Pages:
Author

Topic: Swedish ASIC miner company kncminer.com - page 74. (Read 3050068 times)

full member
Activity: 121
Merit: 100
September 04, 2015, 11:30:57 AM
I usually hang out on LiteCoinTalk, so I'm late to the game, but here it goes...

Questions:

1) Is it worthwhile to replace the cheap VRMs that KNC installed with an upgraded version? 
We can buy them here:
http://avnetexpress.avnet.com/store/em/EMController?action=products&catalogId=500201&storeId=500201&N=0&defaultCurrency=USD&langId=-1&slnk=e&term=BMR4641008%252F001&mfr=ERC&CMP=KNC-Octopart_VSE&c=USD&l=-1

But they aren't cheap, and the minimum order quantity is 65.  Each Titan cube would require 8 VRMs at $30 each.  Group buy?


2) Has anyone successfully refloated any of the components on a Titan cube circuit board?  I have tried to do this, and the solder KNC used has an unusual resin (or epoxy) coating on it.  Nothing would stick to it, and it got powdery when it melted.  Either they used really cheap solder, or solder that was never meant to be fixed?


3) The controller boards are the weakest link.  They eventually need to be eliminated, and the cubes need to be run directly via USB-to-serial. 
   3a) Have we tried offering a bounty to anyone who can provide or reverse-engineer the FPGA code?
   3b) I have cubes that will run reliably on all four dies when they are alone on a controller board, but as more cubes are added, dies begin to fail.  This is not a power issue.  I believe it is the RasPi becoming overloaded and failing to properly handle communication with the controller.  Has anyone been able to run a more powerful RasPi, like a RasPi 2 or some variant?
full member
Activity: 121
Merit: 100
September 04, 2015, 11:28:04 AM
Guys I need your help. Ive looked everywhere for the correct location of the die numbers, and I just cant seem to find them. I need this information badly, as I want to test some new techniques for hardware die repair. If anyone has this information or can find it , I would appreciate it. If you know which die is which, Please use the pic below and edit it and post. Please dont guess. Thanks.

Vegas




 

If you Google pics of the old Neptune(?) boards, someone mapped the cores to the VMs and made a picture of it.  If the boards are the same (and I think they are), the mappings should be the same.
sr. member
Activity: 342
Merit: 250
September 04, 2015, 10:59:10 AM
I dont own a Kill A Watt , but if no one steps up to the plate, then Ill order one. Im sure Steve , MUST have one. LOL

Vegas

lol, actually I do have one ... what was it you needed?  Total watts used at the wall?

I'll see if I can get some stats out to you
legendary
Activity: 1610
Merit: 1003
"Yobit pump alert software" Link in my signature!
September 04, 2015, 10:37:30 AM
We are setup to receive only LTC and BTC , but I can add other payment options if enough people want them. Here are the other coins I can add:

legendary
Activity: 1610
Merit: 1003
"Yobit pump alert software" Link in my signature!
September 04, 2015, 10:24:00 AM
I dont own a Kill A Watt , but if no one steps up to the plate, then Ill order one. Im sure Steve , MUST have one. LOL

Vegas
legendary
Activity: 1078
Merit: 1050
September 04, 2015, 10:22:34 AM
Tail log and current pool/s being hashed on.
legendary
Activity: 2450
Merit: 1002
September 04, 2015, 10:07:23 AM
Im also looking for some wall watts vs webgui watts stats too if anyone has them =)
So I can more accurately estimate wall watts stat in webgui.
legendary
Activity: 1610
Merit: 1003
"Yobit pump alert software" Link in my signature!
September 04, 2015, 09:05:56 AM
The v.99a firmware is polished and clean. We are looking for ideas for future software mods. Im continuing my research and video on the hardware end. I should have that fancy Hakko FR-300 desodering tool in my hand tomorrow, and THAT is when the real fun begins. SO if you guys have any ideas or anything you want added , then now would be a good time to open a discussion about what you want added to the future software mod.

