Scan all your books for a dollar each

I just used 1DollarScan in San Jose to scan a bunch of old yearbooks, text books, and manuals from General Motors I have kept for 35 years. There is no “catch,” it really is a dollar a book, but there are some conditions. To be fair, their definition of a book is 100 pages. So if you have a 101 page book, then its two dollars. If it’s a 199 page book, its still two dollars. A 1001 page book is 11 dollars. It is still the best deal you will ever see. Heck, the Post Office makes more than 1DollarScan if you have to ship the books.

Hiroshi Nakano at 1DollarScan will make high-quality pdf files or jpegs of all your books.

You can use media mail, however, and get a really cheap rate. I like the flat-rate Post Office boxes, and there is always UPS and FedEx ground. If you have a pallet of books maybe freight is the cheapest way. For folks like me that live in Silicon Valley, you can just drop the books off.

Now, if the books are copyrighted, 1DollarScan intends to dispose of them after they scan them, so that there is no copyright issues. You just changed the form of the copyrighted material you already paid for, and the Supreme Court has decided that issue decades ago. If the material is your own, or something like a high-school yearbook, 1DollarScan can return the material, if you pay for the shipping.

I had two big bags of books for 1DollarScan.

Since they count 100 pages as a “set” and every book is at least one “set” it did not take long for me to see that I had over 67 sets. I just stopped counting, since they have a 100-dollar a month platinum deal, where several premium services are included free.

So I want to point out the downsides since you are engineers and analytical. But first, rest easy, because Hiroshi Nakano, the founder of 1DollarScan is also an engineer. He came to Silicon Valley working for a big corporation. After a few years, he returned to Japan. There he noticed similar scanning services growing in popularity, since space is at such a premium in Tokyo. So Nakano returned to Silicon Valley and started 1DollarScan. As you would expect from a fellow engineer, the pricing is rational, the website is clear and it works great, and everything seems too good to be true.

As to those downsides? Well, since he uses very light compression on the pdf files, they are huge. My General Motors Institute yearbook came in at 350MB. That was 242 pages of high-res and mostly images. Because I sprung for the 100-dollar a month deal, the file was named with the title of the book and I can use their “tune up” online service to make smaller pdfs suited for phones or tablets, Kindles or other devices. Alternatively I believe they will provide you with the raw jpeg files, and for an extra dollar, they can do 600dpi jpegs. For me, the pdfs just make more sense, and hundreds of separate jpeg files are too unwieldy to handle.

The only other downside is the OCR (optical character recognition) was not perfect. I had 1DollarScan scan in a big 1960s magazine from the Cleveland Plain Dealer called “Cleveland, a city grows to greatness.” The preface has a small biography of the two authors. The type was tiny and the magazine was 50 years old. The page image is perfect, and you can’t see the OCR errors until you highlight and cut-and-paste the text into a notepad or some other editor. Here is what the OCR produced:

George J. Barmann, coculhor of l{ris work, has been on lhe staff of ihe Plain Dealer since epfember, 1942. He came lo the paper {rom the lllinoir State Journal, in Springfield, where he had gone affergraduation from flre University of lllinoir, in !937. On the Plain Dealer, Barmann spent some time in writing about education. After thal, he did general asignmenf reporting. whish meanr covering almoct the whole range of stories that daily come acro3r the City De*. Barmann, in recenl years, has done chiefly feafurer for the Plain Decler, including a greal many inierview with headline peronaliiies and people of the fheafer. Also, he ha; writlen feature stories aboul lhe Civil War. He lraveled through the Deep South, from New Orleans to Charleston, S.C., and wrole a series of arlicle: on whaf Southerners were thinling in tfii: l00th anniver:ary of ilrat wer. A nalive of Chillicolhe, Ohio, Barmann al*ended Miami Univer:ity ai Orford, O. before lramfering to iournalirm ai lhe Universify of lllinois.

Other fonts came out much better, this was the worse OCR of anything that got scanned. But there is a solution to both the big file size and the OCR accuracy. Based on the advice of analog engineer Walt Jung, I had purchased a copy of ABBYY Finereader 11. I am pretty sure it was under 100 dollars. I used ABBYY to scan in all my loose papers and tax records. I find it far better than TextBridge and other OCR programs, which I also own. ABBYY will take in a pdf file, and re-recognize the text, and save it with much higher compression. So I ran the 1DollarScan pdf into ABBYY and made another pdf file. That file of a 64-page ledger-size book was 9MB instead of 120MB. Here is the ABBYY OCR result of the Cleveland book:

G e o r g e J. Barmann, co­author of this work, has been on the staff of the Plain Dealer since September, 1942. H e came to the paper from the Illinois State Journal, in Springfield, where he had gone after gradua­tion from the University of Illinois, in 1937. O n the Plain Dealer, Bar­mann spent some time in writing about education. A fte r that, he did general assignment reporting, which means covering almost the whole range of stories that d a ily come across the C it y Desk. Barmann, in recent years, has done chiefly features for the Plain Dealer, including a great many interviews with headline personalities and people of the theater. Also, he has written feature stories about the C iv il W a r . H e traveled through the Deep South, from New Orlea ns to Charleston, S.C., and wrote a series of articles on what Southerners were thinking in this 100th anniversary of that war. A native of Chillicothe, Ohio, Barmann attended M ia m i University at Oxford, O., before transferring to journalism at the University of Illinois.

You can see ABBYY was much more accurate, but its problem is that it peppers extraneous spaces in the text. If you searched Google for George Barmann, it would find the 1DollarScan pdf but not the ABBYY pdf. This is because it is trying to line up the highlighted hidden OCR to the image of the text on top of the OCR. Since the font is a bit funky, is hand-typeset and has kerning, ABBYY breaks up words when it adds needless spaces. Both OCR results were a bunch of separate lines that I concatenated above so they would fit this post. Where the ABBYY version has hyphens, those are correct, there were line breaks there. Oh, I know, I can take a screen shot of the pdf images, here:

The 120MB 1DollarScan screenshot has way less image compression, if you click on the image you can see the author’s eyes much more clearly that the image below.

