Holy YouTube clips, Batman! We have tons of clips this week. Chris and our special guest Jeff Keyzer were talking about lots of stuff and it turns out many of them were video based. We can’t wait to see many of Dave’s videos, he’s still traveling around and is still working on editing and uploading many of them. On to the links!
Podcast: Play in new window | Download
Subscribe: Apple Podcasts | RSS
- Shoutouts
- Chiphacker.com has become Electronics Stack Exchange. It’s a really awesome format and a great community so far!
- Discussion Points
-
-
- What would it take to make open hardware chips?
- Jeri Ellsworth has done it! In her garage no less!
- One of our listeners Richard thinks it could be done with laser printers.
- Chip fabs are better places to do that, just like the one Jeff visited (well, not during the taping of this music video):
-
-
- Fabs use some nasty chemicals (and some that aren’t allowed under ROHS) like SF6. But SF6 is fun too!
- Jeff has used the services of James (@laen) who has a great PCB bundling service.
-
- Rants
- No parts? Still? AVRs are hard to find and it hurts small businesses!
- Dumb analysts are sad FPGA lead times are getting better.
- Chris hates it when there aren’t standard pinouts!
Thanks (again) to Jeff from Mighty Ohm for sitting in on The Amp Hour! It was really fun talking to him again and we appreciate it!
Some good ideas in this episode. As far as the printer stuff goes its still percolating. People have done small scale stuff with normal laser printers before. Like the research scientist who was building her own microfluidic devices using laser printed shrink e dinks. http://biopoets.berkeley.edu/davidb/publications/shrinkychips.pdf
The actual masks for various processes are actually printed out to begin with. Mask printers and typesetters get even better resolution than laser printers and perhaps have similar optical layouts. Even the humble dvd player has a laser spot size and precision hovering below 10 microns.
I don’t think things need to go exactly that direction though. Lots of progress is being made on alternative transistor materials. Amorphous silicon and CNT/Graphine have been used with inkjet methods to get pseudo ICs as well at high frequencies. Perhaps that is worth exploring as well.
As far as applications, well the first I would shoot for is chips involving instrumentation. Oscilloscopes and precision meters, various rf analyzers. Much of that would benefit if it could be made on something a bit more integrated. Or using some of the above techniques to make active matrix electroluminescent panels using some of Jerri’s EL techniques.
I think the day of the the micro/nano hobbyist is coming.
The word i’ve heard from Arrow is the silicon shortage was fab problems largely born of the GFC and that it’s industry wide. Fab ramp down and ramp up appears to be like turning an aircraft carrier. People talk about 12 months being a lifetime in technology advancement but that sounds funny when your’e given 6 month lead times. One of our board loaders was recently given a 2013 date for availability. This goes on and still my favourite vendors are spruiking their shiny new products that i know i won’t be able to get my hands on until months from now. Ever watch a product demo and get jealous of the presenter becasue they’ve got one and you can’t get it?
If your vendor sends you a fab change notice for a product you always use, assume it will affect supply and ask them about it.
It’s difficult to pivot a design towards alternative parts where microcontrollers and peripherals and their pinouts are almost always unique with no direct replacement.
If you think a vendor is considering you as a small time player and/or you’re getting no play because you’ve thus far been polite, shake the tree and see if anything falls out. If you’ve got a contract with a date where liquidated damages come in to play, it’s more cost effective to encourage your vendor to help you out than redesign. Even if you just get small quantities from Mouser, Digikey, Farnell or wherever – somehow let them know you’re suffering, because they certainly aren’t going to do anything if you don’t – squeeky wheels and oil. You dont have to be a complete ratbag about it. And make sure you thank them if something comes of it.
—
prediction: fabs completely back to rocking and rolling by June 2012
—
Karl von Moller was talking about shooting for his doco at Australias only? chip fab -Silanna CMS. I checked out the website and was surprised to see pricing. USD$25,000 for the mask and 1 wafer, then USD$8000 for subsequent wafers. I wonder how this compares elsewhere. It’s expensive but it has that feeling of just in reach if your design was worth it and you had or could raise the money to get it made.
The more I hear and see about that documentary, the more excited I get. Will have to look into that fab. Its not a super low price but its lower than I expected.
Ah…MOSIS. Brings back memories of grad school. Many of my grad classmates had their chips go through MOSIS although my own grad chip went through IMEC in Belgium. And it’s not only super big guys, i.e. Intel, that are working at 30nm. I’m doing 28nm design right now and as Jeff says, these small devices are a real pain in the butt to design analog circuits from. The absolute max gate voltage on these devices is less than 0.95V with the nominal being 0.9V and the min our circuits must work to is 0.8V. On top of that, the number of layout restrictions have just shot up exponentially because at that small of a scale, many additional physical phenomena come into play which are not well controlled. So what you could do in the last node is no longer allowed in this node. 🙁
Dang, that’s like having today learn the process all over again. Yuck.
The address you gave for the PCB bundling service is wrong. It should be pcb.laen.org and not .com
Fixed, thanks!
Jeff is back, he needs to be a regular!
Zyvek,
I think you are my biggest fan! Thanks!
USB AVRs are not available, because someone used them in a circuit to hack the PlayStation 3. Within days the already rare USB AVRs were completely gone from the market. Same for every dev or eval board with an USB AVR.
In general, Atmel is in extra trouble, because they have sold and closed fabs as part of their strategy to go fabless. And they fired a lot of people. They now have less design and manufacturing capacities and depend on getting fabrication slots at foundries. That also means if they miss a fabrication slot, because they didn’t get a design ready in time, it is “bad luck, see you at your next slot”.
They changed the contracts with their distributors, including procedures how distributors should keep stock.
They also had massive problems to get their Xmega line into a working state (still lots of Errata). And they are in the middle of a general die shrink. The parts with the “A” suffix are the replacement parts with the shrinked die.
And while they have all these, mostly self-inflicted, problems at hand, the market goes on a roller coaster ride.
Surface mount is here, so deal with it! 😀
Hint: Get several pairs of reading glasses ranging from 1.x to 3x (or more) and you’ll be able to easily work with surface mount.
You can buy lead free solder at Radio Shack
Hint: Get a temperature controlled solder iron (reasonably priced at Sparkfun) with a variety of tips and lots of solder wick. You’ll be soldering lead free in no time.
Remember Technology is always moving forward. Jump on and enjoy the ride.
I’m interested to see if the lead free tin whisker prophecies come true