Vegas
newbie
Activity: 28
Merit: 0
September 04, 2015, 07:37:56 AM
Just looked once again at logs and figured out!!! (It's like a glowing lamp nead my head Smiley )

Try to folow me:

1. I have a problem with ASIC4 connector. In web it does not show temp! And cube not working.
2. The voltage on one pin ov ASIC4 connector is differrent from others connectors.
3. The 1.04 ver log show continious  Die 3-2 restarted
                                                  Restarting bfgminer
                                                  Die 3-0 came UP
4. Ver 2.00 show I2C error

5. What a stupid guy was I!!!! THIS ASIC4 Connector F#CKs BRAIN to the system board!!! That's why nothing is working!!!!





You had the cable around the wrong way?

I am not sure. But if I plug one cube to bad port and one cube to any other port and set 50MHz at Cube on bad port system works continiously without restarts!!!
I need to make some plug for bad port to imitate Cube!!!
Need some advice on that please  Smiley

Here is the official answer on my issue
Quote
Hi,

It is as I stated in my previous email to you: "The only reason a controller board would perform in that way that you have described is if something is not connected correctly, or is causing an interference with the communication between the controller and cubes."

It would appear that there is a short on the controller board which is causing the issue, it could be some debris or something on some point of the circuit which is causing the short, or the main I/O board has failed in some point.

The only way to resolve this would be to change the main I/O board completely.

Med vänlig hälsning | Best regards

Now system is good working if Cube with no power is plugged to bad port. I need some cube imitator to create Smiley
legendary
Activity: 1078
Merit: 1050
September 04, 2015, 07:16:23 AM
Just looked once again at logs and figured out!!! (It's like a glowing lamp nead my head Smiley )

Try to folow me:

1. I have a problem with ASIC4 connector. In web it does not show temp! And cube not working.
2. The voltage on one pin ov ASIC4 connector is differrent from others connectors.
3. The 1.04 ver log show continious  Die 3-2 restarted
                                                  Restarting bfgminer
                                                  Die 3-0 came UP
4. Ver 2.00 show I2C error

5. What a stupid guy was I!!!! THIS ASIC4 Connector F#CKs BRAIN to the system board!!! That's why nothing is working!!!!





You had the cable around the wrong way?
newbie
Activity: 28
Merit: 0
September 04, 2015, 06:09:23 AM
Just looked once again at logs and figured out!!! (It's like a glowing lamp nead my head Smiley )

Try to folow me:

1. I have a problem with ASIC4 connector. In web it does not show temp! And cube not working.
2. The voltage on one pin ov ASIC4 connector is differrent from others connectors.
3. The 1.04 ver log show continious  Die 3-2 restarted
                                                  Restarting bfgminer
                                                  Die 3-0 came UP
4. Ver 2.00 show I2C error

5. What a stupid guy was I!!!! THIS ASIC4 Connector F#CKs BRAIN to the system board!!! That's why nothing is working!!!!



legendary
Activity: 2212
Merit: 1001
September 04, 2015, 05:04:37 AM
Yestarday we  have removed the Titan chip CAP Smiley Today will test it using chip water cooling, to check if the thermal grease was a problem..


ah.....big brass ones you have Smiley

I heard it like this,for some reason  Cheesy

copper member
Activity: 2898
Merit: 1465
Clueless!
September 04, 2015, 04:45:20 AM
Yestarday we  have removed the Titan chip CAP Smiley Today will test it using chip water cooling, to check if the thermal grease was a problem..


ah.....big brass ones you have Smiley
newbie
Activity: 28
Merit: 0
September 04, 2015, 02:05:41 AM
Yestarday we  have removed the Titan chip CAP Smiley Today will test it using chip water cooling, to check if the thermal grease was a problem..
https://i.imgur.com/EMrGqjDs.jpg
legendary
Activity: 2450
Merit: 1002
September 03, 2015, 09:33:27 AM
I tried that. The problem is I have the cap removed and I put copper heatsinks on the dies. The problem is the copper heatsinks is not enough to cool the dies so they overheat and shutdown before I can see the readings. With the cap on, its spreads the heat throughout the cap, and I cannot tell where is heat is coming from. Just hoping someone knew the locations of the die numbers.