The 1DollarScan 120MB pdf run through and re-recognized by ABBYY Finereader 11 is only 9MB, and the text quality is nearly as good. The ABBYY image quality suffers from the higher compression, so you should not erase the original 1DollarScan files.

You can see that the 9MB ABBYY is almost as good for text as the 120MB 1DollarScan pdf, but the image in the 1DollarScan pdf is clearly better. So for things like a yearbook, I definitely will keep the larger 1DollarScan file, and maybe make a ABBYY pdf out of that to send around or post online. I looked into the extraneous spaces in ABBYY and there seems to be no “cure”. I tried making a “tagged” pdf in ABBYY and it is just much bigger and even worse OCR.

OK, so you can see that 1DollarScan is the real deal, here is a photo montage.

Hiroshi Nakano examines the books dropped off for scanning at 1DollarScan.

One side of the warehouse at 1DollarScan is for books waiting to be scanned. Lead times are only a week or two.

Hiroshi Nakano uses this guillotine stack paper cutter to remove the bindings of your books.

Hiroshi Nakano shows the spine of a book that he has cut off with the guillotine stack paper cutter at 1DollarScan.

Here is the workstation where an employee at 1DollarScan feeds several scanners and once, while tending to paper jams and insuring you get the perfect scan.

I should mention that I asked Hiroshi Nakano if I should not show the heart of his operation above with the multiple scanners being fed by his employee. I told him that someone might see it and try to compete with him. Nakano smiled and said, “Nobody can compete with me.” I love the precision and factual nature of my fellow engineers, don’t you? Lets face it, a dollar to scan 100 pages with OCR is pretty remarkable.

Once the books are scanned and the pdfs are posted for your download, 1DollarScan holds your books for two weeks, in case there were any problems. After that, the copyrighted books are recycled, or un-copyrighted materials are sent back to you if you pay shipping.

Hiroshi Nakano from 1DollarScan patiently explained his operation to me and I was assured that it really is true that he can do high-quality scans of your books and magazines for a very low price.

There is real joy in being able to keep all my books in electronic form while dispensing with hundreds, maybe 1000 pounds of paper. Lets see if I can find a picture–

An engineers can collect a lot of paper. There were tax records for my business, project folders for jobs I worked on, letters from college girlfriends, owners manuals, and two big stacks of books you can barely see in the back right corner. The ammo boxes bottom right are full of pictures and negatives.

It took about six months, just scanning in all these loose papers. The hand-written letters from girlfriends I kept as 300dpi jpegs. Same for my hand-written printed notes, the OCR in ABBYY is pointless on handwriting anyway. For pictures, I scanned them at 600dpi, anything finer I could not see any difference on the 47-inch TV I use as a monitor. For negatives and slides I did 2400dpi, which is the same spatial resolution as doing the printed picture at 600dpi. I used ABBYY to make pdf files of any printed materials, including some booklets that I thought 1DollarScan might not want to do. But all the books, yearbooks, magazines, and manuals, well, it was just so nice to send those two big stacks to 1DollarScan and have it taken care of my by some diligent professionals. Lets face it, disk space is nearly free. I have a 2-Terabyte NAS (network-attached storage) at home that can hold all these files with room to spare.

For scanning all my loose documents, I had a Canon laser MF 4890dw on the right. For 11×17 and oversize, a Brother MFC-J6710DW inkjet on the left. For pictures, slides and negatives a Canon CanoScan 8800F back-lit flatbed in the middle.

A standard desk would hold all three of my scanners, The laptop was driving my two TVs, and a wireless keyboard and mouse did the control. Lots of paper towels and Windex to keep the platens clean. It was a monumental job but now it is done. I will keep all my receipts and records on the NAS now. I back it up onto the laptops, and to a SSD (solid-state drive) I keep in the safety deposit box at the bank. I have auto titles and my birth certificate as paper, everything else is virtual. It is heaven. I gave away the two printer-scanners but kept the flatbed so I can do receipts and such as they come in. No more shoe boxes full of receipts for me.

I made three runs to the Sunnyvale dump with the scanned paper. My pal said some companies will let you put your personal stuff in their shredder boxes. Either way, it is great to have all kinds of room, as long as I resist the temptation to fill it up with old test equipment or Sportster parts. And for getting rid of that stuff, you can use flea markets, Craigslist, and eBay.

Stock-engine-test_1969-Chevrolet-327-cu.in.-V-8-42p_ABBYY

Here is a link to a scan I had done by 1DollarScan and then ran through ABBYY Finereader. It went from 28MB to 2.7MB. The booklet a really cool engine test report I had from my student days at GMI. Since Mary Barra, the CEO of GM went to GMI too, hopefully she won’t sic a bunch of high-tone Detroit lawyers on me. I do note there is no copyright symbol on the document. Now the blocky shading on the title is due to the compression in ABBYY. The 1DollarScan document does not have those artifacts. But it is 10 times bigger. Note how the scan is straight and note how they scanned the both sides of the back cover, even when there was no text. You want the whole booklet scanned, they do it. Now they do charge 2 dollars for magazines, and I am not sure they counted this booklet as a magazine or thin book. In any event, that was why it was worth it to get the premium membership for a month. Hiroshi Nakano thought I was being fair and I thought the same about him. More than fair. Part of the premium service is they give the pdf file the name of the book. In this case they carefully typed: Stock engine test, 1969 chevrolet 327 cu.in.v-8, 42p.pdf. Since I am putting it on a web server, I changed the name to my convention: Stock-engine-test_1969-Chevrolet-327-cu.in.-V-8-42p_ABBYY.pdf. When pals talk trash about GM, I whip out this 45-year-old engine test. General Motors knows more about cars and good engineering than all the congressmen and lawyers put together. I still miss being an auto engineer.

