  I made "bars", the same size and with exact location of all the mounting holes on the bars used on the Jupiters, "Super Jup" etc.  
  PM me and I'll send you a set of materials to mount the radiators (w/o all the foam and huge plate covering the VRMs), and risers to keep the board off the base.  One item though, which brand radiator did they use - Arctic or Cooler Master?
  I know i'm a newbie, but if you look on the KnC forum, http://forum.kncminer.com/forum/next-generation/hardware-designs/39700-mounting-hardware, you can see what I put together, as well as everyones input.



  I hope I can help ...

(Sorry the pic is so big  )

This is probably off topic so please forgive me. My KNC Jupiter case is all scratched up and looks like hell. I was wondering if anyone has powder coated their case? I called up a company and got a quote for $60 to do a flat black exterior power coat. I would remove all fans and modules and give them the case. What do people here think about the idea? Please don't hate on me  

found the problem asic...silly me was looking at wrong side...turns out it is a board...while putting together the new setup i kinda lost my balance on step ladder and leaned on one module doh...oh well least it still hashing some...dam me

Hi fubly,

I bought through ebay thermal pads of small thickness 0.15mm, and cut it into 11mm x 11mm pieces with a scissor.

Here it is:
http://www.ebay.com/itm/111341731505?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

Hi tzortz,

how did you glue the little heatsinks?

ambient is room temp 70F  the other 3 look the same...well except the non working one

I bought off ebay , here it is:
http://www.ebay.com/itm/370852348071?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649


Also I bought thermal adhesive pads of 0.5mm and cut in small pieces with scissors.

Whats your ambient?

Also, can you post a screenshot of the advanced page with the vrm temps?

Thanks.

or you can try and do what someone else said is remove the grey pad and use some paste and the metal bracket directly on top of vrms...dunno if the screw will screw all the way down thou...

anyone have a idea of why a controller would all a sudden ramp up volts on vrms? I've tried everything i can think of to try n get this module back up n running...and its not the module cause i swapped it out for another and same problem...all i did was rearrange to a new setup and now i have this problem...tried different plug on psu..tried different psu all together...the spot that is messing up is a new solder connector to make it a 6 mod controller...even redid the solder and no dice...what could i have done to piss it off?

I can get 3 cores to stay on @ 1000mhz and -0.1245...but the 4th will ramp up volt turn off, come back on at half then after a bit ramp back up n turn off etc...that seems low volts for 1000mhz and no cores are off...if i try a higher OC it'll go back to usually 2 ramp up n 2 work fine...it'll also flip which 2 work or ramp...the higher the oc the faster the ramp up n shut off...it was fine at 0x335 n 0.000 volts Sad I just don't get it...it was working fine for few weeks Sad anyone? I'll tip ya if i get working module again...oh even tried different wire too...wish i had another 6 module nov. jup to try it out

also get fault 16 n fault 81 on bertmod sometimes...but i don't see how it could be the module since i've tried a different one

Fixed it. Thanks!

You need to set the NTP time source up in the config, point it to something like pool.ntp.org and it will flip to the right time..

Thanks for the info.... Now I think of it: the plate is heating up from the Tower (and ASCI chip) as well... but removing the plate will also tumble down the Tower >> that's a major DIY project...

With proper cooling and 1.01-padrino FW you should be able to reach 1.5 TH/S. My modules have heatsinks on the VRM's and I have fans blowing on the bottom of each board.

Not sure how to change the date on the stats page, but I just took this screenshot.

that is why...you thought the chips needed the cooling when they don't...the fans blowing on bottom n top will get you 1.5th/s per 6 or close to it...my 6 were doing 1.436th/s cause i had 1 vrm that didn't like to go that high and would drop out...now putting together my new system i messed up my controller board 1 module is all wonky now and the fans blowing in front still get air flow thur heatsink...my chip heatsink isn't even that hot

and wow link for those heatsinks? wish i got those instead...they are huge...if you move those fans down like mine you'd get more then 1.5th/s cause you could push'em faster i bet

Here is how I did it.


I could see a difference but cannot reach 1.5T per 6 boards.
I am on 1.25T.

My ambient is 35C.

ya just picture the copper heatsinks on the bottom...don't really need a picture for that well the heatsinks will help even more on bottom too...spec. with neptunes since they are already oc'd so much it'll help you get the most outta it i think

you can't do both i don't think...ya the plate is alittle hot, but the reason its hot is cause the vrms are screaming hot...i really don't think its sucking up the heat from the vrms..i think its hot just from being around the vrms...oh and those gray things block air flow and trap heat in...and so does the case hope this helps

also i made the mistake of leaving the grey pad on 4 of my modules and only been ran for few weeks like that but you can see the damage it did...the caps are dark...i mean almost black now so i picture in a week or so with all those neptunes heating up, all vrms will be toast and probably break...that grey thing does nothing but heat up the vrms more...or maybe it just transfers the heat to the smd caps...idk but if i was a neptune owner i wouldn't be running them in a knc designed case...they may make a good 28nn chip but they suck at designing cases or heat management

OK, I understand. But the current KnC solution with the large aluminum plate might work as well as the little copper things. If that is the case we only need a better air-flow with Fans positioned the way like in the picture.

The plate is 45 C hot, so cooling it better might help...