Overall my original post on buying surplus industrial automation equipment is still on the mark.  I won’t repeat it again this year; instead, here are some comments based on 3 years of monitoring eBay and adding to my collection.

In Silicon Valley there are only two decent electronic surplus stores: Excess Solutions and Advanced Component Electronics.  There hasn’t been a good local source for mechanical surplus since Triangle Research closed its doors.

On the web, PLCCenter has a great selection, is great for getting an idea of what stuff costs new, but has premium pricing (except for some on-sale items).

eBay is still the best source, but you need to be patient and know what equipment is worth.  In general, I’m willing to pay 10-20% of the original cost, but many eBay sellers try to get 50%, and a few even ask for more than 100%.  In general, “Buy It Now” means “I think my junk is worth a lot”.  Availability is very spotty; some months there’s a lot of interesting stuff, some months there is nothing.

Also, be sure to check condition and return policies.  Many eBay sellers do not have the ability, equipment, or inclination to test industrial equipment, so if it says “as-is”, don’t pay a lot.  Most of the industrial equipment I’ve bought has worked, but I’ve bought a number of AMC and Elmo drives that don’t want to communicate (since the lights blink, I haven’t given up yet; I haven’t had time for extensive troubleshooting).

Don’t forget new equipment; many vendors (including Siemens and Panasonic) have offered somewhat-affordable starter packages including equipment and software (e.g. PLC and programming software).  Some new PLCs are so inexpensive you don’t even need a starter kit: for example, Automation Direct’s Click PLC starts at $69 and the software is free.

Comments on specific equipment:

• Last time I checked, it appears the Cognex Insight smart camera software is now a free download (after registration).  But I recommend verifying this before buying an Insight camera (which will probably cost >$100).
• DVT smart cameras are still often available on eBay, with pricing ranging from $50 (Legend 510 bought at the right time) to $500 or more (color model such as the 542C).
• Galil motion controllers availability is good, with a wide range of pricing (there are many unrealistic sellers).  USB and Ethernet models are more expensive, although if you’re lucky you can buy one for under $250.
• CANOpen interfaces from Kvaser and Ixxat are frequently available for $50-$150.  I’d recommend getting a used Kvaser or Ixxat instead of a new interface from someone else (which will be at least $100 anyway) because they have the best software support.
• Copley CANOpen servo drives are available fairly often; a reasonable price for an Accelnet is $50-$120; the Xenus is more expensive (>$150).  The Accelnets are my favorite servo drive.  I avoid the older models (800-xxxx)  because I can’t find any documentation for them.
• Elmo CANOpen servo drives are frequently available.
• Ethernet Powerlink drives and EtherCAT drives are occasionally available , but the prices typically aren’t reasonable.
• MEI controllers are often available, at a wide price range, but I’ve never seen the software included.  If you don’t have MEI software, don’t buy the board.
• Panasonic PLC’s are frequently available, but in general I think the asking prices are too high.  At least Panasonic now provides a code-sized limited (but still quite useful) free version of FPWin Pro 6.
• Opto 22 I/O controllers, such as the B3000 and LCSX, are frequently available, often at reasonable prices ($50 and up).  Opto 22 PACs are rare and expensive, especially the current models.  Opto 22 I/O module availability is good.
• Wago 750 and Beckhoff K-bus availability is good, and, if you’re patient, you can get them at a reasonable price.
  • The most popular couplers are for DeviceNet, CANOpen ($25-$75), Profibus, and Ethernet (>$100 for 750-842); I’ve also seen Interbus, serial, and EtherCAT.
  • Digital input and output modules are the most common, and cheapest.
  • Analog modules are less common, and more expensive, but if you’re patient, you can get one for <$50.
  • Specialty modules, such as encoder interfaces and stepper drivers, are the least common and most expensive.

I’ve used industrial barcode scanners a number of times, and they work well, especially the raster models.  The laser barcode scanners have a wide scan range and can handle a substantial amount of variation in label position.

My favorite  brand  is Microscan; I’ve also used other brands with good results.  The Microscan QX-870 is a typical  barcode scanner: it has 10 scan lines, can do 300 to 1400 scan/sec, has  a read range of 1″ to 30″, can read all the normal 1D barcodes (UPC, Codabar, Code 39, Code 128, etc)  as well as the PDF417 and Micro PDF417 2D barcodes.

However, most 2D barcodes (such as the popular Datamatrix) need an area sensor.  In other words, you can’t use a laser scanner, you have to use a camera.  Now finding a lens that has a 29″ depth of field (from 1″ to 30″) is challenging.  Of course, the barcode reader could use autofocus, but with a normal lens, that adds a lot of complexity, cost, and still isn’t perfect.  One Microscan 2D barcode reader had a halfway solution: the reader could cycle through a preset set of focus positions until it found a good read.  I wasn’t impressed (although to be fair, I never tried that model).

To make matters worse, Datamatrix codes are often used in direct marked applications; for example, using a laser or ball-peen to create a barcode directly on a metal aircraft part.  Creating lighting that is affordable, compact, and can work on anything from a shipping box to a reflective metal part is hard.

I’ve thought for a long time that liquid lenses (now available from optical suppliers such as Linos) could solve the depth of field problem by allowing affordable and rapid autofocus.  Well, they’re here (and both claim to be “the first”): Cognex has the Dataman 200 series, and Microscan has the QX Hawk with liquid lens and modular zoom.

I think this is a big deal; for example, the QX Hawk claims a read range of 1″ to infinity.  Unfortunately I don’t have any personal experience with either reader, but if I need to read 2D barcodes in the future, I will definitely check both out.

Regression testing tries to verify that software changes do not cause current functionality to fail. A few years ago I wrote software to do regression testing for jobs on a Cognex Insight vision system. Since I do not have a Cognex system at my desk, I cannot give all the juicy details, but I can give an outline.

Machine builders often build a machine to work on a small set of sample parts (say 50). But when it gets into production with many more parts, there are often problems because the production parts show wider variations than the samples (or the manufacturer has made changes).

It’s the same for machine vision – you have to pick one particular part to start designing your machine vision job. Suppose the job works well in production, but somebody wants better results – so you grab a new part and get to work, right? Well, if you’re not careful, you can end up with a vision job that works great for that new part, but not for a typical part, and thus end up worse than you started.

So what I did was save a whole bunch of pictures of good parts and bad parts from production runs, then after I made any changes to the vision job, I ran my part database through the camera, and checked the results.

The Insight cameras use some sort of PowerPC processor, have Ethernet, RS-232 serial, and a bit of digital I/O (e.g. for trigger input). The Insight Explorer user interface software runs on a PC, is written in Java (and works pretty well; note that some newer Cognex products use the .NET framework), and uses a spreadsheet approach to machine vision, which has its pluses (such as simplicity) and minuses (like trying to sequence actions).

I like having the cameras on the network; you can work at your desk with the camera mounted far away. Cognex makes it especially nice by using standard Internet protocols, such as ftp to load and save jobs and pictures (BTW, if you need a free Windows ftp server or client, you should look at FileZilla), and telnet to control the camera.

Using telnet is basically like having a command line interface to the camera. You can do a lot with the camera (load jobs, trigger the camera, insert data, get results, etc), and it’s easy to test out ideas at the telnet command line, then codify them into a program.

I used Python with a free, open source telnet library. At least at the time, there was no free telnet library for .NET, and it didn’t make sense to buy one for this simple application. Then I wrote a Python module to do all the camera control I needed.

To do the regression testing, I wrote a Python program that loaded the desired vision job, then went through the database of pictures, loading each picture, triggering the camera (so it would run the job on the loaded picture), recording the result, comparing the result to the desired result, and then scoring the overall results.

Theoretically, it would be possible to do the same thing using the Insight Explorer software without using a real camera. In that case, you would use a GUI functional testing library (and some good free ones exist for Python) to load the jobs and pictures, then check the results. However, since Insight Explorer is written in Java, the normal GUI testing tools did not work (Java visual widgets aren’t the same as the native Windows ones – OK, maybe if you’re using SWT, but I think they were using Swing). I had limited success automatically inserting keystrokes and waiting, but it wasn’t reliable.

Tony