Hardware security is the only real security

I just came across the epic hack that Wired‘s Matt Honan had perpetrated on him. A hacker added a credit card number to his Amazon account. The next day they called Amazon and said they lost the password. “What is the number of the credit card on the account?” asked the helpful Amazon employee. Once they were in the Amazon account they got into his Google accounts, all helpfully linked by Matt himself, and then the Apple accounts. The hacker was some sociopath kid. He was not interested in money; he just wanted to hurt someone, so he wiped out all the pictures and data on Honan’s phone, computer, and yes, the precious precious cloud. Yes, my precious, one cloud to rule them all.

Just like the Ring in The Lord of the Rings, the cloud can be your worst enemy in the hands of a bad person.

Now initially Honan lamented that he lost all the pictures of his new baby and a bunch of other stuff. The next article showed how he got it all back in a couple days. He says he believes in the cloud even more now. Beats me why he thinks that. If he had not inadvertently left his 1Password account password in his Dropbox on his wife’s computer it might have been much more difficult to recover control of his accounts.

As to all the wiped data, well it was lost forever on the precious cloud, but the nice folks at DriveSavers got his SSD (solid-state drive) in his mac mostly recovered at a cost of $1,690. So since the whole thing gave him half a dozen popular articles to write-up, you could argue getting hacked was the best thing that ever happened to his career. It reminds me of when King Louis XIV’s minister Colbert asked a bunch of writers “What can France do for you?” One shouted back—“Throw us in prison.” It would give them something to write about and the time and solitude needed to write it.

DriveSavers have a full cleanroom to save hacked, damaged, or corrupted hard drives. They can also do forensic hardware analysis on solid state drives (SSDs) as in Matt Honan’s case.

What astonishes me is that this hack happened to a technically astute denizen of San Francisco. Maybe he should move to Silicon Valley, we know a lot about security here and Atmel’s group in Colorado knows even more. Not only did Honan misplace his trust in online accounts and the precious cloud, he kept no secure data backup. He courageously accepts the blame, but also tries to deflect some blame onto Apple and Google. Sorry, your data is your responsibility. Apple and Google quickly closed the social-manipulation hacks the sociopath used, but it is not their job to accept responsibility for your data. That is your responsibility.

This is what we keep harping on here at Atmel. Security is a key pillar in the Internet to Things, and the best security, the only real security, is hardware security. You don’t want these malicious hackers changing your thermostat, or running up your electric bill, or stealing your security camera feeds. Atmel has inexpensive tiny chips you can use to secure these gizmos. Some of our chips use symmetrical authentication. The security chip is programmed with your secret key, and you know the secret key. The microcontroller, and it doesn’t have to be an Atmel microcontroller— it can be anyone’s, sends a random number to the Atmel security chip. The Atmel chip does a mathematical operation on the random number using the secret key, and sends that result back to the microcontroller. The host microcontroller has a local Atmel security chip to do the same mathematical operation on the same random number and then it compares the two results. If they don’t match, the code stops executing. That way no-one can put in bogus code and take over your gizmo. It gives you secure boot and secure downloads and upgrades. You can also use Atmel security chips to verify a battery or accessory is genuine and not some knock-off product.

Atmel’s CryptoAuthentication™ system uses hardware and extreme security to protect your system.

Now since the microcontroller is connected to the Atmel security chips by way of a common SPI port, you might fear a hacker could snoop on the communication and learn the random number sent to the Atmel chips or the mathematical result sent from it to the micro. That’s the beautiful part of this. The micro generates a new random number every time. If the host micro is too small and simple to generate a reliable random number, the tiny Atmel security chip has its own true random number generator (TRNG). So the micro can query the Atmel chip for the number, then query for the result, then do the same operation using the same secret key. So snooping on the serial port will only give you the last serial number and the result. You will have no idea of what the operation was that produced the result. Its like snooping and seeing the number 12 transmitted, but you still don’t know if that was based on 2 time 6 or 3 times 4. Now imagine that problem with numbers hundreds of bits long, and you can see how secure this makes your system.

This USB memory stick has a keypad to unlock it. You can store all your passwords or love letters on it and no one can get in without the code.

So it’s great to have services like 1Password, which is a browser extension combined with a remote server that generates and stores different passwords for all your needs. If, however, you need to use two computers, and who doesn’t, now you get to involve Dropbox so that you can store the master password there so you can get your 1Password even if you are at a Kinkos computer. Thing is, I just feel better with hardware security. In this case, it would be using a USB stick with hardware keypad or fingerprint sensor. Those are great since you don’t need a program on the computer of Surface Pro tablet to run it. You swipe your finger or type in a code and the stick unlocks and you can cut-and paste passwords as you need to. Thing is, there I worry about Windows saving some temporary file. I looked into this a few years ago, and sure enough, even a text file seemed to get cloned somewhere once you opened it off a stick. So the real hardware security is two-factor authentication like you get with an RSA dongle or a YubiKey. Once again, the essential element is a real physical piece of hardware that makes the system secure. I love the YubiKey since it emulates a keyboard, so unless someone infected your computer with a keylogger, there is no record that you used it. And, like the RSA SecurID, even if they do keylog it, the same code never works twice. They are just like that Atmel security chip and just as uncrackable.

The YubiKey is a two-factor authentication system accepted by more and more sites for login. The Nano model is as small as the USB contact pins. Pressing a little button on the device makes it send the one-time log-on code as though it was a USB keyboard.

IR reflow oven for your prototype PCBs

When you use solder paste to assemble your prototype PCB (printed circuit board) you need a stencil or hypodermic needle to apply the paste to the pads on the board. Then you use an IR (infra-red) reflow oven to melt the solder. Scott Fritz, an Atmel IC designer on the third floor, found this neat home-made controller that turns a cheap toaster oven into an IR reflow oven. I assume the name Reflowster is a combination of the words “reflow” and “toaster.”

The Reflowster will do closed-loop control of a cheap toaster over so you can do IR reflow soldering on your prototype circuit boards.

The Reflowster is an Arduino-based controller that that gives you predicable and repeatable heating and cooling profiles to melt the solder paste and connect up all the components on your board. They got their start on Kickstarter, and have actually shipped, so all the Kickstarter people are rewarded. Now the Reflowster folks are starting to offer the product to the general public.

The Reflow controller V3 PRO from PCB POOL in Europe is another product meant to work with a toaster oven.

I have mentioned a similar reflow controller made by the fine people at PCB-POOL in Europe. That article also described how my buddy Wayne Yamaguchi was using a toaster oven a decade ago to make his PCBs. Wayne did not use a controller. He just did a whole bunch of tests until he was satisfied he was getting good whetting and solder fillets on his circuit boards.

The great thing about the Reflowster is that it is a closed-loop controller. It is actually measures the temperature of the oven, and then controls the power to it so that the heating and cooling match the profiles recommended by component makers like Atmel (pdf).

Precise temperature control is needed to do quality lead-free soldering.

While I love, admire, and respect my buddy Wayne Yamaguch’s “theory of experiments” approach, you might really need the Reflowster. If you want to use different ovens, or have changing wall voltage, or the boards you are soldering are different sizes or have a different set of components on them, you want a closed-loop controller. If the chips have a big pad on the bottom, the die-attach-paddle, you need reflow. The other big factor is lead-free solder. Many of the crazy analog engineers I hang out with still use tin-lead solder for prototypes. It looks better, it feels better, and lead solder is more reliable. We also pull the solder off the reel by biting it gently and tugging, so we don’t have to set the soldering iron down. Lead poisoning might explain why we are all crazy. But if you are sane and insist on using lead-free solder, the preciseness of reflow control is important.

Solder paste application is like silkscreening T-shirts. Instead of silk the stencil is stainless steel. Instead of ink you use solder paste. Instead of T-shirts you do printed circuit boards.

Solder paste has its own hassles. You should refrigerate open containers so the little solder balls do not oxidize and change the reflow parameters. I am not sure the same caveat applies to when you use a big hypodermic to apply the solder paste. There the hassle is you have to do it one pad a time. A solder stencil is a thin stainless steel sheet where the PCB fab house has etched through all the areas where there is supposed to be solder. Sometimes called the “cream layer” Its not exactly the solder mask art, but it is pretty close, depending on your particular design. I know you can set up OrCAD 9.2 to do it, and I am sure other CAD packages can make it, or the PCB fab house can create one from your solder mask layer art.

Here is a typical solder stencil from Stencils Unlimited. With one swipe of a squeegee you apply solder paste to the pads for your chips and passive components.

The fine folks at Sunstone used to offer a free stencil, now it looks like they charge a little. Most fab houses can supply one. Assembly houses like Screaming Circuits or Advanced Assembly have the relationships with board houses to they can make your stencil when they assemble your boards. If you are really masochistic, and have a high-powered CO2 laser cutter handy, you can make your own stencils. You can also live in a cave and use flint tools, but I prefer to operate a little higher on the food chain. The LPKF laser mill can make your stencils as well.

This solder stencil from Proto-Advantage lets you apply solder paste for a QFN-32 chip.

There are also hybrid approaches. You can buy cheap solder stencils just for high-pin-count chips on your board. You squeegee the solder paste onto the board for each of those parts, Then you can use a hypodermic for the passive components or hand-solder them after you reflow the big chips.

If you are a big-time engineer on a big-budget project then just contact Screaming Circuits or Advanced Assembly or you local board assembly house (not PCB fab, but board assembly). I know Screaming Circuits can do it all since they have teamed up with Sunstone and Digi-Key. Just send Screaming Circuits the fab Gerber and fab files which they send to Sunstone, the assembly drawing and insert file which they use themselves, and the BOM (bill of material) they order the parts with from Digi-Key. They can do quick-turn and they can ship anywhere in the world.

So the prototyping ecosystem is like this:

• If you are a hobbyist use DIP (dual-inline plastic) chips with 0.1 inch lead spacing and through-hole passive components. You might use surface mount chips on DIP breakout boards.
• If you are a pro-hobbyist or low-budget engineer buy a temperature-controlled Weller soldering iron or a Metcal and a good stereo microscope. Now you can hand-solder surface mount boards. For chips with bottom pads you have to either heat the whole chip with a big soldering iron, use a heat gun, or try to wick the solder in from vias you design in on the backside of the board.
• If you are a hobbyist doing low-volume manufacturing or a medium-budget engineer, go to solder stencils and reflow ovens.
• If you are a hobbyist that hit it big or a big-budget engineer, use the board house to order the parts, get the PCB fab, and assemble and maybe even test your board.

TQM Solutions knows that total quality management means you not only have a mountain of documentation, but that you organize that mountain.

Note that last item. See, as an engineer, your real job is to make a set of documentation so the design can get manufactured by non-engineers and non-technicians and non-hobbyists. Its nice you are a hands-on person. Heck, its critical you are a hands-on person to be a good engineer. But your real responsibility is making sure the CAD files are correct. it might speed things up if you make a first-spin board yourself, and its neat if you make the board on an LPKF mill and you can get parts from the factory floor or Radio Shack or a salvage yard.

I used to design products with parts I found cheap at places like Weird Stuff Warehouse in Silicon Valley. Then one of my designs went to production but Weird Stuff had sold off all the parts I had used. Now I select parts from distributors.

The great thing about using Screaming Circuits and Sunstone and a distributor like Mouser or Arrow or Newark or DigiKey is that you are proving out your documentation. You make it clear to Screaming Circuits that if your pick-and-place insert file has mistakes they tell you, you fix them and they use that file. This way, when they see that the part origin for a DPAK is at the pad and not the part center, they know the vacuum picker cannot pick it up, so you catch that AND FIX IT. If your Gerbers have problems you make sure Sunstone tells you, or you use the free DFM (Design for manufacturing) check offered by Advanced Circuits. Then you FIX THE FILES. Same deal for any BOM mistakes. Make sure somebody tells you so you can FIX THE FILE, and not the text file, you fix the CAD file in OrCAD or Altium or whatever, so it spits out a perfect BOM.

