Xian01
Legendary
Offline
Activity: 1652
Merit: 1067
Christian Antkow
|
|
July 22, 2013, 01:41:45 PM |
|
Wow! Thats a lot of solder balls per chip. Is this amount of connections normally soldered reliably in manufacturing?
Yeah, that density is nucking futs. Raises a red flag.
|
|
|
|
Ytterbium
|
|
July 22, 2013, 01:45:06 PM |
|
Changing the package doesn't mean they changed the chip itself. If you look at the pin-outs on major chips you'll see a lot of them are redundant, extra power lines and stuff like that. I think they use a lot of lines because the connecting lines are so thin - I would imagine putting a high load on an individual line would damage the chip, so they have a bunch of lines to spread out the power load.
The package they're using has space for extra pins, that doesn't mean they're actually going to be connected to anything, or aren't just going to be used as extra power lines.
What? If they change pad-balls they change chip, or have had a design error. Prefer to be a chip change that an error. Not to sound a troll, I am buyer KNC. Who says they changed the number of pads? The balls might not be connected to anything at all. I'm not a semiconductor engineer - but the number of balls isn't necessarily 1-1 with the number of pads on the IC. (oh, and there are cool videos of wire bonding on youtube, you can see how tiny the connecting wires are, they couldn't individually carry a lot of current so you need a lot of lines to carry 250 watts) If you have a 55^2mm package, it may be simpler just to cover the bottom of the package with balls rather then have a custom stencil made or whatever. Or, finally it may simply be that the diagram is incorrect and there will be missing balls on the actual chip. Once you get your miner you can take it apart, heat up the chip and remove it and count the balls if you want. Basically I wouldn't worry that much about the discrepancy, although you could ask them for input since we're all curious. Wow! Thats a lot of solder balls per chip. Is this amount of connections normally soldered reliably in manufacturing?
The way it works, they just set the chip on the board, and heat it up enough for the balls to melt. If you search "bga reball" on youtube there are videos of people taking the chips off and putting them back on at home. Once you put the balls back on the chip, you just have to set it on the board and heat it up for the balls to melt, IIRC.
|
|
|
|
Loredo
|
|
July 22, 2013, 02:08:11 PM |
|
A KnC statement like "Tape-out was executed at 22nd July." would clarify all these speculations. No, it wouldn't. It would likely spawn its own thread, where everyone and their cousin would pontificate over or naively weigh in on whether or not there was enough time to ... blah, blah, blah. Just look at the controversy over each of the bits of information they provided here. The best thing for them to do is what every design/development company routinely does: keep quiet. If you think they're going down, then you should take your marbles and go play somewhere else, expressing your opinion if you want to (opinions are valuable to others, even when their are, in consensus, wrong) If you think they're going to make their deadlines, then chew your fingernails and wait.
|
|
|
|
pixl8tr
|
|
July 22, 2013, 02:13:17 PM |
|
Wow! Thats a lot of solder balls per chip. Is this amount of connections normally soldered reliably in manufacturing?
The way it works, they just set the chip on the board, and heat it up enough for the balls to melt. If you search "bga reball" on youtube there are videos of people taking the chips off and putting them back on at home. Once you put the balls back on the chip, you just have to set it on the board and heat it up for the balls to melt, IIRC. Thanks! Yeah, I am familiar with the process. I have used hot air to reflow some small packages for repair jobs. However, this size BGA is magnitudes away from anything I've experienced. It just seems that there will be have to be either automated or manual x-ray checking of the soldered packages somewhere in the process. Just asking the question from those who know, how difficult is it to solder a BGA this large in a production line manufacturing process.
|
who | grep -i blonde | date; cd ~; unzip; touch; finger; bjobs; uptime; strip;. grab; mount; yes; umount; sleep; brun; Donations: 18ByQvDUmaMKkQbYvUWmnPSu9BWeNxVMoc
|
|
|
titomane
|
|
July 22, 2013, 02:42:55 PM |
|
A KnC statement like "Tape-out was executed at 22nd July." would clarify all these speculations. No, it wouldn't. It would likely spawn its own thread, where everyone and their cousin would pontificate over or naively weigh in on whether or not there was enough time to ... blah, blah, blah. Just look at the controversy over each of the bits of information they provided here. The best thing for them to do is what every design/development company routinely does: keep quiet. If you think they're going down, then you should take your marbles and go play somewhere else, expressing your opinion if you want to (opinions are valuable to others, even when their are, in consensus, wrong) If you think they're going to make their deadlines, then chew your fingernails and wait. It seems that we have to shut up and worship KNC. You believe blindly, others hesitate to see. Both views respectable. You must not say. "likely spawn its own thread, where everyone and Their cousin would pontificate over naively or weigh in on whether or not there was enough time to ... blah, blah, blah." Because this is the thread of KNC.
