Topic: Swedish ASIC miner company kncminer.com - page 12. (Read 3049520 times)

Did KNC mining finally go under?  For about the last 5 or 6 days, there are no more blocks listed as being found by them: https://blockchain.info/blocks/KnCMiner

They normally find about 3 to 5 blocks a day.

Hey KNCfun congrats on getting it to work -- which crystal did you use?  The one from digikey? I got the one from digikey and got someone lined up to swap it next week if that's the right one.

Vegas I still wait to see these titan repair videos that you promised to the public ?
These are common problems that you just discovered   you should knew them by now ....
I already did 34 re-flows to some of my cubes and to friend's cubes that they sent them for repair and most of them they were successful and some not because the little balls lift the contact from it's socket and that means no fixable.
I also did a couple of re-ballngs with custom stencil modified by me,, I have to say that if the little contact is lifted by the small balls then buy buy die .... and you really risk to lift it by yourself ... so re-balling only in special cases !

People don't forget that you was charging them $300-350 for simple repairs that you should charge $50, in common words ripping them off  
I can do these common repairs in 1-2 weeks and charge a normal price ($50-70)...
Lightfoot also has friendly prices  
    
Ps. I was the one that found the mistake in qbert's bridge boards !! So people that got burned from him now they can order the correct crystal and replace it and have finally a working bridge

Yup. The 2.0 requires a different microcode/IPL loaded (Fuck, I'm old) and the Titan 2.00 code doesn't include it. You can recompile the code and load it on a 2.0 Pi, but it's just. easier. to find. a 1.2 Pi.

I've done the reflow with some success, especially on the so-called taco boards, however sometimes the chip connections fail inside the die which is pretty much impossible to reflow. You do need preheat under the board, flux, bring it up to max temps, then use a real air tool at 400c to make it go. I'm sure you have a good IR preheater. And don't sneeze.

The .8v short is the one that sucks rocks; if you made a stencil then you could re-ball the die without the dead die's balls and get 2-3 dies back. Nicely done, post your progress.

Did you figure out how the level drivers for the SPI to die interface work and which ones are replacements yet? That seems to be the source of the board comes up but no dies hash, I have had some success with swapping out the top drivers, but a few still don't work.

Keep posting progress, it's how we help the community.

C

Thanks lightfoot. So, that brings up another question. Im quite positive that the titans only work with a very specific pi, and thats the original model pi that it shipped with. Ive Actually tried it with a pi 2, and it wouldnt boot. Just a blank, bright led on the controller.

Thanks for the info. Ive focused my research heavily on the boards themselves , non stop for the past year. Ive made it my full time job, as im retired.  Here is what I know. Most failures about 80% of the time, is because the wafer gets too hot , and separates from the pcb. Because of the design of the E plate, using 2 screws in the back, and only one in the front, it puts more pressure on the front of the wafer (closest to the fan), and the dies that almost always fail are the first die that gets its voltage from vrm 0 and 1. This die is closest to the exhaust in the back and is easily identified by the small pin hole on the plate.The only repair then is to have it reflowed professionally. Do NOT try a heatgun, because it scorches the wafer, and the many tiny pads, beyond repair, so that when it is lifted professionally, the tiny pads stay with the pcb, and is unrepairable, much like what happens when people use a heat gun on a laptop gpu, it might work for a couple of weeks, then fails. Theres more information , but many people have always wanted to know the ultimate question, and that is: How to fix the .8v dead die issue, and there is the answer. I have had a custom made stencil for the titan along with the professional grade reflow machine. There are still many boards I need to repair and reflow before offering it to the public.

Vegas

Simpler. Just use a 24 gauge wire-wrap wire and run it under the connectors. Works fine, and solves the problem without having to pull the connectors (because the same line always blows)

Just remember to tell the person to junk their Pi and get a new one because that's the root cause.

C

Excellent. So Im going to have this customer ship it to me, and if it is an easy as it looks, Ill try to offer repairs as cheap as possible. Im hoping around $100. If its just removing the connectors, and then sealing the old trace line, and then running new wire, and resoldering connectors, that doesnt seem too difficult. So , Ill have my customer ship me his and see how things go. If its as easy as I think, Ill offer it to you guys for a good price.