Vegas

Have you tried it at 50mhz low voltage?

Not yet, dont know why I didnt try this. Will do today. I badly need to map these dies as I am working on a special fix.

Vegas

Better yet, I would try just turning ON one die at a time, 50mhz low voltage. That way the entire package temperature is lower too =)

Traditionally, if ASICs are laid out anything like standard PCB components, wherever the little dot is, indicates "pin 1" or if there is no painted difference then its assumed to be to the left ... in this case the dot is on the upper left corner of the package according to your pictures. If its laid out logically like standard PCB stuff then that would indicate die 0, then again this is KNC we are talking about so... who fucking knows.
legendary
Activity: 1610
Merit: 1003
"Yobit pump alert software" Link in my signature!
September 03, 2015, 08:08:57 AM
I tried that. The problem is I have the cap removed and I put copper heatsinks on the dies. The problem is the copper heatsinks is not enough to cool the dies so they overheat and shutdown before I can see the readings. With the cap on, its spreads the heat throughout the cap, and I cannot tell where is heat is coming from. Just hoping someone knew the locations of the die numbers.

Vegas

Have you tried it at 50mhz low voltage?

Not yet, dont know why I didnt try this. Will do today. I badly need to map these dies as I am working on a special fix.

Vegas
newbie
Activity: 28
Merit: 0
September 03, 2015, 01:52:19 AM
Im doing this from memory, but I beleive the factory capacitor to be 16v and 1000uf. Has anyone installed a 16v /1500 micro farad to see if it works? If it does work, is there any benefit?

Vegas
I had set 1500 uF. The result is the same as original cubes.. cycle restarts at low versions of firmware (1.02-1.04) and no cubes at higher (I2C com error at 2.0)
Will try to set all new 1000uF but a bit later..
legendary
Activity: 2450
Merit: 1002
September 02, 2015, 11:21:47 PM
I tried that. The problem is I have the cap removed and I put copper heatsinks on the dies. The problem is the copper heatsinks is not enough to cool the dies so they overheat and shutdown before I can see the readings. With the cap on, its spreads the heat throughout the cap, and I cannot tell where is heat is coming from. Just hoping someone knew the locations of the die numbers.

Vegas

Have you tried it at 50mhz low voltage?
legendary
Activity: 1610
Merit: 1003
"Yobit pump alert software" Link in my signature!
September 02, 2015, 07:54:08 PM
I tried that. The problem is I have the cap removed and I put copper heatsinks on the dies. The problem is the copper heatsinks is not enough to cool the dies so they overheat and shutdown before I can see the readings. With the cap on, its spreads the heat throughout the cap, and I cannot tell where is heat is coming from. Just hoping someone knew the locations of the die numbers.

Vegas
sr. member
Activity: 342
Merit: 250
September 02, 2015, 03:41:08 PM
Guys I need your help. Ive looked everywhere for the correct location of the die numbers, and I just cant seem to find them. I need this information badly, as I want to test some new techniques for hardware die repair. If anyone has this information or can find it , I would appreciate it. If you know which die is which, Please use the pic below and edit it and post. Please dont guess. Thanks.

Vegas




 

are you looking for that Ux number? in the pic it shows U3-U7.  on the pcb they run U1-U8

on the advanced settings they run 0-7  ...  for example if zero is running hot, put the heat sink on U1,

is this what you're looking for??  

.... ohhhhh, I bet you got your hakko, have fun with that slick little desoldering gun
Pages:
Jump to: