Here is the latest from Intel. Wondering if this will have any implications or future for audio? Computer industry for sure will benefit, but what makes me worried is that down the line it will be harder and harder to have anything but SMD devices.
I had to break this file in two pages so that I could upload, hope it goes through.
I had to break this file in two pages so that I could upload, hope it goes through.
Haven't fully read either of them (thread/pdf), but it reminds me of http://www.diyaudio.com/forums/lounge/172197-end-diy.html
it's normal that Jaccovity is twisting everything in boobz direction
ok ...... it's also normal that I twist everything in toobz direction
don't be mad at us
No, no... not at all, I really replied to Rodeodave explaining the purpose of my post and not intention of going the path he quoted. While I was typing all your posts came down. It would not make sense for the possible discussion to replicate what is already said in the quoted thread, for what I appreciate he pointed out.
Yes it's nano-tech...which they say will fit 6 million of these transistors in the size of a "period" like this "." at the end of our sentences. So how about a "Monster with 1000x Periods"? That would be 6 Billion transistors!
Intel Reinvents Transistors Using New 3-D Structure
Intel Reinvents Transistors Using New 3-D Structure
Only difference between yuuz and meez is 1 letter (and a hefty discussion who will wear the straight-jacket).
3D instead of planar for small signal trannies.(soldering smt is easier than i thought)
SMD in the reasonable size is OK with the right equipment. But some sizes.... oh man, 0204... When I got my order first I wanted to complain to DigyKey they sent me an empty reel... hahaha. I tried and I tried and absolutely was impossible to do anything by hand on that size. I actually enjoy doing SMD, but small caps I needed to solder in 0204 size, I am telling you no one couldn't do it by hand.
Vlad,
This is really interesting and good work from Intel but it has very little to do with the question of whether or not we can get non-SMD devices. Intel's process technology is used almost exclusively for the microprocessors that go into all of our computers and laptops. Along with the rest of the semiconductor industry, Intel improves performance at a given cost/power by shrinking transistors to ever smaller dimensions. For the last few years, Intel has introduced a new technology node every 2 years. They introduced 45nm transistors in 2007, 32nm transistors in 2009, and now in 2011 they are previewing their 22nm transistor technology.
But getting back to the SMD discussion, Intel long ago went away from chips packaged with leads to SMD parts. This happened roughly in 1998 or 1999 with the Celeron class of processors. So this new technology isn't an indication that leaded parts are going to disappear. That already happened for Intel more than 10 years ago.
---Gary
Hi GaryB,
Really nice to hear from you. Hope to see you again this October.
Thank you for the detailed explanation. Quite amazing and mind boggling what could be done.
What pulled my attention, and thought might be interesting to share, is the statement in the article that this 3D technology is major and first redesign of transistor in 50 years. Having that in mind, I was wondering if there would be any performance major breakthrough related to audio... maybe? Hypothetically thinking, and obviously not knowing the answer, I wondered, more as a afterthought... if there is a dramatical improvement for the audio as the result of new design, if we would be able to even explore it since, as you also explained, Intel's production is certainly geared toward sizes measured in nanometers only. It is pretty bombastic statement - "first major redesign in 50 years", what made me thinking...
Thanks for the nice welcome. I guess I'm not posting often enough if it's a notable event when my name pops up 
. Regarding Burning Amp - I wouldn't miss it for the world. The only thing that would keep me away would be last minute business travel, which can be unpredictable.
Getting back to the Intel announcement, I think Blue mentioned that Intel's transistors are meant for digital logic and memory. The things that make a good digital transistor are good on-off ratio and high drive current at low voltages, typically 1v or lower. The things that we care about for audio, like linearity and being able to swing relatively large voltages (>> 1v), don't matter to Intel. This has caused transistor design for digital devices to diverge from transistor design for analog. So Intel's breakthrough for digital transistors isn't that relevant for analog. They're solving a different problem.
---Gary
. Regarding Burning Amp - I wouldn't miss it for the world. The only thing that would keep me away would be last minute business travel, which can be unpredictable. Getting back to the Intel announcement, I think Blue mentioned that Intel's transistors are meant for digital logic and memory. The things that make a good digital transistor are good on-off ratio and high drive current at low voltages, typically 1v or lower. The things that we care about for audio, like linearity and being able to swing relatively large voltages (>> 1v), don't matter to Intel. This has caused transistor design for digital devices to diverge from transistor design for analog. So Intel's breakthrough for digital transistors isn't that relevant for analog. They're solving a different problem.
---Gary
The current packaging trend for Bottom termination components (qfn etc, ref IPC-7093) is going to affect DIY and home assembly more than any other SMD packaging. These devices are not only small, 0.5mm and smaller pitches, but also have a thermal pad on the bottom of the device, that requires adequate soldering with minimal voids for the device to fuction correctly (you can soon overheat the silicon if this pad is not soldered properly).
These devices have numerous advantages, small, minimal package parasitics for good signal integrity, easy to manufacture, cheep as chips! etc etc.
What they aren't is easy to assemble, even with full SMD lines, reflow ovens, complete control over the solder paste screen (a critical part) X-ray inspection you get problems, so for the home assembler they are going to be a challenge.
The other problem where digital designs are concerned is the mandatory requirement for a proper PCB stack up (ground planes) to handle the ever increasing signal rise times, speeds etc to again achieve the required signal integrity and EMC compatability. The increased pin count density and multiple layer requirements is minimising the number of free design packages that can accomadate these designs.
For simpler analogue designs I dont think its that bad, as we tend to use larger SMD packages due to the power requirements, so for these designs hand soldering is still possible, and the simpler layout requirements maximise the available free design programs available.
FYI some notes on soldering.
http://www.circuitsassembly.com/cms/component/content/article/159/11036-tech-tips
These devices have numerous advantages, small, minimal package parasitics for good signal integrity, easy to manufacture, cheep as chips! etc etc.
What they aren't is easy to assemble, even with full SMD lines, reflow ovens, complete control over the solder paste screen (a critical part) X-ray inspection you get problems, so for the home assembler they are going to be a challenge.
The other problem where digital designs are concerned is the mandatory requirement for a proper PCB stack up (ground planes) to handle the ever increasing signal rise times, speeds etc to again achieve the required signal integrity and EMC compatability. The increased pin count density and multiple layer requirements is minimising the number of free design packages that can accomadate these designs.
For simpler analogue designs I dont think its that bad, as we tend to use larger SMD packages due to the power requirements, so for these designs hand soldering is still possible, and the simpler layout requirements maximise the available free design programs available.
FYI some notes on soldering.
http://www.circuitsassembly.com/cms/component/content/article/159/11036-tech-tips
Further Interesting Components! Metric 0402 chip packages.
http://kurtwhitlock.com/techtalk/BorkesGroves01005.pdf
http://www.smta.org/files/AZ_2008_Process_Characterization_01005_Component_Package.pdf
http://kurtwhitlock.com/techtalk/BorkesGroves01005.pdf
http://www.smta.org/files/AZ_2008_Process_Characterization_01005_Component_Package.pdf
hi Marce,
you ever tried to solder QFN? with my little reflow oven (€20,- toaster oven) it works perfect every time! much easier than any other ic package. when the solder flows the QFN package pushes out all the extra solder without making shorts, where with tsop all extra solder creeps up between the legs making shorts.
you ever tried to solder QFN? with my little reflow oven (€20,- toaster oven) it works perfect every time! much easier than any other ic package. when the solder flows the QFN package pushes out all the extra solder without making shorts, where with tsop all extra solder creeps up between the legs making shorts.
Last edited:
Hi
yes the assembly department solders hundreds every day, they are also x-rayed so we can control void formation. I would reccomend not having the solder squeeze out, it could cause shorts, the reccomended ratio for the thermal pad is 50%-80% taking into account thermal vias etc. i am glad you are having sucsess with them, but it still does not mitigate for DIYers the move to bottom terminated components is problematic as you cannot be a hundred percent of your coverage and whether voids are being formed without x-ray equipement. I am not saying it is impossible, but the trends in SMD components is making it harder for DIYers.
I do boards and assemblies for military communication systems so have to work to the highest standards, and even with full SMD lines we have had problems with these devices durring assembly. The design criteria for PCBs is also more critical with most devices requireing thermal vias and some form of copper plane to remove heat from the exposed thermal pad.
Using a home toaster/oven is not impossible and the first reflow oven I used back in the eighties was basicly a toaster with a controler on the front.
if you would like some information on reflow profiles and their effect on QFN soldering and other documentation regarding these devices PM me.
yes the assembly department solders hundreds every day, they are also x-rayed so we can control void formation. I would reccomend not having the solder squeeze out, it could cause shorts, the reccomended ratio for the thermal pad is 50%-80% taking into account thermal vias etc. i am glad you are having sucsess with them, but it still does not mitigate for DIYers the move to bottom terminated components is problematic as you cannot be a hundred percent of your coverage and whether voids are being formed without x-ray equipement. I am not saying it is impossible, but the trends in SMD components is making it harder for DIYers.
I do boards and assemblies for military communication systems so have to work to the highest standards, and even with full SMD lines we have had problems with these devices durring assembly. The design criteria for PCBs is also more critical with most devices requireing thermal vias and some form of copper plane to remove heat from the exposed thermal pad.
Using a home toaster/oven is not impossible and the first reflow oven I used back in the eighties was basicly a toaster with a controler on the front.
if you would like some information on reflow profiles and their effect on QFN soldering and other documentation regarding these devices PM me.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.