|
|
|
|
Loredo
|
|
July 22, 2013, 02:53:41 PM |
|
It seems that we have to shut up and worship KNC. You believe blindly, others hesitate to see. Both views respectable.
I don't believe blindly, and you or anyone else need not worship KnC. That was not my point. My point is that these threads should not be - but do - become akin to sports talk radio, or AM politics talk. Among the few intelligent, experienced comments are many, many, more from naive, tribal, fans of this "team," or its competitor teams, which lower the signal-noise ratio with post after post of fatuous speculation and bullshit. Sports radio is fun to listen to, sometimes, when caught in traffic, just it was fun to read the debate about the technical details, But there's no money at stake there. Here, there is. That's a big difference.
|
|
|
|
minternj
|
|
July 22, 2013, 02:58:52 PM |
|
But there's no money at stake there. Here, there is. That's a big difference.
which is why you should be thankful that people pay attention to details and notice a specification change from previously disclosed figures. Why is that considered controversial to ask why technical specs have changed after it was supposedly finalized and sent off to the foundry?
|
|
|
|
Ytterbium
|
|
July 22, 2013, 03:07:54 PM |
|
which is why you should be thankful that people pay attention to details and notice a specification change from previously disclosed figures. Why is that considered controversial to ask why technical specs have changed after it was supposedly finalized and sent off to the foundry?
I don't think there's anything wrong with it - we funded their project, and we should be treated like investors (IMO). However, there is a difference between something being concern, and needlessly panicking and freaking out for no reason. I don't think the difference between the stated number of contacts and the diagram is that big of a deal, or a reason to worry about their ability to deliver on spec. Anyway, someone can email and ask, they've been responsive in the past, and it's Monday (Well, past working hours in Sweden). It seems strange to sit around speculating when you could easily just drop them a line.
|
|
|
|
jmaccoin
Newbie
Offline
Activity: 43
Merit: 0
|
|
July 22, 2013, 03:10:28 PM |
|
Packaging usually comes after the chips are made, so I hear....
Wouldn't they only package the devices once they are ready to be shipped?
|
|
|
|
minternj
|
|
July 22, 2013, 03:13:04 PM |
|
Packaging comes after yes but usually designed with the chip as noted by others.
|
|
|
|
pixl8tr
|
|
July 22, 2013, 03:13:34 PM |
|
Packaging usually comes after the chips are made, so I hear....
Wouldn't they only package the devices once they are ready to be shipped? He is speaking to the packaging of the the silicon die into a "chip". That can then be integrated onto a circuit board.
|
who | grep -i blonde | date; cd ~; unzip; touch; finger; bjobs; uptime; strip;. grab; mount; yes; umount; sleep; brun; Donations: 18ByQvDUmaMKkQbYvUWmnPSu9BWeNxVMoc
|
|
|
minternj
|
|
July 22, 2013, 03:16:24 PM |
|
which is why you should be thankful that people pay attention to details and notice a specification change from previously disclosed figures. Why is that considered controversial to ask why technical specs have changed after it was supposedly finalized and sent off to the foundry?
I don't think there's anything wrong with it - we funded their project, and we should be treated like investors (IMO). However, there is a difference between something being concern, and needlessly panicking and freaking out for no reason. I don't think the difference between the stated number of contacts and the diagram is that big of a deal, or a reason to worry about their ability to deliver on spec. Anyway, someone can email and ask, they've been responsive in the past, and it's Monday (Well, past working hours in Sweden). It seems strange to sit around speculating when you could easily just drop them a line. Whos freaking out. This is a discussion board, on a thread that was started by knc . They obviosuly read it since they released the tech specs in response to this thread. A question is being asked. Why not let it be answered instead of being a white knight for knc?
|
|
|
|
titomane
|
|
July 22, 2013, 03:18:59 PM |
|
It seems that we have to shut up and worship KNC. You believe blindly, others hesitate to see. Both views respectable.
I don't believe blindly, and you or anyone else need not worship KnC. That was not my point. My point is that these threads should not be - but do - become akin to sports talk radio, or AM politics talk. Among the few intelligent, experienced comments are many, many, more from naive, tribal, fans of this "team," or its competitor teams, which lower the signal-noise ratio with post after post of fatuous speculation and bullshit. Sports radio is fun to listen to, sometimes, when caught in traffic, just it was fun to read the debate about the technical details, But there's no money at stake there. Here, there is. That's a big difference. You contradict yourself. You do not want to become the sport in a discussion thread. But answer this: No, it would not. It would likely spawn its own thread, where everyone and Their cousin would pontificate over naively or weigh in on whether or not there was enough time to ... blah, blah, blah. A user who has not said any nonsense. If set to your height and you just answered. This really will become a sports discussion.
|
|
|
|
GandalfG
Sr. Member
Offline
Activity: 259
Merit: 250
Dig your freedom
|
|
July 22, 2013, 03:50:32 PM |
|
Packaging usually comes after the chips are made, so I hear....
Wouldn't they only package the devices once they are ready to be shipped? He is speaking to the packaging of the the silicon die into a "chip". That can then be integrated onto a circuit board. Lol, near 3k BGA on PCB :-)
|
Want to say thanks? 16ragydppe9QFRVhrdwEUjgfMS7KCfEFGY
|
|
|
OmegaNemesis28
|
|
July 22, 2013, 04:14:25 PM |
|
Packaging comes after yes but usually designed with the chip as noted by others.
Which is my point, this process is at the end and they can change it if needed especially something so trivial as more balls. Zero worry, only FUD. This. My money is in. They said September, October. I will wait patiently until then. When those months roll around, I will ask questions if I need to.
|
|
|
|
Mota
Legendary
Offline
Activity: 804
Merit: 1002
|
|
July 22, 2013, 04:54:56 PM |
|
+1
|
|
|
|
ASIC-K
Sr. Member
Offline
Activity: 280
Merit: 250
Hell?
|
|
July 22, 2013, 04:58:29 PM |
|
+1
|
|
|
|
rizzman
|
|
July 22, 2013, 05:48:30 PM |
|
Well while they are still engineering this thing it would be nice if they ditched the cube and lined all 4 units up so that the exhaust from the first set of chips isn't used to cool the second set. Even with their spacing offsets it will not be ideal. A linear arrangement of units would be much cooler.
Give them a chance, I would bet good money they are probably going design airflow shaping intakes to separate airflow between units. Probably something like what you would see in a Rack Mount Server
|
|
|
|
Bitcoinorama
|
|
July 22, 2013, 06:11:12 PM |
|
I'll worry more about an issue if, as, and when it appears.
In the mean time this is all hot air and speculation.
They are experienced ASIC and FPGA designers, this is evident. They have a strong reputation, are passionate and competitive, this much is a known. They have worked on much more complex hardware before. This much is provable. I think these guys know what they're doing and a few sceptical forum posts aren't about to give them any 'oh shit why didn't we think of that?!' moments.
Call me a fanboy of you like, but there's a lot of sudden interest with knowledge within chip design from lurkers and new members whilst competitors blatantly sit on the sidelines. Some probably have a good idea what KnC have in mind. Some may not. I think some are trying their hardest to kick up a fuss.
For all intents and purposes those that paid with credit cards and know they are protected should just chill for the next month and enjoy the sun, whilst keeping an eye out for updates as and when they become apparent. That's all that matters...
|
Make my day! Say thanks if you found me helpful BTC Address ---> 1487ThaKjezGA6SiE8fvGcxbgJJu6XWtZp
|
|
|
Kuroth
|
|
July 22, 2013, 06:14:14 PM |
|
Is Josh back in this Thread Trolling again??? Come on Josh, you should spend your time trying to up your BFL shipments from 5 units a day to 10..
Go away Josh..
|
|
|
|
|