Vegas

Yes, you can.


https://bitcointalksearch.org/topic/m.14141421

Ok so even though ive been very quiet here, I have continued my repairs and research on these cubes and controllers. Ive purchased a very high end reball/reflow machine and had a custom stencil made for it specifically for this board. I also have a digital microsope, and the hakko  tweezers, and desoldering gun to remove the pci-e plug, as well as the the best capacitor in circuit test, "the capacitor wizard". So it will just be a few weeks to repair more cubes to full health before I open repairs again. When I do, there will be no cube I cannot fix. But before that I have loads of cubes I still need to repair to full health to be sure, before I open repairs to the public. Anyway, I have done no work at all on the controllers. I know the bridge has a common failure with the traces burning. I have never actually dealt with one before. I had a customer send me pics and asked If I could fix it, and I really could not answer. I know I can easily desolder the female connectors, but then can I run a new wire to connect them, or does the copper trace have to be connected to the board?? Because if the trace line must be connected, that could be a real problem, but I might have some solutions. Anyone know if I can run a new wire not connected to the board, just connecting the female connectors?

Here are his pics.




Vegas

ok.. I talked to a tech at digi-key in live chat...he gave me this link that says it should work (larger version) of the small pad chip we are talking about



http://www.digikey.com/product-detail/en/txc-corporation/9B-24.576MEEJ-B/887-1252-ND/2207672


assuming such does work would it not be easier to just get a protoboard..and trace the leads on the bridge and hack something together like below
with this soldered into the proper lead correctly on this spaghetti hack?



the digikey hack said the small chip replacement is a pain..high heat.....thus maybe just dumb and stupid above would work better .....

this is pretty much the extent of my electronics skills guys....having this idea dumb or not and the bigger cousin of he small chip involved

want to be a 'God' on knc Swedish ASIC miner bitcointalk here?..... hack one of these puppies together and I'll name my cat after you

Let's hope that they will work 

I ordered (2) 24.576MHz from digi-key -- it looks like the one from mouser is physically a little too big

Go ahead and buy a new oscillator
http://www.digikey.com/product-detail/en/abracon-llc/ABM3B-24.576MHZ-B2-T/535-9127-1-ND/675644
http://www.mouser.ee/ProductDetail/ABRACON/ASV-24576MHZ-EJ-T/?qs=GgB7MJ%2fTIsRYewXBpNiXkA%3d%3d

You already know what to do

We can simply hire a lawyer and get at least our money back!
qberty is unbelievable

Let us here know how it goes...I'd at least like to know what to tell an electronics guy around here (or the EE dept at the local univ) for clarity.....I'm gonna have enough problems
I think explaining the crypto mining of scrypt it is used for without adding to the fact I've no idea what/how/etc something like this is fixed....

I did test mine according to how qberty wanted me to ..with a pi pc power adapter to be safe...nada...so then regular test with working psu on a titan etc also nada...

hopefully he is confering with the china clone maker that made the boards for some kinda fix or redress (a guy can hope)

drama crypto always drama

Well, it was kinda a rhetorical question, I'm not counting on Qberty to do anything.  After thrashing Tupsu  & KNCFUN for pointing out some of his little lies,  now he's boxed himself in with some bigger lies ie, (a) he tested them before shipping (b) complainers just want free boards (c) we don't know how to plug them in (d) someone asked "nicely" so he'll make a video (where's the video?) If he'd offer some helpful info on what's wrong & how to fix them, nobody's going to  care much about the other stuff --  then maybe try a batch 3 when the bugs are worked out

I haven't taken mine to an electronics shop yet, maybe next week
 

https://www.facebook.com/qberty

https://twitter.com/qberty1337

https://soundcloud.com/qberty

https://ca.linkedin.com/in/stephen-hardeo-94394842

I doubt he is gonna do a batch 3 unless we get an angle on how to get the batch 2's to work....with me (2 I think?) and the two friends who I told who each got one (2) =4 that is 0.8 BTC BTC gone or at todays rate of 600 bucks that is $480 bucks...I just don't get it thou..his first batch stuff worked just dandy...I'm still running one...


Well it is all down the tubes unless a fix/hack/replace is found...none of them seem to work....

electronic solder skills and or guesses on what is wrong are beyond my 'newb electronics'  skill set ....(damn that chip is small) just saying

You have an angle on what you are gonna try to fix on yours or are you sitting on hands too waiting for qberty to reply and or tell us a probably of a hack/return or whatever..

Want to keep them...need the dang 3 controllers so it is not about the $$$ ..that is a last resort for me....but needed these for some Titan Unicorn Boards

I imagine any old off the shelf surface mount 24.576 MHz crystal oscillator will work, but I dunno for sure. The soldering on mine looks good to me, so not too sure it's a cold solder issue but can't really see the backside of it to tell
 
qberty when are you going to do batch 3?

You first Also where the heck does one get the proper crystal....or did you find it was a cold solder issue?