Now when you send the CAD files to China to get assembled on the cheap, you know the files are correct. Anything less and you are not an engineer, you are an amateur. Proto Express even works with a Chinese partner to ensure you can get cheap-high volume boards that work as well at the Proto-Express boards made right here in Silicon Valley.

The oqo Model 2 used a Via processor. The third model with an Intel Atom never got built since they ran out of money (courtesy Engadget).

This level of diligence and exactness is critical. I worked at oqo, a San Francisco start-up that made the first palm-top computer that ran real Windows OS. The first model was based on the Transmeta “emulated” x86 processor. The second model used a Via chip. Lesson there is never base a business plan on being smarter than Intel. The third prototype never went into production. I had left the company for National Semiconductor, but pals there told me they used an Intel Atom processor and it was a real product that could really work good. But they were running out of money. So I assume in a big rush, they sent the design to the Chinese contract manufacturer. A pal familiar with the company told me oqo had to fly out an engineer to China and there were 1000 ECOs (engineering change orders) to get the design ready for high-volume manufacturing. One thousand mistakes. Now it probably didn’t matter, but its nice to think that if they had scrubbed the CAD files, the fab, the assembly, and the rework documentation through a US quick-turn prototype manufacturer, and fixed most of those mistakes, then maybe they could have gotten that product to market and saved the company.

This is a perfect example of the asymmetric respect problem in engineering. The Chinese manufacturing engineers respected those high-tone former Apple designers at oqo. But the oqo engineers may have thought manufacturing was some triviality and beneath them. Perhaps they thought any idiot should be able to do it. Sorry. Wrong. Dead wrong. Dead just like oqo is today. You need to be every bit as smart, clever, and creative to do manufacturing and test as to do design work. When you take a product all the way to production, you will learn to respect everybody involved. Respect the planners, the clerks, the assemblers, and ALL the engineers. So be a good and respectful design engineer and make sure your CAD files are a good as they can be before you send them out for production. That is your responsibility, not a Flextronics responsibility.

The cover of my mentor Bob Pease’s book Troubleshooting Analog Circuits has one of his “airball” prototypes on the cover. Application engineers like Bob can do proof-of-concept, but don’t try sending this out for high-volume manufacturing.

So like all things, prototyping has an analog continuum to it, There is a place for quick-and dirty hacks. There is place for super-diligence. And there is a whole spectrum of tradeoffs for an appropriate design effort in between those poles. Just don’t do some rush-job today that you just know will bite you a few months later.

[Update] I showed this post to Wayne Yamaguchi and he had this great comment:

“One of the major drawbacks to solder paste is the shelf life.  No matter how you buy it, the container and contents will go bad in about 6 months time, even in the refrigerator.  It’s the flux that ages and slowly solidifies making the reflow consistency different over time.  It’s just a real pain to dispense the paste with a tiny-tip syringe when it it’s fresh, and even worse when it has aged a bit. If I recall correctly, the smallest syringe runs about $50.00.  You can solder a lot with that but if you only make one proto this is an expensive proposition.

“Mine tends to absorb water over time and this makes it pop when reflowing, blowing off chunks of solderpaste in all directions. I can hand-solder 0402 and DFN parts with the soldering iron.  I only need the hot-air station for pads that are not exposed, like power pads and some SMT inductors. I would recommend a really good soldering iron or two, and a general-purpose hot air station if you want to hand-solder small runs of boards. You will need one with a 0.2mm tip or smaller for the leadless and 0402 parts.

“I prefer to still use leaded solder.  It solders at a lower temp and the chemicals are less caustic, unlike the solder flux used for leadless solder.  Unless I have to, I try to use “no clean” flux and occasionally will use Kester 331 (IIRC) for gold-plated pads.

“Good stencils are cut non-vertical.  The edges are beveled so the bottomside is slightly larger than the top, making the solderpaste less resistant to sticking to the stencil.  Hopefully, when you lift the stencil the solderpaste adheres to the PCB and not the stencil.  I’ve never actually seen the process, but, I always imagined  that the paste would not all tranfer.  I guess it works.  Just doesn’t work in my mind.  The bevel is only 5-10 degrees.  Hardly noticeable by the eye, but, I guess it makes a difference.

“Another gotcha will come when you doing rework.  You remove the part in question and/or use solderwick to clean the pads.  If you try and apply solderpaste right away the residual heat will outflow the flux from the syringe tip which is most frustrating as now the tip is full of paste with no flux.  When this occurs it is like concrete and won’t flow.  It is possible to drain more than the tip into the larger part of the syringe tube rendering the whole syringe load bad.  You can unscrew the tip and plunge out the bad material, but, who knows if you now have the right ratio of flux to solder anymore?  If this is the tube from the vendor you can kiss $50.00 down the drain.  If you transferred solderpaste to a smaller syringe the loss is less than $50.00.”

I guess all those years at HP and Agilent, and then being in business for himself gave Wayne a valuable perspective on prototyping. Many thanks to him, and add you own comments below.

Good electronics videos and articles

My buddy Rob Bowers over at Brocade told me about this video channel for home made (aka Maker) electronics projects. It’s produced by Alan “W2AEW” Wolke. You can see by his nickname and video channel name, he is a Ham radio enthusiast. I never got that bug, my projects were more like a wire wrapped around a nail to make an electromagnet.

The video above is what got my buddy Rob excited. He enthused, “Wow electronics for everybody! There may be hope for me. I watched the one on completing the noise source on the Ham It Up! convertor. He builds it, tests the basics, and the shows a simple use case. I feel .031% less stupid. I wanted to know if I should purchase the noise source parts. ‘Yes’ is the answer, after watching this.”

This is the cool thing about the Maker Movement. Rob is not an engineer. He did software QA in the past and now works at Brocade in the IT department. He is technical, but not formally trained. But the Maker movement is about the fun stuff, and the dreary classrooms and boring lectures are dispensed with in favor of learning with a specific objective in mind. It’s all the fun of engineering without the tedium. We invented computers. They can do the tedium, and the math, for that matter.

Electronics enthusiast Alan Wolke at his bench.

You can see from Alan’s bench the passion he has for radio and electronics in general. Any person with a Metcal soldering iron and a Simpson 260 analog voltmeter is OK by me. The extended CRT (cathode ray tube) housing on that scope makes me think it is the 400MHz Tek 2467B, the fast glitch capture version of the Tektronix 2465B. The CRT is longer to add the plates needed for persistence.

Another cool tip from Rob was about Brocade where he works. He told me the labs have vending machines with cables and mice and other day-to-day engineering essentials. The engineers can just swipe their badge into the vending machine, pick out the cable and be on their way, no requisition forms or hassle. What a class outfit.

The good electronics article tip comes from a fellow eFlea attendee. I saw him at the Roasted Bean in Cupertino and he showed me the latest issue of Nuts and Volts magazine.

Nuts and Volts magazine has a ton of good articles about electronics.

Knowing I worked at Atmel, my pal wanted to point out the above article about Arduino by Joe Pardue. Nuts and Volts is a subscription magazine, so you have pay 27 bucks a year for print and digital, or only 20 bucks a year if you don’t want the print magazine.

Even without subscribing, you can download the code samples for the Arduino 101 article, and if you upgrade to the mysterious un-priced “preferred subscriber network” you get access to all the old issues of Nuts and Volts. This is a great complement to Circuit Cellar magazine, which is also a subscription magazine, but for $250 they can also give you a memory stick with every single article they have ever done. I recommend both these magazines since they are aimed at system design. The trade press, where I have worked, is fine to learn about the latest chip or test method. But Circuit Cellar and Nuts and Volts both show you how to hook up the chips, and do the code and everything else to get a working product. They even touch on 3-D printing and the stuff to put your gizmo in an enclosure. No wonder they can charge for a subscription. All they lack is articles about FCC, CE, and UL approvals, and those might happen one day for all I know.

So keep watching those YouTube videos and reading articles, but more importantly, keep hacking on circuits and code. That is the fun stuff that gives real satisfaction and happiness.

The SAM L21 pushes the boundaries of low power MCUs

Atmel just released a new Atmel | SMART ARM-based microcontroller. While the SAM L21 has the performance of a Cortex M0+ core, it also packs a number of ultra-low-power features. The MCU can even do touch sensing for buttons, sliders, and wheels while using extremity low power. One key component is there are five distinct power domains inside the chip. Most lower-power ARM chips, including Atmel’s, simply disable the clock to the various sections. The SAM L21 turns off power to the sections, hence no leakage currents in the thousands of transistors in that section.

The numbers are impressive even without the power cycling. The SAM L21 uses 40uA/MHz, less than half that of the SAM D20, which uses 103uA/MHz. For static power with RAM retention with RTC, the SAM L21 is 4 times better, using 0.9uA instead of 3.8uA like the SAM D20.

The chip is being evaluated against the ULPBench performance metric. Our early testing shows the SAM L21 to be lower power than any of our competitors’ M0+ class chips.

Now, understand that power consumption is no trivial spec and you have to realize “it depends on what the meaning of the word ‘power’ is.” My buddy Dave Mathis is doing a sensor monitor system that is asleep 99.99% of the time, waking up once a day to take a measurement and send it wirelessly to a host. For that you want that low static power consumption. Prior to the SAM L21, that meant that you would look at our AVR 8-bit XMEGA parts. What is important to this application is static power, which is really leakage current in the transistors that make up the CMOS gates. An XMEGA not only has fewer transistors than any ARM part, they are made on a bigger process geometry, which results in less leakage.

What is truly revolutionary about the SAM L21 is that it provides you with 32-bit performance, but since it turns off power to unneeded sections of the chip, there is no leakage current in those sections. The XMEGA is still better for static power, with only 100nA of leakage in RAM retention mode, but it is still an 8-bit chip. If you want to do Internet of things (IoT) where you need a TCP/IP stack, or if you need to do some number crunching, you need an ARM-class 32-bit chip, and that is what the SAM L21 is.

Application engineer extraordinaire Bob Martin explained to me why you might save total power with a more power-hungry chip. He explained that is the deal with sensor fusion hubs. Most hubs have an IMU (inertial measurement unit), a 3-axis accelerometer, and a magnetometer. Now most of these talk SPI (serial peripheral interface) so you can easily read them out with an 8-bit AVR chip. And, Atmel makes the ATWINC1500 radio chip that has the TCP/IP stack inside, so an AVR can talk to it and out the door the data will go, all the way to the Internet. But realize that radio chips use way more power than microcontrollers, even fast big ones. A radio chip needs power to transmit, and that means they take milliamperes of current during transmit. So this is how a sensor fusion hub saves power. Rather than use the radio chip to send the data from each sensor, the ARM-based chip does the math and pre-processing to combine the raw data from all three sensors and then represent the result as a simple chunk of data representing its position in space. Then the radio chip has to send much less data. By using a more powerful microcontroller, you save total power since you reduce the on-time of the radio chip.

My buddy at Google tells me a fellow there has installed Linux on an AVR. He swears it even has a file system. He also notes that it takes a couple hours to boot into a GUI. So that is the other tricky power tradeoff. If you have to leave a little chip slogging away for a long time, you might be better off with a 32-bit ARM-based chip that can wake up, do its thing, and go back to sleep. The same is true for your external peripheral chips. Nick Gray, an analog application engineer points out you might not want the lowest power ADC (analog to digital converter) chip in your design if you are only reading it occasionally. He says what you really care about is how fast the chip wakes up and can take good data. If a fast high-current chip wakes up 10 times faster but uses twice the power, it will still use less energy, less charge, than the slower chip.

Firmware engineers need to be much more familiar with the hardware than Windows or web programmers. My buddy Wayne Yamaguchi has made an LED flasher that lasts for half a year on a tiny 1220 coin cell. You can spend 128 dollars on these flashing cufflinks, but the battery only lasts for 24 hours. What Wayne does is take the AVRtiny10 out of active mode and uses the RTC to wake it up. That is why his flasher lasts 182 times longer. Then again, Wayne has an EE and used to work at HP/Agilent. He studies the Atmel datasheets and then really gets the best out of them with his programming techniques.

Wayne Yamaguchi shows off his blinkie board that uses an ATtiny10 in SOT-6 package.                                                       It runs for 6 months on a 1220 coin cell.

This is the beauty of the SAM L21. If you are a hardware person like me, you know that cycling power to CMOS chips can cause latchup and other problems. That is why it is hard to turn off sections of a chip like the L21 does, rather than just stop the clock. Atmel engineers have done all the hard work to make sure nothing blows up or latches, and they provided the supervisory logic so that all the power cycling is automatic when you are doing touch sensing. I am confident that there is no better chip to do touch sensing for battery powered gizmos that need the processing power for Internet of things applications.

Just realize there is no simple number or even a benchmark that will tell you the power consumption of your particular applications. I can’t find it again, but I recently read an editorial blog where the fellow wanted someone to take all the microcontrollers from all the manufacturers and “just figure out the lowest power one”. Well, that is impossible and the simplistic thinking we expect from bosses, not fellow engineers. There is no one number, there is no one chip that you can simply say is the “lowest power.” It all depends what you are doing with the chip and firmware you write for it. My pal Harry Holt is an application engineer over at Analog Devices. He is an op-amp expert and op amps have 30 or 40 different specs. Harry has a great line he tells the “newbies” to engineering. He says: “There are only three important specs to an op amp.” Harry’s victim gets excited and says, “What are they; tell me, what are the three specs?” Harry smiles and says “That depends on your application.”

Interested in learning more about the SAM L21? Stay tuned for more details. Meanwhile, you can read the latest blog pieces on the ARM-based MCU here.

Video: Vegard Wollan reflects on life and innovation

In the final segment of my interview with AVR microcontroller creator Vegard Wollan, I asked about his background and innovation at Atmel.

In response to my question of how he views his expertise, Vegard noted that he started out as a computer architect and digital designer. It’s simple to see the ease-of-use DNA in the AVR product line when Vegard then noted that he soon saw himself as someone that could make life easy for embedded designers. I think this focus on the customer pervades all of Atmel to this day.

Vegard Wollan reflects on his history of innovation at Atmel.

I went on to ask Vegard what he does in his spare time. His response? Exercising and boating off the beautiful, dramatic Norwegian coastline. I think physical activity is a key thing. In fact, I wish someone had warned me as a young man that engineering has an occupational hazard. You can make a good living sitting at a desk. This was less true when I was an automotive engineer, as I had to go the experimental garage and walk around Ford’s giant complex in Dearborn, Michigan. Nowadays, we all seem chained to a computer, and stuck in a chair all daylong. So, exercise and boating sounds like a great way to stay active and balance our lives a little bit!

As I pictured Vegard sailing around Norway looking at beautiful sunsets, I wondered if that was inspired him to be so innovative. He responded that the primary source of innovation at Atmel is working with a team of creative innovative people. I think this is true in most human endeavors. When I asked my dad why some restaurants had really good service, he noted that good people like to work with other good people. That is why Vegard is spot-on, and quite humble in noting that innovation comes from a team, not any single person.

Want to learn more about the backstory of AVR? You can tune-in to the entire 14-part series here.

Video: Vegard Wollan addresses Internet of Things security

In this video segment from my interview with Vegard Wollan, the co-inventor of the AVR microcontroller, we explore in detail the security problems you need to address as an embedded designer.

Let’s face it, it is obvious that security is a way of thinking. You have to assume bad people are going to try and hack your products. With the oncoming revolution in the Internet of Things, it is important you design the security within, rather than try to tack something on after an exploit.

The co-inventor of the AVR architecture notes that security is essential in embedded systems.

The key thing you have to know is that nothing beats hardware security. This is where the security system is implemented in silicon, storing a secret key, hash algorithms and random-number generator (RNG). Atmel makes both standalone security chips and incorporates the security circuits into some of our microcontrollers including Atmel | SMART ARM-based chips used for smart energy meters. The chips are more sophisticated than a simple IP block. In fact, there are extra layers of metal in the die so that hackers cannot probe the chip without ruining it. Many of the chips also dither the supply current, so a hacker cannot infer the code it is running by observing the tiny variations in supply current as it runs.

Atmel makes symmetrical security chips, where both the chip and the microcontroller code know the secret key, and also asymmetrical security chips, which work like that public and private keys systems you might be familiar with such as PGP and RSA security. And, note that you can uses Atmel’s tiny inexpensive security chips with any microcontroller, 8-bit, 16-bit or 32-bit, including all the micros made by Atmel’s honored competitors.

Interested in more? You can watch the entire 1:1 interview with Vegard here.

Video: Vegard Wollan talks AVR and ARM low-power operation

In this segment of the series, the co-inventor of the AVR microcontroller chip talks about the famously low power that the chips consume.

I had heard that one of the clever things Atmel does to save memory power is that we turn on the memory, fetch four instruction op-codes then turn the memory off again. Now, if there is a branch in these four op-codes that change the program flow, well, we have to turn on the memory and grab another four instructions. But, you can imagine just how often that the chips are executing all for instructions, so that we get those four op codes for the power cost of one fetch.

Vegard Wollan jokes will fellow Norwegian Andreas Eieland [off camera] about divulging the secrets to Atmel’s ultra-low power.

Vegard confirmed that Atmel does this on both the latest AVR and on our Atmel | SMART ARM-based chips. I love this clip since this is where we break the 4th wall as Vegard jokes to the crew that I am giving away too many secrets. I also confirmed that some of our ARM chips have a switching regulator controller built in. For instance, the SAM4L has one switching and one linear regulator built in. Now we don’t put any giant inductors inside the chip, you supply the external inductor, but all the control circuitry is available so you can really minimize the BOM (bill-of-materials).

To allow single-supply operation the ARM-based SAM4L microcontroller has a switching regulator on board, you only need to supply an external inductor.

This is yet another thing that differentiates our ARM-core parts from the competition. Most engineers know how cool and revolutionary the AVR was, but we have applied all the “cool” and more to our ARM-based chips. As Vegard noted, we have many tricks and innovations to sip the lowest amount of power, and that includes having our own processes at our Colorado Springs fabrication facility.

Vegard Wollan on the AVR and ARM cores and peripherals

In the fifth video of the series, I asked the co-inventor of the AVR microcontroller about the progression of the peripherals in the various microcontrollers Atmel offers. Vegard shares that when they invented the first AVR products, the team was concerned with ease-of-use, a clean instruction set that would run C, instructions that ran in a single cycle, and good quality tools.

However, he was just as proud of the peripherals that they then developed for the XMEGA line of AVR 8-bit chips. There, he said the stress was still on low power, but also a set of peripherals that were high performance, robust, strong, effective, and that included analog and digital advanced peripherals. Additionally, Vegard stressed how the XMEGA event system would allow programmers to handle complex events and take action, all without waking up the CPU core in the part.

Vegard Wollan becomes animated when talking about the peripherals in AVR and ARM chips offered by Atmel.

I knew this was cool for the low-power aspect, yet Vegard reminded me that it also allows you to service an interrupt faster and more deterministically — always a good thing in embedded systems. The great news for engineers is that all the cool things Atmel figured out for the XMEGA AVR also went into to the UC3, the 32-bit AVR product lines. Then, we made sure to put these same powerful and flexible peripheral systems into our ARM core-based MCUs. In addition we would add dedicated touch I/O pins and more accurate clocks and references. You can still see the AVR DNA from back in 1990 at the Norwegian University of Science and Technology where the AVR came to life.

What I really loved about Vegard was his humility. Every time I tried to give him credit for the AVR he was sure to remind me that there was a whole team that developed it. And, when I tried to point that the AVR was RISC (reduced instruction set computer) before ARM came out, he told me that he was more proud of the peripherals in all of Atmel’s chips, rather than just the core he invented for the AVR. That’s a good thing to keep in mind.

While using any ARM core will get you the instruction set and header files and open-source tools, Atmel’s ARM chips will also get these great peripherals and the event system to tie them all together, while the CPU sleeps peacefully. A recent article helped me understand Vegard’s Norwegian modesty, but I am sure glad he and his team worked on the AVR and ARM chips.

Vegard Wollan talks AVR chips and tools

While some of my earlier segments with Vegard explored the history of AVR, this video with its co-inventor addresses its product line and the tools one would use to write the firmware for the 8-bit chips.

Vegard touches on the availability of AVR chips in DIP (dual in-line) packages. These larger packages are loved by Makers and hobbyists since they are easy to prototype with. You can solder to the pins without a microscope and it is easy to make changes. They are also well-suited to installing in sockets, so you can replace them, or yank them out and program them in a separate programmer board.

Atmel still makes parts in the older DIP package, loved by hobbyists and Makers alike.

In the interview, Vegard refers to the ball grid array, commonly referred to as BGA by us acronym-loving tech people. BGAs are extremely small, just a little bigger than the silicon die itself. They also tend to transfer heat out of the die effectively, but that is rarely a factor in AVR chips since they are so low power. The headache with BGA chips is that you need an IR reflow oven to solder them on a board. Now, my buddy Wayne Yamaguchi has figured out a toaster oven will get the job done, just don’t toast any bread in it after you put a lead-soldered board into it.

Atmel parts in BGA packages are very small, but take special inspection and rework equipment.

The real headaches with BGA packages are rework and inspection. To replace the chip, you would need a camera mounted hot-air rework station from Metal/OKI; in order to make sure it is soldered correctly would require an X-ray machine (no, I am not kidding) to see that all the balls have sweated onto the pads under the chip. It helps to use gold-immersion finished circuit boards since they tend to be flatter than HASL (hot air solder-leveled) boards. However, if you are making some leading-edge tiny consumer product, all these prototyping and QC hassles are well worth it to get the smallest size possible.

To remove and resolder a BGA on your circuit board, you need to use a high-dollar camera equipped hot-air station like the Metcal Scorpion from Oki.

Vegard confirmed that Atmel uses the AVR 32-bit UC3 core in our touch controllers and mouse controller products. As you will see in the video above, we then went on to discuss Atmel’s legacy of providing really inexpensive demo boards and development tools.

Vegard Wollan smiles with pride as I show him an old demo board I used in 1999.

I also dragged out the actual AVR ICE 200 in-circuit emulator (ICE) I used in 1998, to design a point-of-sale terminal (note I misspeak in the video, calling it an STK200). The remarkable thing was this system would emulate an AVR chip in-circuit, and it only cost 200 dollars, back in an era when Intel Blue-Box 8051 systems were 50 grand.

Vegard Wollan really beams as I describe the 200-dollar Atmel AVR ICE 200, that got my startup off to a fast start in 2001.

To conclude the segment, Vegard Wollan shares how the Atmel Studio 6 integrated development environment is a high-quality software tool to develop your application, and works with AVR 8- and 32-bit parts as well as Atmel ARM-core microcontroller chips. When you add Atmel Gallery, Atmel Spaces, and the Atmel Software Framework (ASF), Atmel Studio becomes an integrated development platform (IDP). And, don’t forget you can get Atmel demo hardware through our distributors or the Atmel Store.